WO2006058738A1 - Système d’évaluation d’image, procédés et base de données - Google Patents

Système d’évaluation d’image, procédés et base de données Download PDF

Info

Publication number
WO2006058738A1
WO2006058738A1 PCT/EP2005/012816 EP2005012816W WO2006058738A1 WO 2006058738 A1 WO2006058738 A1 WO 2006058738A1 EP 2005012816 W EP2005012816 W EP 2005012816W WO 2006058738 A1 WO2006058738 A1 WO 2006058738A1
Authority
WO
WIPO (PCT)
Prior art keywords
pattern
data
database
patterns
location
Prior art date
Application number
PCT/EP2005/012816
Other languages
English (en)
Other versions
WO2006058738A9 (fr
Inventor
Lieven Van Hoe
Bart Verweire
Original Assignee
Lieven Van Hoe
Bart Verweire
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Lieven Van Hoe, Bart Verweire filed Critical Lieven Van Hoe
Priority to US11/791,405 priority Critical patent/US8064663B2/en
Publication of WO2006058738A1 publication Critical patent/WO2006058738A1/fr
Publication of WO2006058738A9 publication Critical patent/WO2006058738A9/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H80/00—ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring

Definitions

  • the present invention is in the field of evaluation of abnormalities in images, image databases and method of data entry.
  • Radiologists are medical doctors who are capable of interpreting images such as those obtained by conventional radiology, ultra-sonography (US), computed tomography (CT), and magnetic resonance imaging (MRI). As these imaging techniques have rapidly evolved technically, more and more anatomical details are available to be assessed non- invasively. For practicing radiologists, it can be difficult to be aware of all possible imaging presentations of all possible diseases. It is estimated that typically a clinical radiologist seeks diagnosis or anatomy assistance on 5-10% of cases daily. The problem is that referring to reference books, web searches, or colleagues interrupts work flow and reduces productivity. Although all information needed by the radiologist to make the correct diagnosis is usually available "somewhere", the radiologist faces three problems with regard to obtaining this information as follows.
  • the information may be fragmented, e.g. there may be basic information about a specific disease in a textbook, while for more detailed information dedicated Internet searches or other books may be required. Thus, even if the user knows where to find the information, finding the complete information he needs may not be straightforward.
  • the radiologist In order to find out to what disease best matches a certain morphologic pattern, the radiologist first needs to have an a priori knowledge about which diseases could cause that pattern, and, secondly, has to find where exactly these diseases are discussed in the book (or other medium), and whether or not the description of imaging findings in this reference text indeed corresponds to his "case". Such a search may be quite straightforward in some cases while it may be quite time-consuming and frustrating in others. Also, even the most recently developed databases on the Internet are organized by location, not by location and pattern.
  • Figure 4 illustrates the steps presently required when a radiologist confronted with an abnormality at a certain location/sublocation and with a particular morphologic pattern wants to identify the corresponding disease(s). For difficult cases, he has to find an appropriate textbook to identify the diseases that may be present with a pattern similar to the one he has identified. During this process, the radiologist creates a list of possible diagnoses (Diagnosis 1 to n), from which the diseases that do not result in the appropriate pattern are eliminated.
  • This list can further be refined by matching the clinical status of his particular patient (e.g., male or female) with the patient-related information provided for the different diseases (e.g. Disease N°1 tends to occur in women). It is clear that this is a quite inefficient process.
  • the radiologist knows that the information he needs is available, he is not able to locate it.
  • the radiologist states that, in such a case, one cannot differentiate between tumor and inflammation.
  • glioblastoma multiforme high-grade glioma
  • ADEM multiple sclerosis
  • abscess abscess
  • the radiologist cannot make a clear diagnosis and is forced to defer making his report until the following morning. His neuroradiobgy colleague will likely be able to tell him whether his preliminary report was somewhat accurate, or not accurate at all.
  • the present invention aims to overcome the disadvantages of the prior art, by providing a system and method for evaluating medical imaging data (and non-imaging data), which follows the natural work flow of the operator, and provides fast, accurate and informative diagnoses.
  • FIGURES Figure 1 Example of a ring-enhancing lesion in the brain with small satellite lesion
  • Figure 4 Shows typical presentation of information in radiology textbooks and other reference materials. The description of morphologic presentations of different diseases
  • Patterns is scattered within the text. Finding relevant information (i.e. finding all diseases that may result in a certain pattern at a certain location) may be a time-consuming task.
  • Figure 5 Steps required when the radiologist consults an existing online database or expert system.
  • Figure 6 Demonstration of the invention's work flow.
  • the possible combinations of patterns and diseases for a given location are represented by the solid circles in the plane of the location.
  • the empty circles correspond to combinations of patterns and diseases that do not occur in practice.
  • Figure 7 Intended work flow for Radiologbal Care. Note the reduced dependency on
  • Figure 9 Demonstration of the invention's work flow.
  • the Radiologist first identifies the location, and the pattern from the medical image. Immediately, the set of possible diseases is reduced, as shown by the solid circles on the intersection between the planes for the given location and pattern.
  • Figure 10 Representation of a case with two patterns (P1 and P2) and locations (L1 and L2). The diagnosed disease is said to be shared amongst the patterns in question.
  • Figure 11 The Expert System automatically presents a reduced list of diagnoses, ordered according to likelihood.
  • Figure 12 Ternary relation between Locations, Patterns and Diseases. This relation is further specified using additional axes Age, Sex, Region. Other axes may be added if necessary.
  • Figure 13 Selection of input parameters corresponds to slicing the multi-dimensional table along the orthogonal axes. This is shown in the dark grey areas of the matrix.
  • Figure 14 Selection of multiple locations/patterns results in multiple slices, and as such more possible diseases. However, the system can automatically give a higher ranking to diseases reoccurring for the specified locations/patterns.
  • Figure 15 System Architecture and major process flows
  • Figure 16 Process Flow for Radiological care, using the invention.
  • FIG 17 Architecture for Online database Access.
  • the grey components are those necessary for the basic set-up, the white components are extensions offering additional functionality.
  • Figure 18 Navigation axes (arrows) and Entry Points (E), towards central database Entity LocationPatternDiseases.
  • Figure 19 The primary use of the database / expert system is to improve radiology work flow. As such, the invention improves the conversion of images to relevant information.
  • Figure 20 The invention relates to the extraction of key parameters out of images and providing these parameters (together with other relevant data such as patient information) to the "black box".
  • the "black box” is designed to convert this information to a clinically relevant output.
  • Figure 22 Extension in which a pattern recognition module is added as a preliminary step. As such, the identification of the abnormality and its location and pattern occurs without human interaction. The resulting information can be sent automatically to the referring physician as a preliminary report.
  • Figure 23 Modification in which the extraction of relevant information out of medical images occurs without human interaction.
  • a potential application could be large-scale screening studies in countries facing a shortage of radiologists.
  • Figures 24 to 28 Example of a blank template suitable for data entry by an expert.
  • Figures 29-43 Examples of partially or fully completed data-entry templates, relevant to patterns located in the liver.
  • Figure 44 Example of a page of a book or output provided by the invention
  • Figures 45 to 49 Examples of a radiologist user interface to the invention DETAILED DESCRIPTION OF THE INVENTION
  • the present invention relates to a method and system for evaluating medical imaging and optionally non-medical imaging data of a subject. Such evaluation assists in the diagnosis of conditions or diseases of the subject.
  • the invention also relates to a computer program implementing the method, a database, and a navigation system for extracting or visualising information from the database.
  • the imaging data relates to at least one abnormality present in one or more medical images of the subject being diagnosed.
  • the abnormality may be characterised by the operator of the invention, for example, in terms of the type of abnormality (the pattem)and/or the location of the abnormality (e.g. the organ, tissue or other location information).
  • the pattern may be, for example, a morphological pattern of the abnormality (e.g. T2 hyperintense, T1 hypointense, peripherally calcified, irregular etc. derivable from one or more medical images of a subject).
  • the pattern may be a non- morphological pattern of the abnormality (e.g. time-intensity curve of contrast uptake, flow curve, muscular contraction pattern etc. also derivable from one or more medical images of a subject).
  • Such data is used as input in a database search to provide a primary diagnosis or list of diagnoses ( Figure 8).
  • One embodiment of the invention is a method and system for assisting with the diagnosis of a patient wherein the location of the abnormality is chosen from a pre-defined selection (e.g. a body location template).
  • a pre-defined selection e.g. a body location template
  • locations may include but are not limited to any of liver, lung, bladder, kidney, brain, spleen, breast, testes, prostrate, colon, stomach, throat, intestine, skin, ovary. Locations may be further restricted by sub-location which refers a position within the location e.g. a part of the brain.
  • the predefined selection may be organised alphabetically, according to location, or organ system etc. A hierarchical selection of locations starting with broad grouping is within the scope of the invention. For example the operator may be presented with a broad group of location categories (e.g.
  • musculoskeletal system central nervous system, musculoskeletal system, circulatory, digestive, respiratory, reproductive, endocrine, urinary
  • subgroups may be selected (e.g. knee, fibula, tibia for the category of musculoskeletal system).
  • the precise location may be selected (e.g. meniscus for the knee).
  • Table 1 shows examples of possible organ systems, locations and sublocations according to a method and system of the present invention. An example of a hierarchical organisation is also depicted.
  • a e oss e organ sys ems, oca ons, su oca ons an erarc ca organ sa on o data entry choices according to a method and system of the present invention.
  • One embodiment of the invention is a method and system for assisting with the diagnosis of a patient wherein imaging data is selected from a pre-defined selection (e.g. a pattern template). It is aspect of the invention that said patterns in the pre-defined selection are divided into categories according to the morphology of the abnormality, uptake of contrast media by the abnormality or function of the abnormality.
  • the pattern may chosen from a predefined selection of morphological patterns (e.g a focal lesion, multiple focal lesions, diffuse disease, abnormal size or anatomy, etc).
  • morphological patterns e.g a focal lesion, multiple focal lesions, diffuse disease, abnormal size or anatomy, etc.
  • it may be chosen from a predefined selection of patterns of contrast uptake (e.g presence or absence, time- related intensity curve, etc). It is another aspect that the patterns in the pre-defined selection may be morphological patterns only.
  • the predefined selection may be organised alphabetically, according to type of scan, according to the pattern group, etc.
  • a hierarchical selection of patterns starting with broad grouping is within the scope of the invention.
  • the operator may be presented with a broad group of morphological patterns (e.g. Focal and Multi-focal lesions, Regional and diffuse disease, Abnormalities in size and congenital disorders) from which subgroups may be selected.
  • the pattern may be selected (e.g. T1 hypo intense lesion, T1 hyper intense lesion, T2 hypo intense lesion, lesions with mixed signal intensity).
  • Examples of pattern groups and patterns defined for the liver location according to the present invention are shown in Table 2 below.
  • An example of a hierarchical organisation of pattern groups and patterns is also indicated.
  • a e xampes o pa ern groups an pa erns, an a erarc ca organsa on o aa entry choices defined for the liver location according to the present invention.
  • the pattern list may also be restricted, based on the input of the location. For example, if the location is the liver, the operator may be presented with a selection of patterns pertinent to the liver. The limited choice of patterns and locations enables the operator to clearly describe the abnormality, and also provides more precisely defined terms for searching the database.
  • the medical images can be obtained by any means of the art.
  • types of medical images include, but are not limited to computer tomography images (CT or CAT scan), positron emission tomography images (PET scan), magnetic resonance imaging scans (MRI), ultrasound images, X-ray images.
  • CT or CAT scan computer tomography images
  • PET scan positron emission tomography images
  • MRI magnetic resonance imaging scans
  • ultrasound images X-ray images.
  • Such images may be combined with computer enhancement methods, computer predictive methods, chemical markers, contrast agents etc., all known to the person skilled in the art.
  • the predefined selection of patterns is further defined by the modality of the medical image.
  • a description of a pattern may thus further indicate the imaging modality e.g. ultrasound, computed tomography, X-ray, magnetic resonance, angiogram or other imaging technique.
  • the non-imaging data relates to other patient information such as the sex of patient, age, ethnicity, immune status and oncological antecedents. Such data may be chosen from a predefined list of choices (e.g. sex (M/F), ethnicity (English, Chinese, American, Polish, Indian, Jewish, Asian, South African, Jamaican, etc)). Such limited choice enables the operator clearly to define the other aspects of the patient. Furthermore, the categorisation of features also provides more clear terms for searching.
  • some non-imaging data may be provided as free text description.
  • the non-imaging data may also used as input in a database search. Examples of types non-imaging data according to the present invention are shown in Table 3 below. An example of a hierarchical organisation of the parameters and possible choices therefrom are also indicated, as are explanatory comments which may be presented to the operator.
  • the user of the interactive expert system will be requested to enter specific non-imaging data only for those combinations of location and pattern where the lists of corresponding diseases vary in function of these non-imaging data.
  • specific non-imaging data For the location "adrenal gland” and pattern "solid lesion”, metastatic disease will be a primary diagnostic consideration in patients with malignant antecedents (particularly lung cancer), while it will only be a secondary consideration in patients without malignant antecedents.
  • a e 3 xamp es o types non- mag ng a a, a erarc ca organ sa on o a a en ry choices and explanatory comments according to the present invention.
  • condition data is also organised into discrete categories.
  • the categories may be organised into hierarchical lists such as, for example, one or more of Neoplastic, Infectious, Inflammatory, Metabolic, Traumatic, Vascular, Ischemic, Degenerative.
  • Each top level category may be subdivided into one or more lower level categories (e.g.
  • Lymphoma Oligodendroglioma, Meningioma, Astrocytoma, Ependymoma, Hemangioblastoma, Chordoma, Craniopharyngioma, Medulloblastoma, Schwannoma, Metastasis, Neurofibroma, Hemangioma, Lipoma, Glioblastoma multiforme, Ganglioglioma, Glioma, Neuroma, Osteoblastoma for the category of neoplastic).
  • Lymphoma Oligodendroglioma
  • Meningioma Astrocytoma
  • Ependymoma Hemangioblastoma
  • Chordoma Craniopharyngioma
  • Medulloblastoma Medulloblastoma
  • Schwannoma Tristasis
  • Neurofibroma Hemangioma
  • Lipoma Glioblastoma multiforme
  • Ganglioglioma Gl
  • the burden of describing the abnormality and non-imaging data by the operator of the invention is alleviated.
  • categorisation allows the database to be searched quickly and accurately, since data therein is already structured into the same discrete categories.
  • the use of discrete categories permits the language or definitions of the input choices to be easily switched.
  • the interface with the operator may be selectable to provide choices such as Latin named categories, English named categories, or a set of synonyms associated with a particular branch of medicine. It also enables the output of the categorised conditions to be provided in a different lexicon. Such flexibility permits the invention to operate at different levels of understanding, with experts having different specialisations, and in different languages.
  • Another embodiment of the present invention is a database, comprising data regarding conditions already diagnosed and associated with at least one abnormality present in a medical image of a subject suffering from the condition.
  • the pattern, the location, and condition associated with the abnormality are categorised in the database according to a discrete selection of patterns, location and condition possibilities.
  • each abnormality may be categorised according to a pattern type (e.g. one of T2 hyperintense, T1 hypointense, peripherally calcified, etc.), according to location (e.g. one of liver, kidney, spleen, etc.) and condition (e.g.
  • the database thus comprises a structured organisation of pre-diagnosed conditions and their associated patterns and locations.
  • the discrete categories are expandable or contractable according to the new methods of imaging, new pattern categories, new imageable abnormalities, new conditions etc.
  • the database further comprises non-imaging information in respect of conditions already diagnosed in a subject suffering from the condition.
  • non-imaging information may be in the form of discrete categories (e.g. sex, race, age, immune status, oncological antecedents) or numerical or textual, non- categorisable information.
  • a database according to the invention is preferably a multidimensional database.
  • a multidimensional database provides at least one dimension for each of condition, location and pattern.
  • a database of the invention comprises all possible combinations of locations, patterns and conditions to which specific attributes and more detailed information are attached. Such organisation can be visualised in Figure 3, which depicts a 3-dimensionaI arrangement of data with one dimension (axes) for locations, patterns and conditions. Each circle in Figure 3 corresponds to a particular combination of a location, pattern and condition. Four diseases are shown (4 planes). Circles of the same size are part of the plane represented by rectangles in the disease-axis, and correspond to the same condition. Not all combinations of any two of these parameters result in a valid combination- an absent circle indicates no data for a particular combination.
  • the database is depicted as a three dimensional array in the Figure 3, it is purely for illustrative purpose. Means for organising multidimensional data in multidimensional databases and database management system are known in the art and any are within the scope of the invention.
  • the number of dimensions in a multidimensional database is at least 3, and may be 4, 5, 6, 7, 8, 9, 10 or more than 10.
  • the number of dimensions depends on the number of patient-related parameters taken into account.
  • the number of dimensions can be variable and optimised for all combinations Location, Pattern Condition, and other relevant parameters.
  • the invention When the operator of the invention identifies the location of an abnormality, the invention is capable of providing a set of possible conditions. In Figure 6, this is visualized as the solid circles on the mesh plane.
  • This first step of knowing the location is similar to the actual working practice of the radiologist. However, under prior art, the radiologist would have to read all available documentation and find out whether there's a matching pattern for each possible condition at the location. That would include the diseases shown as the empty circles on the grey plane. It is not until the information has been read by the radiologist that the disease is classed as relevant or not relevant. The time spent on reading irrelevant documentation corresponds to lost time. The use of a multidimensional database immediately narrows the choices for the operator.
  • an operator provides the method or system with his choice of pattern and location of an abnormality on a medical image, and initiates a search of the multidimensional database, which returns a list of conditions.
  • the search is performed by extracting data from an intersection of the pattern and location planes crossed by the pattern and the location of the abnormality.
  • the data is extracted along the disease or condition axis.
  • Figure 9 illustrates a mode of operating the invention whereby the operator specifies the location and the pattern of an abnormality. Identification of the location reduces the set of possible combinations to those lying in the vertical plane (shaded plane), while identification of the pattern reduces the set to those lying in the horizontal plane (cross- lined plane). The combination of location and pattern results in a reduced set of possible conditions, given by those diseases on the intersection of the two planes.
  • a further advantage of the multidimensional database is evident where simultaneous searching is performed of more than one abnormality.
  • the search is proceeded by extracting disease data from intersections of the pattern and location planes crossed by the patterns and the locations of the abnormalities.
  • the diseases or conditions which are common to the patterns and locations, as well as diseases attributed to the each pattern and location are rapidly provided.
  • a mode of operating the invention in which two patterns and locations are searched is shown in Figure 10.
  • the operator identifies the locations, and the corresponding patterns.
  • the first combination of a location and a pattern, L1 and P1 result in the set of diseases (the diagonally hatched circles) on the intersection of these planes.
  • the second combination, shown (planes L2 and P2) result in another set of diseases (the dot-filled circles), on the intersection.
  • These two sets of diseases have one disease in common, denoted as the check-filled circles in the figure. This condition will be presented to the operator or Radiologist as the most likely condition for the given input parameters.
  • the invention reduces the set of conditions to those diseases on the intersections between the two pattern planes and the location plane, a higher probability can be assigned to the disease that occurs for both patterns.
  • the other diseases corresponding to the dark hatched or dot-filled circles, each occur for only one pattern and location, and are less likely than this shared disease.
  • the multidimensional database may provide additional dimensions for the sex of patient, age, ethnicily, immune status and oncological antecedents, for example.
  • the non-imaging data permits relationships between for example, ethnicity or immune status and the likely condition, to be determined.
  • the non-imaging data may assist with ranking the diagnoses according to probability. For example, if it is known there is a predisposition to a type of liver cancer in men, such information may be used to increase the probability indication of the condition.
  • the probability might be presented as a percentage, fraction, or be used to adjust the placing of the condition in an ordered list, for example.
  • the non-imaging information may also be used to request further information from the operator. For example, where non-imaging information has not been provided by the operator, and the database indicates a relationship between a likely condition and a nonimaging aspect such as age, for example, the system or method may request this information. Such information may be used to change the probability indication of the condition.
  • the use of non-imaging information significantly improves the speed and accuracy of the diagnosis. The generation of long lists of possible diagnoses is avoided; normally such lists have to be refined by the operator and depends on the knowledge and experience of the operator.
  • the use of non-imaging information allows precise questions to be formulated by the invention and provides focused diagnoses.
  • Ranking information is based on the frequency that a particular combination is encountered in practice, for a given combination of non-imaging parameters.
  • a default frequency may be provided. Exceptions to this default frequency are explicitly stored, together with the combination of the non-imaging parameters for which this exception occurs.
  • the default rankings and exceptions may be defined by experts during the data entry process, based on their own experience and available literature data.
  • One embodiment of the present invention is a multidimensional database comprising medical imaging and optionally non-imaging data of subjects, wherein at least one dimension is provided for each of location, condition (or disease), and pattern of an abnormality, and optionally one or more dimensions provided for non-imaging data.
  • the relational links between location and condition (or disease), pattern and optionally non-imaging dala are provided in the database, so that the data can be represented and searched across data planes and intersections.
  • FIG. 12 An example of a multidimensional database is shown in Figure 12.
  • the upper left part of Figure 12 indicates the combination of the location, pattern and disease visualized as a three-dimensional table or matrix.
  • additional dimensions containing non-imaging data associated with one particular location, pattern and disease (LDP1).
  • LDP1 non-imaging data associated with one particular location, pattern and disease
  • the non-imaging data can be viewed as additional dimensions, and, as mentioned above, be used to rank a disease and/or prompt questions to the operator.
  • Figure 13 depicts schematically as cubes, three diseases (top left matrix, dark grey cubes) corresponding to the selected location and pattern. Each of said cubes is associated with further dimensions of age, sex and region (lower-right matrix).
  • Additional information such as age, sex and ethnicity may further influence the ranking of the diagnoses, and may even eliminate some diagnoses from the list.
  • the invention may prompt the operator to enter such relevant information. Therefrom, the probably or ranking of the disease can be established by the invention.
  • the invention allows multiple patterns to be specified, and may automatically assign a higher ranking to those diseases shared by the given patterns.
  • Such method is depicted schematically in Figure 14. This shows a single diagnosis based on the input of two locations and patterns (top left matrix, dark grey cubes). Such diagnosis may also be given a ranking according any correlation between the patient data in respect of age, sex and region (bottom right matrix), as mentioned above.
  • the database is not updated with continued use by the operator. Instead, the imaging and non-imaging data is provided only by experts, and is validated. Thisfeature is described in more detail below.
  • the invention can also be used as a reference medium, comparable to the classical printed or online books. The operator has the possibility to browse the collected information, optionally according tolocation, pattern and/or disease, or non-imaging data.
  • the invention may be provided with a search engine, allowing the user to quickly find a particular item of interest, be it a location, pattern, disease, or other content.
  • a further aspect of the invention is the use of a discriminative value for the combination of patterns and locations.
  • a discriminative value indicates to the method and system whether for the given location, a pattern is discriminative between diseases or is not.
  • Such value is based on the proven significance of the patterns in making a diagnosis. For instance, a focal liver lesion is characterised by a limited number of patterns, of which the operator of the invention will be able to choose. If the operator unknowingly identifies a pattern which has low discriminative value, (e.g. "CT hypodense") he might receive the message that this pattern has a low discriminative value. The invention may then suggest a better approach (e.g. assess pattern of vascularisation).
  • the invention may provide a list of diagnosis, in which the discriminative value is used as a factor to order the list by probability.
  • the database may thus comprise the discriminative value as a further feature of the combined location and pattern dimension.
  • the discriminative value of a pattern at a particular location may be provided by the expert.
  • a system of the invention may be one or more device comprising at least one microprocessor capable of performing a method of the invention.
  • the system may comprise at least one or two networked computers.
  • the operator of the invention may be a specialist or a non-specialist in the field of study.
  • An example of an operator of the invention is a radiologist or a specialist in for example, cardiology, gynaecology, oncology, or other practitioner for whom interpretation of medical images is necessary.
  • an operator of the invention identifies the location and the morphologic appearance (pattern) of an abnormality. All possible relevant combinations of locations and patterns are available in the database. The operator just has to select the appropriate location and pattern. As a result, he is given a focused set of diagnoses. Information on the likelihood of each diagnosis for this particular combination of location and pattern is also given. For some combinations of location and pattern, the system may indicate that additional input is required. Input of this additional (patient- related) information by the user results in an optimised ranking of the different diseases for that particular patient.
  • the flow of input and output according to a method of the invention is shown in Figure 11, in which the system performs a method of the invention.
  • GUI graphical user interface
  • RIATM Radiology Intelligent Assistant
  • the operator is presented with a GUI from which selections are available depending on tie desired search. He can select a search of organs (locations), diseases, patterns or text ( Figure 45, 451). Already selected is the organ system of the abdomen and pelvis (452), from which a list of organs pertinent to each of the upper abdomen, pelvis and both of these is provided (454). In this case, the radiologist has selected Liver (453) as the location. The operator makes further selection of the category of pattern observed (e.g.
  • the focal / multifocal contrast uptake pattern in the liver observed by the operator can be identified from the lists of options (461) presented in Figure 46, from which he selects 'hypervascular, homogenous' (462). Referring to Figure 47, having selected a 'focal and multifocal - hypervascular, homogenous' pattern (471), the interface provides an image of a prototype example (474), a list of diagnoses (472) and general information (473) such as technical remarks, diagnostic values, and a checklist.
  • focal nodular hyperplasia (481) the operator is presented with further selections in Figure 48, such as key discriminative features (482) and general features (483).
  • the first selection provides further information such as specific imaging findings, and specific non-imaging findings.
  • Each discriminative finding is listed, some of which are associated with an additional image.
  • the thumbnail image linked thereto (494) is enlarged (495).
  • the operator is able to make an improved diagnosis of the disease attributable to the pattern observed. Compared to the currently available reference material, being a printed or online reference book, the operator does not have to carry the burden of reading lists of diagnoses, imaging findings, and other material that is irrelevant to the current case.
  • the operator does not need to enter specific search criteria, that already assume a fixed idea about the resulting diagnosis. Instead, the system allows the operator to restrict himself to information that is readily available, and to postpone the interpretation to a later stage.
  • the method or system is integrated within other systems, such as those with which the operator may already be familiar.
  • RIS/PACS and other Applications Currently, most radiologists work with Radiology Information Systems (RIS), Picture Archive and Communication Systems (PACS) or other computerized medical information systems. It is an aspect of the invention that the method, system and/or database has means to be integrated in these applications.
  • RIS Radiology Information Systems
  • PACS Picture Archive and Communication Systems
  • the method or system is provided with a number of standardized services and interfaces, that will allow the invention to be integrated with other applications.
  • an external application may send a message, containing the request, to the invention implemented as a computer program.
  • the invention may respond with a message containing the result.
  • Such message can be captured by the external application, and processed accordingly.
  • the method is executed using a remote computer.
  • a remote computer Such arrangement permits remote requests to be processed, for example, over the internet, using a network and a central server, or any other remote/central server configuration.
  • the method may be integrated locally by providing a local instance of the content database (multidimensional database) and corresponding interfaces.
  • the invention is capable of communicating with an operator using speech or sounds.
  • the' invention may recognize a vocabulary, and may be able to respond to the user. It may, for instance ask for more input, or provide the results of a particular operation.
  • Such speech integration may incorporate control and output of the operator's usual application (e.g. RIS, PACS or another medical information system), depending on the speech capabilities of this type of application. Such integration would result in a complete speech enabled work flow, so increasing the efficiency of the operator.
  • RIS e.g. RIS, PACS or another medical information system
  • the vocabulary used in a certain medical area oten contains several synonyms and different terminologies, while the data available in a content database often only contains a subset of the available vocabulary. An operator may not be able to understand the terms available for input or provided by an output, or may want to use a more familiar vocabulary.
  • the invention incorporates the use of one or more lexicons.
  • lexicons or semantic sources may be built-in or available as "plugs-ins", or as a file of translations, or any other means available to the skilled person. It is an aspect of the invention that such lexicon can be expanded by the operator, and/or that an operator can create their own vocabulary, and link it to the application.
  • Such lexicon provides an advantage that search items can be compared with the available lexicons and semantic libraries. The method may then optimise the search criteria, and use the results from this lexical and semantic comparison as input for the search operation, in order to capture all relevant data.
  • the output may also be translated according to the understanding of the operator.
  • This lexical layer integration is particularly useful for the more classical approach of the search engine.
  • the method of the invention may be capable of using data from sources other than the database. If such external resources are available, the invention may be linked to these resources to extend the information.
  • Such external resources include conventional databases and multidimensional databases.
  • the external sources should be subject to the highest quality criteria. Therefore, a validation of external sources is preferable, and it is even more preferable that external sources are certified.
  • the availability of external sources provides additional flexibility and extensibility to the invention.
  • the invention may thus provide a central integration point for several databases, each built up by a independent providers or partners.
  • Non-interactive formats include the printed form as a book, brochure, or other paper formats.
  • Other examples of non-interactive formats are static electronic information, such as a collection of linked HTML or XML pages, or using other publishing technologies, such as pdf (Portable Document Format).
  • pdf Portable Document Format
  • Interactive formats include the system being integrated in the end user's medical information system such as a RIS, PACS, or EMR (Electronic Medical Record).
  • the operator may interact with the invention using an interface.
  • the interface may be incorporated into a web browser page, a proprietary interface, an interface generated using a database authoring tool etc.
  • the devices providing at least the interface include a mobile phone, a PDA device, an organiser, a desktop computer, a terminal, a networked computer, a system comprising a microprocessor, an input device and a display device.
  • the application of these interactive formats is not limited by the output device.
  • an application may be provided that offers a view to the system.
  • These applications may be further speech enabled.
  • a non-visual, speech enabled application can also be provided as a means to interact with the system.
  • Static database content may be installed locally. Provisions of sufficient disk space to store the content database are known to the skilled person.
  • static content may be made available via a web server.
  • Provisions of networking e.g. wired or wireless connection to an intranet or Internet, use of a web browser capable of working with the standard transfer protocols http and https etc) are known to the skilled person.
  • Dynamic, interactive content may be made available by the invention.
  • Such dynamic content may be provided by way of a connection to the intranet or Internet, and the use of a web browser capable of working with the standard transfer protocols http and https.
  • a method or system is capable of providing an interface for the purpose of browsing data from the multidimensional database. It may comprise means for extracting data according to the searching requests of the browsing user.
  • the interface may permit graphical one, two or three dimensional representations of data, and means for the browsing user to navigate therethrough.
  • the system may be capable of providing a primarily textual online book edition of the database as mentioned bebw.
  • the system may be provided with a search engine allowing the user to specify a number of search criteria.
  • the result set is presented to him, so he can interactively browse any of the results.
  • This engine may be integrated with a Speech Recognition engine to aid the end user in entering the search criteria.
  • the user may be guided by a number of discriminative questions about the problem area which leads to a reduced set of answers, very closely related to the problem. Again, speech recognition may help the end user in answering the necessary questions.
  • a database of the invention comprises all possible combinations of locations, patterns, and conditions to which specific attributes and more detailed information may be attached.
  • the content is provided by a group of experts in image interpretation, that might include, for example Radiologists, assisted by medical doctors having specialised knowledge about particular diseases.
  • a group of experts in image interpretation might include, for example Radiologists, assisted by medical doctors having specialised knowledge about particular diseases.
  • one or more experts from each location or location groups e.g. liver, kidney, colon etc. may provide information to and/or validate the database.
  • a problem with existing systems is that data entry is subjective and can depend on the understanding and language of the operator. Such variation between operators can lead to incorrect entries in the database.
  • the inventors have designed an interactive data entry application which guides the data entry operator. Such guidance may be by asking questions to the operator, the use of choice buttons, pull down menus, used such to limit and make consistent the input of the operator. By using such application, the content is entered in a consistent form, and the reuse of pre-existing data is maximized.
  • an interactive data entry application is organized according to the natural work flow, and to the natural relations between the entities in the database.
  • the data entry operator identifies the location, optionally sub location, pattern, disease, or a combination thereof.
  • the data entry user may add characteristic information towards any type of entity encountered. This characteristic information is divided into several categories, which further enhances the search capabilities of the invention.
  • the characteristic information can consist of short descriptions, long descriptions, figures, with or without captions. Additionally, references to official publications, if applicable, are stored together with the corresponding data. According to an aspect of the invention, the categories are configurable. If, at a certain moment, the operator is required to provide an additional characteristic, such characteristic may be added to the system. To increase the productivity of the data entry users, the data entry application may be speech enabled. Additional interfaces may be provided for import of data in several structured formats (e.g. comma separated text, spreadsheets, XML documents). This interface may extend the solution to also input 3rd party data or data from remote data entry users.
  • structured formats e.g. comma separated text, spreadsheets, XML documents
  • the human body may be divided into basic anatomical locations. For each location, a radiology expert may be asked to create a comprehensive list of patterns and corresponding diagnoses. A standard template can be created such as shown in Figures 24 to 28 and used to enter this information. Note that the information gathered in the early stages may serve to create a structural backbone. Many more details concerning each location, pattern and disease may be added (together with information linked to specific combinations of a location and a pattern, a location and a disease, a pattern and a disease, and a location, a pattern, and a disease) once the basic structure has been defined.
  • a number of basic patterns may be created (see, for example Table 2). Specific types of patterns may be predefined for specific anatomical areas, e.g. bone, heart, hollow organs, lungs, solid organs. The expert may use any of these patterns they consider appropriate. They can also combine existing patterns, make patterns more detailed, or create new patterns if needed.
  • An expert may use his own case material of the organ that has been assigned to him.
  • he can use a handbook or other reference materials providing a collection of pertinent images.
  • patterns in the following order: (a) patterns of morphology, examples of which include focal abnormalities (single or multiple), diffuse disease, and abnormal size & anatomy.
  • the morphological pattern may be something too large or too small, something that is displaced, a lesion or argan with abnormal density, signal intensity or echogenicity, or other morphological patterns considered by the expert important to make a diagnosis.
  • the pattern may be evident by the absence or presence of contrast enhancement, an uptake of a specific contrast media, a contrast enhancement pattern that is homogenous or heterogenous, by exhibiting an evolution of density or signal intensity over time, or other pattern of uptake considered by the expert important to make a diagnosis.
  • a pattern preferably has a limited list of corresponding diagnoses. Three to five diagnoses is most preferable.
  • the pattern may be modified in order to make it clinically more relevant and to reduce the list of diagnostic possibilities. Modifying a pattern may occur in two ways: by combining two different basic patterns (e.g. "hypervascular AND T2 hyperintense") into one, or by adding one or more details to a pattern (e.g. "hypervascular, homogenous"). Once the list of diseases has been reduced, the pattern may be used. In an example, a pattern "focal lesion, hypervascular" located in the liver, provides too many possible diagnoses.
  • the list of diagnoses is still too long, it may be an option to group diseases.
  • the disease group “granulomatous disease” can be used instead of using “sarcoidosis” and “tuberculosis” as separate diseases.
  • all primary malignant diseases can be grouped and called “primary malignant soft tissue tumors” instead of mentioning "squamous cell carcinoma”, “malignant fibrous histiocytoma”, “fibrosarcoma”, “osteosarcoma”, and “rhabdomyosarcoma” as separate entities. By doing so, the output becomes shorter and more relevant.
  • disease groups may only used when the diseases in the group have more or less the same clinical significance. For instance, it would not be wise to group all malignant tumors together, because metastases and lymphoma have a clearly different clinical significance, treatment, and prognosis.
  • a pattern corresponds to only one or two diagnoses
  • the possibility to integrate the pattern into a more general pattern can be considered.
  • the pattern "fatty lesion” in the liver has a very limited differential (mainly lipoma, myelolipoma, both very rare in the liver).
  • patterns refer to single lesions, e.g. "ring-enhancing lesion”. With respect to multiple lesions, the following situations may occir:
  • the multiple lesions may be defined as a separate pattern (e.g. "multiple ring-enhancing lesions")
  • this pattern may be considered to be included in the corresponding pattern for a single lesion
  • a variant pattern may be defined.
  • the pattern "multiple ring-enhancing lesions" in the liver can be defined as variant pattern of "ring-enhancing lesion” because the likelihood of abscess or metastases increases, although the list of diseases remains unchanged.
  • Variant patterns and their significance can be described as a brief entry in the template. If a pattern results in approximately the same differential diagnoses as a pattern already defined (and that has morphologic similarities) it may be mentioned as "variant pattern”.
  • a pattern location is normally the location where a pattern is actually seen, not the location where an expert knows the pattern comes from. For instance, an exouterine fibroma presenting as an adnexal mass corresponds to the pattern "solid lesion" at location adnexa (not uterus). Similarly, intrahepatic splenosis presents a sa focal liver lesion not as a splenic lesion, and should be indicated in the location liver, not spleen.
  • hypodense lesion adjacent to the falciform in the liver has a different significance than a hypodense lesion elsewhere in the liver (usually corresponding to focal steatosis)
  • a specific sublocation can be defined (e.g. Figure 27, 272). If the significance of a particular pattern is different in that specific sublocation, two patterns can be defined: a pattern “hypodense lesion” for the liver in general, and a pattern “hypodense lesion” for this specific sublocation.
  • the list of diagnoses associated with a pattern depends on patient-specific parameters, such as age and sex.
  • Other patient-related parameters that may be relevant in some cases are the oncologic antecedents (known primary tumor or not), the immune status (immune competent or not), and the geographical area where the patient lives.
  • the possibility to define a specific patient subgroup is within the scope of the invention. For example, in Figure 27 (273) the possibility is available to define a specific patient subgroup (e.g. child). If no specific subgroups are defined, the information is for an
  • an expert would define basic, new or combined patterns until all abnormalities found in his collection/practice or in books/journals fit in at least one pattern.
  • the expert would need to account that - the radiologist using the present invention to assist with a diagnosis may have images obtained either with US, CT, MRI, angiography, or conventional X-ray - an incomplete study may have been performed (e.g. unenhanced CT) In all these cases, all abnormalities observed should fit "somewhere" .
  • An expert may be provided with a blank template for data entry according to that shown in Figures 24 to 28, or equivalent which requests similar information from the expert.
  • Data may be entered by the expert using the following steps, or equivalent steps thereof which provide the same data to the database:
  • STEP 1 The name of the expert, the continent in which he lives, the country therein and the body location assigned to him may be entered ( Figure 24, 241).
  • STEP 2 The technique used to detect the abnormality is indicated, for example, by the use of an arrow ( Figure 25, 252) against one of the list of applied techniques (251).
  • a pattern may be entered ( Figure 27, 271).
  • a specific sublocation (272) and/or a specific patient subgroup (273) are entered.
  • High diagnostic value means that this pattern helps in making a clinically useful distinction between different diseases.
  • the goal is to define the patterns in such a way that all patterns have a high diagnostic value but this is not always possible. If 'low' is chosen, the pattern is a "pattern with low diagnostic value".
  • the comment section (275) may be used to explain what should be done next (see also STEP 6).
  • Organ liver
  • pattern "focal lesion, hypodense (unenhanced CT)”.
  • the following text may be inserted "additional imaging should be performed using either contrast-enhanced CT, US, and/or MRI”.
  • a comment is entered if appropriate.
  • any comment can be entered by the expert.
  • the comment can suggest to look for another feature (e.g. pattern of vascularisation) or to obtain additional tests.
  • the pattern has a high diagnostic value, the expert can also choose to give some comment on this (e.g. which diagnoses are unlikely).
  • the comment may also briefly mention technical requirements such as advanced equipment used or particular procedures needed. For example, a CT study of the coronary vessels may require a spiral CT with 64 slices. A technical requirement may be coded, for example "TR".
  • STEP 7 An example image may be added. This is an image (276) of an actual pattern and which typically illustrates this pattern.
  • STEP 8 The names of the diseases most commonly corresponding to this pattern may be entered, in decreasing order of likelihood (277). Typically, entry is provided for five diseases though more or less may be provided. Disease groups are also allowed such as granulomatous diseases, infectious diseases, metastatic diseases, viral pneumonia, mycotic infection (see also above).
  • the disease stage may also be included in the entry (e.g. primary TB, post-primary TB, late-stage TB) as some diseases may have different imaging features during the evolution of the disease. If no hard data concerning the relative frequency of the different diseases as cause of a particular pattern is available, the experience of the expert can be called on. If the expert is unsure about the ranking (1 to 5), he may use his best guess.
  • this disease is the primary diagnosis.
  • step 10 is different from step 11 : a different type of question is answered: - step 10: a pattern is observed, and the expert must consider how likely a particular disease is
  • step 11 a patient presents with a particular disease, and the expert must consider how likely is a particular pattern.
  • disease N°1 a very common disease
  • disease N°2 a very rare disease
  • a particular pattern may be typical for both diseases ('typical' in step 11).
  • disease N°1 can be a common cause of that pattern
  • disease N°2 is an uncommon cause (simply because it is such a rare disease).
  • Key discriminative findings may be entered. These are defined as specific imaging and non-imaging findings that allow this disease to be differentiated from other diseases at the same location and with the same pattern (2710). A brief summary of the frequency and precise significance of these findings (if any) may be entered. Preferably a unique number to each of the key discriminative imaging findings is assigned. For example, example: liver, focal lesion, hypervascular and homogenous - FNH: central scar (not obligate), strong uptake of iron oxide particles, sometimes central feeding artery, usually iso- or hyperdense in portal venous phase. For each key discriminative imaging finding, a typical example (e.g. one or several images) may be provided ( Figure 28). A number (282) should be placed on the images (281) identifying the finding.
  • STEP 13 Variants of this pattern and their significance (if appropriate) can be briefly described (283). In some cases, variants of a pattern result in a slightly modified differential diagnosis. In this case, rather than defining a new pattern, the variant can be described and its significance (e.g. Figure 28, 283, as discussed above).
  • STEP 14 One or more images illustrating key discriminative features is provided by the expert (Figure 28, 281), preferably referenced (282) to a number in the previous slide. For example, in Figure 30, two reference signs (301) find corresponding images in the addendum ( Figure 31 , 311). If needed to avoid confusion, specific diagnosis or technical details can be mentioned as a legend to the figures(s).
  • This database may be open for continuous updating and refinement by the experts. Thus mistakes can be corrected at a later stage.
  • the expert can select input from predefined lists. Examples of selections from which the expert may choose include for example, location, type of abnormality (category), imaging technique corresponding to a description, possible patient subgroups, common/uncommon (step 10), typical/atypical (step 11). Aside from the predefined choices, the expert may also freely define other inputs such as, for example, definition of sublocations, definition of patterns, technical requirements of an imaging technique, key discriminative features, images, comments, and variant patterns. - Examples of completed templates
  • a product will be a series of books.
  • this series one page may be assigned to each pattern (see example page in Figure 44).
  • the system is configured to prevent data being published, either interactively or non-interactively, without specific approval of dedicated Validation users and Approvers.
  • Validation users and Approvers may be different from the data entry users who are responsible for entering the bulk of the data into the database.
  • Such configuration may allow every data item to be checked before it is released as publishable data.
  • the user responsible for validation and/or approval of the data may be presented a list of non-validated modificafons.
  • One aspect of the invention is a system for navigating a multidimensional medical database as mentioned herein.
  • the system comprises a means for accessing a multidimensional database, a means for inputting navigational information, and means for extracting from the multidimensional database a list of conditions, patterns, locations depending on the navigational information.
  • the system may provide navigation tools including interactive displays which make use of alphanumerical characters and graphics to represent the data.
  • the information contained within the system can be browsed along several axes, depending on the known input parameters.
  • the most natural work flow is to provide input about the location and the pattern, but the system will also allow to be browsed as an encyclopaedia for information related to locations or patterns in case the disease is known.
  • a navigation system comprises means to providing results of navigation as one or more of text, numbers, case examples, drawings, computer generated graphics, video clips, or any other type of relevant output.
  • the methods or systems of the invention maybe provided as a computer program held on a computer readable medium, said program comprising computer code for performing the steps of the method or for providing the functionality of the system. Examples of media include an optical disk, tape, magnetic disk, solid-state memory, hard-drive.
  • the program or system may be available for download across a network.
  • a method or system is implemented into a standalone system, for example, as a package on a desktop computer with a screen and input device, on a laptop computer, on a PDA etc.
  • One embodiment of the invention is a device capable of performing a method of the invention.
  • the database may be present on a remote server and a program present on a networked local computer to provide an operator with an interface for interacting with the database.
  • Such interface may by provided by known technologies, for example, displayed in a web page, a proprietary interface, an interface generated using an authoring tool etc.
  • the invention includes any technology which permits the operator to interact with the invention.
  • the invention is capable of displaying a web page on a remote computer. Said web page permits the operator to use the invention. It is an aspect of the invention that the use of the method or system by the operator is recorded by the invention for the purpose of billing the operator or his employer. Such billing systems are known in the art.
  • the invention may provide each operator with an account which is charged according to the use of the invention. Such charging may be according to time, the number of searches, complexity of searched, volume of data transfer, or by license with privilege options, etc.
  • the database is preferably curated by a team of experts in pattern recognition.
  • the database is multidimensional, comprises at least three dimensions, with location, pattern and condition data being stored in least one dimension each.
  • Non-imaging data comprising the sex of patient, age, ethnicity, immune status and oncological antecedents, for example, may be provided in additional dimensions.
  • the database comprises all possible combinations of locations, patterns and conditions to which specific attributes and more detailed information are attached.
  • a database product is capable of providing or interacting with an interface for the purpose of browsing data from the multidimensional database.
  • the interface may comprise means for extracting data according to the searching requests of the browsing user.
  • the interface may permit graphical one, two or three dimensional representations of data, and means for the browsing user to navigate therethrough.
  • the database may be capable of providing a primarily textual online book edition of the database as mentioned below.
  • the database product may be provided with a search engine allowing the user to specify a number of search criteria.
  • the result set is presented to him, so he can interactively browse any of the results.
  • This engine may be integrated with a Speech Recognition engine to aid the end user in entering the search criteria.
  • the database user may be guided by a number of discriminative questions about the problem area which leads to a reduced set of answers, very closely related to the problem. Again, speech recognition may help the end user in answering the necessary questions.
  • One embodiment of the invention is a device comprising the database product or capable of accessing the database product.
  • the multidimensional database product may be provided as a computer program held on a computer readable medium.
  • media include an optical disk, tape, magnetic disk, solid-state memory, hard-drive.
  • the database program or system may be available for download across a network. It may be available under licence or pay-per-use.
  • the present invention closely follows the natural work flow of the Radiologist or medical specialist. Contrary to the current state-of-the-art, the invention allows a considerable reduction in the need for manual intervention, by making an automatic pre-selection based on a number of well-known measurable parameters. As such, the invention facilitates and improves radiology work flow, i.e. the extraction of relevant information from medical images.
  • the invention is structured around a multi-dimensional relationship between location, pattern and disease, and shows further dimensions according to case specific parameters such as age, sex and geographical region. This allows a business intelligence like approach to a content database.
  • the introduction of a ranking of diseases for a particular set of input parameters helps to increase the quality of the diagnoses.
  • the invention enables the user to obtain relevant information after having identified some basic parameters such as Location and Pattern.
  • the invention covers process that enables the transformation of images to diagnoses, and the technical structure of the system is irrelevant to the operator. This is illustrated in Figure 20, where the database / system is indicated as "black box", the functioning of which the operator is unconcerned.
  • the system of the invention has been designed to facilitate and improve natural work flow of the operator.
  • the operator is sometimes referred to as a radiologist, such professional can be any involved in the process of diagnosing a condition on the basis of medical image data.
  • FIG. 15 the architecture of the system of the invention is shown in Figure 15 which is elaborated below.
  • This part of the architecture adds content to the database (1). It is preferably located at the site of the data producer. Preferably, no external operators (i.e. not contractually linked to the producer) would be allowed in this environment. Several roles can be defined, each of which are discussed below.
  • a non-specialist may be dedicated to entering pre-defined sources into the database, via the Data Entry module (6) in the Data Creation environment.
  • the data entered should preferably be validated by a specialist.
  • validation is still another step, the Data Entry operator can be responsible for the bulk import of much of the data.
  • a radiologist may have the same access privileges as a Data Entry operator, but may also allowed to check and validate (7) the data.
  • final validation and approval (8) may be the privilege of an editor or project supervisor. This ensures the high quality of the database content, as any data element is subjected to multiple quality controls.
  • This user may be responsible for the definition of new users, assigning access rights, and for the creation of new categories, and for other administrative tasks using a specific Administrative Module (9), that, for reasons of security, may not be accessible to other Data Entry users or Radiologists.
  • a specific Administrative Module (9) that, for reasons of security, may not be accessible to other Data Entry users or Radiologists.
  • All operations performed by any of these roles may access the database through a common Business Model layer. This layer makes sure that the data remain consistent, even after multiple modifbations.
  • the user may be assisted by a Speech Recognition Engine, both for entering data into text fields, and for navigating throughthe pages.
  • an External Content Provider may provide data.
  • data may be wrapped in a well-known format, e.g. XML (11), which is automatically integrated, using the same Business Logic, into the Content Database.
  • Non-interactive Access refers to fixed layout
  • Interactive Access refers to dynamically changing representations of the data. This is explained in more detail in the next paragraphs.
  • users can obtain a license that allows them to browse online a number of books published using content provided by the invention.
  • Advantages of this type of access are - updates of the database can be part of the license.
  • An online user will be guaranteed to browse the most recent content release.
  • the user may switch between several books on his personal bookshelf, with possibly a maximum number of switches per fixed period.
  • the XML document can be split-up, using an XSLT-style sheet (22) into several linked static html pages, to serve as an online publication of the material.
  • these html pages may be served to several electronic devices such as Workstations, Laptops, PDAs, mobile phones, or other devices (26).
  • Interactive access can add value to the information provided by the method compared to the static resources.
  • Different forms of interactive access are supported by the invention, each of which requires the release of the Content database (1) to an Expert Database (30). This is a copy of the database, in which some technical attributes for monitoring modifications are omitted.
  • the Expert Database may contain several versions, depending on the licensing model.
  • This type of application allows the operator user, a Radiologist (31 ) or specialist in another medical discipline with online access to the content database, to specify a number of search criteria.
  • the result set is presented to him, so he can interactively browse any of the results.
  • This engine may be integrated with a Speech Recognition engine to aid the end user in entering the search criteria.
  • the end user (31) is guided by a number of discriminative questions about the problem area (based on the natural work flow navigation Location, Pattern, Disease), which leads to a reduced set of answers, very closely related to the problem. Again, speech recognition may help the end user in answering the necessary questions.
  • the necessary search criteria are set up, and sent to the producer's site, where a web service (35) listens to the incoming calls, and returns the appropriate answers to the questions.
  • connection between the remote site, and the producer's site may be implemented as a VPN, or using secure socket layer technology (https).
  • All presented solutions may be made available on a number of devices, including workstations, laptops, but also handheld devices suchas PDAs and mobile phones.
  • Such database may be integrated into the device.
  • the technology is not limited to currently available electronic devices. Future electronic devices that may appear on the market, that for example, allow a connection to be made to the Intranet or internet, and are able to display web based content, are appropriate to act as a display device for the invention's content.
  • pattern info e.g. diagnostic value of a specific pattern
  • an intermediate step providing patient-related information to the system.
  • the present invention naturally follows the work flow of the radiologist by: requesting the organ group, location, and sublocation requesting data regarding the pattern, - matching this information to cases in the expert database.
  • the technical architecture for online database access is based on the MVC (Model-View- Controller) design pattern.
  • MVC Model-View- Controller
  • the presentation layer is separated from the view layer, and as such, the presentation logic is split from the business logic. This allows the Model components to be easily reused when other View components are being used to display the model d£a.
  • FIG. 17 An example of a basic implementation of the technical architecture is denoted by the grey components in Figure 17. It consists of a content database (the Data Layer), a set of model components on top of a database access layer or DAO layer (the Model Layer), a set of controller components (the Controller Layer), a collection of view components, for several end user purposes such as data entry, data view, search engine or expert system (the View Layer), and finally a thin client in the form the browser.
  • a technical implementation may comprise one or more of these componente
  • This architecture may most commonly be implemented using the J2EE or .Net platform.
  • the J2EE platform has been chosen.
  • frameworks are available, to enforce the use of the correct design patterns.
  • commercial providers such as IBM, Oracle or Sun offer Integrated Development Environments (IDE) that aid in the rapid development of applications in these platforms.
  • IDE Integrated Development Environments
  • the Model layer may be implemented as a set of Java classes or Enterprise Java Beans.
  • the controller may be based on the Struts Framework, currently the de facto standard for the development of scalable web applications, and the View layer may be implemented using jsp and html pages, also making use of the functionality provided by the Struts framework.
  • the web-container necessary for running the Controller and View components, may be chosen from open source projects, or, for example, from the current high performance application servers from IBM, Oracle, Sun, etc. The choice depends on non-functional requirements such as high availability, performance, etc.
  • the implementation of the web-services for integration with 3rd party systems such as medical information systems may make use of the J2EE platform. This allows integration with the majority of applications conforming to the web-services standards, including those implemented on a Microsoft platform. Integration with Microsoft .Net-based applications can be easily provided. Extensions
  • Lexical and Semantic Integrator Integration with lexical and semantic databases is completely transparent to the end user.
  • the integration may therefore be implemented on the model level, as part of the business logic of the system.
  • Such means for implementing lexical and semantic functions are known to the skilled person.
  • the system may be open to accepting 3rd party data.
  • the most common way of integrating external data is by transforming it into XML, and by capturing the XML data, processing it and integrating it in the content database.
  • An XML integration component may, therefore, be included. As the delegation of the connection to either the internal or an external content database should preferably be completely transparent to the end user, this integration may be performed at the model level, becoming part of the business logic of the system.
  • the database is centred around the notion of Locations, Patterns and Diseases.
  • the model contains the central concept LocationPatternDiseases, which contains all combinations of the three main entities, around which three navigation axes have been defined. These axes allow navigation through the database starting from a particular point of view.
  • the navigation axes are locations, with entry point Organ System - patterns, with entry point Pattern Group diseases, with entry point Disease Group
  • the most common navigation axis will be the location axis, as this will be the main entry point for most search and entry operations.
  • a system implementing a method of the invention as described herein may have one or more of the following modifications: - A modification in which the user navigates through the application using speech, i.e. the user can give specific commands with speech.
  • an intelligent interface between the dictated text (interpreted by speech recognition technology) and the system automatically extracts the most appropriate location and pattern out of the dictated text (using advanced pattern matching techniques, e.g. based on Bayesian Inference and Claude Shannon's principles of information theory) and navigates to the appropriate location/pattern.
  • the radiologist does not have to interrupt his normal work flow at all. Instead, while dictating his/her report, the intelligent interface identifies the combination location/pattern best corresponding to the dictated words, and automatically displays the list of likely corresponding diseases. With one simple action, the user obtains access to further information about this disease.
  • Inputted data may include location and pattern of an abnormality such as a defect in an object.
  • An object can be any such as an engineered object a body part, a building construction, a landscape feature etc. It can be any object susceptible of abnormality for which an abnormality pattern and its location can be characterised, preferably according to discrete categories.
  • a multidimensional database would store lists of diagnoses linked to pattern and location.
  • pattern analysis there is automatic definition of the location and pattern, and identification of the abnormality, so not needing the input of the radiologist (or other user).
  • the radiologist identifies the abnormality and the system defines the location and pattern. Theoretically, this can be performed by pattern recognition or by computing the similarity between a new image and reference images in the database (each linked to a specific location/pattern/disease combination).
  • This extension can be considered as a serial use of two "black boxes", one for identification of the abnormality and pattern analysis, and another one for conversion from image and patient-specific parameters to clinically relevant information ( Figure 22).
  • the Invention is not used primarily as a facilitator of the radiologist's work, but as a parallel or even alternative circuit.
  • the Invention may be used to provide a rapid provisional diagnosis. While typical process times in radiology (time between the actual examination and the arrival of the report where it is needed, i.e. in the hands of the referring physician) are measured in hours, the process time can be reduced to a few minutes or even seconds if the entire process occurs without human interaction.
  • a rapid preliminary report can be sent to the referring physician (e.g. as an SMS to his/here mobile phone or message to another mobile device) allowing immediate action if needed.
  • the invention could be used not only to facilitate radiology work flow but to improve medical care directly. Potential "clients” would then not only be radiologists but all physicians. - Theoretically, it could also be possible that, for some applications, the computer- generated report is sent to the referring physician, or even directly to the patient, without need for supervision by a radiologist ( Figure 23). It can, for instance, be envisaged that certain screening studies conducted on large scale would be performed without human interaction, at least in a first step. Such a set-up could be a necessity, particularly in countries facing a shortage of radiologists.
  • a virtual colonoscopy study could be interpreted by a computer and the result could be sent to the patient's mobile phone (or another device) immediately after the exam, with the additional comment that the patient has to consult his or her physician if the report is not completely normal.
  • One embodiment of the present invention is a method for assisting a physician with making a diagnosis based upon a medical image, the method comprising: providing a location template allowing for a selection of one of a plurality of body locations; receiving at least one selected body location; accessing via a digital computer a table of patterns observed at that body location, the table derived from a digital storage medium storing multi-dimensional data including body locations, patterns, and potential diagnoses; providing a pattern template allowing for selection of one or more patterns from the table of patterns; receiving one or more selected patterns; using the selected one or more patterns to access the storage medium so as to determine a list of possible diagnoses; and outputting the list of potential diagnoses.
  • Another embodiment of the present invention is a method as described above, further comprising receiving patient information.
  • Another embodiment of the present invention is a method as described above, wherein the potential diagnoses are ordered according to likelihood.
  • Another embodiment of the present invention is a method as described above, wherein the potential diagnoses are ordered taking into account patient information.
  • Another embodiment of the present invention is a method as described above, wherein the pattern template provides displayable images of patterns in the table.
  • Another embodiment of the present invention is a method as described above, wherein the patterns are placed in relevant order on the pattern template based upon the selected body location.
  • Another embodiment of the present invention is a method as described above, wherein using the selected one or more patterns to access the storage medium includes locating within a multi-dimensional database the one or more potential diagnoses based upon the one or more selected patterns and locations.
  • Another embodiment of the present invention is a method as described above, wherein locating is achieved by extracting data from the multi-dimensional database at an intersection of pattern and body location planes crossed by the selected body location and pattern.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis based upon a medical image, the system comprising: a processor that outputs one or more templates for display on a display device and receives user selections in response to presentation of the templates on the display device; a digital storage medium containing multi-dimensional data including body locations of a medical image, patterns, and one or more potential diagnoses; wherein the processor outputs one or more diagnoses retrieved from the digital storage medium based upon the user selections.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the multi-dimensional data is stored in a multi-dimensional database.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the one or more templates are stored in memory and are retrieved from memory by the processor.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the processor outputs a body location template that allows for user selection of a location on a human body associated with a medical image.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the processor outputs a pattern template based upon a user selection of a body location and the pattern template allows for user selection of one or more patterns by the user.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the digital storage medium contains a multi-dimensional database and the processor locates the one or more potential diagnoses in the multi-dimensional database based upon user selection of one or more body locations and one or more patterns.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the digital storage medium and the processor may be in communication over a network.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, further comprising: a display device coupled to a computer system for receiving and displaying the output from the processor.
  • Another embodiment of the present invention is a system for assisting a physician with making a diagnosis as described above, wherein the computer system and display device are in communication with the processor over a network.
  • Another embodiment of the present invention is a system for determining a diagnosis based upon a medical image, the system comprising: a processor for receiving a signal related to the medical image containing information regarding location of the medical image within the human body and regarding a pattern within the image; and a database containing a plurality of potential diagnoses wherein each diagnosis is associated with at least one location and one pattern; wherein the processor retrieves one or more diagnoses from the database based upon the location and the pattern received in the signal and provides the one or more diagnoses as an output.
  • Another embodiment of the present invention is a system for determining a diagnosis based upon a medical image as described above wherein the diagnostic database is a multi-dimensional database.
  • Another embodiment of the present invention is a system for determining a diagnosis based upon a medical image as described above wherein the processor outputs a pattern template based upon a user selection of a location and the pattern template allows for user selection of one or more patterns by the user that represent patterns found in the medical image.
  • Another embodiment of the present invention is a further method for assisting a physician with making a diagnosis based upon a medical image, the method comprising: providing a template for making a selection of one of a plurality of locations of a human body and for making a selection of one or more of a plurality of patterns; receiving a signal representative of a selected body location and a selected pattern; and outputting an output signal containing one or more potential diagnoses retrieved from a database that relates body locations, patterns and potential diagnoses.
  • Another embodiment of the present invention is a further method as described above, further comprising receiving a signal representative of patient information.
  • Another embodiment of the present invention is a further method as described above, wherein the potential diagnoses are ordered according to likelihood.
  • Another embodiment of the present invention is a further method as described above, wherein the potential diagnoses are ordered taking into account patient information.
  • Another embodiment of the present invention is a further method as described above, wherein the template provides textual information displayable on a display device regarding patterns typically representative of medical images that are associated with the selected body location.
  • Another embodiment of the present invention is a further method as described above, wherein the output signal is displayable on a display device.
  • Another embodiment of the present invention is a further method as described above, wherein the method is performed on a computer system.
  • Another embodiment of the present invention is a further method as described above, wherein the patterns are ordered on the template based upon the selected location.
  • Another embodiment of the present invention is a further method as described above, further comprising: locating within a multi-dimensional database at least the one or more potential diagnoses based upon the selected pattern and selected body location.
  • Another embodiment of the present invention is a computer program product having computer readable code thereon for use with a computer, the computer program product assisting a physician with making a diagnosis based upon a medical image, the computer code comprising: computer code for providing a location template allowing for a selection of one or more locations of a human body; computer code for receiving a signal representative of at least one selected body location; computer code for providing a pattern template allowing for selection of one or more patterns, wherein the pattern template that is provided is based upon the selection of the body location; computer code for receiving a signal representative of a selected, pattern; computer code for transmitting a signal containing one or more diagnoses based upon the one or more selected body locations and patterns.
  • Another embodiment of the present invention is the computer program product as described above, further comprising: computer code for receiving a signal representative of patient information.
  • Another embodiment of the present invention is a computer program product as described above, wherein the diagnoses are ordered according to likelihood
  • Another embodiment of the present invention is a computer program product as described above, wherein the diagnoses are ordered taking into account patient information.
  • Another embodiment of the present invention is a computer program product as described above, wherein the pattern template provides images of patterns that are associated with the selected location.
  • Another embodiment of the present invention is a computer program product as described above, wherein the transmitted signal is displayade on a display device.
  • Another embodiment of the present invention is a computer program product as described above, further comprising: computer code for ordering the morphological patterns on the pattern template based upon the selected body location.
  • Another embodiment of the present invention is a computer program product as described above, further comprising: computer code for locating within a multi-dimensional database at least the one or more diagnoses based upon the one or more selected patterns and body locations.
  • Another embodiment of the present invention is an alternative computer program product having computer readable code thereon for use with a computer, the computer program product assisting a physician with making a diagnosis based upon a medical image, the method comprising: computer code for providing a template allowing for a selection of one of a plurality of locations of a human body and allowing for a selection of one or more of a plurality of patterns; computer code for receiving a signal representative of a selected body location and a selected pattern; and computer code for outputting an output signal containing one or more potential diagnoses retrieved from a multi-dimensional database that associates body locations, patterns and diagnoses.
  • Another embodiment of the present invention is an alternative computer program product according as described above, further comprising: computer code for receiving a signal representative of patient information.
  • Another embodiment of the present invention is an alternative computer program product according as described above, further comprising computer code for ordering the diagnoses according to likelihood.
  • Another embodiment of the present invention is an alternative computer program product according as described above, further comprising computer code for ordering the diagnoses taking into account patient information.
  • Another embodiment of the present invention is an alternative computer program product according as described above, wherein the template provides images of patterns that are associated with the selected location.
  • Another embodiment of the present invention is an alternative computer program product according as described above, wherein the output signal is displayable on a display device.
  • Another embodiment of the present invention is an alternative computer program product according as described above, further comprising computer code for ordering the patterns on the template based upon the selected body location.
  • Another embodiment of the present invention is an alternative computer program product according as described above, further comprising: computer code for locating within a multi-dimensional database at least the one or more diagnoses based upon the one or more selected patterns and body locations.
  • step (d) extracting from the multidimensional database list of conditions corresponding to the imaging data determined in steps (a) and (b), and (e) providing an evaluation of abnormality using list obtained in step (d).
  • Another embodiment of the present invention is a method as described above wherein the multidimensional database further comprises data regarding the discriminative value of one or more combinations of location and pattern.
  • Another embodiment of the present invention is a method as described above wherein the list of conditions is ranked according to probability.
  • Another embodiment of the present invention is a method as described above wherein a low discriminative value of a selected pattern generates a request to use an alternative approach.
  • Another embodiment of the present invention is a method as described above wherein the multidimensional database comprises one or more additional dimensions corresponding to non-imaging medical data of the subject.
  • Another embodiment of the present invention is a method as described above wherein said non-imaging medical data comprises one or more of patient sex, age, ethnicity, immune status, and oncological antecedents.
  • Another embodiment of the present invention is a method as described above wherein a request is generated to provide non-imaging medical data, when the database indicates such data can adjust the reported probability of a condition.
  • Another embodiment of the present invention is a method as described above wherein the reported probability of a condition is adjusted according to non-imaging data provided.
  • Another embodiment of the present invention is a method as described above, wherein providing at least part of the non-imaging and/or imaging data comprises the use of speech recognition.
  • Another embodiment of the present invention is a method as described above, wherein at least part of non-imaging and/or imaging data is provided by means of an interaction with the method. ⁇ >
  • Another embodiment of the present invention is a method as described above, wherein at least part of the data is provided by means of a speech enabled interaction with the method.
  • Another embodiment of the present invention is a method as described above, wherein the pre-defined selections are comprised in a set of terms represented by a lexicon, wherein the lexicon can be changed according to the understanding or language of the operator.
  • Another embodiment of the present invention is a method as described above wherein the imaging data and non-imaging are provided to the method via another application.
  • Another embodiment of the present invention is a method as described above wherein the list of diagnoses is provided by the method to another application.
  • Another embodiment of the present invention is a method as described above wherein said other application is a Radiology Information Systems and/or Picture Archive and Communication Systems.
  • Another embodiment of the present invention is a method as described above, wherein the method accesses at least one other database.
  • (c) means for accessing a multidimensional database comprising data of patterns, locations, and conditions associated therewith, in which - data of each of patterns, locations, and conditions is comprised in separate dimension(s), and
  • (d) means for extracting from the multidimensional database a list of conditions corresponding to the imaging data determined in steps (a) and (b).
  • Another embodiment of the present invention is a system as described above, further comprising the multidimensional database of step (c).
  • Another embodiment of the present invention is a system as described above wherein the multidimensional database further comprises data regarding the discriminative value of one or more combinations of location and pattern.
  • Another embodiment of the present invention is a system as described above comprising means to rank the list the condition according to probability.
  • Another embodiment of the present invention is a system as described above comprising means to generate a request to use an alternative approach when a low discriminative value pattern is inputted in step (b).
  • Another embodiment of the present invention is a system as described above wherein the multidimensional database comprises one or more additional dimensions corresponding to non-imaging medical data of the subject.
  • Another embodiment of the present invention is a system as described above wherein said non-imaging medical data comprises one or more of patient sex, age, ethnicity, immune status, and oncological antecedents.
  • Another embodiment of the present invention is a system as described above comprising means to generate a request to provide non-imaging medical data, when the database indicates such data can adjust the reported probability of a condition.
  • Another embodiment of the present invention is a system as described above comprising means to adjust the reported probability of a condition according to non-imaging data provided.
  • Another embodiment of the present invention is a system as described above, comprising speech recognition means.
  • Another embodiment of the present invention is a system as described above, comprising interactive means.
  • Another embodiment of the present invention is a system as described above, comprising dictated interactive means.
  • Another embodiment of the present invention is a system as described above, comprising means to change the lexicon used in the pre-defined selection in steps (a) and (b) according to the understanding or language of the operator.
  • Another embodiment of the present invention is a system as described above comprising means to receive the imaging data and/or non-imaging data from another application.
  • Another embodiment of the present invention is a system as described above comprising means to provide a list of diagnoses to another application.
  • Another embodiment of the present invention is a system as described above wherein said other application is a Radiology Information Systems and/or Picture Archive and Communication Systems.
  • Another embodiment of the present invention is a system as described above, comprising means to access at least one other database.
  • Another embodiment of the present invention is a system as described above comprising an architecture comprising one or more of: data layer, - set of model components on top of a database access layer, set of controller components, and set of view components.
  • Another embodiment of the present invention is a system as described above comprising units of information linked to the items location, pattern, and disease, or to specific combinations of these items.
  • Another embodiment of the present invention is a system as described above in which information linked to the item location includes one or more of the following information units: anatomy key facts, non-anatomy key facts, anatomy, anatomic variants, non- anatomy facts.
  • Another embodiment of the present invention is a system as described above in which information linked to the item disease includes one or more of the following information units: description, synonyms, abbreviations, incidence, age/sex distribution, etiology, associated disease, organ(s) typically affected, gross pathology, microscopic pathology, histologic subtypes, clinical presentation, treatment and prognosis, imaging findings: general remarks, preferred imaging test(s) to diagnose this disease, suggested nonimaging test(s) to diagnose this disease.
  • Another embodiment of the present invention is a system as described above in which information linked to a combination of a location and a pattern includes one or more of the following information units: technical remarks, prototype example, diagnostic value of this pattern, diagnostic checklist associated with this pattern, comment variants of this pattern and their significance.
  • Another embodiment of the present invention is a system as described above in which information linked to a combination of a location, a pattern, and a disease includes one or more of the following information units: key discriminative imaging findings, key discriminative non-imaging findings.
  • Another embodiment of the present invention is a system as described above comprising means for an operator to remotely access the system.
  • Another embodiment of the present invention is a system as described above, comprising at least one web server.
  • Another embodiment of the present invention is a system as described above comprising means to generate non-interactive or interactive publications of data in the database.
  • Another embodiment of the present invention is a system as described above comprising means to link the database to an application capable of analysing medical images, identifying an abnormality and / or identifying the pattern and location corresponding to an abnormality, either automatically or semi-automatically, and in which the resulting information provides part or all of the imaging data of steps (a) and (b).
  • Another embodiment of the present invention is a system as described above comprising means to navigating along different navigation axes by choosing either location, pattern, or disease as entry point.
  • - data of each of patterns, locations, and conditions is comprised in separate dimension(s), and - characteristic information on patterns, locations, and conditions is organised into discrete categories, (d) means for inputting navigational information, and
  • (c) means for extracting from the multidimensional database a list of conditions patterns, locations depending on the navigational information.
  • Another embodiment of the present invention is a system as described above further comprising one or more features as described above.
  • Another embodiment of the present invention is a system as described above comprising means for providing results of navigation as one or more of text, numbers, case examples, drawings, computer generated graphics, video clips, or any other type of relevant output.
  • Another embodiment of the present invention is a multidimensional database comprising data of patterns, locations, and conditions associated therewith, in which - data of each of patterns, locations, and conditions are comprised in separate dimension(s), and characteristic information on patterns, locations, and conditions is organised into discrete categories,
  • Another embodiment of the present invention is a multidimensional database as described above wherein data has been validated by at least one medical expert.
  • Another embodiment of the present invention is a multidimensional database as described above further comprising one or more the features as described above.
  • Another embodiment of the present invention is a computer program on a computer readable medium capable of performing a method as described above.
  • Another embodiment of the present invention is a computer program on a computer readable medium capable of providing functionality of a system as described above.
  • Another embodiment of the present invention is a computer program on a computer readable medium capable of providing a multidimensional database as described above.
  • Another embodiment of the present invention is a method for evaluating at least one abnormality in one or more images of an object, comprising:
  • Another embodiment of the present invention is a method as described above wherein a low discriminative value of a selected pattern generates a request to use an alternative approach.
  • Another embodiment of the present invention is a method as described above wherein the multidimensional database comprises one or more additional dimensions corresponding to relevant non-imaging data.
  • Another embodiment of the present invention is a method as described above wherein a request is generated to provide non-imaging data, when the database indicates such data can adjust the reported probability of a condition.
  • Another embodiment of the present invention is a method as described above wherein the reported probability of a diagnosis is adjusted according to non-imaging data provided.
  • Another embodiment of the present invention is a method as described above, wherein the pre-defined selections are comprised in a set of terms represented by a lexicon, wherein the lexicon can be changed according to the understanding or language of the operator.
  • Another embodiment of the present invention is a system as described above, further comprising the multidimensional database of step (c).
  • Another embodiment of the present invention is a system as described above wherein the multidimensional database further comprises data regarding the discriminative value of one or more combinations of location and pattern.
  • Another embodiment of the present invention is a system as described above comprising means to rank the list the condition according to probability.
  • Another embodiment of the present invention is a system as described above comprising means to generate a request to use an alternative approach when a low discriminative value pattern is inputted in step (b).
  • Another embodiment of the present invention is a system as described above wherein the multidimensional database comprises one or more additional dimensions corresponding to non-imaging data of the object.
  • Another embodiment of the present invention is a system as described above comprising means to generate a request to provide non-imaging data, when the database indicates such data can adjust the reported probability of a condition.
  • Another embodiment of the present invention is a system as described above comprising means to adjust the reported probability of a condition according to non-imaging data provided.
  • Another embodiment of the present invention is a system as described above, comprising means to change the lexicon used in the pre-defined selection in steps (a) and (b) according to the understanding or language of the operator.
  • Another embodiment of the present invention is a computer program held on a computer readable medium capable of performing a method as described above.
  • Another embodiment of the present invention is a computer program on a computer readable medium capable of providing functionality of a system as described above.
  • Another embodiment of the present invention is a method for entering data in a multidimensional database as described above by providing an indication of a location and pattern of an abnormality observed in a medical image of a diagnosed subject , and a disease associated therewith.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the pattern category is provided for said pattern, which is the morphology of the abnormality, uptake of contrast media by the abnormality or functional profile of the abnormality.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the group is provided for a pattern of said morphological pattern category, which group is selected from focal abnormalities, diffuse disease or abnormal size and anatomy.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the group is provided for a pattern of said uptake of contrast media pattern category, which group is selected from focal abnormalities or diffuse disease.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the group is provided for a pattern of functional profile pattern category which group is selected from blood flow or muscular contraction.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the modality of the medical image is provided with the pattern.
  • Another embodiment of the present invention is a method for entering data as described above wherein an indication of the organ system, location within said organ and optionally sublocation is provided.
  • Another embodiment of the present invention is a method for entering data as described above wherein the organ system and location is selected from a list such as in Table 1.
  • Another embodiment of the present invention is a method for entering data as described above wherein non-imaging data is additionally provided, corresponding to one or more of age, sex, area, immune status, oncologic antecedents.
  • Another embodiment of the present invention is a method for entering data as described above wherein possible choices of age, sex, area, immune status, oncologic antecedents are selected from the list in Tabfe 3.
  • Another embodiment of the present invention is a method for entering data as described above wherein the number of diseases is between 1 and 5.
  • Another embodiment of the present invention is a method for entering data as described above wherein the list of possible diseases is reduced by combining two or more patterns.
  • Another embodiment of the present invention is a method for entering data as described above wherein the list of possible diseases is reduced by adding one or more details to a pattern.
  • Another embodiment of the present invention is a method for entering data as described above wherein the list of possible diseases is increased by integrating a pattern into a more general pattern.
  • Another embodiment of the present invention is a method for entering data as described above performed by an expert.

Abstract

La présente invention concerne un procédé et un système d’évaluation d’au moins une anomalie dans une ou plusieurs images médicales d’un sujet consistant : (a) à déterminer l’emplacement de chaque anomalie à partir de la sélection prédéfinie, (b) à déterminer le schéma de chaque anomalie à partir de la sélection prédéfinie, (c) à accéder à une base de données multidimensionnelle comprenant des données de motifs, des emplacements et des conditions associées à celles-ci, où la base de données comprend des données de chacun des motifs, des emplacements et des conditions dans des dimensions séparées, et des informations caractéristiques pour des motifs, des emplacements et des conditions sont organisées dans la base de données en catégories discrètes, (d) à extraire de la base de données multidimensionnelle une liste de conditions correspondant aux données d’imagerie déterminées dans les phases (a) et (b), (e) à réaliser une évaluation d’anomalie à l’aide de la liste obtenue dans la phase (d). L’invention concerne également une base de données, programme informatique, un système de navigation dans la base de données et un procédé de saisie de données dans une base de données.
PCT/EP2005/012816 2004-12-02 2005-12-01 Système d’évaluation d’image, procédés et base de données WO2006058738A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US11/791,405 US8064663B2 (en) 2004-12-02 2005-12-01 Image evaluation system, methods and database

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
EPPCT/EP2004/013684 2004-12-02
EP2004013684 2004-12-02
US67767105P 2005-05-04 2005-05-04
US60/677,671 2005-05-04

Publications (2)

Publication Number Publication Date
WO2006058738A1 true WO2006058738A1 (fr) 2006-06-08
WO2006058738A9 WO2006058738A9 (fr) 2007-07-26

Family

ID=35645562

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2005/012816 WO2006058738A1 (fr) 2004-12-02 2005-12-01 Système d’évaluation d’image, procédés et base de données

Country Status (1)

Country Link
WO (1) WO2006058738A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014197669A1 (fr) * 2013-06-05 2014-12-11 Nuance Communications, Inc. Procédés et appareil pour fournir un guide à des professionnels médicaux
WO2015128429A1 (fr) * 2014-02-26 2015-09-03 Grain Ip Procédé et système d'assistance à la détermination d'un état médical
US11183300B2 (en) 2013-06-05 2021-11-23 Nuance Communications, Inc. Methods and apparatus for providing guidance to medical professionals

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6032678A (en) * 1997-03-14 2000-03-07 Shraga Rottem Adjunct to diagnostic imaging systems for analysis of images of an object or a body part or organ
EP1103900A2 (fr) * 1999-11-26 2001-05-30 Medison Co., Ltd. Dispositif de recherche d'images ultrason et transmission d'images ultrason et système de réception
WO2001057777A2 (fr) * 2000-02-04 2001-08-09 The University Of Chicago Procede, systeme et support lisible par ordinateur destines a un poste de travail de recherche intelligente pour interpretation assistee par ordinateur d'images medicales
EP1315125A2 (fr) * 2001-11-20 2003-05-28 General Electric Company Procédé et système pour détecter des maladies pulmonaires
US20040101177A1 (en) * 2002-11-21 2004-05-27 Siemens Aktiengesellschaft Method and system for retrieving a medical picture

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6032678A (en) * 1997-03-14 2000-03-07 Shraga Rottem Adjunct to diagnostic imaging systems for analysis of images of an object or a body part or organ
EP1103900A2 (fr) * 1999-11-26 2001-05-30 Medison Co., Ltd. Dispositif de recherche d'images ultrason et transmission d'images ultrason et système de réception
WO2001057777A2 (fr) * 2000-02-04 2001-08-09 The University Of Chicago Procede, systeme et support lisible par ordinateur destines a un poste de travail de recherche intelligente pour interpretation assistee par ordinateur d'images medicales
EP1315125A2 (fr) * 2001-11-20 2003-05-28 General Electric Company Procédé et système pour détecter des maladies pulmonaires
US20040101177A1 (en) * 2002-11-21 2004-05-27 Siemens Aktiengesellschaft Method and system for retrieving a medical picture

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014197669A1 (fr) * 2013-06-05 2014-12-11 Nuance Communications, Inc. Procédés et appareil pour fournir un guide à des professionnels médicaux
US11183300B2 (en) 2013-06-05 2021-11-23 Nuance Communications, Inc. Methods and apparatus for providing guidance to medical professionals
WO2015128429A1 (fr) * 2014-02-26 2015-09-03 Grain Ip Procédé et système d'assistance à la détermination d'un état médical
CN106104538A (zh) * 2014-02-26 2016-11-09 凡巴茨赛勒艾姆卡有限公司 用于辅助确定医学病症的方法及系统
US9977958B2 (en) 2014-02-26 2018-05-22 Grain Ip Method and system for assisting determination of a medical condition

Also Published As

Publication number Publication date
WO2006058738A9 (fr) 2007-07-26

Similar Documents

Publication Publication Date Title
US8064663B2 (en) Image evaluation system, methods and database
US8117549B2 (en) System and method for capturing user actions within electronic workflow templates
CN108475538B (zh) 用于将第三方应用集成在图像解读工作流中的结构化发现对象
JP4767759B2 (ja) 読影レポート作成装置
US7607079B2 (en) Multi-input reporting and editing tool
Tagare et al. Medical image databases: A content-based retrieval approach
US6366683B1 (en) Apparatus and method for recording image analysis information
US8335694B2 (en) Gesture-based communication and reporting system
Rubin et al. iPad: Semantic annotation and markup of radiological images
US20080117230A1 (en) Hanging Protocol Display System and Method
US20120221347A1 (en) Medical reconciliation, communication, and educational reporting tools
JP2019153250A (ja) 医療文書作成支援装置、方法およびプログラム
CN105074708A (zh) 放射学发现的背景驱动的概要视图
WO2011066222A1 (fr) Interface d'utilisateur destinée à fournir des applications cliniques et des ensembles de données associés sur la base de données d'images
JP2008506188A (ja) ジェスチャ・ベース報告方法およびシステム
JP5098314B2 (ja) 文書作成支援システム、および文書作成支援方法、並びにコンピュータ・プログラム
JP2009238039A (ja) 医用レポートシステム、医用レポート閲覧装置、医用レポートプログラム、及び医用レポート閲覧方法
Deserno et al. Extended query refinement for medical image retrieval
JP2017191457A (ja) レポート作成装置、およびその制御方法
US20080175460A1 (en) Pacs portal with automated data mining and software selection
US11062448B2 (en) Machine learning data generation support apparatus, operation method of machine learning data generation support apparatus, and machine learning data generation support program
ZINK et al. Medical imaging databases: A National Institutes of Health workshop
WO2006058738A1 (fr) Système d’évaluation d’image, procédés et base de données
WO2007050962A2 (fr) Procede et systeme de capture d'actions utilisateur dans des gabarits de flux de travaux electroniques
Valentino et al. Application development environment for advanced digital workstations

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KM KN KP KR KZ LC LK LR LS LT LU LV LY MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU LV MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

DPE2 Request for preliminary examination filed before expiration of 19th month from priority date (pct application filed from 20040101)
121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 11791405

Country of ref document: US

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: COMMUNICATION NOT DELIVERED. NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 69(1) EPC (EPO FORM 1205A DATED 31.08.2007)

122 Ep: pct application non-entry in european phase

Ref document number: 05814019

Country of ref document: EP

Kind code of ref document: A1

WWW Wipo information: withdrawn in national office

Ref document number: 5814019

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 11791405

Country of ref document: US