WO2009060355A1

WO2009060355A1 - Computer-aided diagnosis with queries based on regions of interest

Info

Publication number: WO2009060355A1
Application number: PCT/IB2008/054511
Authority: WO
Inventors: Juergen Weese; Yuechen Qian; Warner R. T. Ten Kate
Original assignee: Philips Intellectual Property & Standards Gmbh; Koninklijke Philips Electronics N.V.
Priority date: 2007-11-08
Filing date: 2008-10-30
Publication date: 2009-05-14

Abstract

A computer-aided diagnosis system for providing image-processing algorithms that are linked to records with the diagnosed images in a medical database is provided. The algorithms are designed to check characteristic lesion properties, e.g. image appearance or border characteristics. The system is configured to process a region of interest in the image with the algorithms to verify whether the actual lesion characteristics correspond to the diagnosis associated to a related stored lesion. A rating of the diagnostic option, e.g. confirmation or rejection, according to the result and usage of the rating to display possible diagnostic options may be determined.

Description

COMPUTER-AIDED DIAGNOSIS WITH QUERIES BASED ON REGIONS OF INTEREST

FIELD OF THE INVENTION

This invention pertains in general to the field of medical encyclopedias. More particularly the invention relates to diagnosis assistance in medical encyclopedias.

BACKGROUND OF THE INVENTION

Medical encyclopedias are emerging on the market to assist physicians and radiologists to read medical images and make clinical decisions. Using such a system, the user may find diagnoses/cases by using anatomy/pathology-based browsing and/or keyword- based search. However, the user is often encountered with a number of possible diagnoses. To narrow down diagnostic options, the user has to follow the instructions/guidelines provided by the system on how to read medical images of the case in question and how to compare them with the found ones in the system.

Online medical encyclopedia systems provide physicians and radiologists with expert knowledge on radiology and help them in differential diagnosis, also referred to as DDx, and treatment planning. In medicine, differential diagnosis is the systematic method physicians use to identify the disease causing a patient's symptoms. The term differential diagnosis also refers to medical information specially organized to aid in diagnosis, particularly a list of the most common causes of a given symptom, annotated with advice on how to narrow down the list. Before a medical condition may be treated, it must be identified. The physician begins by observing the patient's symptoms, examining the patient, and taking the patient's personal and family history. Then the physician lists the most likely causes. The physician asks questions and performs tests to eliminate possibilities until he or she is satisfied that the single most likely cause has been identified. The same also holds for radiologists. When a radiologist reads an image study of a patient he needs to identify abnormalities, measure and classify them, and prepare an imaging finding reports. In case of difficult cases, for instance several DDX options are possible, the radiologist may advise the physician to order more studies. Medical encyclopedias provide information on how to differentiate possible diagnosis. Typically, a user, a physical or a radiologist, may navigate through such an encyclopedia by anatomy and/or pathology and may search diagnosis/cases based on text. Then, the user needs to follow the provided guidelines to read the subject image and compare it with the sample images of the provided cases.

As medical imaging technologies progress, physicians and radiologists are confronted with ever increasingly amount of image data to analyze within a constrained duration of time. Providing assistance for radiologists in quick yet correct decision-making is challenging task. When using a medical encyclopedia the user is encountered with a number of possible diagnoses and finding relevant cases consumes time. The value of medical encyclopedias in a clinical setting is therefore related to the ease and speed of finding the most relevant possible diagnoses.

However, clinical use of commonly known diagnostic encyclopedias is not optimal. Currently, the user of an encyclopedia has to follow guidelines/instructions per diagnosis provided by the encyclopedia. Often the number of relevant diagnosis is large and making DDx is time-consuming and might also be error-prone. Hence, an improved system, method, computer-readable medium, medical encyclopedia, and use allowing for increased flexibility, cost-effectiveness, and time effectiveness would be advantageous. SUMMARY OF THE INVENTION

Accordingly, the present invention preferably seeks to mitigate, alleviate or eliminate one or more of the above-identified deficiencies in the art and disadvantages singly or in any combination and solves at least the above-mentioned problems by providing a system, method, computer-readable medium, and use according to the appended patent claims.

In an aspect of the invention a system is provided. The system comprises a medical database comprising records describing pathologies. The system also comprises a first unit configured to retrieve a region of interest in an acquired image to be analyzed. Furthermore, the system comprises a query unit configured to search the medical database using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases. The system also comprises a second unit configured to retrieve a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases, and process the region of interest in the acquired image using at least one image-processing algorithm of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis of the found set of possible diagnoses and sample cases. In another aspect a method is provided. The method comprises selecting (51) an acquired image, identifying (52) an anatomical region in the acquired image, selecting (53) a record in a database corresponding to the anatomical region in the acquired image, executing (54) at least one image-processing algorithm associated with the record on the acquired image.

In an aspect a computer-readable medium having embodied thereon a computer program for processing by a processor is provided. The computer program comprises a first retrieve code segment for retrieving a user defined region of interest in an acquired image to be analyzed, a search code segment for searching a medical database comprising with records describing pathologies using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases, a second retrieve code segment for retrieving a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases, a process code segment for processing the region of interest in the acquired image using at least one image- processing algorithm of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis in the found set of possible diagnoses and sample cases.

In another aspect a medical database is provided. The medical database comprises a set of sample images of known pathological states, wherein each of the set of sample images of known pathological states is linked to a customized image-processing algorithm for processing an image to detect whether the image comprises the pathological state.

In an aspect a use of the system according to claim 1 for facilitating diagnosis or differential diagnosis of a region of interest in an image dataset. The present invention according to some embodiments has the advantage over the prior art that it improves efficiency of encyclopedia-based diagnostic decision-making.

The present invention according to some embodiments extends the functionality of existing medical encyclopedias and improves their usage in a clinical environment.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects, features and advantages of which the invention is capable of will be apparent and elucidated from the following description of embodiments of the present invention, reference being made to the accompanying drawings, in which Fig. 1 is an illustration showing a system according to an embodiment; Fig. 2 is an illustration showing a selection unit according to an embodiment; Fig. 3 comprises four images 3a to 3e showing an embodiment; Fig. 4 is an illustration of a system according to an embodiment; Fig. 5 is a flow chart showing a method according to an embodiment; and

Fig. 6 is a flow chart showing a method according to an embodiment; and Fig. 7 is a flow chart showing a computer-readable medium according to an embodiment.

DESCRIPTION OF EMBODIMENTS

Several embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in order for those skilled in the art to be able to carry out the invention. The invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. The embodiments do not limit the invention, but the invention is only limited by the appended patent claims. Furthermore, the terminology used in the detailed description of the particular embodiments illustrated in the accompanying drawings is not intended to be limiting of the invention. A problem of prior art is that individual developed image-processing algorithms, which are not linked to diagnostic encyclopedias, are not sufficient to make accurate diagnosis decisions. Accordingly, if there is no correlation between the applied image-processing algorithm and the reference image in the medical encyclopedia, there is no guarantee that the used image-processing algorithm in fact works to be used to diagnose a disease, such as a lesion, or disorder, in the investigated image. At least this would require an extensive and comprehensive standardization of lesion characterization.

The present inventors have realized that as long as the individually developed image-processing algorithms are separated from the medical encyclopedia, they may hardly be used to confirm or rule out a diagnostic option using characteristics, such as image appearance, border characteristics, etc., for example of a lesion in the investigated image, which is also referred to as acquired image throughout this specification.

The following description focuses on embodiments of the present invention applicable to diagnosis of a patient image using medical encyclopedias. An idea of the invention according to some embodiments is to combine image- processing algorithms and medical encyclopedias for facilitating diagnosis, i.e. distinguish between different diseases, or differential diagnosis, i.e. specific characterization of a disease or status of disease progression, of a patient image. By linking image or text-processing algorithms to a medical encyclopedia and applying them to differentiate diagnosis it will facilitate the search of DDx.

Throughout this specification the definition of a medical encyclopedia is a database, i.e. a set of records with an indexing structure, comprising textual or pictorial descriptions, e.g. sample images, of different pathologies with confirmed diagnoses. A medical encyclopedia may also comprise text guidance for facilitating determining of a correct diagnosis of an acquired patient image. The encyclopedia may be stored on a storage medium, such as, but not limited to, a CD, DVD, Tape, hard drive, RAM, ROM, etc.

The present invention according to some embodiments provides a solution to simplify the use of medical encyclopedias and ensure the selection of proper diagnostic options by providing at least one diagnostic image-processing algorithm to at least one case in a medical encyclopedia. In addition to textual instructions or guidelines, the medical encyclopedia in some embodiments is provided with at least one diagnostic algorithm, such as image-processing (IP) or computer-aided detection (CAD) algorithm, linked to at least one diagnostic option. Such algorithms may automatically be executed on the image that the user is investigating, to determine whether certain option is relevant or not. The present invention according to some embodiments improves the process of narrowing down the diagnostic options, thus making the process easier, faster, and optionally automatic. In this way, the user does not need to narrow down the diagnostic options manually. Accordingly, the present invention according to some embodiments has the advantage over the prior art that it improves efficiency of encyclopedia-based diagnostic decision-making.

In an embodiment, according to Fig. 1, a system 10 for facilitating diagnosis is provided. The system 10 may comprise a medical database 11 comprising pictorial and/or textual information regarding pathological or healthy states, such as sample images or guidance texts for facilitating diagnosis of a disease or disorder of a patient. Moreover the system comprises a first processor unit 12 configured to retrieve

12a a user defined region of interest (ROI) on an image of a patient to be analyzed, as is indicated in Fig. 3b. The first processor unit 12 may also be configured to identify 12b a lesion area in the region of interest using a first image-processing algorithm, as is indicated in Fig. 3c. Furthermore, the first processor unit may be configured to identify 12c an anatomical structure in the image, as is indicated in Fig. 3d and 3e, using a second image-processing algorithm. In Fig. 3d the first processor unit has identified the left and right cerebral hemispheres as an anatomical structure. In Fig. 3e the first processor unit has identified the left cerebral hemisphere as an anatomical structure to be queried. The system may also comprise an extraction unit 13 for extracting the anatomical location of the lesion within the lesion area. The location and anatomical structure in which the abnormality appears will help to rule out diagnostic options. For example, brain neoplasm's like germinoma are mainly located the midbrain (pineal gland). Knowing the anatomical structure and location will certainly refine the search space of DDx. In this way, the user may indicate the exact anatomy of interest, e.g. "find diagnosis that exhibit lesions in the midbrain."

For example, an image-processing algorithm may e.g. be used to extract certain image findings, such as "hypo intense mass to white matter on MR Tl weighted images" for a diagnosis "astrocytoma". In this case, one image-processing algorithm may be used to detect hypoinstense on MR Tl weighted images and another image- processing algorithm may be used to detect the "mass" effect. Optionally one image- processing algorithm may be used to classify the brain tissue for MR Tl weighted images, in ordered to know whether the lesion area belongs to white matter or gray matter in normal cases.

Among other typical Tl weighted image findings of anaplastic astrocytoma is "mixed isointense signal to hypointense white matter mass". According to an embodiment image-processing algorithms required for this example could be: a) brain tissue (white matter, gray matter, cerebrospinal fluids) classification image-processing algorithm; b) mass detection image-processing algorithm; c) hypointense detection image-processing algorithm; d) hyperintense detection image-processing algorithm; and e) mixed signal detection. The confidence measures of these 5 algorithms will provide hints on how the system concludes whether the acquired image exhibits the image findings expected from a typical anaplastic astrocytoma case. Image findings may lead to several DXes records in the encyclopedia.

"Hypointense mass" could e.g. be found in diffuse astrocytoma, anaplastic astrocytoma, glioblastoma multiforme, oligodendroglioma, and so on.

In an embodiment, the system comprises a display for presenting a list ordered with respect to the confidence measure. This list may e.g. be used to facilitate diagnosis of the image, and differential diagnosis options based on the list could be presented on the display.

The extraction unit 13 may utilize a third image-processing algorithm to extract the location of the identified lesion area in the region of interest. In some embodiments the extraction unit may extract context information, such as parameters referring to e.g. modality, organ of interest, etc from a session report. The session report may e.g. be a diagnostic image finding reports made by a radiologist, the imaging study order made by the referring physician or Digital Imaging and Communications in Medicine (DICOM) headers of the images. The first, second or third image-processing algorithm may be model-based image segmentation techniques. A general solution to detect anatomical structures and abnormalities does not exist. In this regard, most systems need specialized algorithms to detect and measure abnormalities, e.g. whether the hyper intense signal is heterogeneous or not, affecting gray matter or white matter in the frontal lobes. According to some embodiments, to detect anatomies in the image, a mesh- based image segmentation method may be used. Each organ may be modeled by a mesh model depicting the surface model of the organ. In case of brain, such mesh model may contain parts like left/right cerebral hemispheres, cerebellum and so on. The mesh model may then be adapted to the volumetric image, after which the anatomical regions in the image are identified. Such techniques require initialization to load proper mesh models for the organ under investigation. Usually the anatomical information in the session report (image study orders or image finding reports) is rough and global, not sufficient enough for DDx.

The system further comprises a query unit 14 configured to prepare or retrieve a search query, such as a search mask, that may be used to search the medical database. In Fig. 3e a search query is defined. The query unit may be a part of graphical user interface allowing the user to indicate or refine the search space, such as for example, search diagnosis exhibiting hypo intense signals in the left cerebral hemisphere. The query may e.g. from the user's point of view, look like "search cases that contain hypointense to white matter in the left cerebral hemisphere". In an embodiment the search query may comprise textual input from the user.

In other embodiments the user may use a pointing device, such as a mouse to select presented search query options, provided in the graphical user interface.

In some embodiments the graphical user interface may be run using a user workstation, such as a personal computer (PC), which is configured to send/receive information, such as image information, extracted metadata information, image-processing algorithms to/from a processing device being in connection to a medical encyclopedia. Any type of connection may be possible, such as via a physical connection, a network or wireless. The client workstation may also be configured to access a patient image e.g. from a scanner or via Picture archiving and communication system (PACS).

In some embodiments, the terminology used in the search query is compatible with that set forth by ontology available on the Internet.

In some embodiments the search query may be encoded in standard Unified Medical Language System (UMLS), Foundational Model of Anatomy (FMA), or International Classification of Diseases (ICD-IO) terms and may be used in search for relevant cases in the medical encyclopedia. Other terminologies may also be used, in particular as they become published on the Internet, using exchange languages as set forth for the Semantic Web.

Moreover the query unit 14 may be configured to process the search query in the medical encyclopedia, resulting in a set of possible diagnoses and sample cases, as is indicated in Fig. 4.

The system may further comprise a second processor unit 15 configured to retrieve (15a) a set of selected image-processing algorithms that are associated with the found set of diagnoses and sample cases, and optionally located on a memory 16 or an selection unit 20.

A selected image-processing algorithm may be an algorithm that may be customized from a common IP algorithm and may be adapted to the modality, to the organ and so on. By customized is here meant that the algorithm may be modified particularly for being used on a certain sample image, and the acquired image resulting in parameters that may be used for facilitating diagnosis, or for discrimination between different diagnoses.

Selected image-processing algorithms may be located on a memory and be customized, e.g. by a skilled person within image analysis, to the diagnosed sample images in the medical encyclopedia.

The selected image-processing algorithms associated with the set of diagnoses and sample cases, which are located in the medical encyclopedia, may be stored either in the medical encyclopedia or in a memory connected to the system.

Accordingly, for each sample case in the medical encyclopedia there is provided at least one selected image-processing algorithm applicable to the region of interest of the image that may be used to confirm diagnosis of such a case within the image. In some embodiments the selected image-processing algorithms is designed and assigned to the records of an encyclopedia comprising text and images by a skilled person. Either the selected image-processing algorithms may be integrated into the encyclopedia or being located on an external device being in communication with the encyclopedia.

Moreover the second processor 15 may further be configured to process (15b) the region of interest in the image using the retrieved selected image-processing algorithms to determine at least one diagnosis of the image.

The second processor 15 may also be configured to calculate a confidence measure for each possible diagnosis option of the image.

The second processor may utilize weights for each used image-algorithm in order to calculate the confidence measure, e.g. such as a probability or percentage. For example, if 2 selected image-processing algorithms A, and B are used to process an image for a certain diagnosis option, the resulting confidence measure may be based on partly image- processing algorithm A, and partly on image-processing algorithm B, such as 20% of A, and 80% of B. In this way the second processor 15 may calculate an overall confidence measure for each diagnosis option. The overall confidence indicates how likely it is that a region of interest comprised in the image could be diagnosed with each diagnosis option.

In some embodiments the selected image-processing algorithms is configured to calculate e.g. characteristics of lesion properties such as image appearance, or border characteristics in the image to facilitate subsequent diagnosis of the image.

Moreover, the second processor may verify whether the lesion properties computed based on the acquired image correspond to the lesion properties described in the diagnosed sample image in the medical encyclopedia. Moreover, the second processor may be configured to rate each possible diagnostic option, such as confirmation or rejection, according to the result of the selected image-processing algorithms.

To be able to determine whether a certain diagnosis is relevant or irrelevant for the investigated image, different types of image-processing algorithms may be used. In some embodiments the selected image-processing algorithm may be applied to classify the malignant or benignant state of the lesion in the identified lesion area. In this way the relevance for a certain diagnosis may be determined.

Moreover the selected image-processing algorithms may be configured to check the contrast of the lesion area with respect to the background of the image for determining the level of relevance or similarity, such as for the example regarding determination of hyper intense or hypo intense mentioned above.

The selection or adaptation of IP algorithms to detect anatomical regions may occur during the viewing of the image, if the adaptation runs efficiently. Otherwise it may take place offline, prior to the viewing by the radiologist, e.g. after the image is stored in PACS.

In an embodiment, according to Fig. 2, a selection unit 20 is provided. The selection unit is configured to create, select or adapt an image-processing algorithm that may be used to for diagnosing of an anatomical structure of an image. The selection unit may be connected to a medical encyclopedia 11 comprising already diagnosed sample images.

Optionally the selection unit 20 may store selected image-processing algorithms on a memory 16. Moreover, the memory may comprise not yet adapted image- processing algorithms that may be used by the selection unit to create the selected image- processing algorithms. In an embodiment the selection unit 20 may be used as a stand-alone device being connected to existing medical diagnostic encyclopedias in order to create selected image-processing algorithms for the different medical cases in the medical encyclopedia.

In some embodiment the graphical user interface is configured to enable the user to indicate which anatomy in the acquired patient image that is of interest, e.g. "left cerebral hemisphere", "midbrain", etc. versus "cerebral hemisphere" only. The creation of the search query may thus occur after the anatomical structures have been identified and made visible to the user.

The graphical user interface may also be configured to enable the user to have the possibility to refine an already processed search query. Hence a defined region of interest may be further analyzed and metadata may be extracted and sent to the encyclopedia to refine the search criteria. By refining a search query, e.g. by introducing an additional search parameter, such as sub anatomies or anatomies that can not be segmented, symptoms, demographical data, etc. into the search query the diagnosis detection may be facilitated and a decreased number of potential diagnoses for the acquired patient image may be obtained. In an embodiment, according to Fig. 2, a medical encyclopedia unit 20 comprising the medical encyclopedia 11 and the query unit 14 is provided. Optionally the medical encyclopedia unit 20 may comprise a memory comprising the selected image- processing algorithms that are adapted for the sample images and their respective diagnosis present in the medical encyclopedia. In some embodiments, the graphical user interface may be visualized on a display for presenting processed information, such as determined diagnoses, etc. to a user.

In some embodiments the rating of each diagnosis performed by the second processor may be presented in the display along with each diagnostic option. The physical location of the units of the system according to some embodiments may be varied arbitrarily due to the desired use of the system. For example, in an embodiment the graphical user interface, first and second processor units, the query unit, and adaptation unit may be physically located in a user workstation.

In some embodiments the user workstation may be connected to the encyclopedia and the selected IP algorithms e.g. via Internet, as is indicated in Fig. 4.

Fig. 4 is an illustration showing a system 40 according to an embodiment, and the system functionalities. In this embodiment the medical encyclopedia 41 comprises a medical database 42, a database 43 comprising selected image-processing algorithms for each of the records of the medical database 42, a processor unit 44 for processing a query sent from a user workstation 45, and a processor unit 45 for image-processing 47 of an image of interest using a number of image-processing algorithms corresponding to the set of possible diagnoses resulting from the query processing. A display 46 for presenting the processing results, such as the respective possible diagnosis together with a confidence measure for each possible diagnosis, is also provided. In some embodiments the user workstation may be connected to a scanner or

PACS system to retrieve the acquired patient image, which is to the analyzed.

In other embodiments, an external computing device that is linked to a diagnostic encyclopedia and which processes the acquired patient image using the selected image-processing algorithms to facilitate diagnosis may be provided comprising any of the first or second processing unit, query unit or extraction unit. This embodiment may be advantageous if the computational complexity of the processing is too large, i.e. it takes a too large amount of time, to be performed on the user workstation. In this way the user may obtain faster results by letting the processing to occur in the external computing device.

It should be noted that the present invention according to some embodiments is not limited by the physical location of its units.

In some embodiments the second processor is located within an external device, such a user workstation.

In some embodiment the first processor unit, the extraction unit, the second processor unit, is comprised in an external device, such as a user workstation. In some embodiments the functionality of the first and second processor may be included into one processor.

For ease of explanation a practical example is provided below. For simplicity it is assumed that the medical encyclopedia comprises two sample images, sample image A, and B. Moreover there are in total 5 available image-processing algorithms available for use associated with the two sample images. The selection unit may be configured to select (which may be pre-defined) which algorithms that may be used for the two samples. For example, sample A may be processed by 1, 3, and 5, while sample B may be processed with algorithm 1, 2, 4. Both samples A, and B may thus be processed by algorithm 1. The "selection" of algorithms for each sample is thus defined in the selection unit.

In an embodiment, according to Fig. 5 a method is provided. The method comprises selecting 51 an acquired image. The method may also comprise identifying 52 an anatomical region in the acquired image. Furthermore, the method may comprise selecting 53 a record in a database corresponding to the anatomical region in the acquired image. Moreover the method may comprise executing 54 at least one image-processing algorithm associated with the record on the acquired image.

In an embodiment a method is provided. The method comprises method steps for performing any of the functional steps in some embodiments.

In an embodiment, according to Fig. 6, a method is provided. The method comprises retrieving 61 a user defined region of interest in an acquired image to be analyzed. Moreover the method may comprise searching 62 a medical database comprising a set of sample images of known pathological states using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases. Furthermore, the method may comprise retrieving 63 a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases. Moreover, the method comprises processing 64 the region of interest in the acquired image using at least one of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis in the found set of possible diagnoses and sample cases.

In an embodiment, according to Fig. 7, a computer-readable medium having embodied thereon a computer program for processing by a computer is provided. The computer program comprises a first retrieve code segment 71 for retrieving a user-defined region of interest in an acquired image to be analyzed. Furthermore, the computer program may comprise a search code segment 72 for searching a medical database comprising a set of sample images of known pathological states using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases. Moreover, the computer program may comprise a second retrieve code segment 73 for retrieving a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases. Moreover, the computer program may comprise a process code segment 74 for processing the region of interest in the acquired image using at least one of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis in the found set of possible diagnoses and sample cases.

In an embodiment a use of the system, method or computer-readable medium is provided for facilitating diagnosis of a region of interest in an image dataset. In an embodiment a medical database 41 comprising a set of sample images of known pathological states is provided. Each of the set of sample images of known pathological states is linked to a customized image-processing algorithm for processing an image to detect whether the image comprises the pathological state or not.

Applications and use of the above-described embodiments according to the invention are various and include exemplary fields such as differential diagnosis.

The processor unit may be any unit normally used for performing the involved tasks, e.g. a hardware, such as a processor with a memory. The processor may be any of variety of processors, such as Intel or AMD processors, CPUs, microprocessors, Programmable Intelligent Computer (PIC) microcontrollers, Digital Signal Processors (DSP), etc. However, the scope of the invention is not limited to these specific processors. The memory may be any memory capable of storing information, such as Random Access Memories (RAM) such as, Double Density RAM (DDR, DDR2), Single Density RAM (SDRAM), Static RAM (SRAM), Dynamic RAM (DRAM), Video RAM (VRAM), etc. The memory may also be a FLASH memory such as a USB, Compact Flash, SmartMedia, MMC memory, MemoryStick, SD Card, MiniSD, MicroSD, xD Card, TransFlash, and MicroDrive memory etc. However, the scope of the invention is not limited to these specific memories.

In an embodiment the system is comprised in a medical workstation or medical system, such as a Computed Tomography (CT) system, Magnetic Resonance Imaging (MRI) System or Ultrasound Imaging (US) system. In an embodiment the computer-readable medium comprises code segments arranged, when run by an apparatus having computer-processing properties, for performing all of the functionality steps defined in some embodiments.

The invention may be implemented in any suitable form including hardware, software, firmware or any combination of these. However, preferably, the invention is implemented as computer software running on one or more data processors and/or digital signal processors. The elements and components of an embodiment of the invention may be physically, functionally and logically implemented in any suitable way. Indeed, the functionality may be implemented in a single unit, in a plurality of units or as part of other functional units. As such, the invention may be implemented in a single unit, or may be physically and functionally distributed between different units and processors.

Although the present invention has been described above with reference to specific embodiments, it is not intended to be limited to the specific form set forth herein. Rather, the invention is limited only by the accompanying claims and, other embodiments than the specific above are equally possible within the scope of these appended claims.

In the claims, the term "comprises/comprising" does not exclude the presence of other elements or steps. Furthermore, although individually listed, a plurality of means, elements or method steps may be implemented by e.g. a single unit or processor. Additionally, although individual features may be included in different claims, these may possibly advantageously be combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. In addition, singular references do not exclude a plurality. The terms "a", "an", "first", "second" etc do not preclude a plurality. Reference signs in the claims are provided merely as a clarifying example and shall not be construed as limiting the scope of the claims in any way.

Claims

CLAIMS:

1. A system (10) comprising : a medical database (11) comprising records describing pathologies, a first unit configured to retrieve (12a) a region of interest in an acquired image to be analyzed, - a query unit (14) configured to search the medical database using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases, wherein the system further comprises: a second unit (15) configured to: retrieve (15a) a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases; and process (15b) the region of interest in the acquired image using at least one image-processing algorithm of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis of the found set of possible diagnoses and sample cases.

2. The system according to claim 1, wherein the first unit is further configured to identify (12b) a lesion area in the region of interest using a first image-processing algorithm, and wherein information regarding the lesion area is utilized in the search query.

3. The system according to claim 1, wherein the first unit is further configured to identify (12c) an anatomical structure in the acquired image using a second image-processing algorithm, and wherein information regarding the anatomical structure is utilized in the search query.

4. The system according to claim 1, wherein the first unit is further configured to extract (13) the anatomical location of the lesion within the lesion area using a third image- processing algorithm, and wherein information regarding the anatomical location is utilized in the search query.

5. The system according to claim 1, wherein the search query is user defined.

6. The system according to claim 1, wherein the set of selected image-processing algorithms are customized to calculate characteristic lesion properties such as image appearance, or border characteristics in the acquired image.

7. The system according to claim 1, wherein the customized set of selected image-processing algorithms are modified particularly for being used on a certain sample image within the medical database and the acquired image.

8. The system according to claim 1, wherein the first, second or third image- processing algorithm is a model-based image segmentation technique.

9. The system according to claim 1, wherein the second unit is further configured to rate each possible diagnostic option, such as confirmation or rejection, according to the result of the selected image-processing algorithms.

10. The system according to claim 1, wherein the second unit is further configured to verify whether the lesion properties calculated based on the acquired image correspond to the lesion properties described in the diagnosed sample image in the medical database.

11. The system according to claim 1 , wherein the set of selected image-processing algorithms is configured to check the contrast of the lesion area with respect to the background of the image for determining the level of relevance for a certain possible diagnosis option.

12. The system according to claim 1, wherein the acquired image is a medical image of a patient.

13. The system according to claim 1, wherein the search query comprises textual input from a user.

14. The system according to claim 1, wherein the search query is created by a user using a pointing device, such as a mouse to select presented search query options, provided in the graphical user interface.

15. The system according to claim 1, wherein the query unit is configured to enable the user to indicate which anatomy in the acquired patient image that is of interest.

16. The system according to claim 1, further comprising a display for displaying the sample image, acquired image, or results from image processing.

17. The system according to claim 1, further comprising a display for displaying each confidence measure.

18. The system according to claim 1, further comprising a display for displaying a list sorted according to the confidence measure.

19. The system according to claim 1, wherein the medical database comprises a record containing a set of parameters for respective image-processing algorithm.

20. The system according to claim 1, wherein the medical database comprises a record containing a link to a selected image-processing algorithm or parameters related to a selected image-processing algorithm.

21. The system according to claim 1, wherein said medical database comprises pictorial or textual information regarding the pathologies.

22. A method comprising: selecting (51) an acquired image, identifying (52) an anatomical region in the acquired image, - selecting (53) a record in a database corresponding to the anatomical region in the acquired image, executing (54) at least one image-processing algorithm associated with the record on the acquired image.

23. The method according to claim 22, wherein said identifying is performed by: retrieving (61) a user defined region of interest in an acquired image to be analyzed, said selecting is performed by: searching (62) a medical database comprising records describing pathologies using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases, and said executing is performed by: retrieving (63) a set of selected image-processing algorithms that are customized for the found set of possible diagnoses and sample cases; and processing (64) the region of interest in the acquired image using at least one image-processing algorithm of the set of selected image-processing algorithms, resulting in a confidence measure for each possible diagnosis in the found set of possible diagnoses and sample cases.

24. A computer-readable medium having embodied thereon a computer program for processing by a processor, the computer program comprising: a first retrieve code segment (71) for retrieving a user defined region of interest in an acquired image to be analyzed, a search code segment (72) for searching a medical database comprising records describing pathologies using a search query based on information of the region of interest, resulting in a set of possible diagnoses and sample cases, a second retrieve code segment (73) for retrieving a set of selected image- processing algorithms that are customized for the found set of possible diagnoses and sample cases, a process code segment (74) for processing the region of interest in the acquired image using at least one image-processing algorithm of the set of selected image- processing algorithms, resulting in a confidence measure for each possible diagnosis in the found set of possible diagnoses and sample cases.

25. A medical database comprising a set of sample images of known pathological states, wherein each of the set of sample images of known pathological states is linked to a customized image-processing algorithm for processing an image to detect whether the image comprises the pathological state.

26. Use of the system according to claim 1 for facilitating diagnosis or differential diagnosis of a region of interest in an image dataset.