EP0955887A1

EP0955887A1 - Ultrasound scanning

Info

Publication number: EP0955887A1
Application number: EP97945961A
Authority: EP
Inventors: Gill Imp. College Science Tech. Medicine HEART; Richard I. Imp College Science Tech. Med KITNEY
Original assignee: Imperial College of Science Technology and Medicine
Current assignee: Imperial College of Science Technology and Medicine
Priority date: 1996-11-29
Filing date: 1997-11-27
Publication date: 1999-11-17
Also published as: GB9624970D0; WO1998023210A1; AU5128398A; CN1245408A; CA2272917A1; GB2319841A; JP2001509043A; IL130118A0

Abstract

An ultrasound scanning technique for biological tissue examines the radio frequency signals derived from reflected ultrasonic energy to identify signal parameters indicative of particular tissue histology. Such identification is then used to produce a histological image of the biological tissue. The technique is particularly useful in the examination of blood vessels to identify different types of plaque within the vessel wall. Lipid plaques produce relatively low levels of reflection. Calcified plaques produce high levels of regular reflection. Fibrous plaques may be identified as a highly homogeneous thickened wall area that does not show the characteristics of a lipid plaque. A homogeneity value for the radio frequency signal is calculated for all areas not identified as lipid plaques or calcified plaques and then this homogeneity value is mapped to a colour that is displayed within the histological image.

Description

ULTRASOUND SCANNING

This invention relates to the field of ultrasound scanning. More particularly, this invention relates to the ultrasound scanning of biological tissue in medical applications.

Ultrasound scanning of biological tissue is a widely used diagnostic technique in medicine. Ultrasound scanning is non-invasive, quick, has a very low risk to the subject and requires comparatively inexpensive equipment compared to other scanning techniques. An example of the use of ultrasound scanning is in the examination of blood vessel walls for the detection and characterisation of atherosclerosis.

Diseased blood vessel walls are a major cause of illness. As well as constricting the blood flow, diseased blood vessel walls can release or cause the production of emboli that can become lodged elsewhere completely blocking the blood supply to certain body organs and causing clinical problems such as strokes. Ultrasound scanning may be used to produce transverse or longitudinal cross sectional grey-scale images of blood vessels to identify plaques in the blood vessel walls.

Ultrasound B-mode grey-scale images can be difficult to interpret and in order to assist in this, and to try and obtain a more quantative assessment of the degree of severity of the plaques, blood flow measuring techniques using ultrasound scanning have been developed. An example of this is colour flow mapping that uses Doppler effects to detect the speed of flow of blood through the lumen of a blood vessel. If constrictions/plaque are present in the blood vessel wall, then these tend to disturb the uniform blood flow (e.g. lead to downstream turbulence in the blood flow) that may be more readily identified using colour flow mapping techniques and quantified using Doppler techniques.

Current development effort in ultrasound scanning apparatus and methods for use in such diagnosis is primarily directed toward increasing the frame rate of the image generated by the ultrasound scanner in order to produce a real time smoothly moving representation of the image and the blood flow within that vessel. While such efforts do improve the diagnosis, they suffer from a number of inherent and significant problems.

The plaques within blood vessels can be considered to fall into a number of different characteristic types. A first type is a constriction in the lumen of the blood vessel formed by a fibrous thickening in the vessel wall. Such fibrous thickenings are of a relatively low risk rupture since they are unlikely to break free or lead to thrombosis. Another type of plaque is one in which the thickening contains a calcified region that is relatively hard and brittle. Such calcified regions are thought to present a risk of cracking that in turn may lead to tissue damage of the blood vessel wall, haemorrhage and thrombosis generation. Perhaps the most dangerous type of plaque is a lipid plaque. A lipid plaque within a thickened blood vessel wall consists of a pocket of fatty tissue that is held in place by a fibrous covering (cap). If the fibrous covering ruptures under stress (or for other reasons), then the lipid material is released into the blood flow with potential harmful consequences and the ruptured fibrous cap can itself lead to potential thrombosis. During the healing process a blood clot will form around the ruptured area with the potential for causing emboli or blocking the vessel. It will be seen that thickening of a blood vessel wall can be the result of differing plaque types with very differing risk levels associated with them. These differing risk levels have the result that different treatments are appropriate. However, with an ultrasound scanner, whilst the constriction of the blood vessel and the thickened blood vessel wall may be identified, it is very much more difficult to identify which type of plaque is present. This difficulty is particularly severe between a fibrous plaque and a lipid plaque.

Viewed from one aspect the present invention provides an ultrasonic imaging apparatus for imaging biological tissue, said ultrasonic imaging apparatus comprising: an ultrasonic transducer for emitting ultrasonic sound energy into biological tissue to be imaged and for generating radio frequency signals from received reflected ultrasonic sound energy; and a signal processor for comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology and for generating a histological image that visually distinguishes said one or more regions of different histology.

The present invention takes a different approach to the analysis of the signals generated by an ultrasound scan. The invention recognises that rather than merely using the radio frequency signals from the reflective ultrasonic sound energy to generate grey-scale images (of ever increasing resolution and frame rate), the radio frequency signals contain information embedded within themselves that with careful analysis can reveal histological characteristics of the tissue being scanned. This is not a true histological image in terms of a microscope slide, but is an image that gives and indication of the histological characteristics of the tissue. This histological information is in many cases very much more valuable than any increase in resolution or frame rate. More particularly, the radio frequency signals may be compared by a signal processor with predetermined signal parameters (e.g. amplitude characteristics, peak or trough level characteristics, variance characteristics, phase irregularity characteristics etc) that individually or collectively, and sometimes in combination with characteristics from surrounding areas, identify particular tissue histology. In the above example, the invention may be used to distinguish between normal tissue, lipid plaques, fibrous plaques and calcified plaques within a blood vessel wall. The invention could also be used in the scanning of other biological tissues, e.g. mammography, liver scanning, kidney scanning etc.

The histological image that is produced by the present invention can be displayed independently. However, in preferred embodiments of the invention said signal processor generates a grey-scale image of said biological tissue based upon said radio frequency signal intensity, said grey-scale image being overlain with said histological image.

Overlying the histological image with the grey-scale image allows the information contained within each to mutually support one another for diagnostic purposes in confirming interpretation of the results of the scan.

Interpretation of the histological image is made easier in embodiments of the invention in which said histological image is a colour image with different colours representing different tissue histology or homogeneity. As mentioned before, the present invention may be utilised in the examination of biological tissue of many different types. However, the invention is particularly suited for use in systems in which said biological tissue is an in vivo blood vessel and said regions of different histology include different types of vessel wall plaque.

Within such blood vessel analysis it is found convenient that the signal processor should operate such that a lumen area of an in vivo blood vessel is identified by searching said radio frequency signals for a contiguous region having a signal peak to trough value below a predetermined lumen amplitude threshold and a signal peak value below a predetermined lumen level threshold.

The lumen area search starts from a known point that may be either manually or automatically identified as being within the lumen area. This identification may use or be assisted by ultrasonic blood flow detection. A preferred technique for identifying a lipid plaque is one of searching said radio frequency signals outside of a lumen area for a contiguous area having a signal peak to trough value above a predetermined noise amplitude threshold and below a predetermined lipid plaque amplitude threshold, a signal peak value below a predetermined lipid plaque level threshold, and the presence of a fibrous cap. A preferred technique for identifying a calcified plaque is one of searching said radio frequency signals outside of a lumen area for a contiguous area having a signal peak value above a predetermined calcified plaque peak level threshold, a signal trough value below a predetermined calcified plaque trough level threshold, a variation in signal peak to trough value below a predetermined calcified plaque amplitude variation threshold and a variation in signal period below a predetermined calcified plaque period variation threshold.

A further histological characteristic that may be diagnostically useful can be derived in embodiments in which said signal processor calculates from variations in parameters characterising said radio frequency signals a homogeneity value indicative of the degree of tissue homogeneity within one or more regions of said histological image and controls display within said histological image of said one or more regions for which a homogeneity value was calculated to visually represent said calculated homogeneity value.

In preferred embodiments the homogeneity value is only calculated for those regions not already identified as lumen, lipid plaque or calcified plaque. Fibrous plaques shown as a thickened vessel wall area with a high homogeneity value that is not otherwise identified as a vessel wall plaque. In order to more reliably identify tissue histology said signal processor divides said radio frequency signal into signal areas, at least some signal areas being tested to identify biological tissue histology by reference to their own signal characteristics in combination with signal characteristics of surrounding signal areas. It will be appreciated that the present invention could be embodied as an addon device to an existing ultrasound scanner that has a radio frequency signal and control signal output port that allows access to the radio frequency signals generated by reflected ultrasonic sound energy. Accordingly, viewed from another aspect the present invention provides an ultrasonic imaging apparatus for imaging biological tissue using radio frequency signals generated from reflected ultrasonic sound energy, said ultrasonic imaging apparatus comprising: a signal processor for comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology and for generating a histological image that visually distinguishes said one or more regions of different histology.

Viewed from a further aspect the present invention provides an ultrasonic imaging method of imaging biological tissue, said ultrasonic imaging method comprising the steps of: emitting ultrasonic sound energy into biological tissue to be imaged; generating radio frequency signals from received reflected ultrasonic sound energy; comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology; and generating a histological image that visually distinguishes said one or more regions of different histology.

An embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings in which: Figure 1 schematically illustrates an ultrasonic scanning apparatus for scanning blood vessels;

Figure 2 schematically illustrates a grey-scale image of a blood vessel suffering from a constriction due to vessel wall thickening by a plaque of an unknown type;

Figure 3 illustrates a radio frequency signal along a scan line path through the vessel of Figure 2 with characteristics indicative of a lipid plaque; Figure 4 illustrates a radio frequency signal as in Figure 3 but this time with characteristics indicative of a calcified plaque;

Figure 4 illustrates a radio frequency signal as in Figure 3 but this time with characteristics indicative of a fibrous plaque;

Figure 6 illustrates the sequence of operations of the signal processing portion of the system of Figure 1;

Figure 7 illustrates the histological analysis strategy of the technique set out in Figure 6.

Figure 1 illustrates an ultrasonic scanning apparatus that comprises a conventional ultrasound scanner 2 (e.g. a Sonus 2500 produced by Hewlett Packard) that has an ultrasonic transducer 4 (of a known frequency) that may be used to investigate a blood vessel 6 within the thoracic cavity 8 of a patient. The ultrasound scanner has a radio frequency signal output port that allows access to the raw radio frequency signal data for each scan line. Together with the radio frequency signal scan lines (similar to the raster scan input to a television receiver) there are separate synchronising signals.

The radio frequency signal is supplied via an amplifier stage to a high speed analogue to digital converter 10 which digitises the signal under control of control signal from a control unit and stores it in a buffer memory before passing it to a workstation computer 12 where the digitised radio frequency signal data for a complete signal frame is stored within the RAM of the workstation computer 12.

Once this digitised data has been captured by the workstation computer 12, it may be manipulated therein under software control and the workstation computer functions as a signal processing apparatus. It will be appreciated that the functionality of the analogue to digital converter 10 and the workstation computer 12 could be incorporated within the ultrasound scanner 2 itself to form an integral unit if desired or could be embodied as a separate stand alone device.

Figure 2 illustrates an example grey-scale image of a blood vessel. The blood vessel wall has a thickened portion 14. The lumen 16 may be manually identified by the operator possibly with the assistance of blood flow mapping techniques or maybe automatically identified (e.g. use colour flow mapping to identify a point having the highest velocity blood flow providing this information is passed to the workstation). Whilst the thickened portion 14 of the blood vessel wall is clearly a plaque of some form, it is sometimes difficult for the operator to unambiguously identify which type of plaque is shown.

Figure 3 illustrates the blood vessel of Figure 2 and the radio frequency signal produced along a scan line A, B, C, D, E and F through the blood vessel. In this case the plaque is a lipid plaque. The unthickened blood vessel wall A, B produces a strong and relatively irregular series of echoes in the radio frequency signal. This is followed by a low amplitude portion corresponding to the reflection from the blood within the lumen 16. The fibrous cap of the lipid plaque shows as a narrow area of relatively strong reflection C, D. The lipid body has the characteristic of relatively low reflection in the portion D, E. The final portion is the outer wall E, F.

In practice, the lumen 16 has already been identified by the user and so the portion B, C of the radio frequency signal is already identified and accounted for. In this way, the low amplitude reflection from the lumen is not confused with low amplitude reflection from lipid. As the rest of the radio frequency signal along the scan line is searched the portion D, E has peak to trough values (amplitude) above a noise level but below a predetermined threshold characteristic of a lipid plaque and the absolute value of the peaks is below a threshold indicative of lipid plaques. In this way, the portion D, E can be identified as corresponding to a lipid plaque and corresponding portions in adjacent scan lines may be identified to form a contiguous area within the histological image. Once the lumen and the lipid portion of the scan line have been identified the remaining regions are searched for calcified plaques (as described below) and then subject to a determination of a homogeneity value that is equal to the product of the variance of the peak values and the variance of the peak to trough values. This homogeneity value is then mapped to a corresponding colour range of a colour palette. The lipid region has its own distinctive colour within the colour palette.

It will be appreciated that the threshold values used in the lipid plaque test and other tests are specific to a particular ultrasound scanner and power setting being used. Lookup tables of threshold scaling factors for different power settings may be held within the signal processor and the scanner can undergo calibration to further improve the reliability of the thresholds used. Whilst the absolute values of the thresholds may be scanner and setup specific, the relative values of the threshold are much less so and together provide a good indicator of tissue histology. The threshold values can be indexed to a single value from a lookup table.

Figure 4 corresponds to Figure 3 except in this case the region D, E contains a calcified plaque. The characteristic signal parameters of a calcified plaque are strong and regular echoes. These characteristics may be identified by searching outside of the lumen for a continuous area having signal peak values above a given threshold, signal trough values below a given threshold and a variation in peak to trough values below a threshold in conjunction with a variation in signal period below a threshold. These signal parameters in combination define a highly regular and high amplitude signal. When such a calcified plaque region has been identified it is assigned its own distinctive colour within the histological image and is excluded from calculations of the homogeneity value.

Finally, Figure 5 is similar to Figure 3 except in this case the plaque is fibrous and the area D, E contains fibrous tissue. The area D, E does not meet either of the test criteria that identify a lipid plaque or a calcified plaque and so is not identified as either of these types. Accordingly, the area D, E is subject to the calculation of a homogeneity value together with the portions A, B; C, D and E, F. Thus, in the histological image the thickening will be seen as a relatively homogeneous portion of vessel wall that does not contain any lipid or calcified areas. Figure 6 illustrates the sequence of operation of the signal processor. At step

18 the signal processor captures a single frame of data (post time gain compensation processing) at its maximum dynamic range and reconstructs a grey-scale image. At step 20 the user examines a corresponding grey-scale image of the frame and identifies one point within the lumen. At step 22 the signal processor floods out the lumen area and marks it as lumen by searching away from the identified point until the signal value exceeds a given amplitude threshold indicative of the lumen wall being reached. Scan lines adjacent to the scan line containing the identified point can be searched starting from a point corresponding in distance from the ultrasonic transducer to the identified point.

At step 24 the user identifies the outer edges of the vessel by tracing around these in the reconstructed grey-scale image using a pointing device such as a mouse. Identifying the outer portion of the blood vessel is relatively straight forward to an experienced operator. The outer edges of the vessel are used to terminate the subsequent searches for structures and the calculation of the homogeneity value.

At step 26 those areas outside of the lumen but within the inside edges of the user defined contour are searched for areas (e.g. four signal cycles area treated as an area of sufficient size to be worthy of giving significance to its characteristics) for parameters indicative of lipid histology. If such areas are found, then they are marked within the radio frequency signal data as well as being excluded from further evaluation.

At step 28 a search is carried out for calcified structures and these areas are similarly marked. At step 30 all the remaining areas within the image that are not lumen, lipid core areas or calcified areas are subject to the calculation of a homogeneity value of which the logarithm is taken and then this value used to index a palette of a range of colours. Fibrous tissue is associated with low homogeneity values. At step 32 the histographic images displayed with lipid areas and calcified areas marked in their own distinctive colours and the colours dependent upon the homogeneity value calculated are displayed for the other areas. The lumen is left dark or is assigned its own colour.

Figure 7 illustrates the sequence of operations of Figure 6 in relation to an individual radio frequency scan line. Firstly, the lumen portion B, C is identified by searching away from the given point to find the position at which the signal value exceeds a predetermined threshold. This area is then marked as lumen and excluded from further searches. The second search is conducted through all areas other than the lumen and is looking for signals characteristic of lipid histology. In this case, such signals are found in the region D, E and this region is marked accordingly and excluded from further searching. The third stage is a search for calcified histology signal parameters through all the remaining areas that have not already been marked. In this case, no such characteristic signals were identified. Finally, all the remaining areas are subject to a homogeneity value calculation and marked with a colour value derived from this homogeneity value.

Claims

1. An ultrasonic imaging apparatus for imaging biological tissue, said ultrasonic imaging apparatus comprising: an ultrasonic transducer for emitting ultrasonic sound energy into biological tissue to be imaged and for generating radio frequency signals from received reflected ultrasonic sound energy; and a signal processor for comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology and for generating a histological image that visually distinguishes said one or more regions of different histology.

2. An ultrasonic imaging apparatus as claimed in claims 1, wherein said signal processor generates a grey-scale image of said biological tissue based upon said radio frequency signal intensity, said grey-scale image being overlain with said histological image.

3. An ultrasonic imaging apparatus as claimed in claim 2, wherein said histological image is a colour image with different colours representing different tissue histology or homogeneity.

4. An ultrasonic imaging apparatus as claimed in any one of claims 1, 2 and 3, wherein said biological tissue is an in vivo blood vessel and said regions of different histology include normal tissue and different types of vessel wall plaques.

5. An ultrasonic imaging apparatus as claimed in claim 4, wherein a lumen area of an in vivo blood vessel is identified by searching said radio frequency signals for a contiguous region having a signal peak to trough value below a predetermined lumen amplitude threshold and a signal peak value below a predetermined lumen level threshold.

6. An ultrasonic imaging apparatus as claimed in claim 5, wherein said lumen area search starts from a known point identified as being within said lumen area.

7. An ultrasonic imaging apparatus as claimed in claim 6, wherein said known point is identified using ultrasonic blood flow detection.

8. An ultrasonic imaging apparatus as claimed in claim 4, wherein a lumen area of an in vivo blood vessel is identified by ultrasonic blood flow mapping.

9. An ultrasonic imaging apparatus as claimed in claim 4, wherein a lumen area of an in vivo blood vessel is identified by user input.

10. An ultrasonic imaging apparatus as claimed in claim 4, wherein an outer border of an in vivo blood vessel is identified by user input.

11. An ultrasonic imaging apparatus as claimed in any one of claims 4 to 10, wherein a lipid plaque area is identified by searching said radio frequency signals outside of a lumen area for a contiguous area having a signal peak to trough value above a predetermined noise amplitude threshold and below a predetermined lipid plaque amplitude threshold, a signal peak value below a predetermined lipid plaque level threshold and the presence of a fibrous cap.

12. An ultrasonic imaging apparatus as claimed in any one of claims 4 to 11, wherein a calcified plaque area is identified by searching said radio frequency signals outside of a lumen area for a contiguous area having a signal peak value above a predetermined calcified plaque peak level threshold, a signal trough value below a predetermined calcified plaque trough level threshold, a variation in signal peak to trough value below a predetermined calcified plaque amplitude variation threshold and a variation in signal period below a predetermined calcified plaque period variation threshold.

13. An ultrasonic imaging apparatus as claimed in any one of claims 4 to 12, wherein said signal processor calculates from variations in parameters characterising said radio frequency signals a homogeneity value indicative of the degree of tissue homogeneity within one or more regions of said histological image and controls display within said histological image of said one or more regions for which a homogeneity value was calculated to visually represent said calculated homogeneity value.

14. An ultrasonic imaging apparatus as claimed in claim 13, wherein said homogeneity value is calculated for regions not identified as lumen, lipid plaque or calcified plaque.

15. An ultrasonic imaging apparatus as claimed in any one of the preceding claims, wherein said signal processor divides said radio frequency signal into signal areas, at least some signal areas being tested to identify biological tissue histology by reference to their own signal characteristics in combination with signal characteristics of surrounding signal areas.

16. An ultrasonic imaging apparatus for imaging biological tissue using radio frequency signals generated from reflected ultrasonic sound energy, said ultrasonic imaging apparatus comprising: a signal processor for comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology and for generating a histological image that visually distinguishes said one or more regions of different histology.

17. An ultrasonic imaging method of imaging biological tissue, said ultrasonic imaging method comprising the steps of: emitting ultrasonic sound energy into biological tissue to be imaged; generating radio frequency signals from received reflected ultrasonic sound energy; comparing said radio frequency signals with predetermined signal parameters indicative of different types of biological tissue histology to identify one or more regions of different histology; and generating a histological image that visually distinguishes said one or more regions of different histology.