Dissection apparatus

Abstract

The present invention relates to a dissection apparatus (100) comprising a dissection tool (110) for cutting biological material from an identified region of interest (ROI) in a tissue sample arranged on a slide (130), a light source (160) for illuminating the tissue sample, and an imaging module (150), the imaging module (150) comprising an image sensor (155) and an image processor (153), and configured to acquire a first image of the tissue sample before dissection and to register an annotated boundary (210) of the ROI on the first image, the annotated boundary being identified from a reference image of a corresponding reference sample. The imaging module is further configured to acquire a second image of the tissue sample after dissection, and the image processor (153) is programmed with an algorithm that performs the steps of: detecting the achieved dissection boundary (220) and determining the change in brightness based on a pixel-by-pixel comparison of the first and second images; comparing the detected dissection boundary (220) with a registered annotated boundary (210) to determine at least one indicator representing the achieved quality of the dissection result; and generating a quality report (170) including at least one determined indicator.

Description

[Technical Field]

[0001] The present invention relates to an apparatus for cutting biological material from a sample placed on a flat substrate such as a glass slide. [Background technology]

[0002] For molecular examination in oncology, a portion of a tissue section is collected from a glass slide. Generally, tissue is collected from multiple slides per case to obtain sufficient material for molecular examination. The region of interest (ROI) from which the tissue is collected is indicated by the pathologist on the reference image taken for the specific tissue section. The pathologist draws lines that identify the boundaries of the ROI, usually called annotations. The annotations on the reference image are then transferred to the images of each slide containing further tissue sections cut from the same sample. This is usually called registration of the reference image to the dissection image. After image registration, the dissection tool is controlled to mechanically remove the biological material within the annotated boundaries.

[0003] An example of an automated device for mechanically removing material from a biological sample is disclosed in International Publication No. 2012 / 102779. The device comprises an extraction tool, which has a rotary cutting tip for grinding material from an area of ​​the sample, and a liquid discharge port and a liquid suction port located very close to the cutting tip. The device is configured such that liquid is discharged at the cutting tip, and the ground material and discharged liquid are drawn into the extraction device through the suction port.

[0004] It is also known that biological materials can be removed by scraping using a dissection tool equipped with a scraping blade. An example of a dissection apparatus with such a dissection tool is disclosed in International Publication No. 2022 / 063695. The scraping blade is positioned in the orifice of the dissection tool, and suction is applied during dissection so that the sample material separated by scraping is drawn into the orifice and internal cavity of the tool, where the sample material is collected below a filter element spanning the internal cavity.

[0005] As is understood, the accurate extraction of the region of interest identified by the pathologist from each slide is crucial for the accuracy and quality of molecular examination results. Therefore, the accuracy of dissection results is a key factor in overall process quality control. In automated processes, it is important that the achieved dissection results can be evaluated in an automated and reliable manner so as to be able to identify misalignment issues and / or defects indicating issues with the dissection tool, which may be corrected for future use and determine the usefulness of the extracted material. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] International Publication No. 2012 / 102779 [Patent Document 2] International Publication No. 2022 / 063695 [Overview of the project] [Means for solving the problem]

[0007] The present invention relates to a dissection apparatus comprising a dissection tool for cutting biological material from an identified region of interest (ROI) within a tissue sample arranged on a slide, and equipped with a light source and an imaging module for illuminating the tissue sample, the imaging module comprising an image sensor and an image processor. The imaging module is configured to acquire a first image of the tissue sample before dissection and to register an annotated boundary of the ROI on the first image, thereby allowing the annotated boundary to be identified from a reference image of a corresponding reference sample.

[0008] The imaging module is further configured to acquire a second image of the tissue sample after dissection, and the image processor is configured to acquire a second image of the tissue sample after dissection. The steps include: detecting the achieved dissection boundary based on a pixel-level comparison of the first and second images and determining the change in brightness; The steps include: comparing the detected dissection boundary with the registered annotated boundary to determine at least one indicator representing the quality achieved in the dissection result; • A step of generating a quality report that includes at least one determined indicator of quality. An algorithm is programmed to perform this action.

[0009] In one embodiment, the change in brightness between pixels of the first and second images is determined based on transmission. The light source is then placed below the tissue slide, and the transmitted light is detected by an image sensor. In a further embodiment, the change in brightness is determined based on reflection, the light source is placed above the tissue slide, and the image sensor detects the light reflected from the tissue sample. To enhance contrast, first and second filter elements, such as a crossed polariser, are appropriately provided. The first filter element is placed between the light source and the tissue slide, and the second filter element is placed between the image sensor and the lens of the imaging module.

[0010] After material dissection and removal, the brightness of light detected for a particular pixel in the second image changes. For example, when the optical arrangement configuration includes a cross polarizer, the detected brightness for a pixel in the second image will be lower than the detected brightness for a pixel in the first image having the same coordinates.

[0011] Therefore, the brightness difference between pixels in the first and second images having the same coordinates is determined. Images of unstained, thin tissue sections may have low contrast and large brightness fluctuations, and may contain artifacts, so the brightness change after dissection may be very small and noisy. Therefore, the step of detecting the achieved dissection boundary preferably includes several substeps in which mathematical operations are performed on the determined brightness difference to eliminate noise and mathematically construct the detected boundary.

[0012] In a preferred embodiment, the algorithm performs a substep of normalizing the determined luminance difference between corresponding pixels of the first and second images. Preferably, this is followed by a morphological opening to remove noise from the normalized luminance difference. Next, connected components are detected in the normalized luminance difference, and the extracted areas are identified by thresholding the detected components based on their size. Preferably, filtering is also applied to remove small islands.

[0013] In the following substeps, the edges of the identified cut-out area are detected and selected as the detected dissection boundaries. A suitable method involves applying an edge detector, such as a Laplacian filter, followed by image thresholding, and selecting the larger value. Otsu's thresholding is an example of a suitable thresholding technique.

[0014] Preferably, the determined edges of the identified dissection area are selected as the detected dissection boundary based on their proximity to the annotated boundary, using the maximum distance measurement.

[0015] When an achieved dissection boundary is detected, it is compared to the registered annotated boundary. In some examples, this involves calculating the distance from each pixel on the detected boundary to the nearest point on the annotated boundary. From this calculation, one indicator of the quality of the achieved dissection result that may be determined and reported is the mean deviation with respect to the annotated boundary. This is calculated from the average of the absolute distances from each pixel on the detected boundary to the nearest point on the annotated boundary. If the mean deviation exceeds a predetermined threshold, the generated quality report may further indicate that the achieved dissection result is unsatisfactory, or, if the criteria are met, satisfactory.

[0016] Additionally or alternatively, a quality-determined indicator may be the local maximum deviation of the detected dissection boundary from the annotated boundary. Again, if the maximum distance exceeds a predetermined threshold, the generated quality report may further indicate that the achieved dissection result is unsatisfactory, or satisfactory, if the criteria are met.

[0017] The step of comparing the detected dissection boundary with the annotated boundary may further include detecting the misalignment of the detected dissection boundary with respect to the annotated boundary. Such misalignment indicates a misalignment between the imaging module and the dissection module of the apparatus on which the dissection tool is mounted. When such misalignment is detected, this is appropriately reported along with a proposed corrective action of checking the alignment of the imaging module with respect to the dissection module. Similarly, the size change of the detected boundary with respect to the annotated boundary can be calculated. Such size change may indicate a misalignment of the dissection tool and may be reported along with a proposed corrective action. The image processing algorithm performs a further step of mapping the detected dissection boundary onto the annotated boundary using a best-fit calculation to obtain a relevant transformation matrix. The transformation matrix providing the best fit is then analyzed to detect any misalignment or any change in size of the detected dissection boundary with respect to the annotated boundary, if an annotated boundary exists.

[0018] In other examples, a further indicator of quality determined from a comparison of the first and second images is the percentage of the ROI extracted. Appropriately, the algorithm performs a further step of determining the area of ​​tissue material to be extracted by identifying pixels in the first image that reside within a registered annotated boundary and exhibit a predetermined minimum luminance / intensity level. The area of ​​tissue material to be extracted is determined by identifying corresponding pixels in the second image that exhibit a change in luminance exceeding a threshold, which is set based on the background luminance level. Optionally, the background level may be determined locally.

[0019] Thereafter, the percentage of the excised area within the ROI for the area to be excised is calculated and can be reported as a further indicator of the quality of the achieved dissection result. As will be appreciated, the minimum acceptable percentage of tissue recovery rate from within the ROI can be set, for example, to 97%, and if the calculated recovery rate is less than the minimum value, the generated quality report may further indicate that the achieved dissection result is unsatisfactory, or, if the criteria are met, is satisfactory.

[0020] An advantage of further calculating the percentage of tissue excised from within the ROI is that, even though the detected dissection boundary exactly follows the registered annotated boundary, in cases where the area of tissue within the ROI is not excised, it can be determined that the dissection result is unsatisfactory.

[0021] Pixels of the second image that are outside the annotated boundary and exhibit a luminance change greater than the threshold may also be identified as excised areas. Thus, the quality indicator determined and reported may further include the ratio of the excised sample material outside the ROI to the excised area inside the ROI. This provides an indication of how much the intended excised tissue material is diluted by unintended sample material.

[0022] The quality report may further indicate the presence of any islands of non-excised material when the image processing algorithm detects the edges of such objects. In embodiments where the detected boundary is selected based on proximity to the annotated boundary, if such an object is overly distant from the annotated boundary, such an object is excluded from the calculation of the deviation of the detected boundary from the annotated boundary. For example, when the distance between the object and the annotated boundary is greater than the width of the dissection tool used in the dissection process.

[0023] The generated quality report may include one or more additional outputs.

[0024] In some examples, the report includes a first image of the tissue slide taken before dissection and a second image of the slide taken after dissection. Preferably, the registered annotated boundary is also shown on the first and / or second image.

[0025] In a further example, a second image is shown in which the detected dissection boundaries are superimposed. If at least one determined indicator of quality is the maximum local deviation, and the deviation exceeds an acceptable threshold, the portion of the detected boundary that exceeds the threshold may be further highlighted.

[0026] The quality report may further include, where applicable, an indication that the dissection boundary could not be detected or that the sample material was not removed.

[0027] In addition to comparing at least one determined quality indicator to an acceptable threshold to indicate whether the achieved dissection result is satisfactory, the comparison with the threshold may also be used to assign a quality rating or score to that quality indicator. When multiple quality indicators are determined, it is preferable that each indicator is compared to an acceptable threshold and an overall quality rating for the achieved dissection result is calculated and reported.

[0028] In further development, the step of image registration of the reference image and annotated boundary on the dissection image (the first image on the tissue slide) includes determining the quality of the image registration. One example of an applicable technique is intersection over union, which quantifies how accurately the reference image maps onto the dissection image. This may be assigned a quality assessment and included in the overall quality assessment.

[0029] As can be understood, such comprehensive quality assessment calculations and reporting can provide an indication of whether or not material extracted from a particular slide should be included in subsequent molecular analysis.

[0030] In addition to generating quality reports for each excised tissue slide in a given case, the apparatus of the present invention may also be configured to generate a report for the entire case. Appropriately, the quality assessment calculated for each slide is used to determine the quality assessment for the entire case, which is useful in determining whether the collected sample material corresponds, in principle, to a selected ROI on a reference image.

[0031] Therefore, the apparatus according to the present invention provides automatic determination and reporting of the achieved dissection results, which are extremely valuable in a clinical setting. Furthermore, the automatic determination of dissection results is not limited to the use of any particular type of dissection tool. In some embodiments, the dissection tool comprises a scraping blade, as described in International Publication No. 2022 / 063695. In other embodiments, the dissection tool is equipped with a milling tip, as described in International Publication No. 2022 / 063695. In yet another embodiment, the dissection tool is a laser, and the apparatus is configured for laser capture micro-dissection.

[0032] Other advantages are described in the following detailed description with reference to the attached drawings.

[0033] The present invention will be further described here with reference to embodiments described later. [Brief explanation of the drawing]

[0034] [Figure 1a]This is a schematic side view of a first example of a dissection apparatus and optical arrangement configuration for acquiring images of tissue slides according to the present invention, which includes a dissection tool. [Figure 1b] This diagram schematically shows an alternative optical configuration. [Figure 1c] This figure shows an example of a quality control report generated by the apparatus shown in Figure 1a. [Figure 2a] This figure shows an example of an image of a tissue sample taken up after dissection, with the detected dissection boundary and annotated boundary superimposed on the image. [Figure 2b] This figure shows an example of an image of a tissue sample taken up after dissection, with the detected dissection boundary and annotated boundary superimposed on the image.

[0035] It should be noted that items with the same reference number in different figures have the same structural features and functions or the same signals. If the function and / or structure of such an item has been described, there is no need for a repeated description of it in the detailed description. [Modes for carrying out the invention]

[0036] Pathological diagnostic investigations of biological materials such as tissues and cells form the basis for many treatment decisions, particularly in oncology. For example, genome-based testing is performed to inform treatment options for individual patients diagnosed with cancer. Biological material / tissue may be obtained from biopsy or surgery. Typical tissue processing steps include, for example, formalin fixation and paraffin embedding. Such formalin-fixed, paraffin-embedded (FFPE) biological material is then cut into thin sections that are fixed onto glass slides. Other methods for obtaining and preparing tissue samples are known. For molecular testing, samples are selected from such sections that meet the testing requirements. The pathologist identifies a region of interest (ROI), and the identified ROI on each tissue slide is then cut out to obtain sufficient material for molecular testing.

[0037] The present invention relates to a dicing apparatus for automated dicing of sample material within an identified region of interest. An example of a dicing apparatus according to the present invention is schematically shown in Figure 1a. The apparatus 100 comprises a dicing tool 110 mounted on a robotic arm 120 that is mounted in the apparatus housing 105. The robotic arm 120 comprises a series of actuators for performing the necessary movements during dicing, the movements of which are controlled by the apparatus control unit 140. Appropriately, the robotic arm 120: - A rotary actuator for rotating the tool around a vertical axis R that is perpendicular to the platform 137 of the device supporting the tissue slide 130 to be cut out. - XY stage for translational motion, and, - A Z-stage is provided for vertical motion. The Z-stage may be equipped with hinge bearings for position control to ensure that a constant and precise downward force is applied during the divisor.

[0038] In the illustrated example, the dissection tool 110 comprises a scraping blade 115 for mechanically separating sample material from a tissue slide 130, and the tissue slide 130 is preferably fixed and held in a slide mount 135 to prevent any movement of the slide during dissection. Typically, the slide mount holds multiple tissue slides from the same case and is automatically loaded onto the device's support platform 137. The device is further equipped with a vacuum generator (not shown) that generates an upward airflow at the scraping blade so that the scraped material is drawn into the internal cavity of the dissection tool 110.

[0039] The apparatus further comprises an imaging module / camera module 150 equipped with an image processor 153 configured to acquire and process images of each tissue slide 130. The imaging module has an image sensor 155 and a lens 157, and to facilitate image acquisition, the apparatus further comprises a light source 160 and contrast-enhancing filter elements 151, 152. In the illustrated example, images are acquired by transmitted light, and the light source 160 is positioned beneath the glass tissue slide 130. The first filter element 151 is positioned between the light source 160 and the slide 130, and the second filter element 152 is positioned between the image sensor 155 and the lens 157. Preferably, the first and second filter elements are cross polarizers.

[0040] An example of an alternative optical arrangement configuration is schematically shown in Figure 1b, in which the imaging module 150 is configured to capture an image based on reflected light. The light source 160 is positioned above the slide 130, and the light reflected from the tissue on the slide 130 is detected by the image sensor 155 of the imaging module. Again, first and second filter elements are used to enhance contrast. The first element 151 is positioned between the light source 160 and the slide, and the second element 152 is positioned between the image sensor 155 and the lens 157.

[0041] In an example of a suitable workflow for the dissection process, a batch of tissue slides from the same biopsy case is prepared. The tissue slides may include paraffin-embedded tissue samples, which are then deparaffinized and optionally stained. H&E-stained slides are included for reference and ROI identification. Other tissue slides in the batch may remain unstained. Preferably, the H&E reference slide is scanned by a high-resolution scanner to allow for the identification of tumor cells, and the H&E image is made available to the pathologist for ROI selection and annotation of one or more ROI boundaries. Each tissue slide is then scanned by the imaging module 150, and the captured digital image of each slide is stored. The annotated ROIs are then mapped in batches onto the digital images of each tissue slide. Preferably, the image processor 153 is programmed with an image registration algorithm that automatically transfers the annotated ROI boundaries onto the images of the other tissue slides. The coordinates of the annotations to be transferred are transmitted to the control unit 140 and used to guide the movement of the robot arm 120 and the corresponding movement of the dissection tool, thereby removing material from the ROI on each tissue slide.

[0042] It is important to understand that the region of interest selected by the pathologist must be dissected with high precision and accuracy. In practice, several factors can influence the accuracy of the dissection results achieved. These factors include the accuracy of the reference image registration to the dissection image, the alignment of the imaging module and dissection tool, and the accuracy and precision of the scraping motion and tissue cutting behavior.

[0043] In automated processes, it is crucial to determine and evaluate the quality of the achieved dissection results in an automated and reliable manner to determine whether the collected sample material can lead to successful molecular analysis. The dissection apparatus according to the present invention is configured to analyze the quality of the dissection results by detecting the actually dissected tissue area and comparing it with an annotated area indicated by the pathologist. The detection of the dissected area and its boundaries is based on the principle that removing tissue creates a local brightness difference between pixels in the image of the tissue sample before and after dissection.

[0044] The imaging module 150 is configured to acquire a first image of each tissue slide before dissection and a second image of the slide after dissection is completed. As described above, each slide 130 is fixed and held in the slide mount 135 to ensure that there is no slide movement during dissection and that the first and second images are aligned.

[0045] The image sensor 155 detects the brightness value for each pixel in the first image and for the corresponding pixel in the second image having the same coordinates. The image processor 153 is programmed to determine the brightness difference on a pixel-by-pixel basis.

[0046] Images of unstained, thin tissue sections may have low contrast and high brightness variability. Tissue contrast depends on thickness, imaging modality, and light source and camera settings. Large differences in brightness may exist because local tissue composition can vary considerably. As a result, there is no ideal exposure time. Areas may be over-illuminated or under-illuminated. Using illumination that highlights tissue will result in very small and noisy changes in brightness after dissection in areas containing low-brightness tissue. Images may also be affected by dust or other particles present on the slide before and / or after dissection. It may also occur that a region of interest annotated by the pathologist includes areas where no tissue is present.

[0047] These factors make it difficult to identify the boundaries of the cut-out areas in a robust and accurate manner. Therefore, the image processor 153 is programmed with an algorithm for detecting the achieved dissection areas and boundaries.

[0048] An example of a suitable algorithm performed by the apparatus in Figure 1a includes the following steps:

[0049] In the first step, the pixel-level brightness difference between the first image and the second image is calculated and normalized. I x,y =(Ipr x,y -Ipo x,y ) / Ipr x,y Here, I x,y This is the normalized luminance difference for all locations with coordinates x, y, Ipr x,y This is the pre-dissection brightness value detected for each pixel of the first image having coordinates x, y, IPO x,y This is the luminance value after dissection detected for each pixel of the second image having coordinates x, y.

[0050] In the second step, the area to be cut out is determined. This is, a) I x,y A substep to eliminate noise: morphological opening. b) Substep to detect linked components, c) A substep to threshold the detected components based on their size, and optionally, d) A substep to filter out small islands within the detected components. It should include appropriately.

[0051] In the third step, the edges of the determined and cut-out area are detected. Edge detectors, such as Laplacian filters, are applied to identify areas of rapid brightness changes. Other suitable edge detectors include Prewitt edge detection, Sobel edge detection, or Canny edge detection.

[0052] In the fourth step, image thresholding is applied to select the largest value at the detected edges. A fixed value may be used, or the threshold may be automatically determined by the application of the thresholding technique. Otsu thresholding is one example of a suitable thresholding technique. Other techniques that may be applied include the Moments method or Triangle thresholding.

[0053] The selected large value may be determined as the detected dissection boundary.

[0054] Preferably, the detected boundaries are further selected based on their proximity to the registered annotation boundaries using the maximum distance measurement. This has the advantage of eliminating false positives, as detected edges that are too far from the annotated boundary are highly likely to be errors. The maximum distance measurement may be set based on the width of the dissection tool.

[0055] The apparatus of the present invention is further configured to determine at least one indicator of the quality of the achieved dissection result and to generate a quality control report including this indicator of quality. Preferably, multiple quality indicators are determined and reported. The QC report may be displayed on the apparatus screen and / or other user interface and stored in a laboratory information management system.

[0056] The image processor 153 is further programmed to compare the detected dissection boundary with the registered annotated boundary. The image processing algorithm then performs a further step of calculating the distance of each pixel on the detected boundary to the nearest point on the annotated boundary.

[0057] In one example, the quality indicator determined by calculation is the mean deviation from the annotated boundary, and the mean deviation is determined as the mean of the absolute distance:

[0058]

number

[0059] In a further example, the indicator of quality determined from the calculated distance between the annotated boundary and the detected boundary is the maximum deviation. Appropriately, a histogram of the calculated distances is generated, and the maximum deviation is determined using a cutoff within the histogram to eliminate outliers. A cutoff at the 99th percentile may be used.

[0060] In a preferred embodiment, the image processor is further configured to determine the area of ​​tissue to be cut out from inside or outside the registered annotated boundary. The cut-out area is determined by identifying pixels in the second dissectioned image that show a change in brightness greater than a threshold. The threshold may be determined based on the level of background brightness. One indicator of quality that may be determined and reported is the percentage of the area cut out from the ROI relative to the area of ​​the ROI being cut out.

[0061] The area of ​​ROI to be excised is determined by identifying pixels in the first pre-dissection image that are located within the annotated boundary and exhibit high brightness. Such pixels are identified by setting a minimum value, which may be defined based on the optical arrangement configuration, the type of tissue sample being excised, and the level of background brightness. The area of ​​tissue excised from the ROI is determined by identifying the corresponding pixels in the second post-dissection image that exhibit brightness changes greater than a threshold. The percentage of tissue excised from the ROI is then calculated and reported as a further indicator of the quality of the achieved dissection result. This is useful if an area / island of tissue within the ROI was not excised, even if it shows the minimum deviation in the comparison between the detected dissection boundary and the annotated boundary. For such islands, no brightness change is detected, which affects the calculated percentage of tissue recovery from the ROI. Therefore, depending on the size of any such island, the dissection result may be concluded to be unsatisfactory even if the ROI boundary is precisely excised.

[0062] Similarly, the area to be cut outside the ROI can be determined and reported. The area to be cut outside the ROI is determined by identifying pixels in the image after the second dissection that are outside the annotated boundary and exhibit brightness changes greater than a threshold. This area can be related to the area to be cut inside the ROI and can indicate the degree to which the target tissue has been diluted by unintended sample material. Therefore, the ratio of the determined dissection area outside the ROI to the dissection area inside the ROI is a further indicator of the quality of the achieved dissection result, and this result can be determined and included in the generated report.

[0063] The generated quality control report may also include first and second images of the slide, taken before and after dissection, with annotated boundary overlays. An example of a quality control 170 report is shown in Figure 1c. The report was generated based on the dissection of a deparaffinized tissue sample using a dissection apparatus with a scraping blade, as described with reference to Figure 1a.

[0064] In addition to slide IDs and pre- and post-dissection images with annotated boundary overlays, the determined and reported quality indicators are the mean and maximum deviations from the annotated boundary, as well as the percentage of tissue cut out from within the annotated boundary, i.e., within the ROI. The actual area of ​​material cut out is also reported.

[0065] The report further includes an indication of whether the quality of the achieved dissection results in a given example is acceptable.

[0066] Appropriately, the image processor is programmed with thresholds for each of the determined quality indicators. For example, the threshold for the maximum deviation from the annotated boundary may be set to 0.25 mm. The threshold for the average deviation may be set to 0.1 mm. The threshold for acceptable tissue dissection from within the ROI may be set to a minimum of 97%.

[0067] Each determined quality indicator is compared to its corresponding threshold, and the achieved dissection result is reported as successful when each determined indicator is within the threshold.

[0068] The generated quality report may further include post-dissection images of the slide showing the detected dissection boundaries and annotated boundaries. An example of such an image 200a is shown in Figure 2a. Appropriately, detected boundaries and annotated boundaries are indicated using different colors, for example, red lines for detected boundaries and blue lines for annotated boundaries. In Figure 2a, the annotated boundary 210 is shown by a dashed line, and its predetermined portions are indicated by reference numerals 210a, 210b, and 210c. The detected dissection boundary 220 is shown by a solid line, and the corresponding portions of the detected boundary are indicated by reference numerals 220a and 220b.

[0069] In the illustrated example, the tissue sample is a deparaffinized sample cut using a dissection apparatus with a scraping blade, as described with reference to Figure 1a. A portion of the annotated boundary 210c lies in an area of ​​the actual sample that did not contain tissue. As a result, given that there is no change in brightness before and after dissection, the dissection boundary may not be detected in this area. In the generally lower left region of the image, there is a minimal deviation between the annotated boundary portion 210a and the detected boundary portion 220a. In the generally upper right region of the image, there is a significant deviation between the detected boundary portion 220b and the annotated boundary portion 210b. Since the average deviation exceeds the threshold, the report indicates that the dissection was unsuccessful. The reported image may also include further highlighted areas that exceed the maximum deviation.

[0070] In further development, the step of comparing the detected boundary 220 and the annotated boundary 210 includes identifying the misalignment of the detected dissection boundary 220 relative to the annotated boundary. When the area enclosed by the detected boundary approximately corresponds to the area of ​​the annotated boundary, the misalignment of their respective center points indicates the misalignment of the dissection module relative to the tool module. The generated quality report may include separate instructions for this, along with a suggestion to check the alignment of the tool module and the imaging module. Additionally or alternatively, such identified shifts and proposed corrective actions are reported separately in the maintenance report.

[0071] In other cases, the cut-out area may be larger than the ROI, meaning that the detected boundary 220 may show a size variation relative to the annotated boundary 210. This may indicate tool misalignment. For example, the tool's central axis may be deviating from the perpendicular, or perhaps the cutting edge is misaligned relative to the tool's central axis. If such a size variation is identified, the generated quality report may further include the area of ​​material cut outside the annotated boundary and suggest corrective actions, such as verifying tool alignment. Suggested corrective actions may also be reported separately in the maintenance report.

[0072] To identify the positional or size changes of the detected boundary 220 relative to the annotated boundary, the algorithm is configured to perform a further step of mapping the detected dissection boundary 220 to the annotated boundary 110 in a best-fit operation in order to determine the relevant transformation matrix that establishes a best fit between the two objects. When the detected boundary contains enough features to allow convergence, analysis of the transformation matrix allows for the identification of the positional or size changes of the detected boundary 220 relative to the annotated boundary.

[0073] In the example in Figure 2a, the detected boundary 210 has a "missing section" with respect to portion 210c of the annotated boundary, but contains enough fitting sections, i.e., sufficient features, to enable reliable mapping and transformation matrix determination. In other cases (such as the example in Figure 2b) where the majority of the annotated boundary does not coincide with the detected boundary, it may not be possible to identify the displacement / size change of the detected boundary relative to the annotated boundary.

[0074] Further examples of post-dissection images 200b obtained for different tissue samples are shown in Figure 2b. Again, the tissue samples are deparaffinized samples cut using a dissection apparatus with a scraping blade, as described with reference to Figure 1a. The annotated boundary 210, indicated by the dashed line, also includes portions 210c present in areas of the actual sample that did not contain tissue, where the dissection boundary may not be detected. Where tissue was present, the dissection boundary 220, indicated by the solid line, was detected. This report may further indicate the percentage of the annotated boundary 210 that matches the detected dissection boundary 220.

[0075] In further development, the image processor is further configured to analyze a first image of the tissue slide acquired before dissection in order to detect tissue boundaries. This may be performed after the annotated boundaries have been mapped to the first image. Appropriate image processing algorithms are applied to detect the tissue and its edges and check whether any detected edges lie within the annotated boundaries. This allows for the identification of areas where boundary detection after dissection is not expected. An acceptable threshold may then be set for the percentage of annotated boundaries that match the detected dissection boundaries, and a quality control report includes an indication of whether the results are satisfactory.

[0076] In the example in Figure 2b, the image processing algorithm further detected the edge of a small island 230 of uncut material located within the open portion 210c of the annotated boundary. The presence of island 230 can be seen in the post-dissection image shown in the quality report. The distance between the island and the edge of the annotated boundary was greater than the width of the cutting blade used during dissection, and was therefore excluded from the calculation of the deviation of the detected dissection boundary 220 relative to the annotated boundary 210. As mentioned earlier, the presence of islands also affects the percentage of tissue recovery from ROI and may lead to a determination that the dissection result is unsatisfactory.

[0077] I x,y In cases where no dissection boundary can be detected at all from the image processing of I, this is appropriately included in the quality report.

[0078] For each tissue slide in a particular case, one or more quality indicators are determined and reported. In addition to an indication of whether a particular quality criterion is met, e.g., whether the average deviation between the detected boundary and the annotated boundary is less than a predefined maximum value, a quality assessment may be assigned to each determined quality indicator, whereby an overall quality assessment for the dissection result achieved for a particular slide is calculated. This provides an indication of whether the material cut out from that slide can be reliably used in subsequent molecular tests. As will be understood, the assessments calculated for each slide in a case may be used, in whole or in part, to calculate an overall quality assessment for that case, and the overall quality assessment is appropriately reported in a further report generated for that case.

[0079] As mentioned above, the accuracy / quality of image registration can affect the achieved dissection results. In some embodiments, the apparatus is further configured to determine the quality of the registration of a reference image to a first pre-dissection image of a tissue slide. During the image registration step, the image processor appropriately quantifies the degree of overlap between the dissection image and the reference image by applying techniques such as intersection overunion. Once the annotated boundary is appropriately mapped to the dissection image of the pre-dissection tissue slide, the image processor is further programmed to evaluate the registration quality for the annotated ROI instead of the entire image. Thus, the quality evaluation may be assigned to the image registration of each dissection slide, reported separately in the quality report, and used in determining the overall quality evaluation for a particular dissection slide.

[0080] Examples, embodiments, or optional features, whether indicated as non-limiting or not, should not be understood as limiting the claimed invention. It should be noted that the embodiments described above are illustrative, not limiting, the invention, and that many alternative embodiments can be designed by those skilled in the art without departing from the scope of the appended claims.

[0081] Any reference numerals placed between parentheses in a claim shall not be construed as limiting the claim. The use of the verb “comprise” and its conjugations shall not preclude the existence of elements or steps other than those stated in the claim. The articles “a” or “an” preceding an element shall not preclude the existence of multiple such elements. The present invention may be implemented by hardware comprising several distinct elements and by a appropriately programmed computer. In a device claim listing several means, some of these means may be embodied by the exact same item of hardware. The mere fact that certain measures are listed in different dependent claims shall not imply that combinations of these measures cannot be used advantageously. [Explanation of Symbols]

[0082] 100 Dissection apparatus 105 Device Housing 110 Dissection Tools 115 Cutting blades for dicing tools 120 Robot arm equipped with actuators for moving dicing tools 130 Organizational Slides 135 Slide Mount 137 Support platform for the device 140 Control Unit 150 Imaging Module 151, 152 Contrast Enhancement Filter Elements 153 Image Processors 155 Image Sensor 157 Lens 160 light source 170 Quality Control Report Images of tissue slides after dissection (200a, 200b) 210 Registered Annotated Boundaries 210a, 210b, 210c Certain parts of annotated boundaries 220 Detected dissection boundaries 220a, 220b A predetermined portion of the detected dissection boundary

Claims

1. A dissection apparatus (100) comprising a dissection tool (110) for cutting biological material from an identified region of interest (ROI) in a tissue sample placed on a slide (130), the apparatus comprising a light source (160) for illuminating the tissue sample and an imaging module (150), the imaging module (150) comprising an image sensor (155) and an image processor (153), - Take a first image of the tissue sample before dissection, - Registering an annotated boundary (210) of the ROI on the first image, wherein the annotated boundary is identified from the reference image of the corresponding reference sample. In a dicing apparatus (100) configured to do the following, The imaging module is further configured to acquire a second image of the tissue sample after dissection, and the image processor (153) is configured - A step of detecting the achieved dissection boundary (220) based on a pixel-level comparison of the first and second images and determining the change in brightness, - A step of comparing the detected dissection boundary (220) with the registered annotated boundary (210) to determine at least one indicator representing the quality achieved in the dissection result, - A step of generating a quality report (170) that includes at least one determined indicator. A dissection apparatus characterized by having an algorithm programmed to perform the following.

2. The step of detecting the achieved dissection boundary (220) is: - A step of normalizing the determined brightness difference between corresponding pixels in the first and second images, - A step of detecting connected components in the normalized luminance difference and identifying the extracted area, - The steps include determining the edges of the identified and cut-out area by applying an edge detection filter, - A step of selecting the determined edge as the detected dissection boundary (220) and The dicing apparatus according to claim 1, further comprising:

3. The dissection apparatus according to claim 2, wherein the step of determining the edges of the identified and cut-out area further includes applying image thresholding to identify large values, the identified large values ​​being selected as the detected dissection boundary (220).

4. The dissection apparatus according to claim 2 or 3, wherein the step of detecting the achieved dissection boundary further comprises using the maximum distance measurement to select a determined edge of the identified dissection area as the detected boundary (220) based on its proximity to the annotated boundary (210).

5. The dissection apparatus according to any one of claims 1 to 4, wherein the step of comparing a detected dissection boundary (220) with an annotated boundary (210) includes calculating the distance from each pixel on the detected boundary to the nearest point on the annotated boundary (210).

6. At least one indicator of the achieved quality is - The average deviation of the detected dissection boundary (220) from the annotated boundary (210), calculated from the average absolute distance from each pixel on the detected boundary to the nearest point on the annotated boundary, or - Local maximum distance of dissection boundary detected from annotated boundary The dissection apparatus according to claim 5.

7. The dissection apparatus according to claim 5 or 6, wherein the step of comparing a detected dissection boundary (220) with an annotated boundary (210) further includes calculating the percentage of annotated boundaries that match the detected dissection boundary, and the resulting quality report (180) includes the calculated percentage.

8. Image processing algorithms, - A further step of determining the area of ​​tissue within the ROI to be excised by identifying pixels in the first image that are located within an annotated boundary (220) and exhibit a brightness greater than a predetermined minimum value, A further step of determining the area of ​​tissue cut out from the ROI by identifying a corresponding pixel in a second image that exhibits a change in brightness greater than a threshold, which is set based on the level of background brightness, - Further steps to determine the percentage of organizational area extracted from ROI Execute, The dicing apparatus according to any one of claims 1 to 7, wherein the generated quality report includes a determined percentage as a further indicator of the quality of the achieved dicing results.

9. Image processing algorithms, - A further step of determining the area of ​​material cut out outside the ROI by identifying pixels in the second image that show a change in brightness greater than a threshold and are located outside the annotated boundary (210), - A further step of calculating the ratio of the area outside the ROI to the area inside the ROI. Execute, The dissection apparatus according to claim 8, wherein the generated quality report includes a determined percentage as a further indicator of the quality of the achieved dissection result.

10. The step of comparing the detected dissection boundary (220) with the annotated boundary (210) is: - The detected dissection boundary (220) is mapped onto the annotated boundary (210) using a best-fit operation to obtain the associated transformation matrix, - Analyze the transformation matrix to identify the displacement of the detected boundary relative to the annotated boundary, or the change in the size of the detected boundary relative to the annotated boundary. It further includes, The dissection apparatus according to any one of claims 1 to 9, further configured to report any such identified misalignment, along with a proposed corrective action to check the alignment of the imaging module (150) with respect to the dissection tool (110), and to report any such identified size change, along with a proposed corrective action to check the alignment of the dissection tool (110).

11. The generated quality report (170) - The first image and the second image, including the overlay of the registered annotated boundary (210) - Overlay of the detected dissection boundary (220) on the second image, and, - An indication that the dissection boundary cannot be detected. A dissection apparatus according to any one of claims 1 to 10, further comprising one or more outputs selected from.

12. A dissection apparatus according to any one of claims 1 to 11, wherein the algorithm performs a further step of comparing at least one determined indicator of quality with an acceptable threshold, and the resulting quality report (170) further includes an evaluation of the indicator of quality and / or an indication of whether the achieved dissection result is satisfactory on the basis of the comparison.

13. The dicing apparatus according to claim 12, wherein each of the quality-determined indicators is compared to an acceptable threshold, and the generated quality report (170) further includes an overall quality assessment of the achieved dicing results.

14. The dissection apparatus according to claim 13, wherein the image processor (153) is further configured to determine the quality of the image registration by quantifying the degree of overlap between a reference image and a first image of a tissue slide, and the determined image registration quality forms part of an overall quality evaluation.

Images