Whole slide imaging methods for microscopes

Abstract

A method is provided for operating a microscope scanner (1) including a line scanner (3). Pre-scan data is generated for a target (6) including a sample, wherein generating the pre-scan data includes: obtaining a pre-scan image of the target (6) from the pre-scan; and identifying one or more sample-containing regions (24) and one or more sample-free regions (26) of the target (6) from the pre-scan image. Imaging scans are then performed on the scanned areas of the target (6) including one or more sample-containing regions (24) and one or more sample-free regions (26). The scanning speed of the line scanner (3) is adjusted along the image scanning path based on the pre-scan data, such that the target (6) is imaged at a higher scanning speed in one or more sample-free regions (26) than in one or more sample-containing regions (24).

Description

TECHNICAL FIELD

[0001] The present invention relates to a method of operating a microscope scanner. BACKGROUND

[0002] Whole slide scanners scan pathology or histopathology samples mounted on microscope slides under brightfield viewing conditions. The term "whole slide scanner" is used because the intention is to scan the entire tissue sample on the slide. A pre-scan is typically performed on the whole slide to produce a low resolution "pre-scan image". This image is then analysed to identify one or more areas of interest ("AOI") on the target where sample is believed to exist, and these regions are selectively imaged during a subsequent high resolution imaging scan. Typically, the imaging scan will be performed at a scan resolution of 2000 lpmm to 4000 lpmm (lines per millimetre, where each image acquired by the line scanner corresponds to a single line), or equivalently 51000 dpi to 101000 dpi (dots per inch). The pre-scan will instead be performed at a much lower resolution, typically between 12 lpmm to 61 lpmm (300 dpi to 1500 dpi). The pre-scan can therefore be performed much more quickly (at least an order of magnitude faster) than the imaging scan. As a result, target regions outside the AOI are excluded from the imaging scan to reduce the total scan time.

[0003] If the size of the tissue is the same as or smaller than the size of the pixels in the pre-scan, then good contrast cannot be given in the resulting pre-scan image. As a result, the analysis of the pre-scan image for detecting the presence of tissue can not be able to detect the tissue. This would result in the tissue not being scanned in the high resolution image. To mitigate this problem, pre-scan images of ever higher resolution can be produced to detect even smaller regions of tissue. However, as the resolution of the pre-scan image is increased, it takes longer to perform the pre-scan and it takes longer to analyse the pre-scan image to find regions of tissue.

[0004] The pre-scan image typically has a large depth of field compared to the thickness of the tissue. This is advantageous because no focusing step is required to perform the scan of the tissue, but has the disadvantage that defects on the surface of the slide (such as dust and scratches) are also in focus. This forces the AOI detection algorithm to distinguish between surface defects and tissue. This discrimination between tissue and defects is not trivial and a trade-off must be found between analysis time, how much tissue the AOI detection algorithm misses and how many defects the AOI detection algorithm incorrectly identifies as tissue.

[0005] It is desirable to ensure that the entire sample is imaged during the imaging scan without significantly compromising the speed or efficiency of the scanning process. The present invention is presented in the context of addressing this problem. SUMMARY

[0006] A first aspect of the invention provides a method for operating a microscope scanner comprising a line scanner, the method comprising:

[0007] generating pre-scan data for a target comprising a sample, wherein generating the pre-scan

[0008] data comprises:

[0009] obtaining a pre-scan image of the target from the pre-scan; and

[0010] identifying one or more sample-containing regions and one or more sample-free regions of the target from the pre-scan image, wherein the one or more sample-containing regions correspond to locations on the target where the sample is expected to be present, and wherein the one or more sample-free regions correspond to locations on the target where the sample is expected to be absent; and

[0011] performing an imaging scan of a scan region of the target, the scan region comprising the one or more sample-containing regions and the one or more sample-free regions, wherein performing the imaging scan

[0012] comprises:

[0013] moving the line scanner relative to the target along an image scan path, and using the line scanner to acquire an image of the target at each of a plurality of locations along the image scan path; and

[0014] adjusting a scan speed of the line scanner along the image scan path based on the pre-scan data, such that the target is imaged at a higher scan speed within the one or more sample-free regions than in the one or more sample-containing regions.

[0015] The pre-scan data is analysed to determine regions containing the sample (typically tissue) and regions where the sample is believed to be absent. However, rather than performing the imaging scan only in regions detected as containing the sample, the "sample-free regions" are also imaged as a precaution, as errors can arise from the analysis of the pre-scan data, and it is desirable to ensure that all sample is imaged. Optionally, this can extend to imaging the entire target during the imaging scan. During the imaging scan, the scan speed is increased in the one or more sample-free regions, typically resulting in a corresponding reduction in image resolution in these regions. This enables the sample detection algorithm to set a trade-off between tissue detection and defect detection, such that fewer defects are falsely identified as tissue. Furthermore, even if the sample is imaged at a lower resolution within the conceptual "sample-free regions" than within the "sample-containing regions", the data generated is typically more useful than if no sample-free regions were imaged at all. The increased scan speed in the sample-free region(s) ensures that any additional time loss incurred is minimal.

[0016] Performing the imaging scan preferably further comprises monitoring one or more image parameters from the images of the target acquired along the image scan path, and in response to a change in the one or more image parameters, further adjusting the scan speed of the line scanner within the one or more sample-free zones. By doing so, the scanner can react to the (accidental) presence of a sample within the "sample-free zone". When a sample enters the field of view of the line scanner, it will cause a corresponding change in the monitored one or more image parameters. In response to this change, the scan speed is adjusted, thereby enabling imaging of the sample at high resolution during the imaging scan, regardless of its location on the target. As a result, a higher proportion of the samples on the target can be imaged, which improves the overall usefulness of the scanned image to the microscope user. According to this technique, because the scan speed can be adjusted on-the-fly in response to the monitored image parameter (which is typically related to the amount of detail in the image), the spatial resolution of the scan is not predetermined by the scanner at the point at which the imaging scan starts.

[0017] Typically, the one or more image parameters are monitored to detect the presence of a sample within the one or more sample-free zones, and the scan speed is adjusted so that locations on the target at which a sample is detected are imaged at a lower scan speed than locations on the target at which no sample is detected. This enables a compromise between imaging the entire target at high resolution (which is desirable to capture the entire sample, but is also time-consuming) and imaging only those zones identified as containing a sample from the pre-scan image. Furthermore, if a sample is detected outside of a sample-containing zone, the scanner reacts to ensure that the sample is imaged at a relatively slow speed and, therefore, at high resolution wherever it exists on the target.

[0018] Further adjusting the scan speed based on the pre-scan data preferably comprises imaging the target at a first target speed within the one or more sample-free regions and imaging the target at a second target speed within the one or more sample-containing regions, wherein the first target speed is higher than the second target speed. Thus, further adjusting the scan speed of the line scanner within the one or more sample-free regions preferably comprises reducing the scan speed from the first target speed in response to a change in the one or more image parameters indicative of the detection of the sample. Thus, the sample-free regions can only be imaged at the first target speed if the monitored one or more image parameters are indicative of a first criterion, typically the absence of a sample in the acquired image. If the one or more image parameters are instead indicative of a second criterion, typically the presence of a sample in the acquired image, the scan speed is reduced from the first scan speed to the second target speed or towards the second target speed. Optionally, the sample-containing regions can be imaged at the second target speed if the monitored one or more image parameters are indicative of the first criterion or the second criterion.

[0019] During the imaging scan, the speed of travel can be reduced when the pre-scan data indicates that the edge of the sample is being approached, so that the sample is always imaged at high resolution. To avoid sudden deceleration, this can occur over a region before the tissue edge is reached. Thus, the imaging scan can further comprise reducing the scan speed from the first target speed to the second target speed or towards the second target speed before the line scanner transitions from acquiring images of sample-free regions to acquiring images of sample-containing regions. Advantageously, the transition between speeds is arranged to occur in a sample-free region of the target. Then, during the imaging scan, the line scanner can be moved at a constant speed through the sample-containing regions. This ensures that the sample-containing regions are always imaged at a uniform high resolution. However, the imaging scan can further comprise increasing the scan speed within the one or more sample-containing regions in response to a change in the one or more image parameters indicative of the absence of a sample in the acquired image. This can occur, for example, in the event of an error in identifying sample-containing regions from the pre-scan data. If the monitored one or more image parameters during the imaging scan indicate that a sample is not contained within the region being imaged, the scan speed can be increased to reduce the total scan time.

[0020] Adjusting the scan speed typically comprises decreasing the scan speed of the line scanner in response to the one or more image parameters changing in a first predetermined manner relative to a threshold. For example, the one or more image parameters can change in a first predetermined manner relative to a threshold in response to an image of the sample being taken within the sample free zone. Similarly, further adjusting the scan speed typically comprises increasing the scan speed of the line scanner in response to the one or more image parameters changing in a second predetermined manner relative to the threshold. The threshold is preferably predetermined and stored in memory, for example after calibration of the scanner. The one or more image parameters can comprise a brightness of the image, wherein the scan speed of the line scanner is decreased in response to the brightness of the image decreasing. Similarly, the one or more image parameters can comprise a focus merit value, and wherein the scan speed of the line scanner is decreased in response to the focus merit value increasing. Typically, the focus merit value is a numerical measure of complexity within the image, wherein the greater the value, the greater the detail in the image. An image that is more accurately in focus has a correspondingly higher merit value. The imaging scan preferably comprises calculating a focus merit value at each of the locations along the image scan path, and adjusting the focus height of the line scanner along the image scan path based on the focus merit values.

[0021] During an imaging scan of the entire scan area, images are typically taken at a constant rate and in time. Thus, image capture can not be linked to any particular measurand, but is controlled independently by a timer function. Furthermore, during an imaging scan, the line scanner is typically continuously moved along the image scan path relative to the target. Thus, the width of each pixel (and hence the overall resolution of the image in a particular region of the target) will typically vary depending on the speed at which the line scanner is moved relative to the target at the time the image is captured. Thus, a low resolution image will correspond to an image in which the line scanner moves past more of the target between each successive image capture than for a high resolution image. The spatial resolution of the imaging scan is typically not constant over the image scan path, and will depend on the scan speed at a given location on the target.

[0022] It is particularly desirable that the pre-scan can be performed quickly. When performing the pre-scan, focus data for the target can not be available, so a different camera (typically having a focus depth that exceeds that of the line scanner) can be used for the pre-scan. The focus depth of the pre-scan is preferably in excess of 500 microns (optionally at least 1 mm), whereas the focus depth of each image taken along the imaging path can be less than 3 microns (and typically 1 micron). This typically ensures that the entire sample is within the focus depth of the pre-scan camera. The pre-scan images are typically obtained from a single pass of the pre-scan camera along a pre-scan imaging path, which can be linear or in a raster format. The pre-scan images are typically taken at a constant spatial resolution over the entire pre-scan imaging path.

[0023] Generating the pre-scan data preferably further includes performing an image analysis of the pre-scan image to produce a probability map indicating a likelihood of sample presence at each location on the target, wherein the one or more sample-containing regions and the one or more sample-free regions are selected from the probability map. Adjusting the scan speed of the line scanner along the image scan path based on the pre-scan data can then include adjusting the scan speed according to the likelihood of sample presence at a given location on the target determined from the probability map. The one or more sample-free regions can correspond to locations from the probability map having a low likelihood of containing sample, and the one or more sample-containing regions can correspond to locations from the probability map having a high likelihood of containing sample.

[0024] In addition to the sample-free regions and the sample-containing regions, one or more other regions of the target can be identified for which the probability map is inconclusive as to whether sample is present in the region. For example, generating the pre-scan data can further include identifying one or more inconclusive regions on the target from the probability map having a likelihood of containing sample between the likelihood of the sample-free regions and the likelihood of the sample-containing regions. A different scan speed can be set for this region than for the sample-free regions and the sample-containing regions. For example, adjusting the scan speed based on the pre-scan data can include imaging the target at a first target speed within the one or more sample-free regions, imaging the target at a second target speed within the one or more sample-containing regions, and imaging the target at a third target speed within the one or more inconclusive regions, wherein the third target speed is higher than the second target speed and lower than the first target speed. If the one or more monitored image parameters indicate that sample is present within the field of view of the line scanner at any point through any of the three regions, the scan speed can be set to the target speed associated with the sample-containing regions (in this case, the second target speed).

[0025] A high resolution image of the target is typically generated from a combination of the images acquired during the imaging scan. Generating the high resolution image preferably includes using image interpolation to increase the resolution of the images acquired at the higher scan speeds to correspond to the resolution of the images acquired at the lower scan speeds. Accordingly, the scan speed at which each image was captured during the imaging scan is typically stored in memory. Generating the high resolution image can further include adjusting the interpolation ratio to maintain a constant resolution in the high resolution image.

[0026] The microscope scanner is preferably a brightfield microscope, and the imaging scan is preferably performed in transmission mode for the microscope scanner. The microscope scanner can also be configured to operate in reflection mode. One or both modes can be used as part of the pre-scan. For example, while the transmission mode can be useful in identifying sample-containing regions and sample-free regions, the reflection mode can be used to identify opaque regions of the object, such as where labels can be attached to the object. The opaque regions do not need to be scanned during the imaging scan, so the scan region is preferably selected from the pre-scan image so as to cover substantially the entire object, excluding any opaque regions of the object.

[0027] A second aspect of the application provides a computer program product containing instructions which, when executed using a microscope scanner comprising a line scanner and an object, cause the microscope scanner to perform the method of the first aspect, the object comprising a sample. The second aspect has similar features and advantages as discussed in connection with the first aspect.

[0028] The microscope scanner is preferably a virtual microscope, and the computer program product typically corresponds to a non-transitory computer readable medium, such as a hard drive. The object can take many forms, however preferably the object comprises a biological tissue sample. The object preferably does not exhibit significant variations in surface height, so as to be substantially flat for the pre-scan, which typically has a large depth of focus. For example, the object can be a stained tissue extract that has been sectioned and preserved between flat glass slides. The focus of the line scanner can be adjusted during the high resolution imaging scan, as the surface height of the object can vary by more than the depth of focus used in the imaging scan. BRIEF DESCRIPTION OF DRAWINGS

[0029] Embodiments of the application will now be discussed with reference to the following drawings in which:

[0030] Figure 1 is a schematic diagram of a microscope scanner to be operated in accordance with a first embodiment of the application;

[0031] Figure 2 is a flowchart illustrating a process in accordance with the first embodiment of the application;

[0032] Figure 3 is an example of a pre-scan image taken in accordance with the first embodiment of the application;

[0033] Figure 4 is an example of a pre-scan image taken in accordance with the first embodiment of the application with the scan region identified; and

[0034] Figure 5This is an example of a pre-scanned image obtained in the case of identifying sample-free and sample-containing regions according to a first embodiment of the present invention;

[0035] Figure 6 This is a graph showing the change in scanning speed as the device transitions from a sample-free region to a sample-containing region according to a first embodiment of the present invention.

[0036] Figure 7 This is a graph showing the change in scanning speed when transitioning from a sample-containing region to a sample-free region according to a first embodiment of the present invention;

[0037] Figure 8 This is a graph showing the change in scanning speed when the presence of a sample is detected in a sample-free region according to a first embodiment of the present invention; and

[0038] Figure 9 This is a graph showing the interpolation of data when generating a high-resolution image of a target according to a first embodiment of the present invention. Detailed Implementation

[0039] exist Figure 1 An example of a bright-field microscope scanner 1 is shown. The microscope scanner 1 includes a scanning head 2, which comprises a line scanner 3 in the form of a linear photodetector array. The scanning head 2 also includes an adjustable focusing system 4 and a pre-scan camera 12. A platen is provided.

[0040] 5. A target 6, in the form of a pathological slide containing a tissue sample, is positioned on it. A drive mechanism 7 connects the scanning head 2 to a track 8, allowing the scanning head to move relative to the target 6 as indicated by arrow 9. The track 8 also allows the scanning head 2 to move laterally to image adjacent strips.

[0041] (swath). The microscope scanner 1 is controlled by an electronic controller in the form of a computer system 10. Figure 1 The system also provides coordinate axes for reference in the example below. The vertical axis (z-axis) is aligned with the optical axis of the scanning head 2 and the focal height of the target 6, while the horizontal axis (x-axis) represents the scanning direction (parallel to the optical axis). Figure 1The surface of the target 6 is aligned in the xy-plane, and the lens of the adjustable focus system 4 is movable along the z-axis. In an alternative example, the scanning head 2 remains stationary, while the target 6 is moved in the xy-plane, so as to achieve relative motion between the scanning head 2 and the target 6 during the scanning process. The pre-scan camera 12 can also form a separate unit from the scanning head 2, and operate independently from the scanning head 2. The pre-scan camera 12 can comprise a line scanner, however, the choice is not particularly relevant, as long as a two-dimensional pre-scan image of the entire slide can be obtained using the pre-scan camera 12. In contrast, a line scanner 3 has to be used in the imaging scan following the pre-scan. TDI sensors are not suitable for the imaging scan, nor as 2D sensors operating in 2D imaging mode, as will be discussed later. An example of a suitable microscope scanner 1 is the VENTANA DP200 provided by Roche Diagnostics.

[0042] Reference will now be made to Figures 1 to 9 The operation of the microscope scanner 1 performing the method according to the first embodiment of the application will be described. The method starts in step 100 Figure 2 ), in which a pre-scan is performed to generate pre-scan data comprising a pre-scan image of the target 6. The scanning head 2 is advanced along the track 8, in which the pre-scan camera 12 is used to acquire an image of the target 6 in a single pass along a pre-scan imaging path. The pre-scan camera 12 is configured to acquire an image of the target at a focal depth of 1 mm, which is significantly higher than the thickness of the sample (typically between 2 pm and 10 pm). The focal height of the pre-scan camera 12 is therefore not adjusted during the pre-scan, and the pre-scan image is produced quickly at a relatively low resolution (e.g. 1200 dpi (47 lines per mm)). The pre-scan image provides an overview of the profile of the sample on the target 6, from which one or more areas of interest AOI can be selected for imaging during the first imaging scan.

[0043] Figure 3 An example of a pre-scan image is shown. The opaque area of the target 6 is identifiable, as a label 22 is pasted to the right side of the slide. As is known in the art, this area is identified by operating the microscope scanner 1 in transmission mode and reflection mode during the pre-scan, and comparing the resulting images. The remaining (optically transmissive) part of the target 6 is selected by the computer system 10 as the scan area 20, on which the imaging scan will be performed. This scan area 20 is identified by a herringbone pattern overlaid onto the pre-scan image in Figure 4 .

[0044] At step 101, the pre-scan image is processed using conventional software so that each region in which a sample is visible in the pre-scan image is identified as an AOI, also referred to herein as a "sample-containing region". More specifically, an automated image analysis of the pre-scan image is used to produce a probability map indicating where on the target a sample is expected to be present, from which one or more AOIs are selected. Typically, the boundaries of the AOIs can be selected so as to lie just outside the outer edges of the sample (as identified from the pre-scan image) so that the sample is entirely contained within the AOIs. The pre-scan image can then be displayed to the user using the computer system 10 with the one or more AOIs delineated. Optionally, the user can then manually adjust the boundaries of the one or more AOIs using the computer system if desired. For the purposes of illustration, Figure 5 The pre-scan image of Figure 3 and Figure 4 is shown, but the herringbone pattern is not overlaid on the sample-containing region 24 identified from the pre-scan image. The remaining portion of the scan area 20 (outside the sample-containing region 24) and through which the herringbone pattern extends is referred to herein as a "sample-free region" 26.

[0045] The imaging scan is performed at a much slower speed than the pre-scan, so it is desirable to image only those regions on the target 6 where a sample is present during the imaging scan. Typically, the imaging scan will therefore be performed on only the one or more AOIs identified from the pre-scan data. However, the resolution of the pre-scan image is relatively low, and errors can be made in the AOI detection process. Thus, it is possible that portions of the sample are inadvertently excluded from the one or more AOIs identified. These regions will therefore not be imaged during the imaging scan, which is undesirable. If the imaging scan were instead performed uniformly over the entire scan area 20, the total scan time would be very long, and a large amount of data would be generated, which is slow to process. Therefore, a compromise is made in which the entire scan area 20 is imaged during the imaging scan, however, the target scan speed is faster within the sample-free region 26 than within the sample-containing region 24.

[0046] At step 102 of the method, a seed location is selected within the sample-containing region 24 in the xy-plane (typically at the edge of the sample-containing region 24 in the AOI), and the focal height of the sample is measured at this location using the line scanner 3. The focal height of the adjustable focus system 4 is varied over a predetermined range of focal heights in order to obtain image information from the target 6 with a focal depth of approximately 1 pm at a plurality of different focal positions along the z-axis. The in-focus position is then calculated from the image information at each focal position by using a focus parameter in the form of a focus merit value. Typically, the focus merit value is a numerical measure of the complexity within the image, where the greater the value, the greater the detail in the image. An image that is more in-focus has a correspondingly higher merit value. The focal height with the highest merit value is then stored as the in-focus position for the measurement of the seed location, and the process is repeated for any remaining seed locations identified.

[0047] The image scan path that the scanning head 2 will move through is calculated so that the line scanner 3 images the entire scan area 20, including the sample-containing region 24 and the sample-free region 26. The line scanner 3 is then moved to a start location within the sample-free region 26 and the focal height is set based on the focal height recorded for the in-focus position of the seed location. For example, the focal height at the start location can be the same as the focal height at the in-focus position of the seed location, or the focal height at the start location can be based on an extrapolation of the position.

[0048] The purpose of acquiring images of the no-sample region 26 during the imaging scan is that if a portion of the sample is located in the no-sample region 26 (and thus is not detected by the analysis of the pre-scan image), it can be imaged. However, it is desirable to acquire any images of the sample at a resolution similar to or equal to the resolution of the images acquired within the sample-containing region 24. According to the first embodiment, the computer system 10 therefore monitors one or more image parameters for each image acquired along the image scan path. Each time a sample enters or exits the field of view of the line scanner 3, the monitored image parameter will typically change above or below a threshold value. For example, the focus merit value or the frequency content at each location can be monitored, and each time a sample enters the field of view, the focus merit value or the frequency content will typically rise above a predetermined threshold. Alternatively or additionally, the brightness of the image can be monitored, with a drop in brightness below a brightness threshold (e.g. 90% to 95% of the maximum brightness) indicating that a sample has entered the field of view of the line scanner 3. A clear slide will have a given brightness value with a fixed noise or minimum high frequency response. If the brightness drops below the brightness threshold or the noise or high frequency value increases above a respective threshold, there can be tissue present, and the scan speed can be reduced to increase the scan resolution. Typically, this will be done over one region to eliminate sudden accelerations. In this way, tissue not detected in the pre-scan image can be scanned at high resolution. Once no more tissue is detected, the travel speed can again be increased to reduce the scan time.

[0049] The imaging scan begins at step 103, at which point the scan head 2 moves from a starting position through the image scan path and images are obtained of a plurality of adjacent locations on the target 6. The image scan path is calculated so that the line scanner 3 moves at a first target speed within the no-sample region 26 and at a second target speed within the sample-containing region 24, the first target speed being faster than the second target speed. The spatial resolution of the line scan can be varied by increasing the speed of travel (also referred to as the "travel speed"). If the scan resolution within the sample-containing region 24 is set to 4000 x 4000 lpmm, by increasing the speed of travel, the scan resolution is changed to 4000 x y lpmm, where y can be a number between 4000 lpmm and 1000 lpmm or even lower. Since the no-sample region 26 typically contains only a clear slide, it is not important to scan this region at high resolution. Thus, if the scan resolution is selected to be 1000 lpm, the no-sample region 26 can be scanned at 4 times the speed of the sample-containing region 24. The starting position of the imaging scan is within the no-sample region 26, and thus the line scanner 3 initially moves at the first target speed. However, if a change in one or more image parameters indicative of a sample entering the field of view of the line scanner 3 within the no-sample region 26 is monitored, the scan speed is reduced to the second target speed. Thus, immediate detection of the sample can be made, and the sample is imaged at high resolution wherever it is present on the target 6.

[0050] When the pre-scan data indicates that the edge of the sample is being approached in the imaging scan, then the scan speed can be reduced so that the edge of the sample is captured at high resolution. This is illustrated in Figure 6 The uppermost graph shows, where the uppermost graph indicates two regions of the target 6: the blanked region to the left represents the sample-free zone 26, and the hatched region represents the sample-containing zone 24. Figure 6 The lowermost graph in shows the variation in scan speed as the line scanner 3 transitions from imaging the sample-free zone 26 to imaging the sample-containing zone 24. The image scan path is calculated so that the scan speed is gradually reduced from a first target speed to a second target speed before the line scanner 3 moves from imaging the sample-free zone 26 to imaging the sample-containing zone 24. In this example, the second target speed is half the first target speed. The first image of the sample-containing zone 24 is then obtained at the second target speed, without any sudden deceleration occurring within the zone. Similarly, once the pre-scan data indicates that the edge of the sample-containing zone 24 has been passed, the scan speed is increased back to the first target speed, as shown in Figure 7 This increase in speed occurs over a zone to reduce the rate of acceleration and produce a smooth transition between the different zones. There will be a maximum acceleration / deceleration and a maximum rate of change of acceleration. Taking the example where the target 6 is moving during the imaging scan, and the line scanner 3 remains stationary, to change the speed required to scan a 5kg stage and slide from 1000lpmm to that required for a 4000lpmm scan, it can only be necessary to change 3J, but to do this in a single line time would require 85W. Extending this over 10 lines, the same energy is required, but the power reduces to 8.5W.

[0051] Figure 6 and Figure 7 The vertical dashed lines in the uppermost graph of show the locations at which image capture is triggered. Throughout the imaging scan, the line scanner 3 acquires images at a constant rate. The spatial density of the sampling points (i.e. the locations at which image capture is triggered) therefore increases as the scan speed is reduced from the first target speed to the second target speed Figure 6 ), and then decreases as the scan speed is increased from the second target speed to the first target speed Figure 7 ). Without any interpolation, this variation in scan speed will result in a corresponding variation in image resolution between the sample-free zone 26 and the sample-containing zone 24, as will be discussed later with reference to Figure 9 .

[0052] In the first embodiment, the scan speed is kept constant at the second target speed throughout the sample-containing region 24. However, in other embodiments, the scan speed can be increased if one or more of the monitored image parameters indicate that the sample is not located within the sample-containing region 24. This can occur, for example, if the AOI detection algorithm incorrectly identifies a surface mark on the target 6 as being part of the sample. The depth of focus of the line scanner 3 is typically less than 3 microns, meaning that the surface mark is unlikely to be identified as the sample within the imaging scan. Alternatively, the scan speed can be increased from the second target speed only within the sample-containing region 24 if no sample is detected within a threshold distance (or equivalently, after a threshold number of lines have been acquired). The scan speed can then be increased from the second scan speed to the first scan speed, or alternatively to a value between these two values. For example, the increase in scan speed can be balanced by the probability that the sample is present in a given location or region as determined by the probability map. That is, if the pre-scan analysis gives a good certainty that the sample is present, then the scan speed will not be increased as it is expected that the sample will be imaged within the sample-containing region 24 and it is desirable to image it at the highest resolution once reached. However, if the probability map indicates a relatively low certainty that the sample is present at a given location within the region marked as the sample-containing region, then the scan speed can be increased by a scaling factor. For example, if the certainty that the sample is present at a given location is only 50%, then the scan speed will only be increased by 50% of the possible increase in scan speed.

[0053] In another embodiment, the scan speed is continuously adjusted during the imaging scan as a proportion of the certainty that the sample is found at a given location as determined by the probability map generated from the pre-scan data. Alternatively, if a threshold luminance of 95% is set to indicate a clear slide (which is set to a resolution of 1000 lpmm) and a second threshold luminance of 90% indicates the presence of a sample (which is set to a resolution of 4000 lpmm), then a linear ramp can be used to set the speed for luminance values between this range. For example, if the luminance is measured as 92.5%, then the scan speed can be set to produce a resolution of 2500 lpmm.

[0054] Returning to the first embodiment, once the line scanner 3 progresses from acquiring images of the sample-containing region 24 to again acquiring images of the sample-free region 26, the scan speed is increased to the first target speed (as shown in FIG. 1). If the one or more image parameters monitored indicating the presence of a sample within the field of view of the line scanner 3 change, then the scan speed is again decreased to the second target speed so that the sample is imaged at a suitably high resolution. In this way, if a portion of the sample is not detected by analysis of the pre-scan data, then a high resolution image of it is still acquired. Figure 7 Figure 8 ​An example of detecting a portion of a sample within the sample-free region 26 is shown. The top graph shows the scanning speed as the line scanner 3 moves along the image scanning path. Image parameters in the form of focus evaluation values ​​are calculated from each image acquired along the image scanning path. Figure 8 In the bottommost graph, the focus evaluation value for each image or "row" is marked with an "x" at corresponding positions along the image scan path. When the focus evaluation value is below a focus evaluation value threshold, the scan speed is maintained at a first target speed for continuous image capture. At position "A," the focus evaluation value increases above the threshold, indicating that a portion of the sample at that position (within sample-free region 26) is visible within the image. In response to this change in focus evaluation value, the scan speed is reduced from the first target speed to a second target speed. The line scanner 3 decelerates over the subsequent three rows until the second target speed is reached. The focus evaluation value is approximately constant across the portion of the image scan path extending from position A to position B, indicating that the sample extends through this region. The detection of samples within the so-called sample-free region demonstrates how analysis of pre-scan data alone can provide inaccurate results, and why it is therefore useful for acquiring images of both sample-free and sample-containing regions 26. Although the scanning speed is adjusted in real time in response to changes in monitored image parameters in the first embodiment, in an alternative embodiment, the scanning speed can be set based solely on the analysis of pre-scan data, wherein sampled areas are imaged at a slower rate than non-sampled areas.

[0055] Throughout the imaging scan, the estimated focus level can be calculated using techniques known in the art (e.g., those described in US9116035B2). With the sample within the field of view of the line scanner 3, the focus height of the line scanner 3 is adjusted between each xy point along the image scanning path in the direction of or to the estimated focus position. The target 6 is thus imaged in one or more strips, following a serpentine or raster image scanning path, until the scanned area has been completely imaged.

[0056] In step 104, the individual images acquired by the line scanner 3 are combined to form a high-resolution image of the target. It is desired that the viewing resolution of the output image be uniform for both high-speed and low-speed scanning areas. The resolution of the image in areas scanned at higher scanning speeds can be improved by using interpolation techniques (such as bicubic interpolation) or other image processing techniques known in the art (e.g., as described in Chapter 3 of Numerical Recipes in C: The Art of Scientific Computing; Cambridge University Press; 1992). Therefore, it is possible to scan at a lower resolution but generate an image that appears to be high-resolution. In step 104, the interpolation ratio is appropriately adjusted according to the scanning speed at which the images were acquired to maintain a constant resolution in the output high-resolution image. This will be referred to below. Figure 9 Further discussion.

[0057] Figure 9 It shows Figure 6 The process, however, in this case, the top graph shows the amount of light incident on the line scanner 3. The light level is initially high (approximately 0.9 on the normalized scale shown) and is roughly constant in the region where no sample is found on target 6. Then, as a sample enters the field of view of the line scanner 3 (within sample-containing region 24), the light level drops below 0.1. The light level is monitored from the images captured along the image scan path and the positions of the sampling points marked with circles. The spatial interval between adjacent sampling points is approximately twice that in the region marked "C" (within sample-free region 26) as it is in the region marked "E" (within sample-containing region 24). The region marked "D" shows the area where the scan speed decreases between the first target speed and the second target speed. In the execution of step 104, the image data from the line scanner 3 is interpolated onto a nominal HR grid, such that image data is generated at a constant resolution along the image scan path. In the top graph, the interpolated data is marked with an asterisk. As can be seen, the image data is interpolated between adjacent sampling points in regions C and D. The high-resolution images generated by the imaging scan can then be displayed to the user by the computing system 10 for sample analysis.

[0058] Therefore, an improved method for operating a microscope scanner is provided, which ensures high-resolution imaging of the entire sample without excessively increasing the total scanning time.

Claims

1. A method for operating a microscope scanner comprising a line scanner, the method comprising: generating pre-scan data for a target comprising a sample, wherein generating the pre-scan data comprises: obtaining a pre-scan image of the target from a pre-scan; and identifying one or more sample-containing regions and one or more sample-free regions of the target from the pre-scan image, wherein the one or more sample-containing regions correspond to locations on the target where the sample is expected to be present, and wherein the one or more sample-free regions correspond to locations on the target where the sample is expected to be absent; and performing an imaging scan of a scan region of the target, the scan region comprising the one or more sample-containing regions and the one or more sample-free regions, wherein performing the imaging scan comprises: moving the line scanner relative to the target along an image scan path and acquiring images of the target at each of a plurality of locations along the image scan path using the line scanner; adjusting a scan speed of the line scanner along the image scan path based on the pre-scan data such that the target is imaged at a higher scan speed within the one or more sample-free regions than in the one or more sample-containing regions, monitoring one or more image parameters from the images of the target acquired along the image scan path; and in response to a change in the one or more image parameters, further adjusting the scan speed of the line scanner within the one or more sample-free regions, wherein the one or more image parameters comprise a focus merit value, wherein more accurately focused images have a corresponding higher merit value, and wherein the scan speed of the line scanner is decreased in response to an increase in the focus merit value.

2. The method of claim 1, wherein, adjusting the scan speed based on the pre-scan data comprises imaging the target at a first target speed within the one or more sample-free regions and imaging the target at a second target speed within the one or more sample-containing regions, wherein the first target speed is higher than the second target speed; and wherein the imaging scan further comprises decreasing the scan speed from the first target speed to the second target speed or towards the second target speed before the line scanner transitions from acquiring images of the sample-free regions to acquiring images of the sample-containing regions.

3. The method of claim 1, wherein, monitoring the one or more image parameters to detect the presence of the sample within the one or more sample-free regions.

4. The method of claim 3, wherein the scan speed is further adjusted such that locations on the target where the sample is detected are imaged at a lower scan speed than locations on the target where the sample is not detected.

5. The method of any one of claims 1, 3-4, wherein, adjusting the scan speed based on the pre-scan data comprises imaging the target at a first target speed within the one or more sample-free regions and imaging the target at a second target speed within the one or more sample-containing regions, wherein the first target speed is higher than the second target speed; and wherein further adjusting the scan speed of the line scanner within the one or more sample-free zones comprises decreasing the scan speed from the first target speed in response to a change in the one or more image parameters indicative of detection of the sample.

6. The method of any one of claims 1, 3-4, wherein the imaging scan further comprises: increasing the scan speed within the one or more sample-containing zones in response to a change in the one or more image parameters indicative of an absence of the sample in the acquired image.

7. The method of any one of claims 1, 3-4, wherein, Further adjusting the scan speed comprises decreasing the scan speed of the line scanner in response to the one or more image parameters changing in a first predetermined manner relative to a threshold value.

8. The method of claim 7, wherein, The one or more image parameters change in the first predetermined manner relative to the threshold value in response to an image of the sample being acquired within the sample-free zone.

9. The method of claim 7, wherein, Further adjusting the scan speed comprises increasing the scan speed of the line scanner in response to the one or more image parameters changing in a second predetermined manner relative to the threshold value.

10. The method of any one of claims 1, 3-4, wherein, The one or more image parameters comprise a brightness of the image, and wherein the scan speed of the line scanner is decreased in response to a decrease in the brightness of the image.

11. The method of any one of claims 1, 3-4, wherein, The imaging scan further comprises calculating the focus merit value at each of the locations along the image scan path, and adjusting a focus height of the line scanner along the image scan path based on the focus merit value.

12. The method of any one of claims 1, 3-4, wherein, The images are acquired in time at a constant rate during the imaging scan.

13. The method of any one of claims 1, 3-4, wherein, The line scanner moves through the sample-containing zones at a constant speed during the imaging scan.

14. The method of any one of claims 1, 3-4, wherein, The line scanner moves continuously along the image scan path relative to the target during the imaging scan.

15. The method of any one of claims 1, 3-4, wherein, The focus depth of the pre-scan exceeds 500 microns, and wherein a focus depth of each image acquired along the image scan path is less than 3 microns.

16. The method of any one of claims 1, 3-4, wherein, Generating pre-scan data further comprises performing image analysis of the pre-scan images to produce a probability map indicative of a likelihood of the sample being present at each location on the target, wherein the one or more sample-containing zones and the one or more sample-free zones are selected from the probability map.

17. The method of claim 16, wherein adjusting a scan speed of the line scanner along the image scan path based on the pre-scan data comprises: The scan speed is adjusted according to a likelihood of the sample being present at a given location on the target as determined from the probability map.

18. The method of claim 16, wherein, The one or more sample-free zones correspond to locations from the probability map having a low likelihood of containing the sample, wherein the one or more sample-containing zones correspond to locations from the probability map having a high likelihood of containing the sample, and wherein generating pre-scan data further comprises identifying one or more uncertain zones on the target from the probability map having a likelihood of containing the sample between the likelihood of the sample-free zones and the likelihood of the sample-containing zones.

19. The method of claim 18, wherein, Adjusting the scan speed based on the pre-scan data comprises imaging the target at a first target speed in the one or more sample-free zones, imaging the target at a second target speed in the one or more sample-containing zones, and imaging the target at a third target speed in the one or more uncertain zones, wherein the third target speed is higher than the second target speed and lower than the first target speed.

20. The method of any one of claims 1, 3-4, further comprising generating a high resolution image of the target from a combination of images acquired during the imaging scan, wherein generating the high resolution image comprises using image interpolation to increase a resolution of an image acquired at a higher scan speed to correspond to a resolution of an image acquired at a lower scan speed.

21. The method of claim 20, wherein generating the high resolution image comprises adjusting an interpolation ratio to maintain a constant resolution in the high resolution image.

22. The method of any one of claims 1, 3-4, wherein, performing the imaging scan in a transmission mode for the microscope scanner.

23. The method of claim 22, wherein the microscope scanner is a brightfield microscope.

24. The method of any one of claims 1, 3-4, further comprising selecting the scan zones from the pre-scan image so as to cover an entire target.

25. The method of claim 24, wherein, Covering an entire target excludes any opaque regions of the target.

26. The method of claim 1, wherein, The pre-scan image is obtained from a single pass of a pre-scan camera along a pre-scan imaging path.

27. The method of claim 1, wherein, The focus merit value is a numerical measure of complexity within an image.

28. A computer program product containing instructions that, when executed using a microscope scanner comprising a line scanner and a target comprising a sample, cause the microscope scanner to perform the method of any one of the preceding claims.

Images