Scanning/pre-scanning quality control of slides
A method of analyzing a plurality of histology slides is provided, wherein possible artefacts are detected at an early stage. This is achieved by including a preliminary imaging and image analysis step in the process flow; the step being executed preferably before any diagnostic assessment e.g. image analysis or manual reading is performed by a pathologist or a lab technician. Accordingly, the histology slides reaching the expert for image analysis are of a higher quality, since the slides are ideally free of artefacts. The method thus saves valuable time for the pathologist, since only artefact-free slides will be subject to a detailed analysis. Furthermore, the method minimizes the risk of misinterpretations leading to potentially false diagnoses. A deep learning model, is also disclosed, which is capable of automatically determining whether histopathological images are suitable for diagnostic and/or research assessment. A training of the deep learning model is also disclosed.
1. A method for preparing a plurality of histology slide images for analysis in an inline process, said method comprising:
a) providing a histology slide with a tissue specimen,
b) obtaining an image of at least a part of the tissue specimen on the slide using a first magnification;
c) based on said image, determining whether the slide with the tissue comprises one or more artefact(s) by using a deep learning model trained to recognize the one or more artefact(s), the one or more artefact(s) being an artificial structure or a tissue alteration on the slide as a result of an extraneous factor;
d) based on the determination in c) and further based of a position and/or size of the one or more artefact(s), removing the slide from said inline process if the slide comprises a rejectable artefact, or if no rejectable artefacts are determined, classifying the slide as an accepted slide;
e) obtaining a histology slide image of at least a part of the tissue specimen on the accepted slide using a second magnification, the second magnification being higher than the first magnification, and optionally obtaining (a) further histology slide image(s) of at least a part of the tissue specimen on the accepted slide using further higher magnification(s); and
f) repeating steps a)-e) for each histology slide, thereby preparing the plurality of histology slide images in the inline process.
2. The method according to claim 1 , wherein the removal of a rejectable slide is performed automatically.
3. The method according to claim 1 , wherein the one or more artefacts are selected from the group consisting of dust, tissue folds, glue, broken glass, stain spots, erroneous cutting thickness, missing tissue, holes in tissue, erroneous pen/marker spots.
4. The method according to claim 1 , wherein the deep learning model is trained to recognize different artefact types, such as tissue folds, glue, or holes in the tissue.
5. The method according to claim 1 , wherein the first magnification is below 5×, or less than 2.5×, or around or below 1×, or from about 0.1×to about 0.5X.
6. A system for analyzing a plurality of histology slides, comprising a computer-readable storage device for storing instructions that, when executed by a processor, performs the method according to claim 1 .
7. A method for analyzing a plurality of histology slides in an inline process, said method comprising:
a) providing a histology slide with a tissue specimen;
b) obtaining an image of at least a part of the tissue specimen on the slide using a first magnification;
c) based on said image, determining whether the slide with the tissue comprises one or more artefact(s) by using a deep learning model trained to recognize the one or more artefact(s), the one or more artefact(s) being an artificial structure or a tissue alteration on the slide as a result of an extraneous factor;
d) based on the determination in c) and further based of a position and/or size of the one or more artefact(s), removing the slide from said inline process if the slide comprises a rejectable artefact, or if no rejectable artefacts are determined classifying the slide as an accepted slide;
e) obtaining a histology slide image of at least a part of the tissue specimen on the accepted slide using a second magnification, the second magnification being higher than the first magnification, and optionally obtaining (a) further histology slide image(s) of at least a part of the tissue specimen on the accepted slide using further higher magnification(s);
f) performing an image analysis of at least a part of the histology slide images obtained from the accepted slides; and
g) repeating steps a)-f) for each histology slide, thereby analyzing the plurality of the histology slides in the inline process.
8. The method according to claim 7 , wherein the steps a)-e) are performed for all histology slides before step f) is performed.