A large-scale dataset for mitotic figure assessment on whole slide images of canine cutaneous mast cell tumor

Bertram, Christof A.; Aubreville, Marc; Marzahl, Christian; Maier, Andreas; Klopfleisch, Robert

doi:10.1038/s41597-019-0290-4

Cited by 53 publications

(99 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While this assumption is part of many MC protocols, there is no work known to the authors that provides a general proof for this, especially for a greater variance of tumor types. In fact, for the case of CCMCT, a previous analysis 8 showed no such relationship on the data set used in this work 26 . Also Ki67 hot spots had been found in the periphery in only some cases of human breast cancer 12 .…”

Section: Related Workmentioning

confidence: 52%

“…Our research group built a data set consisting of 32 CCMCT cases, where all mitotic figures have been annotated within the entire tumor area. The data set, including a detailed description of its creation and anonymized forms of all WSI and all annotations, is available publicly 26 . All tissue samples were taken retrospectively from the diagnostic archive of the authors with approval by the local governmental authorities (State Office of Health and Social Affairs of Berlin, Germany, approval ID: StN 011/20), and were originally provided from routine diagnostic service for purely diagnostic and therapeutic reasons, i.e.…”

Section: Methodsmentioning

confidence: 99%

“…The model with the least mean squared error in prediction was chosen for the regression approach. For the object detection-based approach, we used a modestly differing training scheme, which was already described in a previous publication 26 .…”

Section: Regression Of the Mitotic Figure Count Within An Image Patchmentioning

confidence: 99%

See 2 more Smart Citations

Deep learning algorithms out-perform veterinary pathologists in detecting the mitotically most active tumor region

Aubreville

Bertram

Marzahl

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

Manual count of mitotic figures, which is determined in the tumor region with the highest mitotic activity, is a key parameter of most tumor grading schemes. It can be, however, strongly dependent on the area selection due to uneven mitotic figure distribution in the tumor section. We aimed to assess the question, how significantly the area selection could impact the mitotic count, which has a known high inter-rater disagreement. On a data set of 32 whole slide images of H&E-stained canine cutaneous mast cell tumor, fully annotated for mitotic figures, we asked eight veterinary pathologists (five board-certified, three in training) to select a field of interest for the mitotic count. To assess the potential difference on the mitotic count, we compared the mitotic count of the selected regions to the overall distribution on the slide. Additionally, we evaluated three deep learning-based methods for the assessment of highest mitotic density: In one approach, the model would directly try to predict the mitotic count for the presented image patches as a regression task. The second method aims at deriving a segmentation mask for mitotic figures, which is then used to obtain a mitotic density. Finally, we evaluated a two-stage object-detection pipeline based on state-of-the-art architectures to identify individual mitotic figures. We found that the predictions by all models were, on average, better than those of the experts. The two-stage object detector performed best and outperformed most of the human pathologists on the majority of tumor cases. The correlation between the predicted and the ground truth mitotic count was also best for this approach (0.963–0.979). Further, we found considerable differences in position selection between pathologists, which could partially explain the high variance that has been reported for the manual mitotic count. To achieve better inter-rater agreement, we propose to use a computer-based area selection for support of the pathologist in the manual mitotic count.

show abstract

Section: Related Workmentioning

confidence: 52%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Deep learning algorithms out-perform veterinary pathologists in detecting the mitotically most active tumor region

Aubreville

Bertram

Marzahl

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…As of today, few datasets of human mammary carcinoma are available. They provide labels only in small image sections of the whole slide image (WSI) and include up to 1,552 MF annotations [1].…”

mentioning

confidence: 99%

“…Our research group has developed a large-scale, open access dataset with annotations for MF in 32 cases of canine cutaneous mast cell tumors [1]. Entire WSI were completely labeled by two pathologists resulting in 44,800 MF annotations.…”

mentioning

confidence: 99%