Computer-assisted mitotic count using a deep learning–based algorithm improves interobserver reproducibility and accuracy

Bertram CA, Aubreville M, Donovan TA, Bartel A, Wilm F, Marzahl C, Assenmacher CA, Becker K, Bennett M, Corner S, Cossic B, Denk D, Dettwiler M, Gonzalez BG, Gurtner C, Haverkamp AK, Heier A, Lehmbecker A, Merz S, Noland EL, Plog S, Schmidt A, Sebastian F, Sledge DG, Smedley RC, Tecilla M, Thaiwong T, Fuchs-Baumgartinger A, Meuten DJ, Breininger K, Kiupel M, Maier A, Klopfleisch R (2022)

Publication Type: Journal article

Publication year: 2022

Journal

Book Volume: 59

Pages Range: 211-226

Journal Issue: 2

DOI: 10.1177/03009858211067478

Abstract

The mitotic count (MC) is an important histological parameter for prognostication of malignant neoplasms. However, it has inter- and intraobserver discrepancies due to difficulties in selecting the region of interest (MC-ROI) and in identifying or classifying mitotic figures (MFs). Recent progress in the field of artificial intelligence has allowed the development of high-performance algorithms that may improve standardization of the MC. As algorithmic predictions are not flawless, computer-assisted review by pathologists may ensure reliability. In the present study, we compared partial (MC-ROI preselection) and full (additional visualization of MF candidates and display of algorithmic confidence values) computer-assisted MC analysis to the routine (unaided) MC analysis by 23 pathologists for whole-slide images of 50 canine cutaneous mast cell tumors (ccMCTs). Algorithmic predictions aimed to assist pathologists in detecting mitotic hotspot locations, reducing omission of MFs, and improving classification against imposters. The interobserver consistency for the MC significantly increased with computer assistance (interobserver correlation coefficient, ICC = 0.92) compared to the unaided approach (ICC = 0.70). Classification into prognostic stratifications had a higher accuracy with computer assistance. The algorithmically preselected hotspot MC-ROIs had a consistently higher MCs than the manually selected MC-ROIs. Compared to a ground truth (developed with immunohistochemistry for phosphohistone H3), pathologist performance in detecting individual MF was augmented when using computer assistance (F1-score of 0.68 increased to 0.79) with a reduction in false negatives by 38%. The results of this study demonstrate that computer assistance may lead to more reproducible and accurate MCs in ccMCTs.

Authors with CRIS profile

Involved external institutions

Universität Bern Switzerland (CH) Technische Hochschule Ingolstadt Germany (DE) Stiftung Tierärztliche Hochschule Hannover (TiHo) Germany (DE) INDEXX Germany (DE) Michigan State University United States (USA) (US) Roche Pharma Ag Switzerland (CH) Freie Universität Berlin Germany (DE) North Carolina State University United States (USA) (US) Ludwig-Maximilians-Universität (LMU) Germany (DE) SYNLAB UK & Ireland United Kingdom (GB) Animal Medical Center (AMC) United States (USA) (US) University of Pennsylvania United States (USA) (US) Idorsia Pharmaceuticals Ltd Switzerland (CH)

How to cite

APA:

Bertram, C.A., Aubreville, M., Donovan, T.A., Bartel, A., Wilm, F., Marzahl, C.,... Klopfleisch, R. (2022). Computer-assisted mitotic count using a deep learning–based algorithm improves interobserver reproducibility and accuracy. Veterinary Pathology, 59(2), 211-226. https://dx.doi.org/10.1177/03009858211067478

MLA:

Bertram, Christof A., et al. "Computer-assisted mitotic count using a deep learning–based algorithm improves interobserver reproducibility and accuracy." Veterinary Pathology 59.2 (2022): 211-226.

BibTeX: Download