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

Autor:	Bertram CA; University of Veterinary Medicine, Vienna, Austria.; Freie Universität Berlin, Berlin, Germany., Aubreville M; Technische Hochschule Ingolstadt, Ingolstadt, Germany., Donovan TA; Animal Medical Center, New York, NY, USA., Bartel A; Freie Universität Berlin, Berlin, Germany., Wilm F; Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany., Marzahl C; Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany., Assenmacher CA; University of Pennsylvania, Philadelphia, PA, USA., Becker K; University of Veterinary Medicine, Hannover, Germany., Bennett M; Synlab's VPG Histology, Bristol, UK., Corner S; Michigan State University, Lansing, MI, USA., Cossic B; Idorsia Pharmaceuticals Ltd, Allschwil, Switzerland., Denk D; Ludwig Maximilians University, Munich, Germany., Dettwiler M; University of Bern, Bern, Switzerland., Gonzalez BG; Synlab's VPG Histology, Bristol, UK., Gurtner C; University of Bern, Bern, Switzerland., Haverkamp AK; University of Veterinary Medicine, Hannover, Germany., Heier A; IDEXX Vet Med Labor GmbH, Kornwestheim, Germany., Lehmbecker A; IDEXX Vet Med Labor GmbH, Kornwestheim, Germany., Merz S; IDEXX Vet Med Labor GmbH, Kornwestheim, Germany., Noland EL; Michigan State University, Lansing, MI, USA., Plog S; Synlab's VPG Histology, Bristol, UK., Schmidt A; IDEXX Vet Med Labor GmbH, Kornwestheim, Germany., Sebastian F; IDEXX Vet Med Labor GmbH, Kornwestheim, Germany., Sledge DG; Michigan State University, Lansing, MI, USA., Smedley RC; Michigan State University, Lansing, MI, USA., Tecilla M; Roche Pharmaceutical Research and Early Development (pRED), Basel, Switzerland., Thaiwong T; Michigan State University, Lansing, MI, USA., Fuchs-Baumgartinger A; University of Veterinary Medicine, Vienna, Austria., Meuten DJ; North Carolina State University, Raleigh, NC, USA., Breininger K; Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany., Kiupel M; Michigan State University, Lansing, MI, USA., Maier A; Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany., Klopfleisch R; Freie Universität Berlin, Berlin, Germany.
Jazyk:	angličtina
Zdroj:	Veterinary pathology [Vet Pathol] 2022 Mar; Vol. 59 (2), pp. 211-226. Date of Electronic Publication: 2021 Dec 30.
DOI:	10.1177/03009858211067478
Abstrakt:	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.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu