Abstract

A model’s ability to express its own predictive uncertainty is an essential attribute for maintaining clinical user confidence as computational biomarkers are deployed into real-world medical settings. In the domain of cancer digital histopathology, we describe a clinically-oriented approach to uncertainty quantification for whole-slide images, estimating uncertainty using dropout and calculating thresholds on training data to establish cutoffs for low- and high-confidence predictions. We train models to identify lung adenocarcinoma vs. squamous cell carcinoma and show that high-confidence predictions outperform predictions without uncertainty, in both cross-validation and testing on two large external datasets spanning multiple institutions. Our testing strategy closely approximates real-world application, with predictions generated on unsupervised, unannotated slides using predetermined thresholds. Furthermore, we show that uncertainty thresholding remains reliable in the setting of domain shift, with accurate high-confidence predictions of adenocarcinoma vs. squamous cell carcinoma for out-of-distribution, non-lung cancer cohorts.

Safe clinical deployment of deep learning models for digital pathology requires reliable estimates of predictive uncertainty. Here the authors describe an algorithm for quantifying whole-slide image uncertainty, demonstrating their approach with models trained to distinguish lung cancer subtypes.

Details

Title
Uncertainty-informed deep learning models enable high-confidence predictions for digital histopathology
Author
Dolezal, James M. 1   VIAFID ORCID Logo  ; Srisuwananukorn, Andrew 2   VIAFID ORCID Logo  ; Karpeyev, Dmitry 3 ; Ramesh, Siddhi 1 ; Kochanny, Sara 1 ; Cody, Brittany 4 ; Mansfield, Aaron S. 5   VIAFID ORCID Logo  ; Rakshit, Sagar 5 ; Bansal, Radhika 5   VIAFID ORCID Logo  ; Bois, Melanie C. 6 ; Bungum, Aaron O. 7 ; Schulte, Jefree J. 8   VIAFID ORCID Logo  ; Vokes, Everett E. 1 ; Garassino, Marina Chiara 1 ; Husain, Aliya N. 4 ; Pearson, Alexander T. 1   VIAFID ORCID Logo 

 University of Chicago Medical Center, Section of Hematology/Oncology, Department of Medicine, Chicago, USA (GRID:grid.412578.d) (ISNI:0000 0000 8736 9513) 
 Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, USA (GRID:grid.59734.3c) (ISNI:0000 0001 0670 2351) 
 DV Group, LLC, Chicago, USA (GRID:grid.59734.3c) 
 University of Chicago, Department of Pathology, Chicago, USA (GRID:grid.170205.1) (ISNI:0000 0004 1936 7822) 
 Mayo Clinic, Division of Medical Oncology, Rochester, USA (GRID:grid.66875.3a) (ISNI:0000 0004 0459 167X) 
 Mayo Clinic, Department of Laboratory Medicine and Pathology, Rochester, USA (GRID:grid.66875.3a) (ISNI:0000 0004 0459 167X) 
 Mayo Clinic, Divisions of Pulmonary Medicine and Critical Care, Rochester, USA (GRID:grid.66875.3a) (ISNI:0000 0004 0459 167X) 
 University of Wisconsin at Madison, Department of Pathology and Laboratory Medicine, Madison, USA (GRID:grid.14003.36) (ISNI:0000 0001 2167 3675) 
Publication year
2022
Publication date
2022
Publisher
Nature Publishing Group
e-ISSN
20411723
Source type
Scholarly Journal
Language of publication
English
ProQuest document ID
2731308176
Copyright
© The Author(s) 2022. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.