Pathologist Computer-Aided Diagnostic Scoring of Tumor Cell Fraction: A Swiss National Study

Ana Leni Frei; Raphaël Oberson; Elias Baumann; Aurel Perren; Rainer Grobholz; Alessandro Lugli; Heather Dawson; Christian Abbet; Ibai Lertxundi; Stefan Reinhard; Aart Mookhoek; Johann Feichtinger; Rossella Sarro; Gallus Gadient; Corina Dommann-Scherrer; Jessica Barizzi; Sabina Berezowska; Katharina Glatz; Susanne Dertinger; Yara Banz; Rene Schoenegg; Laura Rubbia-Brandt; Achim Fleischmann; Guenter Saile; Pierre Mainil-Varlet; Ruggero Biral; Luca Giudici; Alex Soltermann; Audrey Baur Chaubert; Sylvia Stadlmann; Joachim Diebold; Kristof Egervari; Charles Bénière; Francesca Saro; Andrew Janowczyk; Inti Zlobec

doi:10.1016/j.modpat.2023.100335

Pathologist Computer-Aided Diagnostic Scoring of Tumor Cell Fraction: A Swiss National Study

Mod Pathol. 2023 Dec;36(12):100335. doi: 10.1016/j.modpat.2023.100335. Epub 2023 Sep 22.

Authors

Ana Leni Frei¹, Raphaël Oberson², Elias Baumann², Aurel Perren², Rainer Grobholz³, Alessandro Lugli², Heather Dawson², Christian Abbet⁴, Ibai Lertxundi², Stefan Reinhard², Aart Mookhoek², Johann Feichtinger⁵, Rossella Sarro⁶, Gallus Gadient⁷, Corina Dommann-Scherrer⁸, Jessica Barizzi⁶, Sabina Berezowska⁹, Katharina Glatz¹⁰, Susanne Dertinger¹¹, Yara Banz², Rene Schoenegg¹², Laura Rubbia-Brandt¹³, Achim Fleischmann¹⁴, Guenter Saile¹⁵, Pierre Mainil-Varlet¹⁶, Ruggero Biral¹⁷, Luca Giudici⁶, Alex Soltermann¹⁷, Audrey Baur Chaubert¹⁸, Sylvia Stadlmann¹⁹, Joachim Diebold²⁰, Kristof Egervari¹³, Charles Bénière²¹, Francesca Saro²², Andrew Janowczyk²³, Inti Zlobec²⁴

Affiliations

¹ Institute for Tissue Medicine and Pathology, University of Bern, Bern, Switzerland. Electronic address: [email protected].
² Institute for Tissue Medicine and Pathology, University of Bern, Bern, Switzerland.
³ Medical Faculty University of Zurich, Institute of Pathology, Cantonal Hospital Aarau, Aarau, Switzerland.
⁴ Signal Processing Laboratory 5, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland.
⁵ Labor Dr. Ulm GmbH, Wien, Austria.
⁶ Istituto Cantonale di Patologia, Ente ospedaliero cantonale (EOC), Locarno, Switzerland.
⁷ Institute of Pathology, Medica, Zürich, Switzerland.
⁸ Institute of Pathology, Cantonal Hospital Winterthur, Winterthur, Switzerland.
⁹ Institute of Pathology, Lausanne University Hospital, Lausanne, Switzerland.
¹⁰ Institut of Pathology, University Hospital Basel, Basel, Switzerland.
¹¹ Institute of Pathology, Landeskrankenhaus Feldkirch, Feldkirch, Austria.
¹² Institute of Pathology, Cantonal Hospital St. Gallen, St. Gallen, Switzerland.
¹³ Department of Pathology and Immunology, Geneva University Hospital, Genève, Switzerland.
¹⁴ Institute of Pathology, Cantonal Hospital Thurgau, Münsterlingen, Switzerland.
¹⁵ Labor team w AG, St. Gallen, Switzerland.
¹⁶ Unilabs Bern, Bern, Switzerland.
¹⁷ Pathologie Länggasse, Ittigen, Switzerland.
¹⁸ FMH Pathology, Pathology Department of SYNLAB Switzerland SA, Lausanne, Switzerland.
¹⁹ Institute of Pathology, Cantonal Hospital Baden, Baden, Switzerland.
²⁰ Institute of Pathology, Cantonal Hospital Luzern, Luzern, Switzerland.
²¹ Aurigen, Centre de Pathologie, Lausanne Switzerland.
²² Institute of Pathology and Molecular Pathology, University Hospital Zürich, Zürich, Switzerland.
²³ Department of Biomedical Engineering, Emory University, Atlanta, Georgia; Department of Oncology, Division of Precision Oncology, University Hospital of Geneva, Geneva, Switzerland; Department of Clinical Pathology, Division of Clinical Pathology, University Hospital of Geneva, Geneva, Switzerland.
²⁴ Institute for Tissue Medicine and Pathology, University of Bern, Bern, Switzerland. Electronic address: [email protected].

PMID: 37742926
DOI: 10.1016/j.modpat.2023.100335

Abstract

Tumor cell fraction (TCF) estimation is a common clinical task with well-established large interobserver variability. It thus provides an ideal test bed to evaluate potential impacts of employing a tumor cell fraction computer-aided diagnostic (TCFCAD) tool to support pathologists' evaluation. During a National Slide Seminar event, pathologists (n = 69) were asked to visually estimate TCF in 10 regions of interest (ROIs) from hematoxylin and eosin colorectal cancer images intentionally curated for diverse tissue compositions, cellularity, and stain intensities. Next, they re-evaluated the same ROIs while being provided a TCFCAD-created overlay highlighting predicted tumor vs nontumor cells, together with the corresponding TCF percentage. Participants also reported confidence levels in their assessments using a 5-tier scale, indicating no confidence to high confidence, respectively. The TCF ground truth (GT) was defined by manual cell-counting by experts. When assisted, interobserver variability significantly decreased, showing estimates converging to the GT. This improvement remained even when TCFCAD predictions deviated slightly from the GT. The standard deviation (SD) of the estimated TCF to the GT across ROIs was 9.9% vs 5.8% with TCFCAD (P < .0001). The intraclass correlation coefficient increased from 0.8 to 0.93 (95% CI, 0.65-0.93 vs 0.86-0.98), and pathologists stated feeling more confident when aided (3.67 ± 0.81 vs 4.17 ± 0.82 with the computer-aided diagnostic [CAD] tool). TCFCAD estimation support demonstrated improved scoring accuracy, interpathologist agreement, and scoring confidence. Interestingly, pathologists also expressed more willingness to use such a CAD tool at the end of the survey, highlighting the importance of training/education to increase adoption of CAD systems.

Keywords: artificial intelligence; computer-aided diagnostic tool; digital pathology; interobserver variability; pathology; tumor cell fraction.

MeSH terms

Computers*
Humans
Pathologists*
Switzerland