Liron Pantanowitz1, Gabriela M Quiroga-Garza2, Lilach Bien3, Ronen Heled3, Daphna Laifenfeld3, Chaim Linhart3, Judith Sandbank4, Anat Albrecht Shach5, Varda Shalev6, Manuela Vecsler3, Pamela Michelow7, Scott Hazelhurst8, Rajiv Dhir2. 1. Department of Pathology, University of Pittsburgh Medical Center, Pittsburgh, PA, USA; Department of Anatomical Pathology, University of the Witwatersrand and National Health Laboratory Services, Johannesburg, South Africa. Electronic address: lironp@med.umich.edu. 2. Department of Pathology, University of Pittsburgh Medical Center, Pittsburgh, PA, USA. 3. Ibex Medical Analytics, Tel Aviv, Israel. 4. Ibex Medical Analytics, Tel Aviv, Israel; Institute of Pathology, Maccabi Healthcare Services, Rehovot, Israel. 5. Shamir Medical Center, Beer Yaakov, Israel. 6. KSM Research and Innovation institute, Maccabi Healthcare Services, Tel Aviv, Israel. 7. Department of Anatomical Pathology, University of the Witwatersrand and National Health Laboratory Services, Johannesburg, South Africa. 8. School of Electrical & Information Engineering and Sydney Brenner Institute for Molecular Bioscience, University of the Witwatersrand, Johannesburg, South Africa.
Abstract
BACKGROUND: There is high demand to develop computer-assisted diagnostic tools to evaluate prostate core needle biopsies (CNBs), but little clinical validation and a lack of clinical deployment of such tools. We report here on a blinded clinical validation study and deployment of an artificial intelligence (AI)-based algorithm in a pathology laboratory for routine clinical use to aid prostate diagnosis. METHODS: An AI-based algorithm was developed using haematoxylin and eosin (H&E)-stained slides of prostate CNBs digitised with a Philips scanner, which were divided into training (1 357 480 image patches from 549 H&E-stained slides) and internal test (2501 H&E-stained slides) datasets. The algorithm provided slide-level scores for probability of cancer, Gleason score 7-10 (vs Gleason score 6 or atypical small acinar proliferation [ASAP]), Gleason pattern 5, and perineural invasion and calculation of cancer percentage present in CNB material. The algorithm was subsequently validated on an external dataset of 100 consecutive cases (1627 H&E-stained slides) digitised on an Aperio AT2 scanner. In addition, the AI tool was implemented in a pathology laboratory within routine clinical workflow as a second read system to review all prostate CNBs. Algorithm performance was assessed with area under the receiver operating characteristic curve (AUC), specificity, and sensitivity, as well as Pearson's correlation coefficient (Pearson's r) for cancer percentage. FINDINGS: The algorithm achieved an AUC of 0·997 (95% CI 0·995 to 0·998) for cancer detection in the internal test set and 0·991 (0·979 to 1·00) in the external validation set. The AUC for distinguishing between a low-grade (Gleason score 6 or ASAP) and high-grade (Gleason score 7-10) cancer diagnosis was 0·941 (0·905 to 0·977) and the AUC for detecting Gleason pattern 5 was 0·971 (0·943 to 0·998) in the external validation set. Cancer percentage calculated by pathologists and the algorithm showed good agreement (r=0·882, 95% CI 0·834 to 0·915; p<0·0001) with a mean bias of -4·14% (-6·36 to -1·91). The algorithm achieved an AUC of 0·957 (0·930 to 0·985) for perineural invasion. In routine practice, the algorithm was used to assess 11 429 H&E-stained slides pertaining to 941 cases leading to 90 Gleason score 7-10 alerts and 560 cancer alerts. 51 (9%) cancer alerts led to additional cuts or stains being ordered, two (4%) of which led to a third opinion request. We report on the first case of missed cancer that was detected by the algorithm. INTERPRETATION: This study reports the successful development, external clinical validation, and deployment in clinical practice of an AI-based algorithm to accurately detect, grade, and evaluate clinically relevant findings in digitised slides of prostate CNBs. FUNDING: Ibex Medical Analytics.
BACKGROUND: There is high demand to develop computer-assisted diagnostic tools to evaluate prostate core needle biopsies (CNBs), but little clinical validation and a lack of clinical deployment of such tools. We report here on a blinded clinical validation study and deployment of an artificial intelligence (AI)-based algorithm in a pathology laboratory for routine clinical use to aid prostate diagnosis. METHODS: An AI-based algorithm was developed using haematoxylin and eosin (H&E)-stained slides of prostate CNBs digitised with a Philips scanner, which were divided into training (1 357 480 image patches from 549 H&E-stained slides) and internal test (2501 H&E-stained slides) datasets. The algorithm provided slide-level scores for probability of cancer, Gleason score 7-10 (vs Gleason score 6 or atypical small acinar proliferation [ASAP]), Gleason pattern 5, and perineural invasion and calculation of cancer percentage present in CNB material. The algorithm was subsequently validated on an external dataset of 100 consecutive cases (1627 H&E-stained slides) digitised on an Aperio AT2 scanner. In addition, the AI tool was implemented in a pathology laboratory within routine clinical workflow as a second read system to review all prostate CNBs. Algorithm performance was assessed with area under the receiver operating characteristic curve (AUC), specificity, and sensitivity, as well as Pearson's correlation coefficient (Pearson's r) for cancer percentage. FINDINGS: The algorithm achieved an AUC of 0·997 (95% CI 0·995 to 0·998) for cancer detection in the internal test set and 0·991 (0·979 to 1·00) in the external validation set. The AUC for distinguishing between a low-grade (Gleason score 6 or ASAP) and high-grade (Gleason score 7-10) cancer diagnosis was 0·941 (0·905 to 0·977) and the AUC for detecting Gleason pattern 5 was 0·971 (0·943 to 0·998) in the external validation set. Cancer percentage calculated by pathologists and the algorithm showed good agreement (r=0·882, 95% CI 0·834 to 0·915; p<0·0001) with a mean bias of -4·14% (-6·36 to -1·91). The algorithm achieved an AUC of 0·957 (0·930 to 0·985) for perineural invasion. In routine practice, the algorithm was used to assess 11 429 H&E-stained slides pertaining to 941 cases leading to 90 Gleason score 7-10 alerts and 560 cancer alerts. 51 (9%) cancer alerts led to additional cuts or stains being ordered, two (4%) of which led to a third opinion request. We report on the first case of missed cancer that was detected by the algorithm. INTERPRETATION: This study reports the successful development, external clinical validation, and deployment in clinical practice of an AI-based algorithm to accurately detect, grade, and evaluate clinically relevant findings in digitised slides of prostate CNBs. FUNDING: Ibex Medical Analytics.
Authors: Hongxiao Li; Jigang Wang; Zaibo Li; Melad Dababneh; Fusheng Wang; Peng Zhao; Geoffrey H Smith; George Teodoro; Meijie Li; Jun Kong; Xiaoxian Li Journal: Front Med (Lausanne) Date: 2022-06-14
Authors: Patrick Leo; Sacheth Chandramouli; Xavier Farré; Robin Elliott; Andrew Janowczyk; Kaustav Bera; Pingfu Fu; Nafiseh Janaki; Ayah El-Fahmawi; Mohammed Shahait; Jessica Kim; David Lee; Kosj Yamoah; Timothy R Rebbeck; Francesca Khani; Brian D Robinson; Natalie N C Shih; Michael Feldman; Sanjay Gupta; Jesse McKenney; Priti Lal; Anant Madabhushi Journal: Eur Urol Focus Date: 2021-04-30
Authors: Andi Flory; Kristina M Kruglyak; John A Tynan; Lisa M McLennan; Jill M Rafalko; Patrick Christian Fiaux; Gilberto E Hernandez; Francesco Marass; Prachi Nakashe; Carlos A Ruiz-Perez; Donna M Fath; Thuy Jennings; Rita Motalli-Pepio; Kate Wotrang; Angela L McCleary-Wheeler; Susan Lana; Brenda Phillips; Brian K Flesner; Nicole F Leibman; Tracy LaDue; Chelsea D Tripp; Brenda L Coomber; J Paul Woods; Mairin Miller; Sean W Aiken; Amber Wolf-Ringwall; Antonella Borgatti; Kathleen Kraska; Christopher B Thomson; Alane Kosanovich Cahalane; Rebecca L Murray; William C Kisseberth; Maria A Camps-Palau; Franck Floch; Claire Beaudu-Lange; Aurélia Klajer-Peres; Olivier Keravel; Luc-André Fribourg-Blanc; Pascale Chicha Mazetier; Angelo Marco; Molly B McLeod; Erin Portillo; Terry S Clark; Scott Judd; C Kirk Feinberg; Marie Benitez; Candace Runyan; Lindsey Hackett; Scott Lafey; Danielle Richardson; Sarah Vineyard; Mary Tefend Campbell; Nilesh Dharajiya; Taylor J Jensen; Dirk van den Boom; Luis A Diaz; Daniel S Grosu; Arthur Polk; Kalle Marsal; Susan Cho Hicks; Katherine M Lytle; Lauren Holtvoigt; Jason Chibuk; Ilya Chorny; Dana W Y Tsui Journal: PLoS One Date: 2022-04-26 Impact factor: 3.752
Authors: Leonard M da Silva; Emilio M Pereira; Paulo Go Salles; Ran Godrich; Rodrigo Ceballos; Jeremy D Kunz; Adam Casson; Julian Viret; Sarat Chandarlapaty; Carlos Gil Ferreira; Bruno Ferrari; Brandon Rothrock; Patricia Raciti; Victor Reuter; Belma Dogdas; George DeMuth; Jillian Sue; Christopher Kanan; Leo Grady; Thomas J Fuchs; Jorge S Reis-Filho Journal: J Pathol Date: 2021-04-27 Impact factor: 7.996