MétaCan
Menu
Back to cohort

Digital histopathology-based multimodal artificial intelligence scores predict risk of progression in a randomized phase III trial in patients with nonmetastatic castration-resistant prostate cancer.

2023· article· en· W4379332778 on OpenAlexaff
Felix Y. Feng, Matthew R. Smith, Fred Saad, Pooya Mobadersany, Shaozhou K. Tian, Stephen Yip, Joel Greshock, Najat Khan, Margaret K. Yu, Sharon McCarthy, Sabine Brookman‐May, Tamara R. Todorović, Rikiya Yamashita, Huei–Chung Huang, Timothy N. Showalter, Akinori Mitani, Andre Esteva, Eric J. Small

Bibliographic record

VenueJournal of Clinical Oncology · 2023
Typearticle
Languageen
FieldMedicine
TopicRadiomics and Machine Learning in Medical Imaging
Canadian institutionsCentre Hospitalier de l’Université de Montréal
FundersJanssen Biotech
KeywordsMedicineProstate cancerPlaceboInternal medicineProportional hazards modelRandomized controlled trialBiopsyOncologyAndrogen deprivation therapyCancerUrologyPathology

Abstract

fetched live from OpenAlex

5035 Background: The addition of apalutamide (APA) to androgen deprivation therapy (ADT) improved metastasis-free survival (MFS) and second progression-free survival (PFS2) among patients (pts) with nonmetastatic castration-resistant prostate cancer (nmCRPC) [1]. We applied a previously-reported digital histopathology-based multimodal AI (MMAI) algorithm developed based on 5 phase III randomized trials, validated, and accepted to NCCN guidelines for localized prostate cancer, to evaluate whether MMAI could define risk of progression among nmCRPC pts treated with APA or placebo in the SPARTAN trial. Methods: Pts enrolled in the SPARTAN trial with available H&E-stained biopsy slides from their primary diagnosis were included. H&E slides were digitized. Baseline clinical parameters to generate MMAI scores were Gleason score, age, T stage, and PSA. MMAI scores for distant metastasis (DM) were generated, ranging from 0 to 1. Pts were further categorized into MMAI non-high-risk and high-risk groups using a previously established score cutoff. Kaplan Meier estimates were calculated for PFS2 and MFS; comparisons were performed using log-rank test and Cox proportional-hazards regression for treatment arms and MMAI risk groups. Two-way ANOVA was used to evaluate the interaction between treatment arms and risk groups. Results: The study included 471 pts with 1051 biopsy pathology slides: 311 pts treated with APA, and 156 with placebo. 55 pts were excluded due to missing treatment (n=4) or clinical data (n=49) and inadequate H&E images (n=2), resulting in 273 evaluable APA-treated and 147 placebo-treated pts. 63% of pts were MMAI high risk and 37% MMAI non-high risk. MMAI high risk pts demonstrated significant improvement in MFS with APA (HR 0.19 (95% CI: 0.12-0.29, p<0.005)), but not in PFS2 (HR 0.76 (95% CI: 0.45-1.28, p=0.30)). There was a significant interaction between MMAI risk group and treatment for MFS (p=0.02). Among the placebo-treated cohort, MMAI high risk status was associated with shorter MFS (HR 2.98 (95% CI: 1.72-5.18, p<0.005)) and PFS2 (HR 1.83 (95% CI: 1.09-3.09, p=0.02)). For APA-treated pts, MMAI risk group was not associated with MFS and PFS2. Conclusions: These findings suggest that MMAI may provide prognostic risk stratification for nmCRPC pts and that MMAI high-risk pts may benefit most from treatment with APA. The current results represent the first evaluation of this MMAI classifier in the nmCRPC setting. Prospective validation is warranted. [1] Smith MR, Saad F, Chowdhury S, et al.; SPARTAN Investigators. Apalutamide treatment and metastasis-free survival in prostate cancer. N Engl J Med. 2018;378(15):1408-1418. Clinical trial information: NCT01946204 .

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.007
metaresearch head score (Gemma)0.017
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMetaresearch
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Randomized trial · Consensus signal: Randomized trial
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.233
Threshold uncertainty score0.991

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0070.017
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0020.000
Bibliometrics0.0000.000
Science and technology studies0.0000.001
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.001
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.049
GPT teacher head0.451
Teacher spread0.402 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

Study designRandomized trial
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations2
Published2023
Admission routes1
Has abstractyes

Explore more

Same venueJournal of Clinical OncologySame topicRadiomics and Machine Learning in Medical ImagingFrench-language works237,207