AI and Cancer Research ── Prostate ── 2026-10-10
Prostate cancer: treatment, trials and AI
- Abstract
Accurate patient-level risk stratification of clinically significant prostate cancer (csPCa) on biparametric MRI remains challenging, as most existing approaches require costly lesion-level annotations or have not been evaluated in large temporally separated cohorts. We developed and validated a multimodal deep learning framework combining T2-weighted (T2w), diffusion-weighted (DWI), and apparent diffusion coefficient (ADC) volumes with age, prostate-specific antigen (PSA), and PSA density (PSAD) for patient-level csPCa classification, trained without lesion-level annotations. The study included a retrospective development cohort of 3, 939 examinations and an independent temporal validation cohort comprising 1, 409 prospectively acquired examinations from 13 centers. The model employs three sequence-specific encoders with asymmetric cross-attention fusion and a two-stage transfer-learning strategy. Adding demographic and clinical variables consistently improved performance over MRI-only models, with PSAD emerging as the most informative complementary factor, primarily by improving specificity. The best-performing model—combining MRI, age, PSA, and PSAD with cross-attention—achieved a mean AUC of $$0.765\!\pm \!0.006$$ on the retrospective held-out test set and $$0.741\!\pm \!0.006$$ on the independent temporal validation cohort. Post hoc subgroup analyses suggested broadly stable rank-order discrimination (AUC) across most strata, though fixed-threshold sensitivity and specificity varied substantially, while Grad-CAM maps offered a qualitative indication of more frequent overlap with suspicious regions under cross-attention. These findings support PSAD-informed multimodal patient-level models as a viable tool for refined csPCa risk stratification on biparametric MRI, with potential to reduce false-positive classification of ISUP grade group 1 disease as clinically significant.
Journal IF-equivalent: 5.2 (OpenAlex 2-year mean citedness, value as of 2026-10-04, retrieved 2026-10-05; not the official Clarivate IF)Reference: Dimitriadis A, Kalliatakis G, Osuala R, Kessler D, Diaz O, Mazzetti S, et al. Cross-attention fusion of biparametric MRI sequences with demographic and clinical variables for patient-level classification of clinically significant prostate cancer. Sci Rep. 2026 Oct 8 [Epub ahead of print]. doi:10.1038/s41598-026-71998-x.Checked: Abstract only - Abstract
Importance: The Decipher Prostate Genomic Classifier (GC) and ArteraAI Multimodal Artificial Intelligence (MMAI) platform are widely used prognostic tools for localized prostate cancer (PCa), yet no direct comparison has been performed within the same patient cohort.
Methods: We sought to evaluate the concordance and prognostic value of GC and MMAI across multi-institutional cohorts totaling 688 patients with localized and oligometastatic PCa. These included populations enriched for African American and East Asian patients. GC and MMAI scores were stratified into risk groups based on specimen type (biopsy or radical prostatectomy (RP)). The primary endpoint was distant metastasis-free survival (DMFS), measured from diagnosis to last follow-up or event.
Results: GC and MMAI scores were significantly correlated across multiple cohorts. In Moffitt RP, biopsy, and NCCS cohorts, both MMAI and GC were prognostic alone and after adjusting for clinical variables. Multivariable analysis including both biomarkers and clinical variables showed that MMAI was significant even accounting for GC in Moffitt RP, and borderline in Moffitt biopsy. Analysis of concordant/discordant cases showed that patients who were high-risk for both GC and MMAI did qualitatively worse. In pathway analysis, concordance with GC and MMAI was high for Moffitt RP and oligometastatic patients. Across both biopsy cohorts, while most biological pathways were concordant, there were consistent discordant signaling, metabolic, and DNA repair pathways.
Conclusions: In the first direct patient-level comparison of GC and MMAI in PCa, these biomarkers were found to demonstrate moderate correlation, and both biomarkers demonstrated prognostic value across diverse populations and disease states.
Journal IF-equivalent: 2.2 (OpenAlex 2-year mean citedness, value as of 2026-10-09, retrieved 2026-10-10; not the official Clarivate IF)Reference: Hong BH, Olabumuyi AA, Zhao SG, Trivedi P, Putney R, Katende E, et al. Multi-institutional Patient-level Comparison of Decipher Genomic Classifier and Artera Multimodal Artificial Intelligence (MMAI) in Prostate Cancer. Clinical Cancer Research. 2026 Oct 8 [Epub ahead of print]. doi:10.1158/1078-0432.ccr-26-2269.Checked: Abstract only