1621P - Validation of a digital pathology-based multimodal artificial intelligence (MMAI) prostate biopsy biomarker in a prospective, real-world Swedish prostate cancer (PCa) cohort treated with radical prostatectomy

Background

A MMAI biomarker (ArteraAI Prostate Test) was developed using clinical trial data from North American (NA) men with localized PCa treated with definitive radiation, using digital pathology images and key clinical information to generate prognostic scores. This validation study evaluates the MMAI biomarker in a real-world dataset of non-NA men who underwent radical prostatectomy (RP) for localized PCa.

Methods

The MMAI algorithm used digitized images from diagnostic H&E prostate biopsies at Skåne University Hospital, Malmö, tumor stage, age, and PSA values. The association between the MMAI model, continuously (per SD increase) and categorically (based on preestablished cutoffs), and endpoints of interest were performed with Fine-Gray and cumulative incidence analyses for biochemical recurrence (BCR; two successive PSA measurements ≥ 0.2 ng/mL post-RP) and logistic regression for adverse pathology (AP) at RP (grade group [GG] ≥3, ≥ pT3b, and/or N1). Deaths without events were treated as competing risks for BCR. Men diagnosed as GG <3,

Results

Of 749 men referred for biopsy 2004-2010, 230 underwent RP with available digital pathology slides of whom 143 had sufficient tumor and complete clinical data to generate MMAI scores. The analyzed cohort had a median follow-up of 8.8 years. At diagnosis, median PSA was 7.5 ng/mL, median age 64 years, 29% had GG ≥3, 8% had cT3; 95 men were evaluable for AP at RP. MMAI was significantly associated with BCR (subdistribution HR 2.45 [95% CI 1.77-3.38], p<0.001) and AP at RP (OR 4.3 [95% CI 2.33-9.02], p<0.001). Estimated 5-yr BCR rates for MMAI Intermediate-High vs Low were 25% (95% CI 15%-35%) vs 4% (95% CI 0%-8%), respectively.

Conclusions

We found the prostate biopsy MMAI biomarker, previously shown to be prognostic for distant metastasis and prostate cancer-specific mortality in men getting definitive radiation, to be prognostic for post-RP endpoints: BCR and AP. This biomarker validation study using a real-world cohort further supports the use of MMAI biomarkers in men with PCa outside NA and those treated with RP.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Anders Bjartell, Department of Translational Medicine, Medical Faculty, Lund University.

Funding

Swedish Cancer Society (Cancerfonden #21 1629 Pj), Swedish Scientific Council (Vetenskapsrådet, 2020-02017), and Artera Inc.

Disclosure

A. Bjartell: Financial Interests, Personal, Advisory Board: Accord, Astellas, AstraZeneca, Bayer, Novartis, Pfizer, SAM Nordic; Other, Institutional, Research Funding: Ferring; Financial Interests, Personal, Invited Speaker: IPSEN, Janssen; Financial Interests, Personal, Writing Engagement: Sandoz; Other, Personal, Leadership Role, Research Foundation Chairman: European Association of Urology; Non-Financial Interests, Institutional, Coordinating PI: Astellas, Janssen. V. Liu, M. Tierney, E. Chen, T. Royce, A. Kraft: Financial Interests, Personal, Full or part-time Employment: ArteraAI; Financial Interests, Personal, Stocks or ownership: ArteraAI. M. Sjöström: Financial Interests, Personal, Invited Speaker: Astellas. A.C. Esteva: Financial Interests, Personal, Full or part-time Employment: ArteraAI; Financial Interests, Personal, Stocks or ownership: ArteraAI; Financial Interests, Personal, Leadership Role: ArteraAI. F. Feng: Financial Interests, Personal, Advisory Role: Janssen Oncology, Bayer, Exact Sciences, Bristol Myers Squibb; Financial Interests, Personal, Stocks or ownership: Serimmune, Artera AI. All other authors have declared no conflicts of interest.