Assessing the clinical and biological associations between multimodal artificial intelligence (MMAI) and 22-gene genomic classifier (GC) in localized prostate cancer (PCa).

5114 Background: MMAI and GC are clinically validated prognostic tools for localized PCa, and are incorporated into NCCN guidelines as treatment-decision aids. However, their clinical and biological relationships remain unclear, especially in Asian populations where comparative study is limited. Methods: Patients with newly-diagnosed localized PCa from a single institution in Singapore were enrolled into a prospective protocol (NCT04340024), and underwent image-guided radiotherapy with or without hormonal therapy. All subjects had paired GC and MMAI scores. However, colored marker annotations were present on the H&E slides, which could negatively impact MMAI performance. Associations between GC and MMAI scores with metastasis-free survival (MFS) were evaluated using Cox proportional hazards models. Pearson's correlation was used to assess associations between risk scores and cancer hallmark pathways. Results: 144 men were included, comprising 142 (99%) with NCCN intermediate (int) to high-risk, and 2 (1%) with low-risk disease; median follow-up duration was 78.5 (interquartile range: 68.2–95.3) mo. MMAI and GC scores were correlated ( R = 0.55), but differed in risk stratification; 12/144 (8%), 54/144 (38%), and 78/144 (54%) were MMAI low-, int-, and high-risk, respectively, compared with 51/144 (35%), 32/144 (22%), and 61/144 (43%) by GC in the same order. We observed the largest discordance in the GC low-risk subgroup (n = 51), where 29/51 (57%) and 14/51 (28%) were classified as MMAI int- and high-risk, respectively. Of note, 16/32 (50%) of GC int-risk men were classified as MMAI high-risk. Nonetheless, both tests were prognostic for MFS (MMAI: hazard ratio [ HR ] of 2.6 [95% CI:1.0–6.4], P = 0.035; GC: HR of 3.4 [95% CI:1.5–7.6], P = 0.002), although the difference in HR should be interpreted with caution, as it is attributable to a limited sample size with longer follow-up. Next, we investigated for molecular pathway enrichment of GC and MMAI high-risk groups. Proportions of known prostate cancer molecular subtypes were comparable between the respective GC and MMAI risk-groups (GC high vs MMAI high: PAM50 Basal 51% vs 58%, P = 0.419; Luminal B 46% vs 39%, P = 0.377; PTEN loss 25% vs 22%, P = 0.698; AR lower activity 49% vs 44%, P = 0.512). GC high versus GC low/int tumors however showed positive enrichment in E2F, oxidative phosphorylation, and mTORC1 pathways; negative enrichment in P53 and estrogen pathways. MMAI high was only negatively enriched for fatty acid metabolism. Conclusions: While MMAI and GC provide differing risk categorization, both tests were significantly associated with MFS in men with localized prostate cancer who underwent radiotherapy. Known prostate cancer-associated transcriptomic subtypes were comparable between the respective MMAI and GC risk groups.