Citations

Abstract Background: Accurate staging is essential for treatment selection, prognosis assessment, and trial eligibility. While prior studies attempted automating pathological (p) staging, few address clinical (c) staging and no studies have compared LLM-derived staging with clinician staging during clinical visit and retrospective cancer registry staging. Methods: We developed a multi-agent framework using Gemini-2.0-Flash-001 and GPT-5 to (1) identify reports (2) extract key data, and (3) apply AJCC 8th edition staging criteria. LLM staging was compared with clinician documentation and registry staging in breast cancer patients. Clinician-registry agreement served as the reference standard. Two independent breast oncologists reviewed discordant cases. Results: We analyzed 122 randomly selected breast cancer patients across all three Mayo Clinic sites from 2018 - 2023. LLM performance matched or exceeded human inter-rater agreement for pathological staging (95-99.2% vs. 95-98.4%). Clinical staging was more challenging, with LLM concordance of 73-77.9% for cT and 87-89.3% for cN versus 87.8% and 91.0% clinician-registry agreement. Among 89 patients with clinician-registry concordance, LLM concordance was 98.9% for pT/pN, 79.8% for cT, and 91.0% for cN. Experts favored LLM staging in 27.8% (5/18) of cT and 25.0% (2/8) of cN discordances. Error analysis revealed LLM challenges in handling non-mass enhancements (6 cases), selecting radiologic measurements (2 cases), identifying discrete masses (2 cases), and miscellaneous (3 cases). Conclusion: LLMs achieved human-level performance for pathological staging (98%), supporting human-in-the-loop deployment. For clinical staging (79.8% cT, 91.0% cN), future work must enhance multimodal reasoning and integrate physical examination data. Prospective validation is needed to assess real-world impact. With targeted refinements and oversight, automated staging can transform registry workflows while augmenting clinical decision-making. Citation Format: Arshad Mohammed, Umair Ayub, Pooja Advani, Shakeela W. Bahadur, Amye J. Tevaarwerk, Tufia C. Haddad, Elisabeth I. Heath, Brenda J. Ernst, Ben Zhou, Cui Tao, Sara J. Holton, Karthik V. Giridhar, Irbaz B. Riaz. Automating clinical and pathological staging for breast cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2744.

703 Background: The association of circulating tumor DNA (ctDNA) reduction with treatment response has been investigated across multiple malignancies. This study aims to assess the impact of ctDNA response on survival outcomes in patients receiving enfortumab-pembrolizumab (EV-P). Methods: We identified 32 patients with advanced urothelial carcinoma who received EV-P and had ≥2 ctDNA assessments with ≥1 detectable level. Deep molecular response was defined as >2-log reduction from highest ctDNA detection. Radiographic progression was determined from radiology reports documenting progression by RECIST criteria. Kaplan Meier analyses were performed to compute Hazard Ratios (HR) with 95% confidence intervals (CI) for radiographic progression free survival (rPFS), defined as time from treatment start to radiographic progression or death and overall survival (OS) defined as time from treatment start to death were computed. Log-rank tests were performed to compare PFS and OS in patients with deep molecular response to those without response. A p-value <0.05 indicated a statistically significant association with response. Results: Thirty-two patients were included (median age 71 years; range 39–90). Median follow-up was 10.6 months. At cutoff (10/1/2025), 26 patients were alive and 6 had died. Seventeen (53%) achieved a ≥2-log ctDNA reduction (responders), while 15 (47%) did not (non-responders). Progression-free survival was significantly longer in responders (log-rank P < 0.001); median PFS was 5.9 months (95% CI 2.4–NR) in non-responders and not reached in responders (HR 0.09, 95% CI 0.02–0.44; P = 0.0026). Median overall survival was 12.7 months (95% CI 10.0–NR) in non-responders versus not reached in responders (log-rank P = 0.011; HR 0.11, 95% CI 0.01–0.88; P = 0.037). At cutoff, 94% of responders and 67% of non-responders were alive. Conclusions: Achieving a >2-log decrease in ctDNA during EV + Pembrolizumab therapy was significantly associated with improved progression free and overall survival. These findings support ctDNA kinetics as a potential early biomarker of response warranting prospective validation. Clinical outcomes by >2-log ctDNA reduction status. Endpoint > 2-log drop (n = 17 ≤ 2-log drop (n = 15) Log-rank p Median (mo) ≤ 2-log > 2-log HR (95% CI) P C-index PFS 2 events (12%) 10 events (67%) 0.0002 5.9 (2.4–NA) NR 0.09 (0.02–0.44) 0.0026 0.77 OS 1 event (6%) 6 events (40%) 0.011 12.7 (10.0–NA) NR 0.11 (0.01–0.88) 0.037 0.73

57 Background: Cardiovascular disease (CVD) is the leading non-cancer cause of death among men with prostate cancer (PC), particularly during androgen-deprivation therapy. While CVD risk-prediction models are well established in the general population, barriers to integration within oncology workflows remain poorly defined. Guided by the PRISM and RE-AIM frameworks, this study evaluated multilevel determinants influencing adoption, feasibility, and sustainability of CVD risk-assessment tools in PC care. Methods: A national, cross-sectional survey of 45 physicians (medical oncology 49%, urology 38%, radiation oncology 13%) from 18 U.S. states was conducted. Respondents represented early- (27%), mid- (36%), and late-career (38%) clinicians across academic (87%) and community (13%) settings. Quantitative items mapped to PRISM/RE-AIM domains assessed adaptability, cost/resource feasibility, effectiveness, equity, and sustainability; open-ended responses explored contextual barriers and facilitators. Descriptive statistics summarized quantitative data. Results: Physicians reported moderate-to-high readiness to adopt a CVD risk-assessment tool, emphasizing workflow alignment and EHR integration over algorithmic complexity. Table 1 summarizes domain-specific findings. Ninety-three percent of respondents favored lab-independent models that minimize workflow disruption. Conclusions: Pre-implementation analysis highlighted that structural barriers, rather than algorithmic or statistical limitations, remain the primary obstacles to adopting CVD risk-assessment tools in prostate-cancer care. Physicians emphasized the need for EHR-integrated, non-laboratory-based, and time-efficient tools that align with routine oncology workflows and minimize cross-specialty friction. Addressing these practical and organizational determinants may be pivotal to achieving equitable, scalable, and sustainable cardio-oncology implementation across diverse practice settings. PRISM and RE-AIM constructs evaluation results. Construct Likely/Very Likely (%) Interpretation Perceived Effectiveness 87% Broad confidence in clinical utility Equity in Patient Reach 75% Strong perceived equity in reach Equity in Effectiveness 73% Expected to perform consistently across populations Adaptability to Workflow 71% High compatibility with existing clinic workflows Cost/Resource Feasibility 65% Feasible for resource-constrained settings Likelihood of Adoption 60% Moderate readiness for implementation Equity in Adoption 58% Equitable uptake possible despite resource differences Reach to Target Population 57% Moderate likelihood of reaching most patients Sustainability Over Time 51% Moderate long-term feasibility Consistency of Implementation 42% Need for standardized protocols

567 Background: Selecting adjuvant therapy after nephrectomy for renal cell carcinoma (RCC) requires reviewing scattered electronic health records (EHR) and performing accurate eligibility and risk assessments. Large Language Models (LLMs) may assist physicians by extracting key clinical data and generating evidence-based recommendations between active surveillance and adjuvant pembrolizumab. Methods: We selected RCC patients who underwent nephrectomy at Mayo Clinic in Arizona, Florida, and Minnesota between December 2021 and September 2025. Data extraction from EHRs was performed by a HIPAA-compliant LLM (GPT-5 and Gemini 2.0-Flash-001) and compared to human labels. Data on patients’ demographics, medical history, pathology, laboratory results, and imaging were collected. GPT-5 generated treatment recommendations using a structured prompt that incorporated patients’ features and relevant clinical information that is available to clinicians. Concordance between the human labels and LLM was evaluated for extraction. For treatment recommendations, agreement between the LLM’s output and physician’s documented recommendation in clinical notes was assessed, with disagreements reviewed independently by two blinded genitourinary oncologists. Results: The dataset included 85 patients, median (range) age of 65 (34-84) years; 24.7% were females. Patients who declined adjuvant treatment for personal reasons were excluded. LLM achieved an extraction accuracy of 95% for presenting symptoms, 97% for BMI and ECOG, and > 99% for pathology variables. Concordance between LLM and physician treatment recommendations was 84.7% (72/85). Among 13 discrepant cases, one oncologist agreed with the LLM in 9 cases and the other in 8. The main reasons for disagreement between the physicians and the LLM were that the LLM favored active surveillance for patients with declining renal function, whereas physicians still considered adjuvant pembrolizumab appropriate in such cases. Additionally, the LLM demonstrated hesitancy to recommend pembrolizumab for patients with low performance status. Overall, the 13 discrepant cases were nuanced and challenging, with disagreement observed between the two oncologists themselves in three instances. Conclusions: LLMs demonstrated high accuracy in extracting clinically relevant variables and generating objective, evidence-based adjuvant treatment recommendations for RCC after nephrectomy. Importantly, LLM did not generate any inappropriate or potentially harmful recommendations. Our pipeline can serve as a supportive decision tool to enhance consistency in post-nephrectomy management and assist physicians in choosing between active surveillance and adjuvant pembrolizumab. However, final treatment decisions should remain the responsibility of the treating physician after comprehensive patient evaluation.

360 Background: Adding NHT to androgen deprivation therapy (ADT) improves survival in mCSPC, but their comparative impact on patient-reported QoL and time to deterioration (TTD) across FACT-P domains remains underexplored. Methods: MEDLINE and Embase were systematically searched from each database’s inception through August 21, 2025, to identify Phase III trials assessing NHT treatment in mCSPC. Time to FACT-P (total, physical, emotional, functional, and social/family well-being) deterioration across trials with the latest follow up were extracted. Hazards ratios (HRs) and 95% confidence intervals (CIs) were pooled and mixed treatment comparisons were made using network meta-analysis. P-scores were computed; a higher score indicated better QoL. Results: A total of 4 trials (12 references) with 4,037 patients met the inclusion criteria with median follow up ranging from 14.4 months to 30 months. Median time to QoL deterioration ranged from 11.4 to 12.9 months. In terms of FACT-P overall, darolutamide (DARO)+ADT (HR: 0.76; 95% CI: 0.62-0.93), abiraterone acetate (AAP)+ADT (0.85; 0.73-0.98) significantly improved TTD as compared to ADT alone. DARO+ADT was associated with significant improvement in TTD when compared to APA+ADT (0.75; 0.56-0.98). In terms of FACT-P social/family wellbeing, DARO+ADT significantly improved TTD when compared to AAP+ADT (0.75; 0.57-0.96), and ADT (0.79; 0.64-0.98). Likewise, ENZA+ADT significantly improved TTD when compared to AAP+ADT (0.77; 0.63-0.95), APA+ADT (0.88; 0.61-0.99) and ADT (0.82; 0.70-0.95). However, there was no statistically significant difference between DARO+ADT and ENZA+ADT. In terms of FACT-P physical wellbeing, AAP+ADT was associated with statistically significant improvement in TTD when compared to APA+ADT (0.66; 0.52-0.83), ENZA+ADT (0.74; 0.61-0.90), and ADT (0.75; 0.65-0.87). In terms of FACT-P functional wellbeing, DARO+ADT was associated with statistically significant improvement in TTD when compared to ADT (0.78; 0.63-0.96). The ranking analysis is showed in Table. Conclusions: Novel hormonal therapies significantly delay QoL deterioration in mCSPC, particularly across overall, emotional, and functional domains. Domain-specific analyses suggest darolutamide may preserve social/family and functional well-being, while abiraterone favors physical well-being. Rank table. Rank (P-score) FACT-Ptotal FACT-Psocial/family FACT-Pphysical FACT-Pfunctional FACT-Pemotional AAP+ADT 2 (0.73) 5 (0.18) 1 (0.98) 2 (0.62) 3 (0.52) APA+ADT 5 (0.17) 4 (0.22) 5 (0.07) 4 (0.33) 4 (0.52) DARO+ADT 1 (0.90) 1 (0.88) 2 (0.68) 1 (0.92) 2 (0.66) ENZA+ADT 3 (0.50) 2 (0.84) 4 (0.36) 3 (0.52) 1 (0.67) ADT 4 (0.18) 3 (0.38) 3 (0.42) 5 (0.11) 5 (0.13)

352 Background: Glucagon-like peptide 1 (GLP1) receptor expression in prostate cancer tissue has raised concerns that GLP1 receptor agonists (GLP1A) may influence tumor growth or progression. We compared prostate cancer severity at diagnosis among patients receiving GLP1A versus other glucose lowering therapies. Methods: A retrospective chart review identified patients who received GLP1A or alternative second-line glucose lowering agents (DPP4/SGLT2 inhibitors) and were subsequently diagnosed with prostate cancer. Patients receiving both drug classes were excluded. Baseline demographics (race, ethnicity, age), comorbidities (diabetes, myocardial infarction, peripheral vascular disease) concomitant medications (insulin, metformin, statins), appointments (within past year) and laboratory values (HbA1c, eGFR) at therapy initiation were recorded. Primary outcomes included Gleason grade group and metastatic presentation within three months of diagnosis. Propensity score matching balanced baseline characteristics between groups. Mann-Whitney U tests compared grade groups. Chi-square tests compared metastatic presentation, reporting odds ratios (OR) with 95% confidence intervals (CI). Multivariable logistic regression identified predictors of grade group (low-grade: 1-2; high-grade: ≥3) and metastatic presentation, reporting OR with 95% CI. A p value <0.05 indicated a significant association. Results: Of 260 propensity-matched patients (GLP1A: 130, comparison: 130), 87% were White and 95% were Non-Hispanic/Non-Latino in each group. Median HbA1c was 7% (IQR: 6.4-8) in GLP1A group versus 7.5% (6.2-8.5) in comparison group. Median eGFR was 75mL/min/1.73m 2 (IQR: 61.5-90.0) in GLP1A group versus 60.0 (52.0-80.5) in comparison group. High-grade disease occurred in 33 (25%) GLP1A patients and 37 (29%) comparison patients. Among patients with available imaging (GLP1A: 40; comparison: 32), metastatic presentation occurred in 7 (18%) GLP1A group and 7 (22%) comparison group patients. No significant differences were observed in grade group (GLP1A median: 2; IQR: 1-2 vs. comparison median: 2; 1-3; p=0.44) or metastatic presentation (OR: 1.32; 95%CI: 0.41-4.25; p=0.64) between groups. In multivariable analysis, higher HbA1c was associated with high-grade disease in GLP1A group (OR: 1.83; 95%CI: 1.07-3.35; p=0.03) while higher eGFR was associated with low-grade disease in the comparison group (0.97; 0.93-0.99; p=0.04). Conclusions: GLP1A use was not associated with prostate cancer grade or metastatic presentation at diagnosis compared to alternative glucose-lowering agents. However, the association between elevated HbA1c and high-grade disease in GLP1A patients suggests a potential impact of poor glycemic control on tumor biology. Prospective studies are needed to elucidate the relationship between metabolic dysregulation and prostate cancer severity.

145 Background: Multiple trials have demonstrated addition of androgen-receptor pathway inhibitors (ARPIs) improve survival in patients with metastatic prostate cancer but there is a need to quantify cardiovascular toxicity. Understanding the nature and extent of ARPI-associated cardiovascular adverse events across disease states and toxicity grades is critical to inform treatment decisions and long-term patient safety. Methods: MEDLINE, and EMBASE were systematically searched from database inception through August 30, 2025, for phase III randomized controlled trials evaluating abiraterone, enzalutamide, apalutamide, darolutamide, orteronel in metastatic prostate cancer. Trials were included if they reported CVS toxicity. Meta-analysis using a random-effects model was performed to assess the difference of all-cause and treatment-emergent cardiovascular adverse events between ARPI+ADT and ADT alone. Summary effects were expressed as relative risk (RR) with associated 95% confidence intervals. Results: A total of 21 trials (45 references) with 25,576 patients (mCSPC: 12,537, mCRPC: 13,039) were included in this meta-analysis. Compared to ADT alone, ARPI+ADT significantly increased the risk of all cause, all grade hypertension in both mCSPC (RR 1.60, 1.22–2.10) and mCRPC (RR 1.61, 1.20–2.16) as well as grade ≥ 3 hypertension (RR 1.71, 1.17–2.48) in mCRPC setting. There was an increased risk of all-cause, all grade acute coronary syndrome (RR 2.21, 1.09–4.49) and AV nodal disease (RR 2.53, 1.18–5.43) with ARPI+ADT compared to ADT alone in mCSPC setting. There was increased risk of all cause, all grade, any cardiac events (RR 1.30, 1.10–1.54) as well as grade ≥ 3 any cardiac events (RR 1.70, 1.20–2.42) in mCRPC setting. Risk of all cause, all grade hot flashes (RR 1.47, 1.16–1.85) was also significantly increased with ARPI+ADT when compared to ADT alone. In terms of treatment-emergent adverse effects, a significant increase was observed only for pulmonary embolism with ARPI+ADT when compared to ADT alone (RR 1.66, 1.06–2.60) in mCRPC setting. Conclusions: Current evidence indicates potentially increased risk of hypertension, venous thromboembolic, and ischemic cardiac events with ARPI therapy in metastatic prostate cancer. Proactive CV risk assessment, particularly in patients with preexisting comorbidities or prolonged ARPI exposure, may optimize outcomes and minimize treatment interruptions. All cause adverse events. Adverse Event Disease State All Grade Grade ≥3 Any Cardiac Event mCSPC 0.81 (0.37–1.76) 0.88 (0.16–4.64) mCRPC 1.30 (1.10–1.54) 1.70 (1.20–2.42) Hypertension mCSPC 1.60 (1.22–2.10) 1.43 (0.82–2.52) mCRPC 1.61 (1.20–2.16) 1.71 (1.17–2.48) ACS mCSPC 2.21 (1.09–4.49) 1.77 (0.86–3.61) mCRPC – – Hot Flashes mCSPC – – mCRPC 1.47 (1.16–1.85) 1.46 (0.21–9.97)

144 Background: Androgen receptor pathway inhibitors (abiraterone acetate [AAP], apalutamide [APA], enzalutamide [E], darolutamide [DARO] have become the cornerstone for pts with PCa due to survival benefits. Comparative data on fracture risk among ARPi are limited. This study quantified the risk of bone fractures with ARPi plus androgen deprivation therapy (ADT) and compared fracture risk across different ARPi agents. Methods: MEDLINE and EMBASE were searched from inception through September 1 st 2025 to identify positive phase III trials assessing ARPi in addition to ADT in pts with PCa (m0CRPC, mCSPC, mCRPC). Trials were included if they reported bone fractures in either arm. Pairwise, random-effects meta-analysis was conducted using Paul-Mendel approach. Sensitivity analyses using Hartung-Knapp (HK) adjustment was performed to assess robustness of results. To assess the comparative risk of bone fractures among different ARPi agents, mixed treatment comparisons were made using frequentist network meta-analysis. P-scores were computed to assess relative treatment rankings. Treatment effect was expressed as odds ratio (OR) with 95% confidence intervals (CI). Results: This systematic review included 13 phase III trials with 15468 pts (8890 received ARPi+ADT, 6578 received ADT alone). A total of 596 (6.7%) bone fractures were observed with ARPi+ADT compared to 180 (2.74%) with ADT alone. Patients receiving ARPi+ADT were associated with an increased risk of bone fractures compared to those receiving ADT (OR: 2.40; 95% CI: 1.93-2.98). Sensitivity analyses with HK adjustment showed consistent results (2.40; 1.88-3.06). There was no statistically significant effect modification by PCa clinical setting (p: 0.27). Mixed treatment comparisons showed a statistically significant increased risk of bone fractures with AAP+ADT (2.35; 1.29-4.27), APA+ADT (2.04; 1.47-2.86), and E+ADT (3.13; 2.44-4.00) but not with DARO+ADT (1.30; 0.80-2.08) when compared to ADT alone. Among different ARPi agents, E+ADT was associated with an increased risk compared to APA+ADT (1.54; 1.01-2.33) and DARO+ADT (2.44; 1.41-4.17). Ranking analysis showed that E+ADT was ranked as potentially the least safe treatment (P score: 0.05) in terms of relative fracture risk followed by AAP+ADT (P score: 0.30), APA+ADT (P score: 0.43), DARO+ADT (P score: 0.76). Sensitivity analyses limiting to mCSPC trials only showed consistent results. Conclusions: The risk of bone fractures varies among different ARPi agents, with enzalutamide posing the highest relative risk and darolutamide the lowest. These differences may guide treatment selection, particularly in pts with mCSPC and bone metastases. Prospective studies adjusting for covariates are needed to evaluate bone-protective strategies to mitigate fracture risk particularly in mCSPC patients who are considering treatment with ARPi.

142 Background: LuPSMA is a life-prolonging radioligand therapy that is widely considered to have favorable tolerability. SPPADE symptoms (Sleep disturbance, Pain, Physical dysfunction, Anxiety, Depression, low Energy) are common in prostate cancer, yet data on patient-reported outcomes with LuPSMA in real-world settings remain limited. Methods: The E2C2 study (NCT03892967) evaluates the impact of a collaborative care intervention on patient-centered outcomes and healthcare utilization. This analysis includes mCRPC patients who initiated LuPSMA at Mayo Clinic (June 2022-July 2023) and completed at least one symptom survey within 6 weeks before or 2 weeks after the first LuPSMA treatment (baseline survey). SPPADE symptoms were rated on 0-10 scales; scores of 4-6 indicate moderate and 7-10 severe burden. A composite z-score was derived by standardizing each SPPADE domain to the cohort mean and averaging z-scores across domains per patient. Patients were stratified by composite symptom burden using cohort median and upper-tertile cutoffs. Associations with PSA50 response and overall survival (OS, from start of LuPSMA) were assessed using chi-square and Kaplan-Meier methods. In a longitudinal subgroup of patients completing a baseline survey plus at least one additional PRO survey while receiving LuPSMA (≥2 surveys), scores were compared from baseline to best during-treatment using Wilcoxon and Mann-Whitney U tests. Results: Of 373 patients starting LuPSMA, 51 patients met inclusion criteria. Median age was 71 years, 31% had visceral metastases, and median baseline PSA was 42.1 ng/mL. Overall, 85% of patients reported at least one moderate symptom and 46% at least one severe symptom. The most common moderate-to-severe symptoms were low energy (58%), pain (52%), and sleep disturbance (49%). Using an upper-tertile cutoff of the composite SPPADE score, 17 patients had high and 34 had low/moderate symptom burden prior to LuPSMA. PSA50 response was achieved in 17.6% of patients with high vs 47.1% with low/moderate composite symptom scores (p = 0.08). Median OS was 9.2 vs 16.7 months for high vs low symptom groups (p = 0.15, median split). The longitudinal subgroup included 24 patients with serial ePRO assessments while on LuPSMA. Composite SPPADE scores improved from baseline to the best during-treatment assessment in both PSA50 responders (n=10) and non-responders (n=14), with median composite z-score Δ of 0.35 vs 0.20, respectively (p = 0.68). Conclusions: Among patients with mCRPC receiving LuPSMA, baseline SPPADE symptom burden was high. Though not statistically significant in this underpowered cohort, PSA response and OS were numerically inferior in those with higher baseline burden. These findings suggest that ePRO-captured symptom burden may provide prognostic and patient-centered insights beyond traditional clinical variables. Clinical trial information: NCT03892967 .