Zakharia, Yousef

Abstract Background and Objective Management of von Hippel–Lindau (VHL) disease is challenging due to multisystem involvement and limited systemic treatment options. Belzutifan has shown promising efficacy in clinical trials; however, real-world data remain limited. We aimed to evaluate its effectiveness and safety in a real-world cohort. Methods We conducted a retrospective study of 59 patients with VHL disease treated with belzutifan between August 2021 and January 2026. Objective response rates (ORR) for renal cell carcinoma (RCC), central nervous system hemangioblastomas (CNS HB), and pancreatic neuroendocrine tumors (pNET) were assessed by an independent radiologist. Retinal hemangioblastoma outcomes were evaluated through ophthalmology records by chart review. Results Median follow-up was 29 months. ORR was 27.6%, 39.6%, and 85.7% for RCC, CNS HB, and pNET, respectively. Among 13 patients with retinal hemangioblastomas (20 eyes), 17 eyes (85%) showed improvement. At data cutoff, 47 patients (79.6%) remained on treatment, while 12 (20.3%) discontinued (7 due to adverse events, 3 to disease progression, and 2 to other reasons). The most common adverse event was anemia (86.4%). Grade ≥3 events occurred in 10 patients (16.9%). The procedure rate per patient-year decreased from 0.35 in the 3 years prior to treatment to 0.08 during follow-up (Incidence Rate Ratio, 0.23; 95% CI, 0.12–0.47; P < 0.001) Conclusions Belzutifan demonstrated efficacy across VHL-associated lesions, with favorable safety profile. Most adverse events were grade 1–2. A reduction in procedure rate was observed after treatment initiation.

Background Immune checkpoint inhibitor-based combinations represent the standard of care for advanced renal cell carcinoma. The OMNIVORE trial investigated a response-adaptive strategy, including treatment discontinuation in early nivolumab responders (Arm A) and salvage ipilimumab addition in non-responders (Arm B). Here we report outcomes with extended follow-up. Methods OMNIVORE was a phase II response-adaptive trial in which patients received induction nivolumab monotherapy. Patients achieving a confirmed complete or partial response discontinued nivolumab and entered observation (Arm A), while patients with stable disease or progressive disease received two doses of ipilimumab added to ongoing nivolumab (Arm B). This analysis characterizes long-term overall survival across the full study cohort and durability of response among patients who discontinued nivolumab following an early objective response. Results Of 83 patients who initiated treatment, 12 (14%) were allocated to Arm A and 57 (69%) to Arm B, with a median follow-up among living patients of 59.4 months (range 29.1–85.4 months) in Arm A and 31.4 months (range 4.6–62.6 months) in Arm B. The 3-year overall survival rate from nivolumab initiation was 64% (95% CI 51% to 74%) in the overall cohort, 83% (95% CI 48% to 96%) in Arm A, and 63% (95% CI 47% to 76%) in Arm B. Of the 12 Arm A patients, 6 (50%) remained off nivolumab at 1 year following treatment discontinuation, of whom 5 maintained responses beyond 43 months off therapy, with all remaining alive at last known follow-up, with overall survival ranging from 48.8 to 85.4 months from nivolumab initiation. Of the six patients who resumed treatment, only one achieved a durable complete response remaining on treatment at 83.2 months from nivolumab initiation. All 57 Arm B patients discontinued treatment with a median treatment duration of 3.7 months (range 1–24.8 months) and a median progression-free survival from nivolumab plus ipilimumab initiation of 4.6 months (95% CI 2.7 to 6.5 months). Conclusions With extended follow-up, a meaningful subset of patients achieving an early objective response to nivolumab maintained prolonged treatment-free survival following a short course of treatment. Salvage ipilimumab in nivolumab non-responders demonstrated limited benefit, reinforcing that upfront concurrent dual checkpoint blockade remains the preferred approach. Trial registration number NCT03203473 .

e17005 Background: We previously developed a hybrid framework combining LLMs and rule-based algorithms (RBA) for automated risk stratification in localized prostate cancer (locPCa), achieving 89% accuracy. However, clinical T-stage (cTS) inference from narrative MRI reports remained the primary source of errors. Herein, we evaluated whether ensemble approaches using multiple state-of-the-art LLMs could improve cTS inference and overall NCCN risk stratification accuracy. Methods: This study included patients with locPCa presenting with at least one positive prostate biopsy and MRI report available. Building on our previously validated hybrid framework that extracts structured phenotypic variables (PSA, prostate volume, cTS, Gleason patterns, grade group, positive and examined cores) from unstructured reports using LLMs and applies NCCN-concordant RBA for risk stratification, we focused on improving cTS inference. Four HIPAA-compliant LLMs (GPT-4.1, GPT-5.1, Gemini-2.5-Pro, Gemini-2.5-Flash) were evaluated using structured zero-shot prompts iteratively refined through oncologist feedback. Majority voting was employed to generate consensus cTS predictions. Performance was evaluated using weighted accuracy. Results: A total of 358 patients were included. The median age at diagnosis was 65 years (IQR: 60-69); the majority were White (92%) and non-Hispanic (94%). cTS distribution was T2a (51%), T2c (20%), T3a (16%), T3b (7%), T2b (4%), and T4 (2%). The most prevalent risk category was unfavorable intermediate (38%) followed by favorable intermediate (20%), high risk (19%), very high risk (16%), low (5%), and very low (1%). Individual LLM accuracy for cTS inference ranged from 85-90%, with Gemini-2.5-Pro achieving highest accuracy (90%), followed by GPT-4.1 (88%), GPT-5.1 (87%), and Gemini-2.5-Flash (85%). The four-model ensemble reached consensus in 92% of cases, achieving 93% cTS inference accuracy. When integrated into NCCN risk stratification schema, single-model approaches achieved 89-92% accuracy, while ensemble consensus cases achieved 95% accuracy. Conclusions: The ensemble framework using multiple LLMs significantly improves cTS inference and achieves 95% NCCN risk stratification accuracy in consensus cases (92% of patients), representing a 6-percentage-point improvement over our previous single-model framework (89%). This strategy demonstrates that ensemble methods can overcome individual model limitations for complex clinical tasks and holds promise for reliable AI-assisted risk stratification.

e17003 Background: Accurate ascertainment of prostate cancer (PCa) disease states—biochemical recurrence (BCR), castration-sensitive prostate cancer (CSPC), and castration-resistant prostate cancer (CRPC)—is critical for automated clinical trial enrollment, quality improvement initiatives, and real-world evidence generation. Essential information for disease state ascertainment resides in unstructured EHR data, requiring labor-intensive manual abstraction that is inconsistent and not scalable. We developed an automated hybrid AI framework using large language models to identify PCa disease states from real-world EHR data. Methods: This retrospective study included patients with histopathologically confirmed PCa (2017-2022) and ≥3 years follow-up. Sequential EHR data—including PSA and testosterone levels, clinical notes, pathology and radiology reports—were retrieved across temporal points. BCR was identified using rule-based algorithms (post-radical prostatectomy [RP]: PSA ≥0.2 ng/mL ×2; post-radiation therapy [RT]: nadir + 2 ng/mL ×2). CRPC was determined using androgen deprivation therapy (ADT) timing, PSA kinetics, testosterone < 50 ng/dL, and radiographic progression per PCWG/EAU-ASCO criteria. GPT-4o with structured parametrized prompts extracted metastatic status and radiographic progression. Independent clinicians performed manual annotation for validation. The framework was developed on 10% of data with evaluation on 90%. External validation used a held-out cohort of clinical trial patients. Results: Among 200 PCa patients (median age 66 years [IQR 60-71]; 91% White, 93% non-Hispanic), the framework demonstrated high accuracy: PCa diagnosis (92%), RP dates (94%), RT dates (93%), and ADT initiation (92%). For disease state identification, BCR detection achieved 90% accuracy, castration status 93%, and metastatic state (M0/M1) 87%. External validation on trial-enrolled patients confirmed robust performance: 26/31 (84%) CRPC trial patients and 26/27 (96%) CSPC trial patients were correctly classified, validating the framework’s reliability for clinical trial matching. Conclusions: This hybrid LLM-based framework accurately identifies PCa disease states using real-world EHR data with minimal manual intervention. The system offers a scalable, reproducible approach for automated disease state identification, enabling streamlined clinical trial matching and enhanced real-world data curation in oncology.

Objectives To evaluate whether extracellular volume (ECV) fraction derived from contrast‐enhanced computed tomography (CT) is associated with recurrence risk in patients with treatment‐naïve high‐risk non‐muscle‐invasive bladder cancer (HR‐NMIBC) receiving intravesical gemcitabine/docetaxel (Gem/Doce) therapy. Patients and Methods This retrospective cohort study, conducted at a tertiary referral centre, included 103 patients with HR‐NMIBC treated with intravesical Gem/Doce between 2012 and 2024. All patients underwent pre‐resection contrast‐enhanced CT with an equilibrium phase (~3 min post‐injection). Tumour ECV fraction (%) was calculated using the formula: ECV = (ΔHUtumour/ΔHUaorta) × (1 − haematocrit) × 100, where Δ Hounsfield units (HU) represents the change in attenuation between non‐contrast and equilibrium phases. Regions of interest were manually placed in the tumour and abdominal aorta, and haematocrit values obtained within ±10 days were used. Associations between clinical variables, ECV, and recurrence were evaluated using Cox proportional hazards models, with discrimination assessed by Uno's C‐statistic. Results The median (interquartile range) follow‐up was 17 (10–25) months. High‐grade recurrence‐free survival (HG‐RFS) was 76% at 12 months and 69% at 24 months. CT‐derived ECV was the only variable significantly associated with HG‐RFS (hazard ratio 1.05, 95% confidence interval 1.03–1.08), demonstrating strong discrimination (C‐statistic, 0.81; P < 0.01). A 15% threshold offered the best sensitivity–specificity trade‐off (~70% each) for predicting HG‐RFS at 12 and 24 months. Patients with ECV ≤15% had markedly better HG‐RFS than those with ECV >15% at 24 months (88% vs 41%). Conclusions The CT‐derived ECV is a practical, first‐of‐its‐kind imaging biomarker associated with recurrence risk and useful for risk stratification in patients with HR‐NMIBC treated with intravesical Gem/Doce, a regimen increasingly used in the setting of global Bacillus Calmette‐Guérin shortages and intolerance. ECV can be rapidly quantified from routine contrast‐enhanced CT using standard picture archiving and communication system (PACS) tools, without additional imaging or proprietary software, making it a feasible metric for integration into risk‐adapted management pending large‐scale validation with extended follow‐up.

4614 Background: Enfortumab vedotin (EV) + pembrolizumab (pembro) is the established first-line standard of care for locally advanced/metastatic urothelial carcinoma and represents a novel neoadjuvant and adjuvant treatment strategy for patients with muscle-invasive bladder cancer (MIBC) who are ineligible for cisplatin-based chemotherapy. The randomized phase 3 KEYNOTE-B15/EV-304 study (NCT04700124) evaluates neoadjuvant EV + pembro followed by radical cystectomy plus pelvic lymph node dissection (RC + PLND) and adjuvant EV + pembro vs neoadjuvant chemotherapy followed by RC + PLND in participants (pts) with MIBC who are eligible for cisplatin-based therapy. Methods: Pts with clinical stage T2-T4aN0M0 or T1-T4aN1M0 MIBC (confirmed by central pathology and central imaging assessment) who were eligible for cisplatin-based chemotherapy and RC + PLND were randomized 1:1 to receive either 4 cycles neoadjuvant EV 1.25 mg/kg IV on days 1 and 8 + pembro 200 mg IV on day 1 Q3W, followed by RC + PLND, and adjuvant 5 cycles EV + 13 cycles pembro (EV + pembro arm) vs 4 cycles neoadjuvant gemcitabine 1000 mg/m 2 on days 1 and 8 + cisplatin 70 mg/m 2 on day 1 Q3W, followed by RC + PLND (cis + gem arm). The primary endpoint was event-free survival (EFS) by blinded independent central review. Key secondary endpoints were pathological complete response (pCR) rate by blinded central pathological review and overall survival (OS). Safety was a secondary endpoint; AEs of special interest were based on distinct prespecified lists for each drug. Results: A total of 405 and 403 pts were randomized to EV + pembro and cis + gem, respectively. Median time from randomization to the data cutoff date of October 27, 2025 was 33.6 months (range, 22.5–53.6). Baseline characteristics were generally balanced between groups. EV + pembro significantly improved EFS (median NR vs 48.5 mo; 24-mo estimated EFS rate 79.4% vs 66.2%; HR 0.53, 95% CI 0.41–0.70; 1-sided P <.0001), OS (median NR vs NR; 24-mo estimated OS rate 86.9% vs 81.3%; HR 0.65, 95% CI 0.48–0.89; 1-sided P =.0029), and pCR rate (55.8% vs 32.5%; estimated difference 23.4%, 95% CI 16.7–29.8; 1-sided P <.0001) vs cis + gem. Grade ≥3 treatment-emergent AEs occurred in 75.7% of pts with EV + pembro and 67.2% with cis + gem. Most common grade ≥3 drug-related AE of special interest for EV was skin reactions (14.1%); most common grade ≥3 AE of special interest for pembro was severe skin reactions (13.9%). Conclusions: Neoadjuvant and adjuvant EV + pembro significantly improved EFS, OS, and pCR rate compared with neoadjuvant gem + cis in pts with MIBC who were eligible for cisplatin-based chemotherapy. The safety profile of EV + pembro was consistent with prior experience with the combination. These results support EV + pembro as an effective perioperative treatment option in this setting. Clinical trial information: NCT04700124 .

4559 Background: Baseline QOL has been linked to survival in mRCC therapy clinical trials (Cella, 2024); however, there is a paucity of prospectively collected real-world QOL data across front-line management strategies. We used data from ODYSSEY to examine the relationship between baseline QOL and early death. Methods: ODYSSEY is a multi-site, prospective observational study of 468 US pts with mRCC. Pts must have mRCC (any histology); have no prior systemic therapy (ST) or be within 3 months of starting ST; and follow up at a PCORnet study site. Exclusion criteria include treatment for cancer except mRCC. The primary objective is to determine patterns of change in QOL and symptom burden of pts with mRCC. Early death was defined as death within 12 months post-enrollment. All pts with non-missing FKSI-19 or FKSI-DRS at enrollment were included. Cumulative event probabilities as a function of time were stratified by either < or ≥ median QoL value and plotted as a step function vs time since enrollment to create cumulative event curves. To determine association between QoL at enrollment and early death, separate Cox proportional hazards models (unadjusted and adjusted) for time-to-first-event as a function of baseline QoL were fit. The adjusted model controlled for the following baseline covariates: IMDC risk (favorable vs intermediate or poor), histology (clear cell vs non). Results: As of 1/5/26, 468 pts were enrolled; 425 had non-missing FKSI-19 or FKSI-DRS at enrollment (329 immediate ST, 96 deferred ST [DST]). ST included 129 IO-IO, 108 IO-TKI, 92 Other. DST included 57 Active Surveillance, 22 Local Therapy Planned, 17 Other/Missing. 209 and 190 pts have < median FKSI-19 (score = 59) and FKSI-DRS (score = 29), respectively; 216 and 235 pts have ≥ median FKSI-19 and FKSI-DRS, respectively. Those below the median have lower rates of nephrectomy, worse IMDC risk and KPS, shorter time since RCC diagnosis, and higher rates of bone metastasis (all p<0.05). Cumulative event probabilities for death at 12 months are 21% and 7% in those < and ≥ median baseline FKSI-19, respectively; and 21% and 7% in those < and ≥ median baseline FKSI-DRS, respectively. For FKSI-19, the unadjusted hazard ratio (HR) for early death is 0.55 (95% CI: 0.43 – 0.70; p<0.001) and adjusted HR is 0.65 (0.49 – 0.85; p = 0.002) per 1 SD higher score. For FKSI-DRS, the unadjusted HR for early death is 0.57 (95% CI: 0.46 – 0.71; p<0.001) and adjusted HR is 0.66 (0.51 – 0.86; p =0.002) per 1 SD higher score (higher score indicates better QOL). Conclusions: Across management strategies in our ODYSSEY cohort of systemic therapy naïve mRCC patients, better baseline QOL was associated with better survival. This is to our knowledge the largest reported real world mRCC cohort with prospectively collected PROs. Future analyses will focus on association of outcomes with longitudinal PROs. Clinical trial information: NCT04919122 .

Adoptive cellular immunotherapy (ACT) has revolutionized hematologic malignancies, yet translation to solid tumors has historically been limited. This landscape shifted significantly in 2024 with US Food and Drug Administration approvals of lifileucel, a tumor-infiltrating lymphocyte (TIL) therapy for advanced melanoma, and afamitresgene autoleucel, an engineered T-cell receptor (TCR) therapy for synovial sarcoma. These approvals mark the clinical arrival of ACT for solid tumors and highlight a rapidly expanding therapeutic landscape. Beyond these indications, multiple ACT platforms including TIL, TCR, chimeric antigen receptor (CAR) T-cell, CAR-natural killer, and CAR-macrophage therapies are under active clinical investigation across diverse solid tumor indications. Clinical experience to date has defined key barriers to efficacy, including impaired tumor trafficking, antigen heterogeneity, immunosuppressive tumor microenvironment, and limited cellular persistence. In parallel, rapid advances in cellular engineering are reshaping the field, with the development of armored constructs, logic-gated and multiantigen targeting strategies, innate immune–based platforms, and novel manufacturing approaches including allogeneic and in vivo cell engineering. As ACT enters clinical practice for select solid tumors, distinct toxicity profiles and logistical requirements necessitate careful patient selection and multidisciplinary coordination. Early biomarker testing, timely referral to specialized centers, and familiarity with evolving toxicity management frameworks are increasingly critical. Here, we seek to provide a practice-oriented framework for understanding emerging ACT platforms, clinical data, toxicity considerations, and implementation strategies relevant to contemporary solid tumor oncology care.

Abstract Background: Clear cell renal cell carcinoma (ccRCC), metastatic castration-resistant prostate cancer (mCRPC), and glioblastoma (GBM) are treatment-refractory tumors marked by high TGF-β1, a master suppressor of antitumor immunity and antigen presentation. Pharmacologic methylseleninic acid (MSA) potently inhibits TGF-β1, PD-L1, and VEGF without significant off-target toxicity. FN14, the TWEAK receptor and a metastasis driver, is selectively overexpressed in these malignancies, making it a high-specificity CAR-T target. Methods: We developed second-generation anti-FN14 CARs with CD28 or 4-1BB costimulation and CD3ζ signaling and lentivirally transduced them into human CD4+/CD8+ T cells. Cytotoxicity against FN14+ ccRCC, mCRPC, GBM lines and patient-derived tumors was measured using LDH release and xCELLigence assays. Effector function was assessed by intracellular cytokines, phenotyping, and multiplex analysis. MSA was tested at physiologic concentrations to evaluate effects on CAR-T viability, metabolism, function, and tumor susceptibility. Results: Anti-FN14 CAR-T cells mediated rapid, antigen-specific lysis across all tumor types, achieving &gt;50% specific lysis at an E:T ratio of 1:1. MSA (2.5-5 µM) reduced tumor PD-L1 and TGF-β by &gt;40% and improved CAR-T metabolic activity, enhancing IL-2 and IFN-γ secretion. Manufacturing CAR-T cells with MSA (2.5 µM) increased cytotoxicity and persistence versus controls (p &lt; 0.01). MSA’s effects were dose- and schedule-dependent at clinically compatible concentrations without detectable off-target toxicity. Conclusion: MSA augments anti-FN14 CAR-T cytotoxicity, metabolic fitness, and sustained effector function by suppressing tumor-derived TGF-β1 and PD-L1 and directly enhancing CAR-T activity. As a clinically relevant TGF-β1 inhibitor and active seleno-L-methionine metabolite, MSA represents a readily translatable strategy to reduce CAR-T exhaustion and improve efficacy in TGF-β-rich solid tumors. Citation Format: Gloria B. Kim, Obed B. Amissah, Casey R. Ager, Yousef Zakharia, Youcef M. Rustum. Methylseleninic acid enhances anti-FN14 CAR-T cell effector function and redirects cytotoxicity against TGF-β-rich kidney, prostate, and brain tumors [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 6496.

LBA630 Background: Enfortumab vedotin (EV) + pembrolizumab (pembro) is the established first-line standard of care for locally advanced/metastatic urothelial carcinoma and represents a novel neoadjuvant and adjuvant treatment strategy for patients with muscle-invasive bladder cancer (MIBC) who are ineligible for cisplatin-based chemotherapy. The randomized phase 3 KEYNOTE-B15/EV-304 study (NCT04700124) evaluates neoadjuvant and adjuvant EV + pembro followed by radical cystectomy plus pelvic lymph node dissection (RC + PLND) vs neoadjuvant chemotherapy followed by RC + PLND in participants (pts) with MIBC who are eligible for cisplatin-based therapy. Methods: Pts with clinical stage T2-T4aN0M0 or T1-T4aN1M0 MIBC (confirmed by central pathology and central imaging assessment) who were eligible for cisplatin-based chemotherapy and RC + PLND were randomized 1:1 to receive either 4 cycles neoadjuvant EV 1.25 mg/kg IV on days 1 and 8 + pembro 200 mg IV on day 1 Q3W, followed by RC + PLND, and adjuvant 5 cycles EV + 13 cycles pembro (EV + pembro arm) vs 4 cycles neoadjuvant gemcitabine 1000 mg/m 2 on days 1 and 8 + cisplatin 70 mg/m 2 on day 1 Q3W, followed by RC + PLND (cis + gem arm). The primary endpoint was event-free survival (EFS) by blinded independent central review. Key secondary endpoints were pathological complete response (pCR) rate by blinded central pathological review and overall survival (OS). Safety was a secondary endpoint; AEs of special interest were based on distinct prespecified lists for each drug. Results: A total of 405 and 403 pts were randomized to EV + pembro and cis + gem, respectively. Median time from randomization to the data cutoff date of October 27, 2025 was 33.6 months (range, 22.5–53.6). Baseline characteristics were generally balanced between groups. EV + pembro significantly improved EFS (median NR vs 48.5 mo; 24-mo estimated EFS rate 79.4% vs 66.2%; HR 0.53, 95% CI 0.41–0.70; 1-sided P <.0001), OS (median NR vs NR; 24-mo estimated OS rate 86.9% vs 81.3%; HR 0.65, 95% CI 0.48–0.89; 1-sided P =.0029), and pCR rate (55.8% vs 32.5%; estimated difference 23.4%, 95% CI 16.7–29.8; 1-sided P <.0001) vs cis + gem. Grade ≥3 treatment-emergent AEs occurred in 75.7% of pts with EV + pembro and 67.2% with cis + gem. Most common grade ≥3 drug-related AE of special interest for EV was skin reactions (14.1%); most common grade ≥3 AE of special interest for pembro was severe skin reactions (13.9%). Conclusions: Neoadjuvant and adjuvant EV + pembro significantly improved EFS, OS, and pCR rate compared with neoadjuvant gem + cis in pts with MIBC who were eligible for cisplatin-based chemotherapy. The safety profile of EV + pembro was consistent with prior experience with the combination. These results support EV + pembro as an effective perioperative treatment option in this setting. Clinical trial information: NCT04700124 .