Browse Articles

Neuropathy following chimeric antigen receptor T-cell therapy: Incidence and associated risk factors.

e23448 Background: Neurotoxicity associated with chimeric antigen receptor T-cell (CAR-T) therapy, particularly immune effector cell–associated neurotoxicity syndrome (ICANS), is well described and affects up to 20-70% of patients. However, data on the development of peripheral and cranial neuropathies post CAR-T are limited. We aimed to evaluate the incidence of neuropathy post CAR-T and identify associated risk factors. Methods: We conducted a retrospective cohort study using the TriNetX database (2014–2025). We included adults (age ≥18y) who received any one of the six CAR-Ts (Tisagenlecleucel, Axicabtagene, Brexucabtagene, Lisocabtagene, Idecabtagene and Ciltacabtagene) and had no pre-existing neuropathy within 1 y prior to receiving CAR-T. Neuropathy occurring within 3 months and 1 y post–CAR-T were identified using ICD codes for peripheral or cranial neuropathies. Diabetic, hereditary, and paraneoplastic neuropathies were excluded. We then compared CAR-T recipients who developed neuropathy to those who did not. Univariate analyses were performed to assess associations with demographics, malignancy type, CAR-T product, prior medications, comorbidities and ICANS. Results: Among 2472 CAR-T recipients without pre-existing neuropathy, 357 (14.4%) developed neuropathy within 1 y, including 175 (7%) cases occurring within 3 months. Of the 357 patients, 286 (80%) developed peripheral neuropathy and 71 (20%) developed cranial neuropathy; facial nerve involvement being the most common cranial neuropathy (n = 52, 73%). Patients who developed neuropathy had a mean age of 62.8 y, were predominantly male (64%), non-Hispanic (79%) and White (79%). In univariate analysis, when compared to CAR-T recipients who did not develop neuropathy, we identified factors significantly associated with neuropathy. These included- primary cancer diagnosis of multiple myeloma (34% vs 25%, P < 0.001 ), and solitary plasmacytoma (6% vs 3%, P = 0.02 ); use of CAR-T Ciltacabtagene (14% vs 8%, P < 0.001 ) ; use of cisplatin(8% vs 5%, P < 0.01 ), oxaliplatin (11% vs 7%, P < 0.01 ), gemcitabine (14% vs 8%, P < 0.001 ), carfilzomib (12% vs 6%, P < 0.001 ), talquetamab (6% vs 3%, P = 0.02 ); comorbidities like alcohol use disorder (6% vs 4%, P = 0.03 ), primary hypertension (65% vs 56%, P < 0.01 ), obesity (32% vs 25%, P < 0.01 ), hyperlipidemia (46% vs 37%, P < 0.01 ), and CKD (25% vs 19%, P < 0.01 ). Development of ICANS (18% vs 13%, P < 0.01 ) was also significantly associated with subsequent neuropathy. Conclusions: Neuropathy occurs in 14.4% of patients within 1 y of CAR-T, with half of the cases within 3 months. Peripheral neuropathy predominates, and facial nerve involvement is the most common cranial neuropathy. Risk is higher among patients with multiple myeloma, those receiving Ciltacabtagene and those who develop ICANS. Early neurologic surveillance may be beneficial, particularly in high-risk patients during early post CAR-T period.

Phase I/IIa dose finding study of triplet regimen of relatlimab (RELA) ipilimumab (IPI), and nivolumab (NIVO) in first-line therapy of metastatic melanoma (MM) (TRINITY).

TPS9603 Background: Approved immune checkpoint blockade (ICB) agents for the treatment (tx) advanced, unresectable metastatic melanoma (MM) include NIVO (anti-PD1) alone or in combination with IPI (anti-CTLA-4) or RELA (anti-LAG3) and have transformed outcomes for patients with this disease. While doublet combinations induce higher rates of durable disease control and modestly improved survival vs single agent, higher doses of IPI induce higher objective response rates (ORR) also increase grade ≥3 immune-related adverse events. Notably, there is no established dose-response relationship for NIVO or RELA. IPI has a demonstrated role in expanding the T-cell receptor (TCR) repertoire, modulating suppressive T cell populations. NIVO+RELA regulates exhaustion of activated T cells, improving effector function. Together, these suggest potential synergy with triplet ICB. Results reported in RELATIVITY-048 (NIVO 480 mg Q4W + RELA 160 mg Q4W + IPI 1 mg/kg Q8W) demonstrated impressive efficacy (ORR 59%) and seemingly improved progression free and overall survival (PFS, OS) over previously reported doublet tx. The RELATIVITY-048 study design did not include dose escalation (DE) of IPI. Our team seeks to identify the optimal dose and schedule if IPI to combine with NIVO+RELA, identify the recommended phase II dose (RP2D) for this triplet ICB treatment and, ultimately, seeking to maximize the clinical benefit while maintaining a toxicity profile comparable to approved doublet regimens. Methods: In this single center, investigator initiated, PhI/IIa study, our team is evaluating triplet ICB (NCT06683755) in untreated, unresectable MM pts. Pts will receive NIVO 480mg + RELA 160mg IV Q4W (FDA approved dose). IPI DE will begin at 0.5mg/kg Q4W, escalating to 2mg/kg Q4W x4 induction doses. Pts will receive maintenance tx with NIVO+RELA Q4W at the approved dose. The maximum tolerated dose and the RP2D will be determined using a Bayesian optimal interval design, accruing 12-18 pts in DE. An additional 12-18 pts will be enrolled at RP2D to determine the ORR (primary objective) by RECIST 1.1. Secondary objectives include PFS, OS, and correlative analysis. Non IPI containing prior adjuvant or neoadjuvant tx is allowed if the last dose has been >6 months. Pts with asymptomatic brain metastasis on ≤10mg corticosteroids are allowed. Safely biopsiable lesions are required for pts enrolled in the PhII portion. By optimizing the dose and schedule of IPI in combination with NIVO+RELA, we seek to maximize clinical benefit while maintaining a toxicity profile comparable to approved regimens. This study is currently accruing at MD Anderson Cancer Center. Clinical trial information: NCT06683755 .

Time to transplant by sociodemographic group following MELD 3.0 implementation in hepatocellular carcinoma.

e23208 Background: The MELD 3.0 policy was implemented in July 2023 to improve mortality prediction and address allocation inequities through updated creatinine parameters and sex-based coefficients. Although designed to enhance equity and efficiency, its impact on hepatocellular carcinoma (HCC) candidates - who rely heavily on exception points - remains unclear. We examined whether MELD 3.0 improved time to transplant equitably across sociodemographic groups. Methods: Using UNOS/OPTN data, we retrospectively analyzed 5,896 adult HCC candidates listed during symmetric 21-month periods before (n = 3,435) and after (n = 2,461) MELD 3.0 implementation. Time to transplant was assessed using adjusted restricted mean failure time (RMFT). Difference-in-differences models were used to estimate differential policy effects by sex and race/ethnicity. Results: Overall, MELD 3.0 was associated with shorter wait times for transplantation. Compared with the pre-implementation era, adjusted RMFT decreased by 102.4 days for male candidates and 93.3 days for White candidates ( p < 0.001). However, subgroup effects were heterogeneous. A new sex-based disparity emerged, with female candidates experiencing a 16.6-day longer wait time than male candidates ( p < 0.001), whereas no difference existed pre-implementation. Racial disparities also evolved unevenly. The Black–White gap in wait time widened from 30.7 days pre-MELD 3.0 to 67.9 days post-implementation ( p < 0.001). In contrast, the Asian–White gap narrowed and was no longer statistically significant (37.6 to 16.3 days, p = 0.18). Hispanic candidates experienced partial improvement, though a significant gap persisted (15.9 days, p < 0.001). Conclusions: MELD 3.0 improved overall time-to-transplant efficiency for HCC candidates but was associated with heterogeneous equity effects. Despite system-wide gains, a new disadvantage emerged for women and racial disparities for Black candidates widened, while disparities for Asian candidates were mitigated. These findings highlight the need for continued policy refinement to ensure equitable access to transplantation.

Outcomes of radical radiotherapy in de novo low-burden metastatic hormone-sensitive prostate cancer vs prostate radiotherapy alone.

5097 Background: The current standard of care (SoC) for metastatic hormone-sensitive prostate cancer (mHSPC) is androgen deprivation + systemic therapy. Oligometastatic HSPC, however, represents a unique disease entity following the STAMPEDE trial, where de novo mHSPC patients with a CHAARTED-defined low disease burden (DN-mHSPC-LV) had a hazard ratio of death of 0.68 after prostate radiotherapy (PRT). Despite the SABR-COMET results of a survival benefit with stereotactic RT on metachronous oligometastatic sites of predominantly prostate patients, radical RT (RRT) in the DN-mHSPC-LV setting has not been studied. This study analysed the outcomes of such patients receiving RRT to all disease sites vs PRT alone. Methods: DN-mHSPC-LV patients receiving RRT to the prostate and all oligometastatic sites (Arm A) vs PRT alone (Arm B) in a Hong Kong tertiary hospital between 1 st January 2012 and 1 st January 2023 were reviewed. Outcomes of progression-free survival (PFS), overall survival (OS) and treatment-related toxicity (TrT) were analysed using the Kaplan-Meier and multivariable Cox-proportional hazards models. Results: 127 eligible patients (60 in Arm A & 67 in Arm B) were identified. The median numbers of oligometastatic sites of 1-3 and = > 4 were balanced between the two arms. The median Gleason score was 8 in both arms. Median T/N/M staging was mostly balanced but more patients in Arm B had N1 disease. Presenting PSA (ng/mL) was higher in Arm A (38.5) than Arm B (27.9). Arm A received up to 76Gy/38# to the prostate + a boost of up to 67.5Gy or stereotactic RT to the oligometastatic sites. Arm B received 55-60Gy/20# to the prostate alone. Patients were started on long-term androgen deprivation therapy. Median follow-up was 40.2 months and 30.3 months in Arms A and B, respectively. Median PFS was not reached (NR) in either of the treatment arms. On multivariate analysis, 5-year PFS was 86.1% vs 50.3% in Arms A and B, respectively ( P = 0.0173; HR 0.213; 95% CI 0.059-0.761). OS data were not yet mature and there was no difference in TrT between the arms. Conclusions: To our knowledge, this is the first study comparing radical RT against the SoC of prostate RT alone in DN-mHSPC-LV. Radical RT was well tolerated and is significantly cheaper than next-generation hormone agents in prostate cancer. 5-year PFS was significantly better with radical RT vs prostate RT alone. PFS is an established surrogate for OS in prostate cancer. This study therefore justifies prospective studies on RRT for DN-mHSPC-LV. More importantly, as prospective trials with an OS endpoint in prostate cancer will take close to a decade to mature, this study provides interim evidence to support radical treatment in this group of patients, in whom cure or long-term control may be achieved, where prospective trials are anticipated to take much longer to complete.

A phase II, exploratory trial of cadonilimab (AK104) plus gemcitabine-cisplatin (GC) as a conversion therapy for patients (pts) with unresectable or systemic treatment–failed biliary tract cancer (BTC).

e16177 Background: PD-(L)1 inhibitor combined with GC is an established first-line treatment for unresectable BTC. However, most pts still experience disease progression, and no universally accepted second-line regimen exists. Surgery remains the only potentially curative option. Cadonilimab is a tetrameric bispecific antibody both targeting PD-1 and CTLA-4. This study evaluates the efficacy and safety of AK104 plus GC as a conversion therapy for unresectable or systemic treatment-failed BTC pts. Methods: This ongoing single-center phase II trial enrolls two cohorts. Cohort A (treatment-naïve, unresectable) received up to 8 cycles of AK104 (10mg/kg) + GC (Q3W), with imaging every two cycles to assess resectability. Pts converted to resectable undergo surgery followed by AK104 or combined with GC (combination cycles at investigator’s discretion); unresectable pts switch to AK104 monotherapy—all until PD. Primary endpoints were R0 resection rate and OS. Cohort B (prior systemic-therapy failed) received AK104+GC (investigator-determined cycles) followed by AK104 maintenance, with OS as the primary endpoint. Secondary endpoints for both cohorts were PFS, ORR, DCR and safety. Results: As of Nov 2025, 31 pts were enrolled (14 treatment-naïve, 17 systemic therapy-failed), all with ECOG PS 1. Baseline characteristics for treatment-naïve pts: ICC 28.6%, ECC 42.8%, GBC 28.6%, stage III–IV 85.7%; for systemic therapy failed pts: ICC 11.8%, ECC 41.1%, GBC 47.1%, stage III-IV 94.1%, with 94.1% having received prior ICIs plus chemotherapy. Median maximum tumor diameter was ~32 mm in both cohorts. Among 22 evaluable pts (11 per cohort), the overall ORR was 36.4% and mPFS was 4.8 mo (95% CI 2.6-NE); mOS was not reached at a median follow-up of 5.5 mo (95%CI 4.3-11.2). In treatment-naïve unresectable pts, ORR was 54.5% with an R0 conversion rate of 18.2% (2/11), and median PFS and OS were not reached. In systemic therapy-failed pts, ORR was 18.2%, mPFS was 3.3 mo (95% CI 2.0-9.7), and mOS was 6.9 mo (95% CI 4.1-NE). Grade ≥3 TRAEs occurred in 48.4% (15/31) pts. The most common any grade TRAEs were decreased lymphocyte count (55%), decreased white blood cell count (42%), and anemia (42%). No treatment-related deaths occurred. Conclusions: Cadonilimab combined with GC showed promising anti-tumor activity and manageable safety in both unresectable and systemic treatment-failed BTC pts. Longer follow up is required to further evaluate the efficacy and safety. Clinical trial information: MR-31-24-050538. Efficacy outcomes. Evaluable pts n, (%) treatment-naïven=11 systemic therapy-failed n=11 Alln=22 ORR 6（54.5） 2（18.2） 8（36.4） PR 6（54.5） 2（18.2） 8（36.4） SD 3（27.3） 3（27.3） 6（27.3） PD 2（18.2） 6（54.5） 8（36.4） DCR 9（81.8） 5（45.5） 14（63.6） R0 conversion rate 2（18.2） / / mPFS, mo (95%CI range) NR 3.3 (95% CI 2.0–9.7) 4.8 (95% CI 2.6–NE) mOS, mo (95%CI range) NR 6.9 (95% CI 4.1–NE) NR

Addition of 18F-DCFPyL PSMA PET to MRI to avoid unnecessary biopsy in men on active surveillance: Final results from a phase II diagnostic trial (PROMPT).

5125 Background: Multiparametric MRI (mpMRI) is used in active surveillance (AS) for prostate cancer (PC) but has limited specificity. This prospective trial evaluated whether combining 18F-DCFPyL PSMA PET with mpMRI improves the detection of clinically significant PC (csPC) in patients on AS and reduces unnecessary biopsies prompted by false-positive mpMRI. Methods: In this IRB-approved, single-arm Phase II study, 91 patients on AS underwent PSMA PET and mpMRI on a dedicated PET/MRI scanner followed by systematic and targeted biopsies. PET and mpMRI were interpreted independently. The primary outcome was the diagnostic accuracy of mpMRI versus the combined PSMA PET/MRI for csPC. Results: Among 246 targeted lesions, the area under the curve (AUC) for combined PSMA PET/MRI was (0.77) compared to mpMRI alone (0.70) (p=0.2). PSMA PET/MRI showed improved specificity (64.6% vs. 27.5%) with a comparable negative predictive value (89.1% vs. 89.7%). Of 57 csPC lesions, 6 (10.5%) were uniquely identified by the combination. Half of the biopsies for false-positive mpMRI or PET lesions (122/246; 50%) encompassing 17/91 (19%) of patients could have been avoided with combined imaging. Management changed in 9% of patients based on PET/MRI. Conclusions: In this first-in-field trial the addition of PSMA PET to mpMRI with dedicated PET/MRI imaging enhances sensitivity in detecting csPC and may reduce the need for unnecessary biopsies in AS patients. Incorporating PSMA PET/MRI into clinical practice could refine surveillance protocols and support biopsy omission in select men with negative imaging. The single health-system nature of the trial may limit the generalizability of our findings. Diagnostic accuracy of the mpMRI, PSMA PET/MRI, and for the combination, assessed using the “AND” operator. a Parameter MRI, PIRADS≥3 PSMA-PET, Likert≥3 MRI, PIRADS≥4 PSMA-PET, Likert≥4 MRI PIRADS≥3 ANDPET Likert≥3 MRI PIRADS≥3 ANDPET Likert≥4 MRI PIRADS≥4 ANDPET Likert≥3 MRI PIRADS≥4 ANDPET Likert≥4 Accuracy 41.9% 48.0% 63.8% 63.8% 66.7% 72.8% 78.5% 79.3% Sensitivity 89.5% 84.2% 71.9% 73.7% 73.7% 64.9% 63.2% 57.9% Specificity 27.5% 37.0% 61.4% 60.8% 64.6% 75.1% 83.1% 85.7% PPV 27.1% 28.7% 36.0% 36.2% 38.5% 44.0% 52.9% 55.0% NPV 89.7% 88.6% 87.9% 88.5% 89.1% 87.7% 88.2% 87.1% Number of lesions (% of 257) 188 (76) 167 (68) 114 (46) 116 (47) 109 (44) 84 (34) 68 (28) 60 (24) a The detection of clinically significant cancer was considered on per-lesion analysis. MRI = magnetic resonance imaging; PET = positron emission tomography; PI-RADS = Prostate Imaging Reporting and Data System; PSA = prostate-specific antigen; PSMA = prostate-specific membrane antigen. PPV= positive predictive value, NPV = negative predictive value.

The impact of body composition on cardiac function in women with early breast cancer taking cardiotoxic chemotherapy.

12021 Background: Anthracyclines and HER2-directed therapies have improved survival in early breast cancer (EBC), but cardiotoxicity remains a concern for long-term survivors. Cardiovascular disease may ultimately exceed breast cancer as a cause of death, highlighting the need for better understanding of therapy-related cardiac dysfunction. Body composition abnormalities are linked to adverse cardiovascular outcomes but are not well captured by body mass index are understudied. We evaluated longitudinal associations between baseline CT-derived body composition and echocardiography measures of cardiac structure and function. Methods: This is a retrospective analysis of women with stage I–III EBC treated with anthracyclines and/or trastuzumab at the University of Pennsylvania. Eligible patients had a pre-treatment and ≥1 follow-up echocardiogram and pre-treatment abdominal CT or PET-CT imaging for body composition assessment. L3-based segmentation quantified skeletal muscle mass, muscle density (SMD), subcutaneous and visceral adiposity (SAT, VAT), and adipose tissue density (SATD, VATD). Core-lab echocardiographic outcomes included systolic function (LVEF, GLS) and diastolic function (E/e′). Using multivariable-adjusted generalized estimating equations (GEE) models, we evaluated longitudinal associations between baseline body composition (per 1-SD increase) and two complementary outcomes for each metric: (1) absolute values over time (overall burden) and (2) change from baseline (within-participant worsening). Results: Across 182 participants, the median age was 49.5 years (Q1-Q3 41–60), HER2-positive disease was present in 64 (35.4%), estrogen receptor–positive disease in 114 (62.6%), and triple-negative disease in 37 (20.4%). Anthracycline-based therapy was administered to 150 (82.4%), radiation therapy to 140 (76.9%), and mastectomy was performed in 106 (58.6%). Baseline hypertension, hyperlipidemia, and diabetes were present in 56 (30.8%), 43 (23.6%), and 12 (6.6%), respectively. In multivariable-adjusted GEE models, higher baseline VAT was associated with a greater decline in LVEF over time (−2.1 LVEF points per 1-SD increase; 95% CI −4.1 to −0.1; p=0.04). We next evaluated diastolic function, where higher baseline VAT was associated with a greater increase in E/e′ over time (β 0.38 per 1-SD increase; 95% CI, 0.01 to 0.74; p=0.04). In models evaluating absolute E/e′ over time, higher VAT (β 0.51; 95% CI, 0.11 to 0.91) and SAT (β 0.37; 95% CI, 0.06 to 0.59), and lower SMD (β −0.34; 95% CI, −0.66 to −0.02) and VAT density (β −0.36; 95% CI, −0.67 to −0.05) were associated with higher E/e′. No adjusted associations were observed for longitudinal change in GLS (all p>0.05). Conclusions: Baseline visceral adiposity was associated with worsening in longitudinal changes in systolic and diastolic function during cardiotoxic breast cancer therapy.

Pseudoprogression after neoadjuvant immunotherapy for resectable NSCLC: Incidence, PET/CT patterns, and differentiation.

8032 Background: Neoadjuvant immunotherapy combined with chemotherapy is now the standard of care for resectable non-small cell lung cancer (NSCLC). However, evaluating treatment response remains challenging due to the unique response pattern of pseudoprogression associated with immune checkpoint inhibitors (ICIs), which may lead to unnecessary cancellation of curative surgeries. This study aimed to characterize pseudoprogression and investigate strategies for distinguishing pseudoprogression from true tumor progression in the neoadjuvant ICI setting. Methods: This retrospective study included 283 patients with resectable NSCLC who underwent neoadjuvant ICI therapy (n=235) or targeted therapy (n=48) between January 2021 and December 2023. PET-CT scans were acquired at baseline and before surgery. Radiological and metabolic responses were evaluated using RECIST 1.1 and EORTC criteria, respectively. Pathological responses were assessed according to IASLC guidelines. Multivariate logistic regression and receiver operating characteristic analyses were performed to identify predictors of pseudoprogression and evaluate diagnostic performance. Results: Among 283 patients (ICI 235, TKI 48), 9.2% (26 patients) were classified as having progressive disease (PD) by RECIST 1.1, all in the ICI cohort. According to EORTC on PET-CT, progression increased to 25.1% (71 patients). Despite radiologic progression, all 26 patients underwent R0 resection; pCR and MPR rates were 53.8% and 11.6%. Among patients without radiologic progression, pCR and MPR were 30.0% and 19.1%, with no differences in R0 rate or pathological response (p = 1.00 and 0.06). Among 71 patients with progressive metabolic disease (PMD) on PET/CT, 69 (97.2%) achieved R0 resection; 27 (38.0%) achieved pCR and 12 (16.9%) achieved MPR, with similar outcomes by PMD status (p = 1.00 and 0.53). In both the ICI cohort and overall population, primary tumor progression was associated with lower MPR rates than lymph node or distant progression (16.7%, 61.2%, and 100%; p = 0.001) and worse EFS and OS. In the TKI cohort, imaging-defined progressive or stable disease was associated with lower MPR and worse survival. Pseudoprogression occurred in 7.2% by CT, increasing to 16.2% with PET-CT, predominantly nodal. A ≥53.1% reduction in primary-tumor SUVmax identified pseudoprogression (AUC 0.865; sensitivity 73.8%; specificity 88.6%). A multivariable model combining ΔSUVmax, longest diameter change (%), smoking status, and histology showed excellent discrimination (AUC 0.935). Conclusions: Pseudoprogression is common in resectable NSCLC treated with neoadjuvant ICI therapy, presenting as nodal or distant lesions without primary tumor progression. A multifactorial model incorporating primary tumor SUVmax changes provides a robust tool for distinguishing pseudoprogression and guiding clinical decision-making.

Virtual multiplex immunofluorescence characterization of spatial immune architectures and tertiary lymphoid structures in high-grade serous ovarian cancer.

5585 Background: GigaTIME ( Cell , 2026) is a deep learning–based virtual multiplex immunofluorescence (mIF) model that has been validated using TCGA datasets and enables spatial tumor immune microenvironment (TIME) profiling directly from H&E whole-slide images (WSIs). This study aims to evaluate the feasibility of GigaTIME for characterizing the spatial TIME of high-grade serous ovarian cancer (HGSOC) using public datasets. Methods: We analyzed UBC-OCEAN and CPTAC-OV HGSOC H&E WSIs using GigaTIME to generate virtual 21-channel (e.g., DAPI, CK, CD3, CD8, CD20, and Ki67) mIF pixel-level inferred probability maps. Exploratory analyses of 5×5 stitched regions of interest (ROI, 2560×2560 pixels ≈ 1.28×1.28 mm) were performed to detect tertiary lymphoid structures (TLS) via density-based spatial clustering of applications with noise (DBSCAN), identified as CD20⁺ clusters (major axis ≥100 μm) adjacent to CD3⁺ zones, with maturity defined by intra-cluster Ki67⁺ proliferation. Tumor-immune interaction metrics were analyzed in ROIs using Group Degree Centrality (GDC, fraction of tumor cells contacting immune cells) and Cluster Co-occurrence Ratio (CCR, observed/expected contacts). CPTAC-OV HGSOC H&E WSI–derived virtual mIF features were correlated with bulk proteomics using Spearman’s rank correlation. Integrated K-means clustering was performed on the combined cohorts after Z-score normalization. Results: The UBC-OCEAN cohort (N=195) exhibited an immune-inflamed phenotype with high tumor-immune engagement (mean GDC [mGDC] 0.25) and structural aggregation (mean CCR [mCCR] 1.41), whereas CPTAC-OV (N=222) displayed an immune-excluded phenotype with minimal contact (mGDC 0.05) and near-random distribution (mCCR 1.03). While total TLS density was comparable (UBC-OCEAN 3.92 vs. CPTAC-OV 3.60), UBC-OCEAN exhibited a twofold higher density of mature TLS compared to CPTAC-OV (1.83 vs. 0.77). Integrated clustering (k=2, silhouette=0.71) revealed that High-TLS (N=39) tumors paradoxically exhibited significantly lower mGDC (p=0.016) and mCCR (p=0.001) despite higher TLS density, indicating a spatially segregated and organized immune architecture. In CPTAC-OV validation, virtual CD3 intensity showed a significant positive correlation with bulk proteomics (N=17, ρ =0.547, p=0.023), supporting the biological validity of the T-cell-based spatial metrics. Conclusions: GigaTIME is feasible for evaluating comprehensive spatial TIME including TLS and demonstrates the heterogeneous immune landscape of HGSOC, providing spatial insights that are not captured by bulk proteomics alone. This warrants the utility of virtual mIF-based profiling as a scalable tool to identify distinct immune phenotypes for personalized immunotherapy at minimal cost.

Global affordability of the ASCO Annual Meeting: A longitudinal analysis over 5 years.

e23160 Background: International oncology conferences shape scientific careers, guideline adoption, and global research networks. We systematically quantified how affordability of attending the ASCO Annual Meeting (AM) varies across world regions/income groups using income-equivalent metrics. Methods: We conducted a longitudinal, multi-component cost analysis of ASCO AM from 2021–2025 for physicians across all World Bank regions and income strata. Total direct costs included travel, accommodation, registration, visa fees, and foregone clinical income from five missed working days. Travel costs were modeled using representative regional hub cities; U.S. airfare data were obtained from U.S. Department of Transportation reports, while international airfare trends were adjusted using the U.S. CPI Airline Fares Index (2025 baseline). Accommodation and daily expenses were estimated using Choose Chicago hotel performance data and U.S. GSA per diem rates. Registration fees were obtained from publicly available sources, with missing values estimated by linear interpolation. Affordability was assessed using two complementary metrics: the Cost-to-Income Ratio (CIR), expressed as income-equivalent time in months, and a GNI-based Affordability Ratio (AR_GNI). Physician incomes were estimated using World Bank GNI per capita adjusted with WHO Global Health Workforce Statistics–based wage multipliers. Results: ASCO attendance required > 10-fold more income-equivalent time for LMIC than HIC physicians. Mean CIR was 0.44 for HIC physicians (≈0.4 m of income) versus 4.35 for LMIC physicians, exceeding four months of income. When expressed as income-equivalent time by region, total attendance costs corresponded to 0.2 months of physician income in North America, compared with 11 months in South Asia and over 25 months in Sub-Saharan Africa. Regional and income-stratified CIR and AR_GNI values are summarized in Table 1. Registration fees accounted for less than 15% of total costs across most regions, indicating that income-adjusted registration alone does not address affordability barriers. Conclusions: International oncology meetings are essential for professional development and knowledge exchange, and current ASCO affordability measures contribute meaningfully to access. However, overall attendance costs differ substantially across regions and income groups, reflecting the influence of travel-related and opportunity costs beyond registration fees. This analysis provides data to inform continued initiatives to enhance equitable global participation. AR_GNI and CIR by world regions and incomes. Region/Income Level AR_GNI CIR expressed as Income-equivalent time (month) North America 1.09 0.2 Europe/Central Asia 1.98 0.5 East Asia/Pacific 4.45 1.4 Latin America/Caribbean 5.08 1.5 Middle East/North Africa 7.96 2.7 South Asia 16.81 11.2 Sub-Saharan Africa 33.30 25.3 HIC 1.99 0.44 LMIC 8.3 4.3

National perspectives on artificial intelligence use in cancer care: Confidence, adoption, and barriers.

e13648 Background: Artificial intelligence (AI) in cancer care offers opportunities to reduce administrative burden, improve efficiency, and support care delivery, yet it prompts questions around safety, equity, and clinical judgment. To assess real-world perspectives, the Association of Cancer Care Centers (ACCC) conducted a national survey examining how oncology professionals engage with AI, their perceived value, and the barriers to adoption. Methods: ACCC, with guidance from an expert committee, created a national online survey of multidisciplinary staff at US cancer programs from May and August 2025. The 26-item survey assessed experiences and perceptions of AI, organizational adoption and governance, and implementation barriers and facilitators. Completed responses were analyzed using descriptive and stratified statistics in Stata 18, with qualitative data examined using rapid inductive thematic analysis. Results: Respondents (N=168) from 36 states included 61% in care delivery and 39% in administrative/operational roles, primarily from community (46%) and NCI-designated or academic (44%) programs. Care delivery staff were less confident than administrative/operations respondents in describing AI use (21% vs 4% “not at all [confident]”), discussing benefits of AI use (18% vs 7%), and critically evaluating AI systems (26% vs 13%). Administrative/operations respondents reported higher confidence contributing to AI implementation (45% vs 32%) and a higher likelihood to adopt/expand AI for treatment planning (52% vs 34%), clinical trial matching (49% vs 32%), and prior authorization (52% vs 35%). Rural respondents were more likely to express concern about declines in patient-provider communication (63% vs 22% suburban and 30% urban), as were those without AI experience (54% vs 24%). Academic respondents rated improving clinical decision-making/diagnostic accuracy, and establishing performance benchmarks and evaluating AI systems, as higher motivators for integration (0.82 vs 0.41; 0.53 vs 0.26), whereas community respondents rated patient engagement higher (0.36 vs 0.13). Respondents with AI experience reported a higher likelihood to adopt/expand AI for chatbots and virtual assistants (52% vs 16%), real-time alerts (35% vs 14%), personalized patient education (46% vs 27%), and individualized patient navigation (38% vs 11%). Qualitative responses reinforced these findings, highlighting stakeholder engagement, evidence of effectiveness, training, integration, and governance as critical for AI implementation. Conclusions: The results reveal broad AI use despite limited AI tools and governance implementation. Respondents were confident acknowledging AI’s limitations but less confident in evaluating its utility, highlighting opportunities for ACCC to support adoption, governance, and AI integration focused on efficiency and patient outcomes.

Validation of a droplet digital PCR test for monitoring ESR1 ctDNA variants in blood.

e13066 Background: Estrogen receptor 1 (ESR1) mutations represent a critical mechanism of acquired resistance to aromatase inhibitor (AI) therapy in hormone receptor-positive (HR+) breast cancers. These mutations may occur in 20-40% of patients during treatment. Recent results from the landmark SERENA-6 Phase III trial demonstrated the potential clinical utility of ctDNA-guided treatment switching. This study represented the first global Phase 3 trial to demonstrate the potential clinical utility of using ESR1 ctDNA monitoring for treatment guidance and management of breast cancers. Methods: We have validated a droplet digital PCR (ddPCR) test for the detection of ESR1 variants in plasma-derived cfDNA. The ESR1 ddPCR test targets the 11 most prevalent ESR1 variants, approximately 97% coverage of ESR1 mutations in breast cancer. Analytical sensitivity studies were performed using synthetic DNA controls diluted into wild-type background DNA at varying input concentrations. Serial dilutions were tested to establish the limit of detection (LOD) across different DNA input amounts, with variant allele frequencies (VAF) ranging from 0.025% to 0.2%. Clinical validation is being conducted in a blinded study comparing performance of the new assay versus two standard of care NGS panels. Results: Our studies demonstrated robust analytic performance of the ESR1 ddPCR assay across multiple input concentrations. With 30 ng cfDNA input, we achieved analytical sensitivity ranging from 0.025-0.5% VAF, potentially enabling sensitive detection of low-level ESR1 mutations that may be present in early resistance. When DNA input was reduced to 10 ng, reflecting real-world clinical sample limitations, the assay maintained excellent sensitivity of 0.025-0.1% VAF. The assay also demonstrated high specificity with no false positive calls in wild-type controls. We observed assay sensitivity (0.2%-9.9%) and specificity (100%) in interim analysis of clinical specimens (n = 31). Conclusions: We have successfully developed a highly sensitive and specific ddPCR test for the detection of multiple ESR1 variants in plasma. With sensitivity as low as 0.025% VAF using 10 ng of cfDNA and 100% clinical concordance with reference NGS assays, this approach represents a robust technology for monitoring cfDNA in liquid biopsies. Importantly, the affordability of PCR-based testing and its suitability for decentralized implementation offer a practical path for broader global adoption of ctDNA-guided treatment strategies, as highlighted by emerging clinical utility data.

Mapping the Reactivity of the C═C Bond of Cyclic Enol Ether Derivatives of Sugars: Nucleophilicity Parameters of Glycals

ABSTRACT Glycals are unsaturated sugar derivatives constituting a large family of polyhydroxylated synthons which are widely exploited as synthetic intermediates in the synthesis of natural products and as final molecules for biochemical applications. We report the first investigation in the understudied area of structure–reactivity relationship and quantitative mapping of reactivity in endo ‐ and exo ‐glycals compounds. For quantifying the nucleophilicity of the C═C double bond of a series of endo ‐ and exo ‐glycals, we performed kinetic investigations of their C─C bond formation with reference electrophiles of known electrophilicity parameters. Fast spectroscopic techniques experiments (stopped‐flow and laser‐flash photolysis) enabled us to determine the rate constant of hundreds of reactions, which allows the first comprehensive mapping and structure–reactivity analysis of the nucleophilic reactivity of this wide family of cyclic enol ethers. Quantum–chemical calculations corroborate with kinetic investigations and highlight the crucial role of ring strain variations to explain the relative nucleophilicity of endo ‐ and exo ‐glycals. We examined also the influence of substitution and showed that alkoxy substituents decrease nucleophilicity through inductive effects and hyperconjugation.

Corrigendum to “Heat generation and Joule dissipation influence on Magnetohydrodynamic Cu- H2O and Al2O3-H2O nanofluid convection with nanoparticle volume fraction and ramped and isothermal wall temperature: A finite element approach” [Next Nanotechnol. 8 (2025) 100261]

Transportation access as a structural determinant of quality of life and symptom burden among cancer patients.

1536 Background: Growing evidence links the built environment to cancer outcomes, with NCCN recommendations emphasizing the importance of transit access in cancer care. Cancer patients, especially at advanced ages, are vulnerable to inadequate transit access due to frequent medical appointments and treatment-related physical and financial toxicity. This study examined the relationship between public transit access, quality of life (QoL), pain, and fatigue among adults with cancer. Methods: We analyzed data from 31,126 adults with cancer enrolled in the NIH All of Us cohort, which is designed to include populations historically underrepresented in biomedical research. Cancer cases were identified using ICD and SNOMED codes. Multivariable ordinal logistic regression was used to assess associations between distance to public transit and self-reported QoL, pain, and fatigue adjusted for age, sex, race/ethnicity, income, education, and insurance. Similar models examined sociodemographic predictors of transportation-related healthcare delays. Results: Decreased public transit access was associated with worse QoL, pain, and fatigue (Table 1). Transportation-related delays in care were more likely among older adults aged 65+ (aOR=0.30, p=0.02) compared to 18-25 year olds; women (aOR=1.13, p=0.03); unemployed individuals (aOR=1.13, p<0.001); and those with lower income (aOR=7.60, p<0.001). Participants identifying as Black (aOR=1.65, p<0.01), Hispanic (aOR=1.27, p<0.01), or multiracial (aOR=1.73, p<0.01) also reported higher odds of transportation-related delays. Conclusions: Reduced public transit access is associated with worse quality of life, symptom burden, and care delays among adults with cancer. Providers can utilize transportation coordination as a tool for improving patient outcomes, particularly in the functional domains of pain and fatigue and among underserved populations. Further, policymakers should identify and address gaps in public transit access, yielding downstream benefits for population health, quality of life, and survivorship among cancer patients. Consideration for transportation should be integrated into treatment plans for all cancer patients to reduce financial toxicity and treatment costs while improving outcomes. Adjusted odds of worse health outcomes by decreased transit access: Adjusted association between public transit access and pain, fatigue, and quality of life among adults with cancer. a Outcome aOR (95% CI) b P-value Low Quality of Life 1.09 (1.00, 1.19) 0.04* Moderate/High Pain 1.12 (1.06, 1.18) <0.001**** Moderate/High Fatigue 1.14 (1.08, 1.20) <0.001**** a Adjusted for age group, sex, race/ethnicity, employment, income, insurance. b Public transit stop >15 minutes away. *p < 0.05, ****p ≤ 0.001.

Real-world characteristics, treatment patterns, and outcomes of renal cell carcinoma at a high-complexity center in southwestern Colombia.

e16548 Background: Real-world data on renal cell carcinoma (RCC) in Latin America are limited, despite the widespread adoption of tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs) in metastatic disease. Differences in access to treatment, patient characteristics, and clinical practice may influence outcomes compared with clinical trial populations. This study aimed to describe clinical characteristics, treatment patterns, and outcomes of patients with RCC at a Colombian referral center. Methods: We conducted a retrospective cohort study including patients with histologically confirmed RCC diagnosed between 2014 and 2024 at Fundación Valle del Lili, Cali, Colombia. Demographic, clinicopathologic, and treatment-related variables were collected. Outcomes included overall survival (OS), progression-free survival (PFS), and disease-free survival (DFS). Survival outcomes were estimated using the Kaplan-Meier method, and univariable Cox proportional hazards models were used to evaluate factor associated with OS. Results: A total of 1,064 patients met inclusion criteria. Median age was 60 years (IQR: 52-69) and 59.4% were male. Clear cell RCC was the most frequent histological subtype (72.8%, n = 775). At diagnosis, 86.3% (n = 918) had localized disease, of whom 13.3% (n = 122) developed recurrent disease during follow-up, while 13.7% (n = 146) presented with de novo metastatic disease. Among these 268 patients with advanced disease, 227 (84.7%) received systemic therapy. First-line regimens consisted of TKI monotherapy (52.8%), ICIs plus TKI (27.7%), and dual ICIs (17.1%). Among patients progressing on first-line therapy, 124 (54.6%) received second-line treatment, predominantly TKI monotherapy (45.2%) and ICI monotherapy (21.8%). Median OS was 31.1 months (95% CI 18.2-40.9) in patients with de novo metastatic disease. Among patients with metastatic or recurrent disease who received systemic therapy, median PFS was 41.9 months (95% CI 32.4-55.8). Median DFS in patients with localized disease was 156.7 months (95% CI 137.0-176.7). In univariable analysis, age ≥65 years (HR 2.29, 95% CI 1.66–3.17, p < 0.0001) and ECOG performance status 2–4 (HR 3.20, 95% CI 1.89–5.42; p < 0.0001) were associated with worse OS. Conclusions: This large real-world cohort provides a comprehensive characterization of patients with renal cell carcinoma treated at a high-complexity referral center in southwestern Colombia. The observed heterogeneity in clinical presentation, treatment patterns, and survival outcomes highlights the challenges of RCC management in routine practice and underscores the value of locally generated real-world evidence to inform clinical decision-making in Latin American populations.

Sociodemographic disparities and clinical outcomes in chimeric antigen receptor (CAR) T-cell therapy: A real-world study at a diverse academic cancer center.

e23435 Background: Chimeric Antigen Receptor T-cell (CAR-T) therapy is an effective treatment for relapsed/refractory hematologic malignancies. However, disparities in access, toxicity burden, and outcomes may exist for patients from minoritized and underserved backgrounds. This study evaluated real-world CAR-T therapy outcomes at a single academic cancer center, focusing on socioeconomic and demographic predictors of treatment response and toxicity. Methods: We performed a retrospective cohort study of 101 adult patients who received commercial CAR-T therapy (axicabtagene ciloleucel, brexucabtagene autoleucel, ciltacabtagene autoleucel) at USC Norris Comprehensive Cancer Center between 2022–2025. Demographic, geographic, clinical, and treatment-related variables were extracted from REDCap. Multivariable logistic regression and chi-square/Fisher’s exact tests were used to assess associations between sociodemographic variables—including race/ethnicity, insurance status, Social Vulnerability Index (SVI), and distance from the transplant center—and key outcomes: treatment response, toxicity (CRS/ICANS), ICU admission, and delays to infusion. Results: The cohort had a median age of 62 years; 63% were male and 38% identified as Hispanic. Diffuse Large B-Cell Lymphoma (DLBCL) (44%) was the most common diagnosis, followed by Multiple Myeloma (22%) and Acute Lymphoblastic Leukemia (ALL) (21%). Insurance coverage included commercial (49%), Medicare (27%), and Medicaid (23%). In multivariable analyses, female sex was significantly associated with lower odds of achieving a treatment response compared to males (OR 6.69 for males, p = 0.038). Public insurance (Medicare or Medicaid) was associated with increased risk of severe neurotoxicity (ICANS grade ≥3, p = 0.012) and ICU admission ( p = 0.035). DLBCL diagnosis was significantly associated with reduced odds of response ( p < 0.05), highlighting disease-specific variation in CAR-T efficacy. No statistically significant associations were found between outcomes and SVI, race/ethnicity, or distance to the transplant center. Time from referral to infusion also did not significantly differ by ethnicity, insurance type, or SVI quartile. Conclusions: In this single-center cohort, female patients and those with DLBCL had significantly lower odds of response to CAR-T therapy. Public insurance was associated with higher neurotoxicity and ICU utilization. These findings underscore potential disparities in CAR-T outcomes based on sex, insurance status, and disease subtype, even within an academic setting. Larger, multi-center studies are warranted to confirm these associations and inform equity-focused strategies in cellular therapy delivery.

Machine learning/artificial intelligence in liquid-biopsy–based cancer screening.

e22576 Background: Early diagnosis is vital for cancer prognosis, yet high-level screening evidence (Grade A/B) currently exists only for breast, cervical, lung, and colorectal cancers. Liquid biopsy has emerged as a promising alternative, but gaps remain in prediction accuracy and Tumor of Origin (TOO) identification. While ML/AI is established in radiology, its role in liquid biopsy remains under-recognized. Integrating ML/AI with multi-omics data offers significant potential to elevate positive predictive values and bridge the gap between technological potential and clinical application. Methods: We systematically identified over 50 commercial and developmental cancer screening kits via technical files, peer-reviewed publications, and clinical reports. Test features were extracted to evaluate diagnostic performance. We categorized underlying ML/AI algorithms based on their architectural framework, data integration capacity, and clinical validation metrics. Results: No multi-cancer early detection (MCED) kits have received full regulatory approval; most are currently accessible only as Laboratory Developed Tests (LDTs). Evaluation reveals that market-leading products still rely on preliminary frameworks, such as Logistic Regression (LR) or Random Forests (RF). While robust, these models lack the complex feature extraction capabilities of deep learning. A key finding is the prevailing technical conservatism in the diagnostic sector. Mature products prioritize interpretability and robustness over complex "black-box" models, as regulatory bodies (e.g., FDA, NMPA) favor transparent decision boundaries and predictable performance. Consequently, the primary hurdle is developing advanced AI that meets these rigorous standards for clinical transparency. Conclusions: The diagnostic sector remains a technologically conservative niche, lagging in cutting-edge AI integration. Bridging this gap requires deep synergy between biomedical researchers, kit developers, and algorithm engineers. Transitioning from traditional models to advanced, yet interpretable, AI is essential to unlock the full potential of liquid biopsy for next-generation cancer screening. List of selected cancer screening kits and their algorithms. Dx Country Target Algorithm Galleri US Methylation Penalized Multinomial Logistic Regression Shield US Methylation + Fragmentomics + DNA Mutation LR Freenome US Multi-omics LR / SVM / Ensemble Delfi US Fragmentomics LR / LASSO / PCA OverC CN Methylation LR / SVM PanSeer CN Methylation LR / LASSO Mercury CN Fragmentomics XGBoost / GLM / GBM / RF Genecast CN Methylation, Fragmentomics KNN / SVM / LR / GBDT / RF GeneSolutions SG Methylation LR / LASSO

Molecular characterization and treatment patterns of lung cancer in solid organ transplant (SOT) recipients.

e20699 Background: SOT recipients are at increased risk of malignancy due to immunosuppression. Lung cancer is among the most common post-transplant cancers, yet molecular characteristics and treatment patterns remain poorly defined. As targeted therapies shape non-small cell lung cancer (NSCLC) care, understanding tumor genomics is essential. Methods: We conducted a single-institution retrospective study of SOT recipients diagnosed with lung cancer 2007-2025. Demographic, transplant, histologic, molecular, treatment, and survival data were abstracted. Molecular analyses focused on NSCLC, including PD-L1 tumor proportion score (TPS) and next-generation sequencing (NGS). Overall survival (OS) was assessed by molecular and treatment subgroups. Results: 85 SOT recipients developed lung cancer (73 NSCLC - 43 adenocarcinoma [adenoCa], 30 squamous). Median age at diagnosis was 66 years, median time from transplant to diagnosis was 5.4 years, and median follow-up was 9 months. Lung was the most common transplanted organ (33%), followed by kidney (28%) and heart (22%). 53 NSCLC patients had TPS assessed, and 49 had NGS. TPS did not differ significantly by histology; most tumors had low TPS ( < 1%) (44% adenoCa vs 60% squamous), while intermediate TPS (10–50%) was more frequent in adenoCa (25% vs 15%). Actionable mutations were identified more often in adenoCa patients (42% vs 3%, p = 0.0001), most commonly KRAS G12C (15%) and classical EGFR mutations (14%). There was no significant difference in OS between patients with actionable versus non-actionable mutations (1-year OS 68% vs 52%). In contrast, OS differed significantly by treatment strategy (p < 0.0001), with 1-year OS of 50% for chemoimmunotherapy, 43% for targeted therapy and 14% for chemotherapy. Conclusions: NSCLC after SOT has a molecular landscape comparable to the general population, supporting comprehensive molecular profiling. However, TPS was lower and co-mutation patterns differed, suggesting immunosuppression may affect tumor biology. Despite frequent actionable mutations, survival was driven more by treatment strategy than molecular status, highlighting potential underuse of targeted approaches, low uptake of immunotherapy despite strategies to improve post-transplant safety, and the need for further study on transplant-specific tumor biology. Actionable mutations and systemic therapy patterns by stage of NSCLC. Characteristics Stage I (n=34) Stage II (n=7) Stage III (n=7) Stage IV (n=29) Total (n=77) p value N (%) N (%) N (%) N (%) N (%) Actionable Mutation 0.2 No 22 (64.7%) 6 (85.7%) 7 (100%) 23 (79.3%) 58 (75.3%) Yes 12 (35.3%) 1 (14.3%) 0 (0%) 6 (20.7%) 19 (24.7%) Systemic Therapy 0.0005 Chemotherapy 2 (5.9%) 3 (42.9%) 2 (28.6%) 16 (55.2%) 23 (29.9%) Chemoimmunotherapy 0 (0%) 0 (0%) 0 (0%) 2 (6.9%) 2 (2.6%) Targeted Therapy/Chemo 4 (11.8%) 1 (14.3%) 1 (14.3%) 7 (24.1%) 13 (16.9%) None 28 (82.4%) 3 (42.9%) 4 (57.1%) 4 (13.8%) 39 (50.6%)

Alpha-fetoprotein as a biomarker of response to first-line chemoimmunotherapy in metastatic hepatocellular carcinoma.

e16247 Background: Atezolizumab plus bevacizumab (Atezo/Bev) and durvalumab plus tremelimumab (Durva/Tremi) are preferred first-line regimens for patients with unresectable or metastatic hepatocellular carcinoma (HCC). Comparative real-world data to guide treatment selection remains limited. Elevated AFP is a known poor prognostic marker; however, its role as a predictive biomarker for differential treatment benefit remains understudied. Methods: We conducted a retrospective cohort study using the TriNetX database to identify adults with metastatic HCC who received first-line Atezo/Bev or Durva/Tremi after January 2022 with AFP > 1000 ng/ml. Propensity score matching was used to balance baseline demographics, liver disease characteristics, and comorbidities. Primary outcome was overall survival (OS). Secondary analyses evaluated treatment-related toxicities and outcomes in patients with high baseline AFP. Adverse events were identified using diagnosis codes and laboratory abnormalities within 3 months of treatment initiation. Results: Before matching, 712 patients treated with Atezo/Bev and 296 treated with Durva/Tremi met inclusion criteria. After propensity score matching, 269 patients per cohort were included, with well-balanced baseline characteristics. In the matched cohort, the survival probability in the atezo/bev vs durva/tremi arms was as follows- at 1 year- 50.38% vs 56.54% (p = 0.494); at 3 years-22.78% vs 27.29% (p = 0.542); at 5 years- 22.78% vs 27.29% (p = 0.542). Across sensitivity analyses the results consistently showed a trend towards improved survival with durva/tremi, though differences were not statistically significant. Toxicity profiles differed between regimens but were consistent with existing literature. Conclusions: In this real-world analysis of first-line therapy for metastatic HCC with high AFP, Durva/Tremi demonstrated a trend towards improved overall survival, albeit not statistically significant. This hypothesis-generating analysis raises the possibility that AFP may be clinically informative in decision-making for the first line of treatment. These findings need to be further verified in larger studies.