Browse Articles
Discover research articles across all indexed journals
The Liver We Share
A medical student's first encounter with a hepatectomy comes full circle when he becomes a living liver donor for his wife with metastatic pancreatic cancer, reshaping his understanding of medicine as they navigate the uncertainty of her diagnosis together.
CLEAR-ACCESS: Barriers and facilitators to accessing symptom support among culturally and linguistically diverse cancer patients receiving treatment.
e13520 Background: Patients from culturally and linguistically diverse (CALD) backgrounds experience inequities in cancer outcomes, partly due to barriers in accessing timely care for treatment-related side effects. Symptom support such as Symptom and Urgent Review Clinics and access to Clinical Nurse Consultant-led care provide early intervention and can reduce unplanned emergency department presentations and admissions, yet CALD patients are under-represented amongst service users. This study examined the barriers and facilitators influencing access to symptom support for CALD patients receiving systemic anti-cancer therapy. Methods: A qualitative study was conducted across two Australian cancer centres. Focus groups and semi-structured interviews were undertaken with CALD patients, caregivers and community representatives from Vietnamese, Mandarin and Greek-speaking backgrounds as well as healthcare professionals (HCPs). Data were collected in participants’ preferred language. Transcripts were analysed using inductive reflexive thematic analysis. Themes were mapped to Levesque et al.’s framework for patient-centred access to healthcare. Results: Thirty-five participants took part: patients (n = 20), caregivers (n = 4) and community representatives (n = 7) and HCPs (n = 4). Among patient, caregiver and community representatives, 26% (10/35) identified as Chinese, 14% (5/35) as Greek and 46% (16/35) as Vietnamese. The median age of patients, caregivers, community representatives and HCPs were 61 years (51-79), 58 years (51-65), 54 years (37-69) and 47 years (38-62) respectively. Four interrelated themes influenced access to and engagement with symptom support: (1) Availability of family facilitate access : relatives were central in contacting services but reliance on them created delays when they were unavailable or patients wished to avoid burdening them; (2) Limited awareness of available services and supports reflected the challenges of language-discordant, information-dense encounters with inconsistent interpreter access (3) Limited availability of accessible, in-language information impairs healthcare navigation and (4) Patient confidence in communication with HCPs shape engagement: trust and shared language with general practitioners supported symptom discussion, often positioning primary care as a preferred access point. Conclusions: Access to symptom support among CALD patients was shaped by the knowledge of the services and how understandable and reachable those services were as well as how confidently participants could engage with them. These insights directly inform ADVANCE-ACCESS, a co-design study across four Australian health services focused on developing interventions to improve access to symptom support services for CALD cancer patients.
Palliative care engagement, advance care planning, and healthcare utilization in hepatobiliary malignancies.
e24046 Background: Although advance care planning (ACP) is central to goal-concordant oncology care, its implementation in clinical practice is often absent or delayed, particularly in hepatobiliary cancers, where symptom burden is high and there are complex care needs. We examined how palliative care involvement relates to ACP documentation and healthcare utilization in patients with hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA). Methods: We conducted a retrospective cohort study of adults with advanced HCC or CCA evaluated at a tertiary hepatobiliary oncology clinic (2023-2025). The primary exposure was palliative care involvement (ever vs never) with referral timing categorized as early or late based on the median time from first oncology visit. Primary outcomes were advance directive documentation and any code-status update after the initial oncology visit. Group comparisons used Chi-squared/Fisher’s exact tests and Wilcoxon rank-sum tests for continuous variables. Results: Among HCC patients (n = 252), those seen by palliative care had higher rates of ACP documentation, including advance directives (69.6% vs. 32.5% p < 0.001) and code-status updates (60.9% vs. 13.2% p < 0.001). Patients who received palliative care had higher total hospital days (median 14.2 vs 6.45 p < 0.001) and a greater number of hospitalizations (median 2.5 vs 2.0, p < 0.001), while ICU admission rates, ICU days, and emergency department visits did not differ significantly between groups. Among patients who received palliative care, later referral was associated with a higher likelihood of hospitalization (97.1% vs 82.6%, p = 0.009) and a longer time from first oncology visit to hospitalization (median 159.0 vs 34.6 days, p < 0.001). Among patients with CCA (n = 122), palliative care involvement was associated with higher rates of advanced directives (73.7% vs 43.5% p = 0.002) and code-status updates (56.6% vs 17.4% p < 0.001). Patients seen by palliative care were more likely to be hospitalized (92.1% vs 60.9%, p < 0.001) and had more total hospital days (median 19.0 vs 8.0 days, p = 0.015), while other utilization metrics were similar. Conclusions: Across HCC and CCA, palliative care engagement was strongly associated with improved ACP documentation following the first oncology visit. Higher hospitalization rates among patients seen by palliative care likely reflect referral bias towards those with greater symptom burden and complexity. These findings overall support the early integration of palliative care alongside oncology care to normalize ACP and ensure alignment with patient goals.
Practice disparities in palliative care utilization across different solid malignancies in the United States: Geographic trends and impact on hospital outcomes.
e24090 Background: Palliative care is a specialized approach for patients with life limiting illness. Utilization disparities reflect differences in disease trajectory, symptoms, practice patterns, access, and socioeconomic or cultural factors. Research is needed to improve care delivery and promote equity. While racial disparities have been studied for some cancers, gaps remain across common tumors and the impact of geographic disparities. Methods: The National Inpatient Sample Database (2016–2020) was analyzed using ICD-10 codes: GI (colon, rectal, anal, gastric, esophageal), GU (kidney, bladder, prostate), lung, breast, and pancreatic cancers. Cohorts were stratified by palliative care use. Primary outcomes included healthcare costs (TOTCHG), hospital length of stay (LOS), and discharge to nursing home (NH) services. Multivariate regression, adjusted for confounders, assessed the impact of palliative care on these outcomes and geographic disparities (Northeast as reference). Results: Palliative care use was highest in lung, followed by pancreatic, GI, breast, and GU cancers. Utilization was higher among Whites (lung 66%, pancreas 69%, GI 65.7%, breast 64%, GU 72%) and lower among Black (15–20%) and Hispanic patients (8–10%) (p < 0.001). Lung cancer had the shortest LOS and lowest TOTCHG; other cancers had longer stays and higher charges. Mortality was highest in GU cancers and lowest in lung cancer. Regionally, use was highest in the South and lowest in the West. Discharge to NH was more frequent in the South and less in the West. LOS > 1 week was less likely in the Midwest and West; TOTCHG was lower in the Midwest, higher in the West. Mortality was higher in the West and lower in the South and Midwest vs the Northeast. Conclusions: Disparities in palliative care utilization, mortality, and resource use exist across cancer types, racial groups, and regions. Low use in the West was associated with higher mortality, while earlier integration in other regions correlated with lower mortality and costs. Findings underscore the need for timely, equitable integration of palliative care to improve outcomes and optimize resources. Geographic disparities in outcomes among palliative care users. Mortality (aOR, 95% CI) Discharge to NH (aOR, 95% CI) Length of Stay > 7d (aOR, 95% CI) Total Hospital Charges (Adjusted β, USD, 95% CI) Midwest: 0.77 (0.69–0.86), p <0.001 Midwest: 1.57 (0.57–4.28), p =0.374 Midwest: 0.90 (0.73–0.87), p =0.008 Midwest: −$19,114 (−$27,875 to −$10,352), p <0.001 South: 0.63 (0.57–0.70), p <0.001 South: 3.00 (1.33–6.79), p =0.008 South: 0.93 (0.87–1.00), p =0.054 South: −$4,868 (−$13,678 to $3,940), p =0.279 West: 1.25 (1.10–1.42), p <0.001 West: 0.68 (0.61–0.76), p <0.001 West: 0.90 (0.83–0.97), p =0.007 West: +$37,623 ($27,002–$48,245), p <0.001 Adjusted for demographics, cardiac comorbidities, and hospital characteristics.
A meta-analysis and systematic review of cardiac adverse events in patients with HER2-altered non–small cell lung cancer (NSCLC) enrolled in clinical trials.
e24026 Background: HER2 inhibitors showed meaningful clinical activity in HER2-altered NSCLC. HER2 inhibitors-related cardiotoxicity remains inconsistently reported in NSCLC trials unlike in breast cancer trials. Cardiotoxicity can have a significant effect on patients’ morbidity and mortality. We conducted this meta-analysis to determine the incidence of cardiotoxicity with HER2-inhibitors in NSCLC. Methods: We performed a systematic search on MEDLINE, Embase, CENTRAL, Scopus, and Web of Science from inception to 10/10/25. Phase I-III clinical trials enrolling adults with HER2-altered NSCLC treated with HER2 inhibitors were included. The primary outcome was the pooled incidence of cardiotoxicity. Prespecified assumptions were applied for unreported cardiotoxicity outcomes, including conservative handling of missing data, zero-event imputation when appropriate, and classification of absent grade 5 events as no treatment-related cardiac deaths. Pooled incidence estimates were generated using random-effects meta-analyses of proportions with logit transformation and restricted maximum likelihood estimation in the R metafor package. In exploratory analyses, oncologic outcomes were compared between arms with and without reported grade ≥3 cardiotoxicity using nonparametric tests. Results: 48 arms from 39 trials were eligible, including 2035 patients. Most arms (79.2%) and patients (71.4%) were from phase II clinical trials. 19 (48.7%) trials reported cardiotoxicity. Patients were treated with tyrosine kinase inhibitors (TKIs) on 20 (42%) arms, monoclonal antibodies (mABs) on 15 (31%) arms, and antibody-drug conjugates (ADCs) on 15 (31%) arms. The incidence of severe grade cardiotoxicity was 5.1% (95CI, 2.4–10.3; I², 0%). No grade 5 cardiotoxicity was reported. The incidence of key cardiac events did not significantly differ by drug class. Pooled ORR was 34.6% (95CI, 29.3–40.3). ORR differed by drug class (p = 0.03), with the highest observed with ADC (44.5%, 95CI 35.0–54.4), followed by TKI (34.2%, 95CI 26.2–43.3), and mAB (30.3%, 95CI 24.3–37.0). Median OS was 13.8 (IQR 10.1–15.0) months. There were no significant differences in OS, PFS, and ORR between arms with versus without reported grade ≥3 cardiotoxicity. Conclusions: While not reported in all trials, the incidence of severe cardiotoxicity was rare in NSCLC patients treated with HER2-directed therapy. It was not associated with poor outcomes. Incidence of key cardiac adverse events. Cardiac adverse event arms, n Pooled incidence, % (95CI) I², % Cardiomyopathy 41 9.6 (3.8–14.1) 0 LVEF decline 42 9.2 (4.4–18.4) 39.1 Atrial fibrillation 41 5.5 (2.3–12.7) 0 Cardiorespiratory arrest 41 5.3 (0.7–29.7) 0 QTc prolongation 41 3.1 (5.7–10.4) 41.6 Acute heart failure 41 1.3 (4.1–12.0) 0 Abbreviations: 95CI, 95% confidence interval; LVEF, left ventricular ejection fraction.
Updated efficacy and safety of SHR-1501, an IL-15RαFc superagonist, with or without Bacille Calmette Guerin (BCG) for high-risk non-muscle invasive bladder cancer (NMIBC): A phase 1/2 study.
4626 Background: SHR-1501, an IL-15 agonist fusion protein composing of a humanized antibody Fc region fused with IL-15 and IL-15Rα sushi domain, demonstrated promising efficacy, well tolerance, and acceptable safety in alone or in combination with BCG in patients with BCG-naive and BCG-unresponsive high-risk NMIBC ( ASCO 2025 ). Here, we report the updated results of this phase 1/2 study (NCT05410730). Methods: In the dose-escalation phase 1a and 1b parts, SHR-1501 monotherapy (200, 400, and 600 μg) or SHR-1501 (600 μg) in combination with BCG (120 mg) was administered to patients with high-risk NMIBC. In the phase 2 part, patients with BCG-naive NMIBC (cohort A), BCG-unresponsive NMIBC carcinoma in situ (CIS; cohort B), and BCG-unresponsive high-grade Ta/T1 NMIBC without CIS (cohort C) were enrolled to receive SHR-1501 (600 μg) plus BCG (120 mg). During the induction phase, all patients received weekly intravesical study treatment for 6 weeks. In the maintenance phase, instillations were administered weekly for three weeks at months 3, 6, 12, 18, and 24 following the initial induction dose. Primary endpoints were dose-limiting toxicity (DLT), maximum tolerated dose (MTD), and recommended phase 2 dose in phase 1a and 1b parts; and was complete response (CR) rate for cohort B and 12-mo disease-free survival (DFS) rate for cohorts A and C in phase 2 part. Results: As of Oct 31, 2025, 112 patients were enrolled (n=8 in phase 1a; n=6 in phase 1b; n=30, 25, and 43 in cohorts A, B, and C in phase 2). The median follow-up duration was 23.0 months (range 3.5-25.6) in patients with BCG-naive NMIBC, 6.5 months (range 2.6-21.2) in patients with BCG-unresponsive NMIBC CIS, and 13.5 months (range 2.5-23.1) in patients with BCG-unresponsive high-grade Ta/T1 NMIBC without CIS. In cohort B, the overall CR rate was 80.0% (20/25), the median DFS was 12.0 months (95% CI 6.0-NR). The 12-mo DFS rate was 90.3% (95% CI, 72.8-96.8) in patients with BCG-naive NMIBC and 62.7% (95% CI, 44.9-76.1) in patients with BCG-unresponsive high-grade Ta/T1 NMIBC without CIS. The 18-mo DFS rates were 90.3% (95% CI, 72.8-96.8) and 58.2% (95% CI, 39.6-72.9), respectively. Treatment-related adverse events (TRAEs) and grade 3 TRAEs occurred in 90 (86.5%) and 19 (18.3%) of 104 patients with SHR-1501 + BCG. The most common TRAEs were urinary tract infection (62.5%) and pollakiuria (35.6%). No TRAEs led to death. Conclusions: This updated analysis confirms the promising efficacy and manageable safety profile of SHR-1501 monotherapy or in combination with BCG in BCG-naive and BCG-unresponsive high-risk NMIBC patients. Two randomized, controlled phase 3 trials are underway in both BCG-unresponsive and BCG-naïve, high-risk NMIBC populations, with the recommended dose of 600 μg SHR-1501 plus BCG. Clinical trial information: NCT05410730 .
S2433: Randomized phase III study of second-line chemotherapy with or without panitumumab for KRAS wild type, locally advanced, or metastatic pancreatic adenocarcinoma.
TPS4264 Background: Pancreatic ductal adenocarcinoma (PDA) is characterized by driver mutations in KRAS , TP53 , CDKN2A , and SMAD4 . KRAS wild type (WT) tumors, present in 5-10% of PDA, often harbor alternative mitogen-activated protein kinase (MAPK) pathway drivers or other targetable molecular alterations (e.g., microsatellite instability [MSI-high], DNA damage repair mutations), enabling targeted therapy in approximately 30% of cases; however, most lack actionable drivers and require a tailored treatment approach. EGFR is overexpressed in 30%-70% of PDA. Prior studies of EGFR-targeted therapy demonstrated modest overall survival (OS) benefit in unselected populations, with retrospective analyses suggesting benefit confined to KRAS WT PDA. Recent data showed survival benefit from the anti-EGFR antibody nimotuzumab when added to 1 st line gemcitabine vs gemcitabine alone in KRAS WT PDA. These findings provide a strong rationale to evaluate anti-EGFR mAb in combination with multi-agent chemotherapy in KRAS WT PDA in the United States. Methods: This randomized (1:1) phase III study will enroll 94 pts to investigator choice 2 nd -line chemotherapy – 5 fluorouracil (5FU) plus irinotecan or nanoliposomal irinotecan following 1 st -line gemcitabine-based therapy, or gemcitabine plus nab-paclitaxel following 1 st -line 5FU-based therapy – with or without panitumumab. Eligibility requires KRAS WT and BRAF V600E WT status by tumor tissue-based next-generation sequencing and no known mutations in PTEN , NRAS , EGFR extracellular domain exons 1-16, no amplifications of HER2 and MET , and no gene fusions of RET , NTRK1 , and ALK . One prior line of chemotherapy for locally advanced or metastatic PDA is permitted. Pts with cancers harboring molecular alterations (e.g., MSI-high, ROS fusion) may have received prior targeted therapy, and past maintenance therapy with a PARP inhibitor does not count as a line of therapy. The primary endpoint is OS; secondary endpoints include safety, progression-free survival, overall response rate, disease control rate, and duration of response. Quality of life will be assessed using the Functional Assessment of Cancer Therapy – General (FACT-G), Functional Assessment of Chronic Illness Therapy (FACIT) Item GP5, and selected patient-reported outcome (PRO)-CTCAE items. Blood and tumor samples will be collected for correlative studies. Standard eligibility criteria apply. Assuming panitumumab improves median OS from 6.3 to 12.6 months, 84 eligible patients provide 90% power with a one-sided alpha of 0.05. Kaplan-Meier methods will estimate OS, with early stopping for efficacy or futility. This study is open to accrual (NCT06998940). Clinical trial information: NCT06998940 .
Quantum mechanics-based multi-tensor AI/ML as predictor of patients' overall survival, gene targets, and drug responses from their glioblastoma tumors' whole genomes.
3020 Background: The drug failure rate has increased to ~95%, despite the growth in targeted therapies. As clinical trials demonstrated, a targeted gene alone does not predict whether patients have longer life expectancy in response to the drug. As studies with model organisms showed, the effect of the drug, and the mechanisms underlying it, depend on the entire multi-ome. But multi-omic data are small-cohort, noisy, and high-dimensional, i.e., extremely difficult to model. Methods: We have developed our artificial intelligence and machine learning (AI/ML) to overcome these challenges [doi: 10.1073/pnas.0530258100, 10.1158/1538-7445.AM2025-CT227]. We demonstrated our algorithms in the unsupervised modeling of, e.g., whole genomes of 85 astrocytoma patients. Mechanistic interpretation showed that the modeling blindly removed batch effects, separated normal demographic variations, and discovered a disease-specific genome-wide pattern of DNA copy-number alterations. This pattern was used to derive an actionable predictor of patients’ overall survival (OS) and gene targets to sensitize their tumors. We computationally validated both the predictor and the modeling in federated studies of mutually-exclusive sets of 59–251 patients. The modeling repeatedly discovered a representation of the predictor in every study, across astrocytoma grades II, III, and IV, i.e., glioblastoma (GBM), patients. We experimentally validated the predictor in a clinical trial of 79 GBM patients, initially retrospectively, and, in a four-year follow up, also prospectively [doi: 10.1063/1.5142559, 10.1145/3624062.3624078, 10.1200/JCO.2024.42.16_suppl.e14028]. In all the cohorts, the predictor, with 75–95% concordance with OS, was more accurate than all standard-of-care indicators. With 100% reproducibility among Complete Genomics, Illumina, and Ultima whole-genome sequencing, and > 99% when including Affymetrix and Agilent DNA microarrays, the predictor was also the most precise. Results: Here, we describe functional genomic experimental validation of both a predicted gene target and the predicted tumors’ responses to the targeting. Guide RNAs were designed and a lentiviral CRISPR-Cas9 all-in-one vector was utilized to knock out the modeling-predicted target METTL2A . Knockout validation at the protein level was performed using Western blot. Knockout in the patient-derived GBM cell lines U-87 MG and U-118 MG resulted in significantly attenuated cell viability and proliferation. The level of attenuation was significantly different between the cell lines, consistent with their whole genome-based predicted responses. Conclusions: Our quantum mechanics-based multi-tensor AI/ML solved the 75-year-old problem of correctly predicting — patients’ OS, drug responses, and gene targets — from their GBM tumors' whole genomes.
Response to HiDAC/mitoxantrone/venetoclax (HMV) in patients with relapsed/refractory (R/R) acute myeloid leukemia (AML).
6527 Background: Outcomes for patients (pts) with R/R AML are dismal, especially after failure of Venetoclax (Ven)-based regimens. Effective strategies are urgently needed particularly in the absence of targetable mutations. An early-phase study of HMV demonstrated a 75% remission rate in pts with R/R AML, however the majority of these pts were Ven-naïve (Ruhnke et al, EHA 2024). We aimed to study a real-world cohort of pts with R/R AML treated with HMV. Methods: A retrospective cohort of 16 pts with R/R AML treated with HMV at the University of Chicago (2020–2025) was analyzed. Clinical and disease characteristics were collected; European Leukemia Net (ELN) 2022 criteria were used for risk stratification and response assessment. Overall response rate (ORR) was defined as complete response (CR) + CR with partial hematologic recovery (CRh) + CR with incomplete hematologic recovery (CRi). Overall survival (OS) was estimated using Kaplan-Meier methods. Results: Sixteen pts with R/R AML were included; 2 pts had ALL with lineage switch to AML. Median age at diagnosis was 54.5 years. Amongst the 14 pts with AML at initial diagnosis, 28.6%, 14.3%, and 50.0% had favorable, intermediate, and adverse ELN 2022 risk, respectively. Our cohort of pts was heavily pre-treated with 56.2% receiving ≥3 lines of therapy prior to HMV; two pts had received prior allogeneic hematopoietic stem cell transplant (allo-SCT). The ORR was 43.8%; in addition, 6.3% attained morphologic leukemia-free state (MLFS), 6.3% obtained a partial response (PR), and 43.7% had no response/not evaluable (NR/NE). Among pts that received prior intensive AML chemotherapy (n=8), the ORR was 50.0% with HMV. The ORR was 41.6% in pts with prior Ven exposure (n=12). Median OS from HMV initiation was 5.1 months (95% CI: 2.2-9.2) with two deaths within 30 days of HMV initiation; 62.5% of patients received subsequent therapy and 37.5% proceeded to allo-SCT. Two patients are alive in remission at time of data analysis; both underwent allo-SCT after HMV. Conclusions: In this heavily pretreated R/R AML cohort, HMV produced encouraging response rates and facilitated transition to allo-SCT in a subset of pts. These data warrant further evaluation of HMV as an R/R option even in pts with prior Ven exposure. Patient characteristics and efficacy outcomes. Median Age at Time of Receiving HMV (N=16) 54.5 years (Range 21-66) Gender (N=16) Male 69%; Female 31% # of Lines of Therapy Received Prior to HMV (N=16) 1 Line 25%; 2 Lines 18.8%; ≥3 Lines 56.2% Best Response to HMV (N=16) CR 31.2%; CRh 6.3%; CRi 6.3%; MLFS 6.3%; PR 6.3%; NR/NE 43.7% Best Response to HMV in Patients Who Received Prior AML-Based Intensive Chemotherapy (N=8) CR 25%; CRh 12.5%; CRi 12.5%; MLFS 0%; PR 0%; NR/NE 50% Best Response to HMV in Patients Who Received Prior Venetoclax (N=12) CR 25%; CRh 8.3%; CRi 8.3%; MLFS 0%; PR 8.3%; NR/NE 50%
A phase 1 study of the safety, pharmacokinetics, and pharmacodynamics of intermittent oral PRTX007, an IRF7-biased TLR7 agonist prodrug.
e14606 Background: Systemically active toll-like receptor 7 (TLR7) agonists have demonstrated immune-activating potential but are often limited by inflammatory toxicities. PRTX007 is an orally administered small-molecule TLR7 agonist prodrug designed to selectively activate IRF7-driven interferon responses while minimizing pro-inflammatory cytokine activation. Study PRTX007-002 evaluated an intermittent weekly dosing regimen to increase the magnitude of interferon-pathway engagement following favorable safety and pharmacodynamic findings with every-other-day dosing in a prior Phase 1 study. Methods: PRTX007-002 was a randomized, placebo-controlled Phase 1 study in healthy volunteers. Fifteen subjects received either oral PRTX007 750 mg (n = 12) or placebo (n = 3) on a 3-days-on/4-days-off schedule for two weekly cycles plus one additional dose (seven total doses). Safety, pharmacokinetics (PK), and pharmacodynamic (PD) responses were assessed. Results: PK properties were consistent with prior clinical experience, with no accumulation of PRTX007 or PRX034 upon repeated dosing. Of the 15 subjects enrolled ten (66.7%) subjects experienced TEAEs considered related to study intervention. There were no serious TEAEs. A total of 3 of 15 (20.0%) participants discontinued study drug due to on-target, interferon-like symptoms which resolved within 24 hours with conservative supportive care (2 events each of headache and inflammation, and 1 event of chills, all in the PRTX007 750 mg treatment group). Median peak IFN-alpha levels in PRTX007-treated subjects were 17.1 pg/mL (96.14% CI: 3.0-51.1; p < 0.001 vs baseline), with 92% and 58% exhibiting greater than 10-fold and greater than 100-fold increases, respectively. In all three placebo control subjects IFN-alpha levels remained below the lower limit of quantification. Regarding pro-inflammatory cytokines, IL-6 excursions from baseline were transient and modest in magnitude and the levels of TNF⍺ and IL-1β for all participants remained in the normal range. Interferon-associated PD markers moved in concert in the PRTX007 treatment group, demonstrating interferon-pathway biomarker induction with increasing magnitude over consecutive dosing days and re-induction upon subsequent dosing cycles, consistent with reversible IRF7-biased plasmacytoid dendritic cell activation. Conclusions: Intermittent weekly oral dosing of PRTX007 demonstrated a favorable safety profile with a controlled statistically significant induction of interferon-pathway PD markers in healthy volunteers, with minimal pro-inflammatory cytokine activation. These findings informed dosing regimen selection for a subsequent Phase 2 study PRTX007-003 evaluating PRTX007 in combination with pembrolizumab as neoadjuvant therapy for resectable stage III melanoma. Clinical trial information: 12624000981527.
First results from the phase 1/2 LINKER-AL2 trial of linvoseltamab (LINVO) in patients (pts) with relapsed or refractory (RR) systemic light chain (AL) amyloidosis.
7502 Background: Systemic AL amyloidosis is caused by deposition of misfolded amyloid fibrils produced by clonal plasma cells in vital organs. With no currently approved therapies, there is a need for new targeted agents eliciting rapid normalization of involved free light chain (iFLC), which is linked to improved organ function recovery and overall survival. Here are the first results of LINVO (human BCMA×CD3 bispecific antibody) in RR systemic AL amyloidosis from the Phase 1 portion of the Phase 1/2 LINKER-AL2 trial (NCT06292780). Methods: Eligible adults with RR systemic AL amyloidosis and ≥1 prior therapy received intravenous step-up doses of LINVO 5 mg and 25 mg (Cycle [C] 0 Days 1 and 8, respectively) followed by 28-day cycles of full subcutaneous (SC) doses of 80 mg or 240 mg administered once weekly in C1–C2 and then once every 4 weeks in C3–C12. The Phase 1 primary endpoint was incidence of dose-limiting toxicities (DLTs) within 28 days after the first full dose. Key secondary endpoints were rates of hematologic complete response (hCR; by FLC ratio and absence of AL), very good partial response or better (≥hVGPR), objective response (hOR), time to hCR, and safety. Results: As of data cut-off (DCO) August 25, 2025, 20 pts (median age, 64 years; prior daratumumab therapy, 60.0%) received LINVO 80 mg (n=7) or 240 mg (n=13), with a median follow-up of 9.6 (range 1.1–25.0) weeks. No DLTs or ICANS occurred. All pts had ≥1 treatment-emergent adverse event (TEAE; Grade (Gr) ≥3, 55.0%). One pt with prior cardiac disorder discontinued (Gr 5 ventricular fibrillation, unrelated to LINVO). Most common TEAEs were CRS (50.0%; Gr 1, 35.0%; Gr 2, 15.0%) and infusion-related reactions (45.0%; Gr 1, 25.0%; Gr 2, 20.0%), most during step-up dosing. LINVO elicited rapid responses (Table): hOR rates were 100% (80 mg: ≥hCR, 57.1%) and 92.3% (240 mg: ≥hCR, 38.5%), which are expected to deepen with longer follow-up. Most pts reached iFLC <20 mg/L (80 mg: 100%; 240 mg: 84.6%). Conclusions: LINVO demonstrated generally manageable safety and rapid, meaningful clinical activity, with most pts achieving hOR and iFLC <20 mg/L. These results support continued evaluation of LINVO in RR systemic AL amyloidosis. Longer follow-up data, including organ responses, are being evaluated. Clinical trial information: NCT06292780 . Investigator-assessed hematologic responses (ISA criteria). Response, n (%)* LINVO 80 mg (n=7) LINVO 240 mg (n=13) † Median follow-up, weeks (range) 12.1 (6.9–25.0) 6.7 (1.1–15.3) hOR 7 (100) 12 (92.3) hCR 4 (57.1) 5 (38.5) ≥hVGPR 7 (100) 12 (92.3) Low dFLC ‡ 0 1 (7.7) dFLC <10 mg/L 7 (100) 11 (84.6) iFLC <20 mg/L 7 (100) 11 (84.6) Median time to hCR, weeks (range) 3.2 (2.1–18.3) 3.1 (1.0–7.1) *Unless stated otherwise. † Not evaluable, n=1. ‡ If baseline dFLC ≥20 mg/L and <50 mg/L, response other than hCR reached with dFLC <10 mg/L. dFLC, difference between iFLC and uninvolved FLC; ISA, International Society of Amyloidosis.
Postoperative pancreatic fistula after resection of pancreatic neuroendocrine tumors: Incidence, predictors, and clinical impact.
4177 Background: Postoperative pancreatic fistula (POPF) remain a major source of morbidity after pancreatic surgery. Data specific to pancreatic neuroendocrine tumors (pNETs) are limited, although these neoplasms differ biologically and surgically from pancreatic adenocarcinoma, potentially influencing POPF risk and outcomes. This study evaluated the incidence, predictors, and clinical impact of POPF in a well-characterized cohort of patients undergoing resection for pNETs. Methods: All patients who underwent resection for histologically confirmed pNETs between 2010-2019 at a tertiary center were retrospectively analyzed. Clinicopathological parameters and surgical procedures were assessed for association with POPF occurrence and severity according to the International Study Group on Pancreatic Fistula (ISGPF) criteria. Statistical analyses included χ², Fisher’s exact, Spearman correlation tests, and multivariable regression analysis; significance was defined as α = 0.05. Results: Among 106 patients (44% female, mean age 61 (± 13.5) years), 85% received a formal pancreatic resection, 15% an atypical resection. More than half (53%) of pNETs were located in the pancreatic tail. The majority (64%) of tumors were G1, only 4% G3, with 72% of diseases in early stages UICC I or II. 12% of pNETs were functional and exclusively insulinomas. 53 patients (50%) developed a POPF, predominantly Grade B (60%), followed by Grade A (34%) and Grade C (6%). No surgery- or POPF-related mortality was observed. POPFs were most frequent after distal pancreatectomy for tail lesions (60% vs. 23% body vs. 17% head; p = 0.065), and in functional pNETs (77% vs. 46% in non-functional pNETs; p = 0.072). Distant metastases were inversely associated with POPF (M0 54% vs. M1 28%; p = 0.043). A nonsignificant trend toward higher POPF rates was observed after atypical resections (73% vs. 48%; p = 0.095). Higher POPF grade was associated with female sex (p = 0.016), and tumor proliferation index (p = 0.046); female sex remained independently associated with POPF grade in multivariable regression analysis (beta = 0.334, p = 0.044). Patients with POPF exhibited lower long-term mortality during follow-up (8% vs. 29%; p = 0.009) and lower mean preoperative chromogranin A levels (82.8 ng/mL vs. 167.8 ng/mL; p = 0.038). Conclusions: POPF occurred in half of all pNET resections, most commonly of Grade B severity. Tumor location, functionality, and absence of metastases were associated with higher POPF risk, and female sex with higher grade POPF. The association of POPF with lower long-term mortality likely mirrors the favorable tumor biology and localized disease profile of patients selected for curative surgery. This underscores the long-term oncological benefit of surgical resection in pNETs and provides relevant insights for postoperative risk stratification and perioperative management.
Real-world analysis of targeted tumor genomic profiles in prostate cancer among Black and White men within the MedStar Health Network.
e22523 Background: Prostate cancer (PC) outcome disparities between White and Black men result from a combination of biologic and socioeconomic factors. Utilizing next-generation sequencing (NGS) is key to precision oncology. This study explores the association between targeted tumor genomic profiles, disease biology, and social determinants of health. Methods: This is a real-world prospective cohort study of men diagnosed with advanced PC who received standard of care (SOC) between 2023 to 2025. NGS was performed on tissue biopsy samples at Caris Life Sciences as part of SOC. Data on disease biology at diagnosis (stage, Gleason score, PSA), NGS prior to first-line systemic therapy, and national area deprivation index (ADI) by quartile was analyzed using Chi-square test. Results: A cohort of 87 men (52.8% Black, 43.6% White, 3.4% Other) were included. Eleven men did not have molecular data available. The association between Gleason score (low/intermediate [6-7] vs. high [8-10]), stage (I-II, III, IV), and PSA (low, intermediate, high) compared to national ADI quartile (Q1-Q4) was not statistically significant (p = 0.16, 0.92, 0.58). Clinically, Black patients presented with significantly higher PSA levels at diagnosis (p = 0.04), but there was no significant association between stage and race (p = 0.97). Fifteen targeted mutations were seen in DNA damage repair genes (DDR) and oncogenic signaling with seven appearing in black and five in white males. All targeted mutations were observed in patients with high Gleason scores. Pairwise chi-square analysis found no significant difference between race and ADI quartile versus targeted mutation status (p = 0.98, p = 0.66), however a statistically significant difference was observed between PSA and targeted mutation status (p = 0.004). Sixty-one mutations were found in tumor suppressor genes (52.5% White and 47.5% Black). Fifty-two mutations were found in androgen receptor and prostate lineage genes (48.1% White and 48.1% Black). Forty-two mutations were found in DDR genes (26.2% White and 73.8% Black). Twenty-four mutations were found in oncogenic signaling genes (37.5% White and 62.5% Black). Twenty-one mutations were seen in epigenetic/transcriptional genes (38.1% White and 57.1% Black). Conclusions: This study combines early genomic profiling with disease biology to inform risk-reductive treatment strategies. Our study uniquely utilizes a racially balanced prospective cohort; however, the limited sample size limited our scope to a descriptive analysis. Targetable genomics were observed in higher Gleason scores and diverse mutations profiles including DDR, oncogenic signaling, and transcriptional mutations had higher frequency in Black patients. Identifying population-specific genomic differences, including high-risk alterations, may reveal therapeutic opportunities to reduce clinical outcome disparities.
Utilization and outcomes of minimally invasive resection in patients with neuroblastoma and Wilms tumor in a national cohort.
10043 Background: Tumor resection is a cornerstone of therapy for neuroblastoma (NB) and Wilms tumor (WT). Minimally invasive surgical approaches such as laparoscopic or robotic tumor resection hold promise to reduce surgical morbidity while maintaining excellent oncologic outcomes. This study aims to determine the utilization and outcomes of laparoscopic and robotic resection of NB and WT among children in a national cohort, hypothesizing that utilization is increasing over time and outcomes are not inferior to open resection. Methods: A retrospective cohort analysis of patients <18 years old in the 2020-2023 National Cancer Database who underwent resection of NB or WT was performed. Surgical approach trends were assessed via Cochran-Armitage tests. Conversions to open surgery, length of stay (LOS), and readmissions were compared between approaches using Kruskal-Wallis, Chi-square, and Fisher’s exact tests. Results: Among 1,262 children who underwent resection of NB (n=581) or WT (n=681), open procedures were most common (83.4%), followed by laparoscopic (13.9%), and robotic (2.7%). Utilization of each approach was stable over time aside from increased robotic resection of NB from 0% to 4% (p=.024). Approach did not vary by patient age (Table 1). Median WT size and lymph node yield were similar between approaches (both p>.05). In NB, laparoscopic procedures were performed on smaller tumors than open or robotic (p<.001). There were no robotic conversions to open; 11.4% of laparoscopic procedures were converted without change over time or difference between NB and WT (both p>.05). Median LOS was greater for open procedures than laparoscopic or robotic (p<.001). Readmission rates (p=.64) and margin status (p=.63) were similar between approaches. Outcomes did not vary by diagnosis. Conclusions: Utilization of minimally invasive approaches for NB or WT resection remains low despite noninferior or improved outcomes. Efforts to safely increase utilization are needed, especially for robotics, so that survivorship can be improved through reduced surgical morbidity. Differences in patient characteristics and outcomes by surgical approach. Characteristic or Outcome Open Laparoscopic Robotic p value Age in years (median, IQR) 3 (1, 4) 3 (1, 5) 3 (1, 7) 0.08 % female 53% 53% 44% 0.59 NB tumor size in mm (median, IQR) 72 (50, 101) 47 (35, 71) 70 (51, 86) <0.001 WT tumor size in mm (median, IQR) 110 (75, 140) 115 (93, 150) 93 (54, 125) 0.10 % metastatic disease 33% 30% 15% 0.07 % neoadjuvant therapy 35% 27% 22% 0.014 % conversion to open n/a 11% 0% 0.05 Negative margins 86% 83% 96% 0.63 # Lymph nodes assessed in WT (median, IQR) 5 (3, 8) 6 (3, 10) 4 (3, 9) 0.62 LOS in days (median, IQR) 5 (4, 7) 4 (2, 6) 4 (1, 5) <0.001 % readmission 15% 14% 9% 0.64
Infectious complications in mantle cell lymphoma patients treated with CAR-T therapy: A real-world TriNetX analysis.
e19088 Background: Chimeric antigen receptor T-cell therapy has significantly improved outcomes for patients with relapsed or refractory mantle cell lymphoma. However, infectious complications remain a major source of morbidity following CAR-T therapy. Real-world data comparing infections in CAR-T–treated versus non–CAR-T–treated mantle cell lymphoma patients remain limited. Methods: We conducted a retrospective cohort study using the TriNetX Global Collaborative Network. Adult patients with mantle cell lymphoma and documented infectious diagnoses were identified and stratified based on prior CAR-T exposure. Two cohorts were analyzed: patients who developed infections after CAR-T therapy and patients with mantle cell lymphoma who developed infections without CAR-T exposure. Demographics, infection types, healthcare utilization, and patient arrival rates across healthcare organizations were evaluated. Results: The CAR-T–associated infection cohort included 20 patients from 10 healthcare organizations, while the non–CAR-T infection cohort included 202 patients from 38 healthcare organizations. Patients in the CAR-T cohort were predominantly older and male. Infections across both groups included bacterial, viral, and opportunistic pathogens, with sepsis representing a common and clinically significant complication. Although the CAR-T cohort was smaller, it exhibited a high burden of severe infections, including gram-negative sepsis and viral infections. Patient arrival rates were lower in the CAR-T cohort, reflecting the specialized nature of CAR-T therapy delivery. Conclusions: In this real-world analysis, mantle cell lymphoma patients who developed infections following CAR-T therapy represented a small but high-risk population with substantial infectious morbidity and healthcare utilization. These findings highlight the need for enhanced infection surveillance, preventive strategies, and early intervention in CAR-T–treated mantle cell lymphoma patients.
TP53 mutation adoption of stromal-adhesion/neuronal-like programs as a driver of stress adaptation and acute myelogenous leukemia cell survival.
6545 Background: TP53 -mutated AML is associated with early relapse and poor survival, highlighting the need to define p53-driven transcriptional programs that may reveal targetable therapeutic vulnerabilities. Systematic deconvolution of the TP53-associated transcriptome provides a critical framework to uncover dependencies associated with drug resistance. Methods: After IRB approval, 311 AML patients were screened; 93 (30%) with available NGS were included (32 TP53-mutated, 61 TP53-wild type). Bulk transcriptome data were available for 11 TP53+ and 13 TP53-WT cases. Overall survival was analyzed using Kaplan–Meier and multivariable Cox regression. Transcriptional modules predictive of TP53+ AML were identified using Random Forest and pathway enrichment via KEGG/Reactome with confirmation by ToppGene. Public single-cell RNA-seq datasets from the MIT Single Cell Portal were queried to assess whether low TP53 expression states recapitulate bulk TP53-mutated transcriptional programs. Results: OS was 168 and 624 days (d) in TP53+ and TP53 WT cases (HR = 4.2, p = 0.02). Only CK retained independent effect for OS (HR = 0.33, 95% CI = 0.13–0.82, p = 0.014) when accounting for TP53 mut status and age. TP53 mut status imprinted a deep transcriptomic signature (AUC = 0.94). Top upregulated genes included FHL2 (p = 4.0×10−6, FDR = 0.004), HMGA2 (p = 1.4×10−3, FDR = 0.02), ARHGEF12 (p = 5.2×10−3, FDR = 0.04), RYR3 (p = 2.1×10−3, FDR = 0.03), LRRC7 (p = 7.7×10−3, FDR = 0.05), ALDH1A1 (p = 1.4×10−3, FDR = 0.02). Top downregulated genes included CSF1 (p = 4.7×10−4, FDR = 0.007), CD34 (p = 1.2×10−3, FDR = 0.01). To uncover pathway-level shifts co-opted by TP53+ AML, we contrasted pathway-level projection against TP53 WT cases. Upregulated pathways included adhesion / stromal dependence (p = 0.0004; FDR = 0.0028), neurotrophin / calcium signaling / Ca2+-coupled survival (p = 0.0006, FDR = 0.0021). However, downregulated pathways included cytokine-cytokine receptor interaction (p = 4.7×10−4, FDR = 0.007) and myeloid differentiation (p = 1.2×10−3, FDR = 0.01) [Fig. 2]. Interestingly, in lineage-resolved single-cell reference, low TP53 GE was visually colocalized with perivascular/stromal cells exhibiting high FHL2 and THY1 (CD90) expression. This data reconfirms prior data suggesting that reduced TP53 function favors niche-adapted perivascular/MSC-like reprogramming (THY1-high) that protects “resistant AML clones” (Mizuno et al., ASH 2024). Conclusions: These findings unveil that TP53+ AML co-opts leukemic blasts toward a stromal-adhesive and neuronal-like state, allowing survival under metabolic and therapeutic stress. Reversing stromal adhesive / neuronal-like rewiring may attenuate TP53 + blasts hijacking on vascular niche that maintains relapse potential.
“Divergent paths in the right colon”: A 24-year analysis of appendiceal cancer versus cecal cancer mortality in the United States (1999-2023) and age and sex disparities.
e15712 Background: Despite their anatomical proximity, appendiceal and cecal cancers represent distinct biological entities with divergent incidence patterns. While right-sided colon cancer mortality has generally benefited from screening advances, appendiceal cancer remains largely occult. We analyzed US mortality trends over 24 years to quantify the widening gap between the dramatically rising appendiceal cancer against the backdrop of stable cecal cancer mortality, specifically examining the impact of early-onset disease and sex disparities. Methods: We utilized the CDC WONDER database to identify cancer-specific mortality for appendiceal (ICD-10: C18.1) and cecal (ICD-10: C18.0) malignancies from 1999 to 2023. Age-adjusted mortality rates (AAMR) per 100,000 population were calculated. Joinpoint regression analysis was performed to calculate the Average Annual Percent Change (AAPC). Cohorts were stratified by sex and age to assess disparities in early-onset mortality. Results: A total of 21,120 deaths were identified (Appendix: 11,971; Cecum: 9,149). From 1999 to 2023, overall appendiceal AAMR increased significantly (AAPC: +4.814, CI: 4.09-5.54; p<0.01), whereas cecal mortality remained stable (AAPC: 1.17, CI: -2.07-4.54; p>0.05). The divergence was most pronounced in the early-onset cohort (<50 years): appendiceal mortality rose sharply (AAPC: [+4.71, CI: 3.42-5.12; p<0.001), significantly outpacing cecal trends. Sex-stratified analysis revealed that while men had higher absolute mortality rates in both groups, the rate of increase for appendiceal cancer was steeper in men (AAPC: 5.008, CI: 4.31-5.71; p<0.01). Conversely, cecal mortality in older adults (≥50) showed a stable trend (AAPC: 0.05, CI: -1.68-1.89; p>0.05), while appendiceal cancer revealed a stark increase (AAPC: 5.91, CI: 4.90-6.92; p<0.01) reflecting screening efficacy not seen in appendiceal cohorts. Conclusions: This 24-year analysis confirms a significantly rising mortality burden for appendiceal cancer, sharply contrasting with the stability of cecal cancer, challenging the practice of grouping these entities in epidemiological reporting. The alarming rise in early-onset appendiceal mortality highlights a critical public health gap, as current colorectal screening guidelines fail to detect these malignancies. Distinct clinical vigilance and research into environmental or biological drivers of appendiceal carcinogenesis are urgently warranted. Cecal vs appendiceal cancer mortality trend 1999-2023. Average Annual Percent Change (AAPC) Cohort Lower Endpoint Upper Endpoint AAPC Lower CI Upper CI Test Statistic~ P-Value~ Cecum-Malignant neoplasms-3 JP 1999 2023 1.1787 −2.0782 4.5439 0.7019 0.482725 Appendix-Malignant neoplasms-1 JP 1999 2023 4.8145* 4.0938 5.5402 13.3572 <0.000001 * = Statistical significance at p<0.05.
Development of an analytics dashboard to measure infusion treatment initiation timelines.
e23250 Background: Prior studies demonstrate that longer time from cancer diagnosis to treatment initiation is associated with worse survival outcomes. However, these analyses largely rely on national datasets, limiting institution-level visibility into care delivery processes and the ability to isolate infusion specific delays. As a result, cancer centers lack integrated tools to systematically measure time to infusion treatment (TTI) across referral, scheduling, and infusion workflows. To address this gap, we sought to develop an automated, scalable EMR-based analytics dashboard. Methods: We conducted a retrospective health services analysis by creating an EMR-integrated Tableau analytics platform to automate measurement of referral-to-scheduling and scheduling-to-treatment intervals. Metric definitions and visualizations were iteratively refined with physicians and multidisciplinary stakeholders to ensure clinical relevance. The dashboard supports stratified analyses by time period, cancer type, patient characteristics, infusion duration, regimen, and provider. Descriptive analyses included patients initiating chemotherapy between July 2024 and December 2025. Results: A total of 2,024 curative intent referrals to the infusion center were recorded. Time from referral entry to infusion was ≤14 days for 49% of referrals and ≥30 days in 17%. Among patients starting systemic therapy in 2025, median TTI varied by malignancy, ranging from 10 days in BMT to 18 days in gynecologic oncology (Table 1). Process-level analysis showed referral-to-scheduling intervals were relatively consistent across services (median 3-5 days), whereas the scheduling-to-treatment intervals were more variable (median 5-12 days). Reliability of TTI varied widely, with initiation within ±5 days of the planned start date ranging from 38% to 94% across teams. Conclusions: Development of this analytics platform establishes a reproducible, institution-level framework for measuring TTI and addresses a critical gap between national datasets and actionable, local process metrics. Strengths include process orientation, scalability across all oncology teams, clinician-facing transparency, and incorporation of real-time data. Limitations include incomplete capture of externally administered infusions, and heterogeneity in provider selection for planned start date. This platform is being extended to radiation oncology, surgical oncology, and new patient access to support broader evaluation of timely cancer care. BMT Breast GI GU Gyn H&N Hema Skin Thoracic TTI, median days (n= # patients) 10 (n=64) 14 (n=259) 17 (n=215) 17 (n=134) 18 (n=124) 12 (n=36) 14 (n=29) 13 (n=113) 14 (n=91) Referral entry to infusion scheduled, median days 3 4 4 4 4 3 5 4 4 Infusion scheduled to treatment start, median days 7 8 11 12 10 5 8 8 7 % starting within ±5 days of planned 66% 64% 55% 82% 55% 38% 60% 94% 45%
Long-term outcomes of CNS lymphoma patients treated with glucarpidase.
2091 Background: High-dose methotrexate (HD-MTX) is the backbone of potentially curative therapy for central nervous system lymphoma (CNSL) but requires inpatient admission for hydration and monitoring to minimize toxicity. Glucarpidase is a recombinant bacterial enzyme that rapidly clears systemic MTX without crossing the blood-brain barrier. It is FDA-approved as a rescue agent in the setting of MTX toxicity. Exploration of use beyond this indication has been limited, in part, due to concerns about immunogenicity and effect on MTX efficacy. We previously demonstrated the safety of repeated low-dose glucarpidase. Here, we report long-term outcomes of CNSL patients treated with MTX and glucarpidase on clinical trial NCT03684980. Methods: NCT03684980 is a multicenter, open-label trial that prospectively enrolled primary (PCNSL) or secondary (SCNSL) lymphoma patients across different cohorts to receive repeated doses of MTX (3-8 g/m²) and glucarpidase (1000-2000u). Patients who received ≥ 6 of 8 planned doses were included. Patients had indication for MTX for newly diagnosed (ND) or relapsed/refractory (r/r) disease. There were no restrictions on post-MTX therapy or consolidation. Objective response rate (ORR), progression-free survival (PFS) and overall survival (OS) were assessed. Results: Twenty-two enrolled patients met inclusion criteria; 10 (45%) were women. Median age was 70 years, and median Karnofsky Performance Status was 80. PCNSL accounted for 20/22 (91%) of cases; 5/22 (23%) were treated for r/r disease. Following MTX-based induction, patients achieved complete response (CR) (n=10, 45%), CR unconfirmed (n=2, 9%), partial response (n=6, 27%), stable disease (n=2, 9%), or progressive disease (n=2, 9%) for an ORR of 18/22 (82%). Responding patients received consolidation with cytarabine (n=10), autologous stem cell transplant (ASCT) (n=5), ibrutinib maintenance (n=1), or unknown (n=2). Median follow-up was 55.5 months. Median PFS and OS were not reached overall. PFS and OS were not reached for PCNSL, versus 7.3 and 21.0 months for SCNSL. PFS was not reached for ND patients versus 12.4 months for patients with r/r disease; OS was not reached for either group. Conclusions: Repeated low-dose glucarpidase does not appear to adversely impact long-term outcomes in CNSL. Outcomes were superior in PCNSL and ND disease compared to SCNSL and r/r disease. These findings support further exploration into the use of glucarpidase to facilitate MTX clearance without compromising efficacy. Clinical trial information: NCT03684980 .
Impact of social vulnerability on end-of-life care in hematologic malignancies.
12036 Background: Patients with hematologic malignancies frequently receive intensive inpatient care near theend of life. While disparities in end-of-life (EOL) care are well described in oncology, how cumulative social vulnerability shapes treatment intensity and palliative care involvementamong hospitalized patients with hematologic cancers remains unclear. We examined the association between social vulnerability burden and inpatient EOL care patterns using a nationally representative database. Methods: We performed a retrospective cohort study using the 2021–2022 National Inpatient Sample of adult hospitalizations with hematologic malignancies. A composite hematologic social vulnerability index (cSVI-Heme) was constructed using validated proxies available in NIS: minoritized race/ethnicity, Medicaid or uninsured status, and lowest ZIP-code income quartile. Hospitalizations were classified as low (0), moderate (1), or high (≥2) vulnerability.The primary outcome was high-intensity EOL care (invasive mechanical ventilation, cardiopulmonary resuscitation, or acute dialysis). Secondary outcomes included palliative care utilization and in-hospital mortality. Associations were assessed using multivariable logistic regression and inverse probability of treatment weighting, adjusting for demographics, comorbidity burden, acute organ failure, sepsis, malignancy subtype, and hospital characteristics. Results: The cohort included 202,802 hospitalizations, of which 54.2% were low, 31.0% moderate, and 14.8% high vulnerability. Overall, 5.9% involved high-intensity EOL care, increasing stepwise across vulnerability tiers (5.1% vs 6.5% vs 7.9%; p < 0.001). After adjustment, compared with low vulnerability hospitalizations, moderate vulnerability was associated with higher odds of high-intensity EOL care (aOR 1.25, 95% CI 1.19–1.30), and high vulnerability with substantially higher odds (aOR 1.50, 95% CI 1.42–1.58), demonstrating a clear dose–response relationship. Findings were confirmed in propensity-weighted analyses.This gradient was driven primarily by acute dialysis (3.1% vs 4.6% vs 6.2%; p < 0.001), while rates of mechanical ventilation and CPR were similar. High vulnerability was independently associated with higher in-hospital mortality (aOR 1.20, 95% CI 1.12–1.28), whereaspalliative care utilization did not increase after adjustment. Conclusions: Among hospitalized patients with hematologic malignancies, greater social vulnerability is associated with higher-intensity end-of-life care and increased mortality, with a clear dose–response relationship. These differences are not accompanied by greater palliative care involvement, highlighting a mismatch between treatment intensity and supportive care. Integrating social vulnerability into inpatient hematologic oncology care may support earlier goals-of-care discussions and more equitable EOL care.