Citations

3624 Background: Increasing chemotherapy completion and reducing chemotherapy adverse events (AE) are critical to improve survival after colon cancer diagnosis. Sex, as a biological variable, may impact chemotherapy treatment differently, and thus further investigation is needed to examine sex differences in chemotherapy completion and adverse events among patients with colon cancer. Methods: Among an NCI-sponsored trial conducted among patients with stage III colon cancer (CALGB/SWOG 80702), all patients received standard adjuvant chemotherapy FOLFOX (fluorouracil, leucovorin, and oxaliplatin). To signal chemotherapy completion, we utilized relative dose intensity (RDI) calculated as the ratio of the delivered dose intensity to planned dose intensity; and reduced RDI (RDI <85%) was considered as a clinically significant deviation from standard FOLFOX. From clinicians’ records of NCI’s Common Terminology Criteria for Adverse Events (CTCAE), we primarily focused on clinically significant AE such as neutrophils decrease, nausea, platelets decrease, hypertension, peripheral neuropathy, diarrhea, fatigue, gastritis, creatinine increase, gastric ulcer, myocardial ischemia, and cerebral ischemia; and severe AE was defined as the occurrence of any above AE with CTCAE grade ≥3. Using multivariable logistic regression that adjusted for body surface area (BSA) and other clinicopathological confounders, we estimated adjusted odds ratios (OR) for the associations of sex with reduced RDI and severe AE. Results: Of 2201 patients, mean (standard deviation [SD]) age was 60.9 (10.9) years, 1019 (46.3%) were female, 1750 (79.5%) were White, 172 (7.8%) were Hispanic, 964 (43.8%) experienced reduced RDI, and 1156 (52.5%) had severe adverse events. Compared to males, females were at significantly higher risks of experiencing reduced RDI (OR [95% CI]: 1.57 [1.27-1.93], P <0.001) and severe AE (OR [95% CI]: 1.73 [1.41-2.12], P <0.001). Conclusions: Our findings suggested females are more likely than men to experience reduced RDI and severe AE during colon cancer chemotherapy. Clinical Impact: In the era of precision medicine, sex (as a biological variable) should be considered in optimizing colon cancer chemotherapy to improve completion and reduce toxicities. Clinical trial information: NCT01150045 .

e23020 Background: Precision medicine studies must overcome the challenges of identifying patients with uncommon molecular targets and small cohort sizes for correlative studies. BASECAMP-1 (NCT04981119) is a prescreening study that uses a single assay to address both problems by: 1) efficiently screening potentially eligible patients for subsequent clinical trials of Tmod logic-gated chimeric antigen receptor T-cell (CAR T) therapies; and 2) providing a large dataset available for translational studies. Methods: Initially, patients were identified for BASECAMP-1 by consenting and genotyping all potentially eligible patients for human leukocyte antigen (HLA)-A*02 heterozygosity. To improve screening efficiency, we co-developed a patient-matching program with Tempus AI, Inc (Tempus), in which Tempus informs the principal investigators of next-generation sequenced (NGS) patients with HLA-A*02 heterozygosity within their health system who may be eligible for BASECAMP-1. Patients are screened using the Tempus xT platform, which uses genomic and RNA sequencing (RNAseq) analyses to determine gene mutations and expression. Enrolled patients with colorectal (CRC), non-small cell lung, ovarian (OVCA), and pancreatic cancers (PANC) were analyzed for tumor purity, mutation distribution, HLA-A*02 loss of heterozygosity (LOH) status, and RNAseq to validate the data available for future clinical analyses. Results: As of January 3, 2025, 80 participants have been enrolled in BASECAMP-1. Through individual screening methods, 1684 patients were screened at 13 study sites, of which 30 patients had tumor-specific HLA-A*02 LOH and were enrolled (~1 eligible patient identified per 56 screened). The patient-matching program identified 282 patients with HLA-A*02 LOH at 12 sites; of these, 50 consented to the BASECAMP-1 study (~1 patient enrolled per 6 identified), demonstrating higher efficiency than the individual patient screening method. Correlative data from xT were available for 360 patients with germline HLA-A*02 and provided insights into mutational distribution and target expression. LOH status was not associated with tumor-specific differences in missense mutational distribution. RNAseq data demonstrated higher CEACAM5 expression in CRC and higher MSLN in PANC and OVCA vs other tumor types, while EGFR was consistently expressed with reduced variance across tumor types, confirming these as potential target antigens for Tmod CAR T interventional trials EVEREST-1, EVEREST-2, and DENALI-1. Conclusions: Identification and enrollment of HLA-A*02 LOH patients into BASECAMP-1 for logic-gated CAR T therapeutic studies were accelerated by the patient matching program. Concurrent xT analysis provides opportunities to augment analyses from robust propensity-matched comparisons, response correlation studies, and other translational discovery. Clinical trial information: NCT04981119 .

Abstract Inotuzumab ozogamicin (InO) is an antibody-calicheamicin conjugate with high efficacy in lymphoid malignancies. It targets the B-cell surface protein CD22, which is expressed in most B-cell acute lymphoblastic leukemia (B-ALL) cases, albeit with variable intensity. However, factors governing CD22 expression and thus leukemia sensitivity to InO remain incompletely understood. Using multiomic characterization of 196 human B-ALL samples, coupled with ex vivo InO sensitivity profiling, we showed that early leukemia differentiation arrest at the pre–pro-B stage is associated with resistance to InO. Screening of 1639 transcription factor genes identified early B-cell factor 1 (EBF1) as a key regulator of CD22 expression (false discovery rate of 7.1 × 10−4). When comparing the assay for transposase-accessible chromatin with sequencing profiling results of the most InO-sensitive and -resistant cases (50% lethal concentration &lt;10th vs &gt;90th percentile, n = 18), the binding motif for EBF1 was strikingly enriched in regions with differential open chromatin status (P = 8 × 10−174). CRISPR interference targeting EBF1 binding sites at the CD22 locus led to an ∼50-fold reduction in cell surface CD22 expression and, consequently, an ∼22-fold increase in InO resistance in ALL cell lines. Interestingly, within BCR::ABL1 ALL, we observed intrasubtype heterogeneity linked to EBF1 transcriptional downregulation (P = 1.1 × 10−15) and/or somatic alteration (P = .004), which led to reduced CD22 expression (P = 8.3 × 10−11) and ex vivo and in vivo resistance to InO. Collectively, these findings point to the direct impact of EBF1 on CD22 expression during B-cell development, which, in turn, contributes to interpatient variability in InO response, even within the same subtype of B-ALL.

e13656 Background: Evidence-based medicine is crucial for clinical decision-making, yet studies suggest that a significant proportion of treatment decisions do not fully incorporate the latest evidence. Large Language Models (LLMs) show promise in bridging this gap, but their reliability for medical recommendations remains uncertain. Methods: We conducted an evaluation study comparing five LLMs’ recommendations across 50 clinical scenarios related to multiple myeloma diagnosis, staging, treatment, and management. The evaluation included three general-purpose LLMs (OpenAI o1-preview, Claude 3.5 Sonnet, Gemini 1.5 Pro), one retrieval-augmented generation (RAG) system (Myelo), and one agentic workflow-based system (HopeAI). General-purpose LLMs generated responses based solely on their internal knowledge, while the RAG system enhanced these capabilities by incorporating external knowledge retrieval. The agentic workflow system extended the RAG approach by implementing multi-step reasoning and coordinating with multiple tools and external systems for complex task execution. Three independent hematologist-oncologists evaluated the LLM-generated responses using standardized scoring criteria developed specifically for this study. Performance assessment encompassed five dimensions: accuracy, relevance, comprehensiveness, hallucination rate, and clinical use readiness. Results: HopeAI demonstrated superior performance across accuracy (82.0%), relevance (85.3%), and comprehensiveness (74.0%), compared to OpenAI o1-preview (64.7%, 57.3%, 36.0%), Claude 3.5 Sonnet (50.0%, 51.3%, 29.3%), Gemini 1.5 Pro (48.0%, 46.0%, 30.0%), and Myelo (58.7%, 56%, 32.7%). Hallucination rates were consistently low across all systems: HopeAI (5.3%), OpenAI o1-preview (3.3%), Claude 3.5 Sonnet (10.0%), Gemini 1.5 Pro (8.0%), and Myelo (5.3%). Clinical use readiness scores were relatively low for all systems: HopeAI (25.3%), OpenAI o1-preview (6.0%), Claude 3.5 Sonnet (2.7%), Gemini 1.5 Pro (4.0%), and Myelo (4.0%). Conclusions: This study demonstrates that while current LLMs show promise in medical decision support, their recommendations require careful clinical supervision to ensure patient safety and optimal care. Further research is needed to improve their clinical use readiness before integration into oncology workflows. These findings provide valuable insights into the capabilities and limitations of LLMs in oncology, guiding future research and development efforts toward integrating AI into clinical workflows.

7547 Background: Minimal residual disease (MRD) has been recently accepted by the Food and Drug Administration (FDA) as an endpoint for accelerated approval in Multiple Myeloma. However, emerging data from recent trials were not included in previous analyses. While literature-based meta-analyses on the correlation between MRD and approved clinical outcomes typically require extensive manual review, leveraging AI with expert-in-the-loop validation can efficiently generate reliable evidence from comprehensive clinical studies with up-to-date outcomes. Methods: An AI-assisted framework was developed to identify relevant studies and extract critical information via two independent objectives. The first objective examined trial-level associations, modeling treatment effects on MRD and clinical endpoints across patient populations using weighted least squares, with association strength measured by coefficients of determination (R²) and 95% confidence intervals (CIs). The second objective analyzed individual-level associations using synthetic individual patient data (SynthIPD) generated from the published Kaplan-Meier plots and summary statistics of patient subgroups. Results: AI-assisted screening identified eligible studies (>50 patients per treatment arm) reporting progression-free survival (PFS), overall survival (OS), and MRD-negative complete response rates (MRD-CR rate) using multi-parameter next-generation flow cytometry or sequencing methods (sensitivity threshold ≥10⁻⁵), expanding previous analyses from 15 to 20 two-arm studies. Trial-level analysis demonstrated an R² of 0.69 (95% CI 0.50–0.89) for PFS log hazard ratio versus MRD-CR rate log odds ratio. Analysis of synthetic individual data from Kaplan-Meier curves using a novel digitization method yielded a global odds ratio of 7.28 (95% CI 5.60–8.95) for individual-level correlation between MRD-CR rates and PFS outcomes. Conclusions: This study validates MRD-CR rate as an endpoint for accelerated approval in MM through rapid AI-assisted literature review and synthetic individual patient data. The findings demonstrate moderate correlation between MRD-CR rate and median PFS at both trial and individual levels, consistent with previous literature but incorporating additional eligible studies. The novel SynthIPD approach presents an efficient alternative to traditional data-sharing methods while maintaining analytical robustness. These results align with current Oncologic Drugs Advisory Committee (ODAC) surrogacy analysis methods and support the utility of MRD assessment in MM clinical trials.

5048 Background: Ra-223, an alpha-emitting radionuclide, is the first agent of its kind approved for the treatment of mCRPC. Although a pivotal phase 3 study evaluated its short-term safety, there is a need to investigate its long-term safety. The REASSURE study prospectively examined the long-term safety of Ra-223, including secondary primary malignancies (SPMs), in a large patient (pt) population enrolled across Europe, the United States, Israel, and Latin America. Methods: We report final analyses (data cut-off Oct 24, 2024) of REASSURE (NCT02141438), a global, noninterventional study (enrolment 2014–2017). Primary outcomes were the incidence of SPMs, short- (30 days) and long-term (7 years) safety events, and bone marrow suppression (BMS) management in pts who had ≥1 Ra-223 dose. Secondary outcomes included overall survival (OS). Results: Analyses included 1472 pts; median follow-up was 17 months (range 0.3–95.4). Median age was 73 years and 80% of pts had an ECOG PS of 0/1. In evaluable pts, median alkaline phosphatase, prostate-specific antigen, and lactate dehydrogenase levels were 133 U/L, 59 ng/mL, and 266 U/L, respectively. Overall, 81% of pts had bone-only metastases at baseline; 19% of pts had metastases in the bone plus other sites (mostly lymph nodes). Prior treatments included abiraterone (48% of pts), enzalutamide (39%), docetaxel (39%), and cabazitaxel (9%). Pts received a median of 6 Ra-223 doses; 67% received ≥5 doses. SPMs occurred in 2% of pts (25 SPMs in 24 pts). Of these pts, 16 (67%) and 1 (4%) had received prior or concomitant radiotherapy, respectively. Overall, 3% of pts had drug-related serious adverse events >30 days after completing Ra-223. Fractures were reported in 10% of pts and were less common in pts with (7% of 605) than without (12% of 867) concomitant BHA use. During Ra-223 and up to 30 days after the last dose, there was no notable difference in the incidence of abnormal platelet counts between pts with (3%) or without (2%) prior chemotherapy; similar findings were seen for abnormal neutrophil counts (5% and 5%, respectively). BMS treatments, assessed from the start of Ra-223, were more common in pts who had received prior taxanes (38%) than in those who had not (26%). The most common life-prolonging therapies received after Ra-223 were docetaxel (18%), enzalutamide (15%), abiraterone (11%), and cabazitaxel (11%). Median OS was 15.6 months (95% CI, 14.6, 16.4); pt subgroups that survived the longest will be characterized and presented. Conclusions: This real-world safety analysis of pts with mCRPC is the longest follow-up of a radiopharmaceutical reported to date and supports the well-established favorable safety profile of Ra-223. The incidence of SPMs was low. The rate of fracture was low, especially in the presence of BHAs. Prior taxane chemotherapy use had no impact on hematological toxicity. Clinical trial information: NCT02141438 .

e13741 Background: The NCCN guidelines for first-line (1L) regimens in CLL have evolved over the past decade, shifting from ibrutinib (ibr) and chemoimmunotherapy (CIT) to second-generation BTK inhibitor and venetoclax combinations. We evaluated real-world use of preferred 1L treatment (tx) since 2016 in CLL patients (pts) in routine care to identify differences in prescribing based on race/ethnicity and practice type. Methods: This retrospective observational study utilized the US nationwide Flatiron Health electronic health record-derived de-identified database. Eligible pts had confirmed CLL and initiated 1L tx between 01/01/16-07/31/24. Primary outcome was receipt of preferred 1L tx defined by NCCN guidelines in four time periods by race/ethnicity (Hispanic, White, Black, Asian/Other). Odds ratios (ORs) were estimated using logistic regression, additionally adjusting for age, sex, ECOG performance status, IGHV, del17p/ TP53 mutation status, time period, and practice type. Results: A total of 7528 pts were included. Compared with White pts (n = 5472), Black pts (n = 640) were younger (median age at 1L: 68 vs 71). More Black and Hispanic (n = 290) pts were treated at community practices (86% vs 80% White). Of those pts tested, more Black pts had unmutated IGHV than White (77% vs 56%). Presence of del17p/ TP53 mutation was similar across races/ethnicities. The proportion of pts receiving preferred 1L tx based on the NCCN Guidelines significantly differed by race/ethnicity (Table) ( P = 0.0021). The proportion of Hispanic pts treated with preferred 1L tx was significantly lower than White (OR = 0.61; 95% CI: 0.47-0.79), but Black was similar to White (OR = 1.07; 95% CI: 0.89-1.30). Updates to NCCN guidelines were significantly associated with pts receiving preferred 1L tx by practice type ( P = 0.0005). In 2016-2018, 44% of community practices and 55% of academic centers adopted targeted therapies (TTs); in 2019, ibr use increased in both practices (77% vs 68%, respectively). Adherence to preferred tx improved across practices in 2020 but decreased with the prioritization of second-generation therapies. After the approval of zanu, use of TTs was 71% in community practices and 74% in academic centers. Conclusions: Inequities in pts with CLL receiving preferred 1L tx suggests disproportionate use of CIT and ibr by race/ethnicity. Use of preferred TTs also differed by practice type and time period, with increased adoption after pivotal trials. Time period % Pts receiving preferred tx by race a White Black Hispanic Asian/Other 2016-2018:Preferred tx ibr 46% 50% 43% 45% 2019(v2.2019: Jan 2019)Preferred tx ibr 76% 76% 67% 79% 2020-Jun 2022(v4.2020: Feb 2020)Preferred tx acala, VO, ibr 84% 84% 69% 83% Jul 2022-Jul 2024(v2.2023: Aug 2022)Preferred tx acala, VO, zanu 72% 66% 56% 73% VO, venetoclax + obinutuzumab. a By listing of NCCN preference (preferred vs other) among non-del17p and age >65.

3040 Background: EVEREST-2 is a first-in-human, phase 1/2 trial to assess the safety and efficacy of A2B694, an autologous, logic-gated Tmod CAR T therapy targeted to MSLN, which is normally expressed in the mesothelium and can be upregulated in many solid tumors. A2B694 is designed to overcome challenges of on-target, off-tumor toxicity that have limited other MSLN-targeted approaches by combining a CAR-activating receptor targeting MSLN with a blocker CAR that recognizes HLA-A*02, to distinguish between normal and tumor cells (Tokatlian, et al. J Immunother Cancer. 2022). Methods: Adults with recurrent unresectable, locally advanced, or metastatic cancers with MSLN expression who have progressed after standard-of-care therapy are eligible for EVEREST-2. Enrollment to EVEREST-2 and collection of T-cells occurs through the ongoing prescreening study BASECAMP-1 (NCT04981119). When clinically appropriate, A2B694 is manufactured from cryopreserved T cells, and patients undergo lymphodepletion before A2B694 infusion. The dose-escalation phase is evaluating the safety and tolerability of A2B694 to identify the recommended phase 2 dose (RP2D). Dose escalation was started at 1x10 8 Tmod positive cells (dose level [DL] 1) and will increase up to 14x10 8 in combination with low-dose IL-2 (DL 5). The dose-expansion phase will confirm RP2D and collect biomarker data to further characterize A2B694. Results: As of January 15, 2025, 5 participants (median age: 60 years; range, 50-84) have enrolled on EVEREST-2 and received A2B694 at DLs 1-2; participants had ovarian cancer (n = 3), pancreatic cancer (n = 1), and non-small cell lung cancer (n = 1) and had received a median of 4 prior lines of therapy (range, 1-7). Lymphodepleting chemotherapy was well tolerated with no significant cytopenias observed. The most common adverse events were lymphopenia (7 [14.6%]) and decreased appetite (6 [12.5%]). There were no dose-limiting toxicities, cytokine release syndrome, nor related neurotoxicity. One participant was admitted to the hospital for decreased appetite. No long-term toxicities have been noted up to 7.5 months post-infusion. Of the participants who have received A2B694, 5 were efficacy evaluable at DLs 1-2. A2B694 was detected post-infusion in the peripheral blood in all patients. Additionally, A2B694 was detected in an abdominal tumor biopsy from a patient with pancreatic cancer 42 days post-infusion. Conclusions: The logic-gated approach was successful at reducing toxicity seen with prior MSLN-targeted CAR T therapies, and A2B694 showed successful CAR T expansion and tumor infiltration. The maximum tolerated dose has not been reached, and results from the dose-escalation phase continue to determine the RP2D. Clinical trial information: NCT06051695 .

e16408 Background: A2B530 is a logic-gated CEA-targeting Tmod CAR T therapy being evaluated in the EVEREST-1 (NCT05736731) phase 1/2 study. Tmod CAR T therapy addresses challenges of on-target, off-tumor toxicity that have limited application of CAR T therapy to solid tumors (Tabernero, et al. J Clin Oncol. 2017); through use of a CEA-targeted CAR-activating receptor combined with a blocking receptor recognizing HLA-A*02, A2B530 can distinguish between tumor and normal cells (Sandberg, et al. Sci Transl Med. 2022). Methods: Patients with unresectable locally advanced or metastatic solid tumors expressing CEA and tumor-associated HLA-A*02 LOH are enrolled through the ongoing prescreening study BASECAMP-1 (NCT04981119). Eligible patients undergo leukapheresis and, when clinically appropriate, A2B530 is manufactured for the EVEREST-1 study. Patients undergo lymphodepletion before A2B530 infusion. The phase 1 dose escalation will evaluate safety, tolerability, manufacturing feasibility, and biomarkers starting at 1x10 8 Tmod positive cells (dose level [DL] 1) up to 6x10 8 cells and higher doses of lymphodepleting chemotherapy (LDC; DL 5). After the recommended phase 2 dose has been established, dose expansion will confirm safety and activity. Results: As of January 15, 2025, 14 total participants were enrolled on EVEREST-1 and had received A2B530 infusion at DLs 1-5, 4 (29%) with pancreatic (PANC) and 10 (71%) with colorectal cancer. Median age was 57 years (range, 34-76), with a median of 2 prior lines of therapy (range, 1-6). LDC was well tolerated: neutropenia nadired at days 7-10, with recovery in all patients, and no significant episodes of prolonged cytopenia. There were no dose-limiting toxicities, grade > 3 serious adverse events (AEs), or episodes of related neurotoxicity. Most gastrointestinal AEs occurred early and were grade 1 or 2. One patient in DL 2 (2x10 8 cells) experienced possible grade 2 cytokine release syndrome (CRS), which started on day 15 and resolved with supportive therapy within 2 days; later the CRS diagnosis was confounded by a positive blood culture from the participant’s central line. The low rate of CRS is consistent with the low levels of serum cytokines measured up to 28 days post-infusion. Central response assessment has been completed for the 6 participants in DLs 1-2, with a partial response observed in 1 of 3 (33%) of participants with PANC. Peak expansion shows a possible dose response. Conclusions: Treatment with A2B530 was well tolerated, showing potential clinical efficacy and a potential dose response, although these are initial results. Maximum tolerated dose has not been reached, and dose escalation continues to determine the optimal phase 2 dosing of this treatment regimen. Clinical trial information: NCT05736731 .

12011 Background: Patients (pts) with high HCT-Comorbidity Index scores (HCT-CI of ≥3), older age (≥65 years), and/or who are frail per gait speed (<0.8 meters/second) have increased morbidity and mortality after allo-HCT compared to younger and healthier counterparts. We report the phase II analysis (PIIA) and interim phase III primary outcome analysis (PIIIA) of a seamless phase II/III prospective, randomized clinical trial conducted at 11 transplant centers to test new approaches to improve quality of life (QOL) in this population. Methods: Phase II compared specialist-administered supportive and palliative care (SPC), patient-administered management of comorbidities (MC, e.g. physical exercise, stress reduction, etc), both (SPC+MC), and usual care (UC) for change in Functional Assessment of Cancer Therapy – Bone Marrow Transplantation (FACT-BMT) QOL scores from baseline to day 90 (D90). Pts deceased prior to D90 were assigned FACT-BMT score=0. The winning phase II arm moved forward versus UC in phase III. Results: PIIA was done after enrolling 35 pts to each of the 4 study arms. Calculating the difference between FACT-BMT scores on D90 minus baseline, excluding missing data, indicated that only SPC resulted in a small improvement in QOL compared to either MC or SPC+MC (Table); hence SPC was the winning phase II arm. The PIIIA was conducted after enrolling 158 SPC pts and 153 UC pts. The SPC and UC arms were well balanced with median age (both 68 years), HCT-CI ≥3 (49% vs 55%), frailty per gait speed (13% vs 12%), female sex (41% vs 34%), non-white race (9.5% vs 11.5%), and Hispanic or unreported ethnicity (7.7% vs 5.7%), respectively. After median follow-up of 362 days, 45 pts had died in each arm. The mean QOL difference between D90 and baseline was 2.83 for SPC and 4.27 for UC (difference of differences, -1.44, 95% CI of difference, -5.03 to 2.14, p=0.43). Fitting a generalized linear model that included baseline, D30, and D90 values and testing the null hypothesis that the slope of scores differs between SPC and UC resulted in p=0.85. Using the Kaplan Meier method, no difference in survival was observed across arms (HR 1.13 [0.75-1.73]). Conclusions: Specialist-administered palliative care showed no meaningful improvement in QOL, nor a survival advantage, compared to usual care in frail, older and comorbid allo-HCT recipients, resulting in the cessation of this Phase II/III trial. Analysis of the whole patient population for primary and secondary outcomes is in progress. Clinical trial information: NCT03870750 . Change in FACT-BMT (D90 value minus baseline value) comparing the 4 arms of phase II. Group Mean difference (sd) Median difference (range) SPC (n=28) -2.93 (28.52) 0.25 (-102 to 38.22) CM (n=18) -18.37 (33.80) -9.50 (-98.33 to 11) SPC+CM (n=27) -15.76 (32.69) -7.67 (-90 to 32)