Citations

LBA514 Background: Preclinical and clinical data show that antibody-drug conjugates (ADCs) such as T-DXd may synergize with immunotherapy. Rilvegostomig (R) is a monovalent, Fc-reduced, dual checkpoint bispecific IgG1 monoclonal antibody against PD-1 and TIGIT receptors. R + T-DXd was evaluated as neoadjuvant therapy for up to 4 cycles. Methods: I-SPY2.2 is a phase II neoadjuvant trial with three neoadjuvant sequences (Blocks A/B/C). All HER2-negative subtypes were eligible for R + T-DXd in Block A. Predicted responders at the end of Block A or B could undergo surgery; otherwise, they continued to Block B +/- C. Block B pts received paclitaxel +/- carboplatin +/- pembrolizumab and Block C pts doxorubicin + cyclophosphamide (AC) +/- pembrolizumab, HR+ immune+ received IO in Blocks B/C. Results are analyzed by receptor and response predictive subtypes (RPS), that include expression-based immune, and luminal signatures with hormone receptor (HR) and HER2 status. Some pts were randomized to a control regimen (skipping R + T-DXd in block A and starting in block B). To minimize risk of interstitial lung disease (ILD) that could lead to treatment discontinuation, pts underwent pulmonary function tests and high-resolution chest CTs every 6 weeks. Abnormalities were reviewed in real time. The primary endpoint was pathologic complete response (pCR), within each RPS and HR/HER2 signature, with estimated pCR (EpCR) rates compared to subtype-specific block A goal rates and control pts in those groups. Other endpoints included safety and treatment discontinuation. Results: 105 pts were treated with R + T-DXd in block A, while 31 pts were randomized to the control regimen. Block A EpCR rate was 57% in the 12 HR+Immune+ pts vs a pre-specified goal of 15% (posterior probability, PP > .99). In 35 HR-Immune+ pts EpCR was 52% vs a pre-specified goal of 40% (PP = .87). No other groups (Immune-) had a PP > .85 for their respective pCR goal rate. HER2-IHC was 0 (27.6%), 1+ (49.5%) and 2+ (22.9%) in the treatment arm vs. 37.7%, 37.7%, and 24.7% in the control arm, with no difference in distribution between arms in immune subtypes. The most common adverse events in Block A were fatigue (84.8%), nausea (81.0%) and alopecia (58.1%). ILD was the most common adverse event of special interest in Block A and overall, occurring in 12 pts (11.4%; 7G1, 4G2, 1G3) during Block A and in 3 pts (G1, 2.9%) in other Blocks. 8 pts discontinued T-DXd + R due to ILD and proceeded to the next Block. Conclusions: When considering the full treatment strategy including block B and C regimens, experimental strategies had similar EpCR rates vs. controls in immune+ subtypes but block A rates exceeded pre-defined goal thresholds. This infers that treatment beginning with R + T-DXd has equivalent efficacy to standard of care for immune+ pts, but 64% pts with pCR skip chemotherapy (Block B) and 97% avoid AC (Block C). R + T-DXd participants avoided persistent ILD. Clinical trial information: NCT01042379 .

Background: There is growing support for high-efficacy disease-modifying therapy (DMT) in multiple sclerosis (MS), but escalation (ESC) approaches remain common. Objective: To describe decision-making in a pragmatic trial of early high-efficacy treatment (EHT) versus ESC. Methods: DELIVER-MS is a multi-center, pragmatic, randomized controlled trial (RCT) with a parallel observational study (OBS), which enrolled treatment-naïve people with RRMS in 31 UK/US sites. Primary outcome was as follows: 36-month brain volume loss according to initial treatment approach (EHT vs. ESC). Stepwise multivariable logistic regression was used to predict participation in RCT versus OBS and choice of EHT versus ESC within the OBS cohort. Results: In total, 816 people with MS were enrolled. Participants declined randomization due to preference for a particular DMT (85%), efficacy concerns (20%), and safety concerns (9%). RCT versus OBS participation was associated with lower relapse rate ( p = 0.043) and greater brain parenchymal fraction ( p = 0.002). Among 374 in the OBS cohort, 125 (33%) chose ESC and 249 (67%) chose EHT. People commencing EHT had higher education attainment ( p < 0.001) and relapse rate ( p = 0.025). Conclusion: Baseline DELIVER-MS data demonstrate that participants with milder disease are more likely to participate in RCT. The choice of EHT versus ESC was associated with demographic factors and disease activity. Clinical trial registration: NCT03535298

e18574 Background: ND has historically been recommended to reduce infectious complications in cancer patients (pts). While the adoption of ND has declined over the last decade, recent studies - particularly in pts undergoing hematopoietic stem cell transplantation (HSCT) - have raised uncertainty regarding the safety of LD. Herein, we performed a meta-analysis of studies directly comparing ND with LD in pts with cancer. Methods: We conducted a systematic search of MEDLINE, Embase, and Cochrane databases from inception through January 2026. Comparative retrospective and prospective studies were included if they were comparing ND with LD in adults or children with cancer. Primary outcomes were major infections and bacteremia/fungemia. Secondary outcomes were pneumonia, diarrhea, and all-cause mortality. Random-effects meta-analyses were performed to calculate pooled risk ratios (RR) with 95% confidence intervals (CI). The review followed PRISMA 2020 guidelines. Results: Fifteen studies enrolling 4,320 patients were included in the primary analysis. Six studies enrolled only adult patients, three were restricted to pediatric populations, and six included mixed adult and pediatric cohorts. Across these studies, ND was not associated with a significant reduction in major infections (RR 1.02, 95% CI 0.88–1.18; p=0.795), bacteremia/fungemia (RR 0.97, 95% CI 0.78–1.22; p=0.805), pneumonia (RR 0.98, 95% CI 0.59–1.63; p=0.933), diarrhea (RR 1.12, 95% CI 0.96–1.31; p=0.159), or all-cause mortality (RR 0.94, 95% CI 0.77–1.15; p=0.632). Sensitivity analyses restricted to RCTs showed consistent results. Subgroup analysis of hematological/HSCT studies (9 studies; 1,769 pts) also showed no differences across outcome. Analysis restricted to RCTs in hematologic/HSCT pts (6 studies; 855 pts) showed no differences in major infections, pneumonia, diarrhea, or mortality; however, ND was associated with a lower risk of bacteremia/fungemia (RR 0.67, 95% CI 0.47–0.95; p=0.023). Heterogeneity across studies was low to moderate. Conclusions: ND was not associated with reductions in major infections, pneumonia, diarrhea, or mortality compared with LD. A reduction in bacteremia/fungemia was observed in RCTs of hematologic malignancy and HSCT pts. Overall, these findings suggest that routine ND implementation is unnecessary for most cancer pts, although cautious dietary practices may remain reasonable in selected high-risk hematologic malignancy/HSCT populations.

9030 Background: AI use is ubiquitous among HOT and healthcare workers, but with variable base knowledge and comfort level. Its indiscriminate use may be problematic. To address this educational gap, we developed a WS for HOT to assess their use of AI and educate them on how to employ it in MedEd and patient-centered CP based on LSP. Methods: A faculty member developed a WS on how to use AI in MedEd and CP based on LSP of analogy, contrasting cases, elaboration, generation, and question driven learning. The WS was for HOT, but faculty could join. It had 3 sections: introduction to basic concepts of AI; principles of the Health Insurance Portability and Accountability Act and protected health information; AI in MedEd; and AI in CP. Sections had real use case examples of AI, for participants to interact, ask questions, and share their personal use cases. HOT were invited to compare and generate examples during the WS, based on their educational and clinical experience and needs. Participants’ AI use and self-perceived and objective knowledge about it were assessed in a pre activity survey, a post WS survey, and a post 8 week survey to evaluate long term learning. Results: A 2 hour WS with 9 HOT and 2 faculty was completed successfully. 10 participants completed the pre WS survey, and 11 completed the post WS and the post 8 week surveys. 100% of them used generative AI tools before the WS, while 10% had prior training in AI. 90% used AI in personal life and MedEd, 70% used it in CP, and 30% in academic research. Table 1 shows participants’ self-assessed knowledge of AI, comfort level in using AI for MedEd and CP, self-assessed preparedness to implement AI in MedEd and CP, and objective knowledge. 8 weeks post WS, 91% of participants reported increased AI use in CP, leading anywhere from slight to significant decreases in time spent on documentation tasks for 91% of them. In MedEd, 73% reported increased AI use, and 82% described slight to significant improvement in the quality of time dedicated to it. 82% described the WS as extremely relevant to their daily responsibilities, and 91% reported confidence navigating ethical and privacy risks of AI in HemOnc. All participants were interested in having a similar WS in the future. Conclusions: An LSP based WS about AI increased the self-perceived and objective knowledge of HOT about using this tool in MedEd and CP. It also increased HOT comfort level, preparedness, and efficiency using AI. With the ubiquitous use of AI by HOT, ongoing education about it focused on supporting high-quality patient care should be a part of fellowship curriculum. Pre WS % Post WS % Post 8 week % Moderately/Very knowledgeable about AI 50 91 100 Comfortable using AI in MedEd 60 82 82 Prepared to implement AI in MedEd 30 91 73 Comfortable using AI in CP 40 73 82 Prepared to implement AI in CP 0 91 73 Correct answers on objective assessment 53 68 71

8549 Background: Despite two decades of anti-angiogenic therapy in NSCLC, no predictive biomarker identifies which patients benefit. As VEGF-targeted combinations advance in clinical development, patient selection biomarkers remain a critical unmet need. QVT (Quantitative Vessel Tortuosity) Score is an automated imaging biomarker measuring chaotic tumor vasculature from baseline CT scans and has been shown to be associated with immune checkpoint inhibitor (ICI) outcomes. We hypothesized that QVT Score could identify patients with chaotic angiogenesis tumors and greater sensitivity to anti-angiogenic therapy. Methods: We analyzed a subset of 334 patients from the SWOG S0819 (NCT00946712) trial with treatment-naïve stage IV non-squamous NSCLC receiving carboplatin/paclitaxel with (56%) or without (44%) bevacizumab. QVT Scores were derived from baseline CTs based on radiologist-defined tumor annotations and radiomic features of tumor vasculature (e.g. vessel twisting, curvature, and branching). Overall survival (OS) was evaluated as the primary endpoint using Cox proportional hazards models with QVT × bevacizumab interaction terms. Bevacizumab was non-randomized in S0819 (physician/patient discretion): to address this, we employed doubly robust estimation combining inverse probability of treatment weighting (IPTW) with multivariable covariate adjustment for age, performance status, histology, smoking history, and stage. Results: The QVT Score × bevacizumab interaction was significant (p=0.007, doubly robust estimation) and stayed consistent across sensitivity analyses (p=0.017 multivariable; p=0.047 IPTW), suggesting differential treatment benefit based on pre-treatment vascular phenotype. QVT Score was strongly OS-associated without bevacizumab (HR=3.03, p=0.003) but not with bevacizumab (HR=1.71, p=0.079), consistent with mitigation of high-risk biology. Bevacizumab benefit increased across QVT quartiles (Table 1): patients in the highest quartile had a 60% reduction in mortality (HR=0.40, p<0.001), while patients in the lowest quartile showed no OS benefit (HR=0.89, p=0.64). Conclusions: Despite unfavorable prognosis on standard-of-care therapies (chemotherapy, ICIs), we found that patients with elevated baseline QVT Score may benefit from the addition of anti-angiogenic agents. As VEGF-targeted combinations including bispecific antibodies enter development, radiomic biomarkers may play a crucial role in enriching trial populations and enabling mechanistic longitudinal monitoring. These findings warrant prospective validation in independent randomized clinical trials. Bevacizumab treatment effect by QVT score quartile. QVT Score Quartile N Bevacizumab effect (Hazard Ratio) P Q1 84 0.89 (0.56 -1.42) 0.64 Q2 83 0.70 (0.44-1.12) 0.14 Q3 83 0.54 (0.34-0.85) 0.01 Q4 84 0.40 (0.24-0.65) 0.0003

8534 Background: Immunotherapy (IO) has transformed treatment for non–small cell lung cancer (NSCLC), but not all patients (pts) benefit, underscoring the need for predictive biomarkers to guide IO versus chemotherapy (Ch). CheckMate 227 (CM227; NCT02477826) showed a survival benefit of first-line nivolumab plus ipilimumab (nivo+ipi) over Ch in stage IV NSCLC. CD8⁺ T cells are key mediators of antitumor immunity and are associated with IO benefit. We developed an artificial intelligence (AI)–based pipeline (VIGOR-CD8) that predicts spatial CD8⁺ immune signatures from routine baseline H&E whole-slide images using histopathology foundation model embeddings (H-Optimus-0) and virtual gene expression modeling (HE2Gene), and evaluated its ability to identify pts who do and do not derive IO benefit over Ch in CM227. Methods: 1,598 pts with advanced NSCLC were analyzed, including multi-institutional retrospective cohorts (n = 487; 65 for patch-level CD8⁺ prediction and overall survival (OS), 422 for patient-level OS) and a blinded CM227 validation subset (n = 1,111). For orthogonal validation, 86,470 H&E patches were co-registered with quantitative CD8⁺ immunofluorescence. H-Optimus-0 and HE2Gene immune-related embeddings were used to train a random forest classifier to predict patch-level CD8⁺ probability. Patch-level probabilities were aggregated into a patient-level CD8⁺ signature and dichotomized into biomarker-positive (B⁺) and biomarker-negative (B⁻) groups by the training-set median. Cox models assessed the impact of VIGOR-CD8 on OS. In CM227, prognostic and predictive utility were evaluated using treatment-specific analyses; investigators were blinded to outcomes, and models were trained on independent, non-overlapping cohorts. Results: VIGOR-CD8 was associated with longer OS in the testing cohort (n = 422; HR 0.68, 95% CI 0.53–0.87, p = 0.00163) and CM227 (n = 1,111; HR 0.8, 95% CI 0.67–0.96, p = 0.016), irrespective of treatment type and PD-L1 expression. Among pts with evaluable PD-L1, B⁺ pts treated with nivo+ipi had superior OS versus Ch (n = 617; HR 0.72, 95% CI 0.58–0.90, p = 0.003), while in B⁻ pts there was no significant OS difference between treatment arms (n = 130; HR 1.18, 95% CI 0.79–1.75, p = 0.436), supporting a predictive rather than purely prognostic role for VIGOR-CD8. Conclusions: An AI-derived CD8⁺ immune signature from routine baseline H&E slides was associated with favorable OS in CM227 and predicted differential benefit from nivo+ipi versus Ch. VIGOR-CD8 may help identify advanced NSCLC pts most likely to benefit from first-line dual IO, but further validation in independent and prospective trials is warranted.

625 Background: I-SPY 2 is a multicenter phase II adaptive platform trial evaluating novel agents for stage II-III high-risk breast cancer using response-predictive subtypes (RPS) and a sequential neoadjuvant treatment design of up to three blocks (A/B/C). Investigational regimens are given in Block A, followed by response-guided escalation to standard-of-care (SOC) therapy (tx) in subsequent blocks or early surgery for responders. Topoisomerase I (TOPO1) inhibition plus PARP inhibition may enhance DNA damage and tumor response. DAN-222 is a tumor-targeted polymeric camptothecin conjugate designed to improve the therapeutic index of TOPO1 inhibition. Based on favorable phase 1 safety and preliminary activity in metastatic breast cancer, DAN-222 plus niraparib was evaluated as Block A tx in HER2− disease. Notably, the neoadjuvant regimen used a higher dose of niraparib (200 mg daily) than the prior phase 1 metastatic study. Methods: Patients (pts) with HER2− breast cancer were randomized to DAN-222 plus niraparib for up to 12 weeks in Block A. Response was assessed by serial MRI and biopsy. Inadequate responders escalated early to SOC tx in Blocks B/C. Primary endpoint was pathologic complete response (pCR). Secondary endpoints included residual cancer burden, safety, early discontinuation, and treatment-strategy performance versus subtype-specific SOC tx controls. Given established PARPi sensitivity in BRCA mut disease, the key objective was to assess clinically meaningful activity of DAN-222 plus niraparib in BRCA wt pts. Accordingly, a prespecified Bayesian interim futility analysis focused on BRCA wt and tested the probability of the true pCR rate after Block A alone exceeded 15% after 50 pts. Results: Fifty pts enrolled from Sep to Dec 2024. In BRCA wt (n=47), estimated Block A pCR was 3%, with posterior probability of 0.02 that that true pCR exceeded 15%, meeting futility criteria and prompting arm closure. Across all blocks, pCR rate was 57% (8/14) for HR-HER2−; 11% (4/36) for HR+ HER2−, largely after SOC tx. 2/3 BRCA mut pts achieved pCR. Treatment strategies initiating with DAN-222 plus niraparib underperformed subtype-specific controls. 68% discontinued Block A early, mainly due to toxicity. Grade ≥3 cytopenias were common: neutropenia (n=25; 50%), anemia (24%), and thrombocytopenia (18%). Cystitis occurred in 36% (1 grade ≥3). Conclusions: Neoadjuvant DAN-222 plus niraparib showed insufficient activity and no treatment-strategy benefit in BRCA wt HER2− breast cancer, underscoring challenges for TOPO1–PARP inhibitor combinations. Although not supported for further development in this setting, these results highlight the strength of I-SPY 2 in rapidly enrolling patients and quickly distinguishing ineffective from promising therapeutic strategies to accelerate progress toward improved outcomes for patients with curable breast cancer. Clinical trial information: NCT01042379 .

3592 Background: Chemotherapy-induced peripheral neuropathy (CIPN) is a frequent dose-limiting toxicity from oxaliplatin that impairs quality of life in gastrointestinal (GI) cancer patients (pts). While strategies to limit exposure exist, their impact on survival remains debated. We evaluated whether reduced oxaliplatin exposure (ROE) maintains survival while mitigating toxicity. Methods: We conducted a systematic review and meta-analysis (PubMed, Embase, Scopus, Web of Science, EBM Reviews) of studies comparing ROE vs standard-duration oxaliplatin-based chemotherapy in GI cancers. ROE was defined as planned limitation of oxaliplatin exposure, including shortened duration (3-4 months), maintenance-based de-escalation, or stop-and-go strategies. Primary endpoints were disease-free survival (DFS) in curative settings, progression-free survival (PFS) in palliative settings, and overall survival (OS). Secondary endpoints included grade ≥3 CIPN and grade ≥3 adverse events (AEs). Random-effects models used Hartung-Knapp-Sidik-Jonkman correction. Heterogeneity was assessed via I² and Cochran’s Q. Results: 24 studies, comprising 15 randomized controlled trials (RCTs) and 9 retrospective cohorts, were included (N = 30,404 pts; ROE: 13,348; control: 17,056). Settings included adjuvant colorectal cancer (CRC; n = 14), metastatic CRC (mCRC; n = 5), gastric cancer (GC) (adjuvant n = 1, metastatic n = 3), and metastatic pancreatic cancer (PDAC) (n = 1). ROE strategies included shortened duration (n = 14), maintenance (n = 8), and stop-and-go (n = 2). In adjuvant CRC, ROE did not compromise DFS (HR 1.07; 95% CI, 0.98-1.16; p = 0.12; I² = 21.7%), including in high-risk subgroups (HR 1.14; 95% CI, 0.93-1.40; p = 0.14; I² = 38.9%); OS was also not different between groups (HR 1.00; 95% CI, 0.94-1.07; p = 0.89; I² = 0%), including high-risk pts (HR 1.01; 95% CI, 0.93-1.10; p = 0.68; I² = 0%). In mCRC, ROE showed no detriment to PFS (HR 1.11; 95% CI, 0.99-1.25; p = 0.07; I² = 0%) or OS (HR 1.06; 95% CI, 0.92-1.23; p = 0.33; I² = 0%). Similarly, in metastatic gastric cancer, PFS (HR 0.82; 95% CI, 0.34-1.99; p = 0.44; I² = 77.3%) and OS (HR 0.90; 95% CI, 0.67-1.21; p = 0.26; I² = 0%) were comparable. Findings remained concordant after restricting the analysis to only RCTs. Pooled analysis was not feasible for PDAC and adjuvant GC studies, but individual study findings were consistent with the above. ROE significantly reduced the risk of grade ≥3 CIPN (OR 0.32; 95% CI, 0.20-0.53; p < 0.001; I² = 80.6%) and overall grade ≥3 AEs (OR 0.65; 95% CI, 0.50-0.86; p = 0.01; I² = 59.9%). Conclusions: Across GI malignancies, ROE was not associated with inferior DFS, PFS, or OS and was associated with lower rates of severe neurotoxicity and high-grade AEs. These findings support planned limitation of oxaliplatin exposure as a strategy to improve the therapeutic index of oxaliplatin-based chemotherapy across disease settings.