Citations

TPS649 Background: While advancements in early-stage breast cancer (ESBC) treatment have improved survival, disease recurrence and therapeutic resistance remain significant challenges. Current biomarkers guide initial treatment, yet novel diagnostics are required to improve risk stratification and recurrence prediction. Circulating tumor DNA (ctDNA) offers a non-invasive method for detecting minimal residual disease (MRD), potentially enabling the identification of molecular relapse before radiographic progression. However, deeper validation of ctDNA kinetics across the treatment continuum is needed to inform precision oncology strategies in ESBC. Additionally, the logistical complexity of tissue-dependent assays presents a barrier to scalability. In order to address these gaps, the purpose of the GEMINI-BREAST study is to evaluate a methylation-based, tissue-free ctDNA assay (Tempus xM) and integrate this with comprehensive clinical and multi-omic data to determine prognostic utility for invasive disease-free survival (iDFS) and long term outcomes in patients with ESBC. Methods: GEMINI-BREAST (NCT07211178) is a prospective, non-interventional, multi-center study enrolling up to 900 participants with ESBC treated with curative intent. Eligible patients enroll into one of three analytical cohorts (n ~ 300 each): Cohort 1 enrolls pre-neoadjuvant therapy (high-risk HR+/HER2- [Stage II-III], HER2+ [Stage II-III], or TNBC [Stage I-III]); Cohort 2 enrolls post-surgery/pre-adjuvant therapy (same subtypes, no evidence of disease); and Cohort 3 enrolls long-term survivors (high-risk HR+/HER2- [Stage II-III], ≥ 5 years since diagnosis). Longitudinal blood samples for ctDNA analysis and archival tissue for NGS are collected at standard-of-care intervals, including, as applicable: pre-neoadjuvant, neoadjuvant on-treatment (weeks 3, 6, 12) and post-treatment, post-surgical/pre-adjuvant, adjuvant on-treatment (every 3 months), end of definitive therapy, surveillance for up to 5 years (every 3-6 months for Cohorts 1/2; every 6-12 months for Cohort 3), and at progression/recurrence. The primary objective is to evaluate the association between ctDNA dynamics (Cohort 1) or MRD status (Cohort 2/3) and iDFS, with secondary endpoints assessing overall survival, lead time to recurrence, and the predictive value of ctDNA clearance for pathological complete response. Recruitment is currently ongoing at U.S. sites. Clinical trial information: NCT07211178 .

TPS4642 Background: Although first-line (1L) EV + pembrolizumab has improved outcomes for patients with la/mUC, most experience disease progression, highlighting an unmet need for novel therapies. TROP2 is a glycoprotein highly expressed in several epithelial tumors, including UC. Dato-DXd is a TROP2-directed antibody–drug conjugate comprising an anti-TROP2 monoclonal antibody, a tetrapeptide-based cleavable linker, and a topoisomerase I inhibitor payload (DXd), designed to induce selective tumor-cell death and reduce systemic exposure to the payload. Dato-DXd has demonstrated durable efficacy and a manageable safety profile in patients with la/mUC both as monotherapy (TROPION-PanTumor01 [NCT03401385]) and in combination with immunotherapy (TROPION-PanTumor03 [NCT05489211]). We describe a Phase 2/3 trial comparing the efficacy and safety of Dato-DXd + platinum CT (carboplatin or cisplatin) vs gemcitabine + platinum CT in participants with la/mUC with disease progression on or after EV + pembrolizumab. Methods: TROPION-Urothelial03 (NCT07129993) is a global, multicenter, randomized, open-label, Phase 2/3 trial. In Part A (Phase 2), ~60 participants will be randomized 1:1 to receive Dato-DXd 4 or 6 mg/kg + platinum CT. In Part B (Phase 3), ~570 participants will be randomized 1:1 to receive Dato-DXd at the recommended Phase 3 dose (determined using Part A data) + platinum CT vs gemcitabine + platinum CT. Randomization in Part A is stratified by assignment of platinum CT (carboplatin vs cisplatin), and in Part B, by assignment of platinum CT, presence vs absence of liver metastasis at screening, and time to progression on 1L EV + pembrolizumab (<6 vs ≥6 months). Study treatment continues until disease progression, unacceptable toxicity, withdrawal of consent, death, loss to follow-up, or other reasons per protocol. Radiographic tumor assessments occur every 6 weeks for 54 weeks, then every 12 weeks. The primary endpoint of Part A is objective response rate (ORR) by investigator per Response Evaluation Criteria in Solid Tumours, version 1.1 (RECIST 1.1). The dual primary endpoints of Part B are progression-free survival (PFS) by blinded independent central review (BICR) per RECIST 1.1 and overall survival (OS). Secondary endpoints include duration of response by investigator per RECIST 1.1 (Part A), and PFS by investigator per RECIST 1.1 and ORR by BICR and investigator per RECIST 1.1 (Part B). In Part A, ORR will be summarized with 2-sided 95% exact confidence intervals using the Clopper–Pearson method by trial group. In Part B, PFS and OS will be compared between trial groups using a log-rank test stratified by randomization stratification factors, with hazard ratios from the stratified Cox regression model. Clinical trial information: NCT07129993 .

TPS1147 Background: Due to dramatic improvements in neoadjuvant and adjuvant HER2-directed therapy (tx), most patients (pts) with early HER2+ breast cancer are cured. As a result, over half of pts newly diagnosed with HER2+ metastatic breast cancer (MBC) now present with de novo stage IV disease. Anti-HER2 tx has also significantly extended survival for pts with HER2+ MBC, with a subgroup of exceptional responders alive many years (yrs) after diagnosis. However, the paradigm for HER2+ MBC remains non-curative, and pts receive tx indefinitely with significant toxicities and costs. The SAPPHO study is investigating whether an intensification approach of sequential, non-cross resistant anti-HER2 tx followed by tx discontinuation is associated with long-term disease control in pts with HER2+ MBC. Methods: SAPPHO is an open-label, phase II, single-arm trial testing a sequential regimen of non-cross resistant, HER2-targeted tx with curative intent in pts with de novo HER2+ MBC. Eligible pts must have biopsy-proven, de novo MBC with high HER2 expression (3+ by immunohistochemistry). Pts with brain metastases are eligible upon receipt of local tx. Treatment consists of an induction regimen (trastuzumab-pertuzumab-taxane [THP] x 4 cycles, followed by trastuzumab deruxtecan [TDXd] x 6 cycles, followed by trastuzumab emtansine [TDM1]-tucatinib x 4 cycles), followed by a maintenance regimen (HP-tucatinib for 1 yr). Given results from DESTINY-Breast09, the sequence of TDXd-P x 6 cycles followed by THP x 4 followed by TDM1-tucatinib x 4 can be chosen as alternative induction tx per investigator and patient choice. Pts who remain progression-free after completing maintenance will stop all anti-HER2 tx. Endocrine tx will be continued for pts with hormone receptor+/HER2+ tumors. Tumor specimens from breast and a metastatic site are collected at baseline and the end of induction. Serial plasma samples for ctDNA analysis are collected at baseline, during treatment, and during follow-up. The primary endpoint is the probability of being progression-free and off anti-HER2 tx 4 yrs from the start of induction. With a sample size of 72 pts, the study is designed to have 91% power to reject the null hypothesis that the probability of being off anti-HER2 tx and progression-free is less than 24%, with an alternative hypothesis of 40%. Key secondary endpoints are overall survival, overall response rate by modified RECIST 1.1 after induction, and safety. Correlative endpoints include the relationship between ctDNA dynamics and outcomes. Patient-reported outcomes will be analyzed, including quality of life, illness intrusiveness, financial toxicity, anxiety, distress about cancer progression, perception of benefit and risk of progression. SAPPHO began enrollment in Q3 2024, and the study is open at 4 US sites within the Translational Breast Cancer Research Consortium (NCT05721248). Clinical trial information: NCT05721248 .

e15509 Background: Large language models (LLMs) are increasingly explored for oncology decision support, yet comparative evidence evaluating medical-specific versus general-purpose LLMs using expert-reviewed clinical cases is limited. We compared a medical-specific LLM (Google MedGemma 27B) with a general-purpose LLM (ChatGPT-5) using gastrointestinal (GI) and genitourinary (GU) oncology vignettes. Methods: This cross-sectional study evaluated 32 de-identified oncology vignettes (26 GI, 6 GU) created by a U.S. board-certified medical oncologist and informed by real-world cases. Vignettes included text-only information on the clinical context, description of pathology, tumor staging, biomarkers, and relevant comorbidities. Both LLMs were prompted to provide National Comprehensive Cancer Network(NCCN) guideline-based treatment recommendations, monitoring strategies, maintenance therapy considerations, and references, along with a self-reported confidence score (0–100). Board-certified U.S. medical oncologists independently reviewed outputs using a structured instrument; two disease specific expert oncologists evaluated each case. Outcomes included correctness and comprehensiveness (1–10 scales), hallucinations (factually fabricated or unsupported statements), NCCN traceability, potential patient harm, and clinician trust. Associations between accuracy and comprehensiveness were assessed using correlation coefficients. Inter-rater reliability was assessed using a two-way random-effects intraclass correlation coefficient (ICC). Results: ChatGPT-5 demonstrated higher mean correctness (8.65 vs 4.68) and comprehensiveness (8.68 vs 5.22) scores compared with MedGemma 27B. Accuracy and comprehensiveness were strongly correlated overall (Pearson r = 0.93, p < 0.001), with significant model-specific correlations (ChatGPT-5 r = 0.86, p < 0.001; MedGemma r = 0.77, p < 0.001). Hallucinations occurred less frequently with ChatGPT-5 (12.3% vs 65%), and NCCN-traceable recommendations were more common (75.4% vs 36.7%). Responses judged as potentially harmful occurred in 18.5% of ChatGPT-5 outputs versus 70% of MedGemma outputs. Clinician trust favored ChatGPT-5 (84.6% vs 15%). Inter-rater reliability improved with rater averaging (ICC(A,k): correctness 0.72; comprehensiveness 0.64). Conclusions: In expert-reviewed GI and GU oncology cases, ChatGPT-5 demonstrated higher clinical accuracy, comprehensiveness, and guideline alignment than a medical-specific LLM; however, a substantial rate of hallucinations and potentially harmful recommendations persisted across models. These findings underscore that, despite performance differences, current LLMs require rigorous clinician oversight and should not be used as standalone decision-support tools in oncology.

e12707 Background: Anthracycline-based chemotherapy remains a cornerstone of breast cancer treatment but is limited by dose-dependent cardiotoxicity. Sodium-glucose cotransporter-2 inhibitors (SGLT2i) have demonstrated cardioprotective effects across diverse populations. Preclinical models suggest they may mitigate anthracycline-induced cardiotoxicity. However, clinical data evaluating this association remain limited. Methods: We conducted a retrospective cohort study using the TriNetX real world data platform, including adults with breast cancer who received anthracycline within one year of diagnosis. Patients were stratified by SGLT2i use within one year of diagnosis into the SGLT2i group and the non-SGLT2i group. Propensity score matching (PSM) was performed to balance baseline demographics and comorbidities. Outcomes included 3-year all-cause mortality, heart failure, myocardial infarction, atrial fibrillation, mean left ventricular ejection fraction (LVEF), mean Troponin I, and mean brain natriuretic peptide level (BNP), analyzed using risk ratios (RR) and Kaplan-Meier survival analysis. Results: After PSM, 2912 patients were included in each group, with a mean follow-up of 630.3 days in SGLT2i group and 756.7 days in non-SGLT2i group. Mean age was 61.5 years vs 62.3 years (SGLT2i group vs non-SGLT2i group), with more than 98% of females in both groups. Most patients were White (49.0% in SGLT2i group and 48.6% in non-SGLT2i group), followed by African American (23.7% vs. 23.4%), unknown races (14.0% vs 14.0%), and Asian (9.5% vs. 9.8%). At baseline, the SGLT2i group had lower mean LVEF (53.4% vs 61.9%), higher mean Troponin I (0.992 vs 0.0849), and higher mean BNP levels (778 vs 240). SGLT2i group was associated with significantly lower 3-year all-cause mortality (Hazard ratio 0.52, 95% CI 0.44–0.61). Heart failure incidence (9.5% vs 6.0%; p = 0.0005) was significantly higher in the SGLT2i group. Mean LVEF in SGLT2i group was significantly lower (51.7% vs 58.9; p < 0.0001) and mean BNP level was significantly higher in SGLT2i group (685.5 vs 383.0; p = 0.0336). No significant differences were observed in myocardial infarction (3.9% vs 3.4%; p = 0.4459), atrial fibrillation (3.4% vs 4.5%; p = 0.0799), and mean troponin I (0.085 vs 0.113, p = 0.5191). Conclusions: In this real-world cohort, SGLT2i use was associated with lower 3-year all-cause mortality despite higher rates of heart failure. Notably, patients receiving SGLT2i had worse baseline cardiac function yet demonstrated more favorable trends in cardiac biomarkers, including smaller declines in LVEF, reductions in BNP, and lower mean troponin I levels, suggesting a slower progression of cardiac dysfunction. These findings support a potential cardioprotective and survival benefit of SGLT2i in patients with breast cancer receiving anthracycline therapy and warrant confirmation in future studies.

Abstract Background Upstream regulators of the homologous recombination (HR) repair pathway are promising targets for overcoming temozolomide (TMZ) resistance. We investigated whether pharmacologic inhibition of Retinoblastoma Binding Protein 4 (RBBP4)/p300-mediated HR activity by CCS1477 or NEO2734 sensitizes glioblastoma (GBM) to TMZ. Methods Lentiviral-mediated shRNA was used to silence RBBP4 and EP300 (p300). A fluorescence-based multiplex flow-cytometric host cell reactivation assay was used to measure DNA repair activity, while TMZ-induced double-strand breaks (DSBs) were assessed using a comet assay and γ-H2AX foci formation. Cytotoxicity was monitored using an Incucyte device, and survival analysis was performed to evaluate efficacy in orthotopic tumor models. Drug distribution was assessed using liquid chromatography-mass spectrometry (LC-MS/MS) and acetylation of lysine 27 of histone H3 (H3K27Ac) immunofluorescence for target inhibition. Promoter occupancy was determined using chromatin immunoprecipitation. Results GBM patient-derived xenograft tumors with high RBBP4 expression showed elevated levels of 6 HR genes (RAD51, RAD50, BRCA1, BARD1, BRIP1, and FIGNL1), with RAD51 and BRIP1 correlating strongly with RBBP4 and p300. Silencing RBBP4 or p300 decreased HR and microhomologous-mediated end joining (MMEJ) activity and delayed repair of TMZ-induced DSBs. CCS1477 and NEO2734 suppressed RAD51 and MYC in a concentration-dependent manner, prolonged γ-H2AX foci, and enhanced TMZ sensitivity. Both compounds inhibited H3K27Ac and were detectable in orthotopic tumors by LC-MS/MS and significantly extended survival alone and in combination with TMZ. Conclusion These findings suggest that the RBBP4/p300-axis is a key regulator of HR-mediated repair of TMZ-induced DSBs, and inhibition by either CCS1477 or NEO2734 may be beneficial as monotherapy, but further studies are needed to determine the benefit of combining these agents with TMZ.

9557 Background: High-dose interleukin-2 (IL-2) is administered following tumor-infiltrating lymphocyte (TIL) therapy, yet cardiotoxicity frequently limits dose delivery. The clinical significance of peri-dose troponin elevations during IL-2 remains poorly defined. We evaluated troponin kinetics during IL-2 dosing and their association with significant cardiac events and IL-2 interruption. Methods: We retrospectively analyzed patients with metastatic melanoma treated with TIL therapy followed by inpatient high-dose IL-2 at Mayo Clinic Rochester (N = 24; 87 IL-2 doses). High-sensitivity troponin and ECGs were obtained before and after each IL-2 dose. Clinically significant cardiac events were defined as QTc prolongation > 480 ms, arrhythmias (excluding isolated sinus tachycardia), new ECG abnormalities, vasopressor requirement, or cardiology consultation. Dose-level outcomes included cardiac events and subsequent IL-2 dose interruption. Paired troponin changes were assessed using Wilcoxon signed-rank testing. Dose-level associations were evaluated using clustered logistic regression to account for repeated IL-2 doses within patients. Results: 24 patients received 87 IL-2 doses (median 3 [IQR 2 - 5]). Median age was 63 years (IQR 54 - 70), and 58% were male. Baseline troponin median was 11 ng/L (IQR 6 - 21); 38% had baseline troponin > 14 ng/L. Post-dose troponin exceeded 14 ng/L in 16/24 patients (67%) and 48/87 doses (52%). Median delta troponin was 0 ng/L (IQR −1 to 6; range −24 to 133), with post-dose levels higher than pre-dose values ( p = 0.02). Clinically significant cardiac events occurred in 33% of patients; no post-IL-2 reductions in left ventricular ejection fraction were observed. At the dose level, cardiac events occurred only when post-dose troponin exceeded 14 ng/L (15/48 [31%] vs 0/39 doses; p < 0.001). In clustered logistic regression, post-dose troponin ≥15 ng/L was independently associated with cardiac events (OR 9.6, 95% CI 1.5 - 60.6; p = 0.02) and subsequent IL-2 dose interruption (OR 3.4, 95% CI 1.1 - 10.7; p = 0.04). When modeled continuously, each 1-unit increase in log-transformed post-dose troponin was associated with higher odds of cardiac events (OR 3.3, 95% CI 1.6 - 6.7; p = 0.001) and IL-2 interruption (OR 2.8, 95% CI 1.6 - 4.9; p < 0.001). Baseline troponin correlated with peak post-dose troponin (ρ = 0.44, p = 0.04). Conclusions: Peri-dose troponin elevations are common during post-TIL high-dose IL-2; however, higher post-dose troponin independently identifies clinically significant cardiac toxicity and predicts IL-2 interruption after accounting for repeated dosing within patients. Baseline troponin modestly predicts peak elevations, supporting standardized, risk-stratified troponin monitoring and prospective evaluation of toxicity-adaptive IL-2 delivery strategies in TIL therapy.

8598 Background: Immune checkpoint inhibitors (ICI) have become a cornerstone of treatment for lung cancer. Glucagon-like peptide-1 receptor agonists (GLP-1 RA) have been associated with improved outcomes in cancer patients in prior studies; however, their impact on outcomes among lung cancer patients receiving immunotherapy remains limited. Methods: We conducted a retrospective cohort study using the TriNetX real-world data platform. Adult patients with lung cancer treated with ICIs, including PD-1, PD-L1, or CTLA-4 inhibitors, were included. Patients were stratified based on exposure to GLP-1 RA. Propensity score matching (1:1) was performed to balance baseline demographics and comorbidities, including type 2 diabetes mellitus and body mass index. Primary outcomes included all-cause mortality, intensive care unit (ICU) admission, and hospitalization. Secondary outcomes included immune-related pneumonitis, colitis/enteritis, cachexia, and major adverse cardiovascular events (MACE). Results: After matching, 2,013 patients were included in each group, with a median follow-up of 372 days in the GLP-1 RA group and 383 days in the non-GLP-1 RA group. Mean age was 66.8 and 67.2 years, respectively, and approximately 50% of patients were male. Most patients were White (77.2% vs 71.8%), followed by Black (14.1% vs 14.0%) and Asian (2.1% vs 1.8%). GLP-1 RA use was associated with a significantly lower all-cause mortality rate compared with non-use (35.7% vs. 53.4%; p < 0.0001), with a hazard ratio of 0.65 (95% CI, 0.59–0.71). Significant reductions were also observed in ICU admissions (22.8% vs 29.0%; p = 0.0001), hospitalizations (64.1% vs 71.8%; p = 0.0001), pneumonitis (5.2% vs 6.8%; p = 0.0457), and MACE (19.9% vs 25.3%; p = 0.0026). No significant differences were observed in colitis/enteritis (9.3% vs 9.4%; p = 0.9704) or cachexia (8.0% vs 9.8%; p = 0.0650). In a subgroup analysis of lung cancer patients not receiving ICI, GLP-1 RA use was similarly associated with lower mortality, ICU admission, hospitalization, and MACE, without differences in immune-related toxicities. Conclusions: GLP-1 RA use among lung cancer patients was associated with improved survival, reduced healthcare utilization, and fewer selected adverse events in patients receiving ICI therapy. These associations likely reflect established metabolic and cardiovascular benefits of GLP-1 RA and may also involve antitumor mechanisms suggested in prior studies. Residual confounding related to healthcare access and socioeconomic factors cannot be fully excluded and should be considered when interpreting these findings. Prospective studies are warranted to validate these observations and clarify underlying biological mechanisms.

7543 Background: Delayed neurotoxicity (DNT) after ciltacabtagene autoleucel (cilta-cel) is associated with severe morbidity and mortality occurring in 10-15% of patients. A high post-infusion peak absolute lymphocyte count (ALCpeak ≥3x10 9 /L; ALC hi ) is correlated with an increased risk of DNT (25-30%). We previously examined prophylactic dexamethasone (PPX DEX) administered for 3 days when ALC hi was reached and found no reduction in ALC expansion, DNT incidence or severity. Here, we report the use of prophylactic cyclophosphamide (PPX CTX) on ALC kinetics and DNT. Methods: We conducted a retrospective analysis of patients (pts) treated with cilta-cel at Mayo Clinic (09-12/2025; n =60). Pts with ALC hi received CTX 500 mg IV/PO once (n = 22). Outcomes and ALC kinetics including ALCpeak, ALC D28 (ALC D28 ), and area under the curve D1-D28 (ALC AUC ) were compared with PPX DEX cohort (12/2024-08/2025; n =50) and a historical cohort without PPX (04/2022-11/2024; n =116). Results: Pre-lymphodepletion (LD) demographics were similar across historical ALC hi , PPX DEX, and PPX CTX groups. Pre-LD CRP was higher in the historical ALC hi cohort; however, baseline ferritin, CRS, and ICANS, and disease burden did not differ significantly between groups. There were no statistically significant differences in time to ALCpeak or ALC kinetics among the 3 cohorts. Among pts with ALC < 3×10⁹/L treated during the PPX CTX study period, 2 developed cranial nerve palsy (IEC-CNP; 2/38, 5%). Both presented with unilateral facial nerve palsy and were treated with corticosteroids. One pt resolved symptoms by one month, while the second had persistent symptoms at 2 weeks and remains on follow-up. Among ALC hi patients in the PPX CTX cohort, DNT occurred in 4/22 pts (18%), including 1 case of IEC-CNP (5%) and 3 with parkinsonism (IEC-PKS, 14%). Within the PPX CTX cohort, baseline characteristics and post-infusion CRS or laboratory parameters did not differ between patients with and without DNT. Among patients who developed DNT, ALC kinetics were similar across PPX CTX, PPX DEX, and historical cohorts. PPX CTX was well tolerated, with 2 cases of grade 1 nausea reported. The incidence of DNT did not differ significantly between the PPX DEX (14/61, 23%) and PPX CTX (4/22, 18%) cohorts. However, DNT presentations in the PPX CTX cohort appeared milder and were responsive to treatment. Two of three patients (67%) presented with mild symptoms (one with dynamic dysphasia; second with flat affect) that responded to DEX. A third patient presented with typical parkinsonian features (hypomimia, bradykinesia, rigidity) and demonstrated clinical improvement following high-dose CTX. Conclusions: PPX CTX was well tolerated. While it did not significantly impact ALC expansion kinetics or DNT rates, first few cases of IEC-PKS suggest that the presentation may be milder and amenable to less toxic treatment. Larger cohorts are needed to refine optimal prophylactic strategies.