Browse Articles
Discover research articles across all indexed journals
Efficacy and safety by tumor location in the phase 3 PANOVA-3 trial of Tumor Treating Fields (TTFields) with gemcitabine/nab-paclitaxel in locally advanced pancreatic adenocarcinoma.
e16367 Background: Outcomes of unresectable, locally advanced pancreatic adenocarcinoma (LAPAC) remain poor, with therapy based mainly on data from trials in metastatic disease. TTFields use alternating electric fields to disrupt cancer cell proliferation. The phase 3 PANOVA-3 trial (NCT03377491) demonstrated that TTFields with gemcitabine/nab-paclitaxel (GnP) significantly improved overall survival (OS; HR 0.82) and pain-free survival (HR 0.74) vs GnP in patients with unresectable LAPAC. We report an analysis of efficacy and safety in PANOVA-3 by tumor location. Methods: Patients with newly diagnosed LAPAC were randomized 1:1 to receive GnP with or without concomitant TTFields. Kaplan-Meier methodology was used for time-to-event analyses. Response evaluations were based on RECIST V1.1. OS and pain-free survival were analyzed post-hoc without stratification in patient subgroups with tumors in the head, body, and tail of pancreas. Differences between treatment arms were compared using a 2-sided log-rank test. Results: Tumors were located in: head of pancreas, n=163 and 158 in the TTFields + GnP and GnP arms, respectively; body of pancreas, n=81 and 79; tail of pancreas, n=9 and 19. Baseline characteristics were generally similar in the TTFields + GnP and GnP arms in each subgroup, with the exception of gender (male: head 51.5% vs 41.1%; body 45.7% vs 48.1%; tail 88.9% vs 52.6%). Median OS was 15.8 (12.7, 18.0) vs 12.9 (11.5, 15.4) months in the TTFields + GnP and GnP arms in the head of pancreas subgroup; 18.6 (95% CI: 15.7, 22.5) vs 16.4 (13.1, 20.1) months in the body of pancreas subgroup; and 13.9 (3.0, 17.1) vs 14.9 (11.5, 18.7) months in the tail of pancreas subgroup (all P>0.05). Median pain-free survival was 10.2 (6.6, 16.5) vs 7.6 (5.9, 9.5) months, 18.7 (11.0, NE) vs 9.3 (6.6, 16.3) months (P=0.041), and 5.4 (1.2, NE) vs NE (4.5, NE) months, respectively, in these subgroups (P>0.05 unless otherwise stated). The type and incidence of device-related adverse events (AEs) were comparable. Conclusions: In this post-hoc subgroup analysis with small sample sizes, median OS and pain-free survival were consistent with the primary analysis of the overall population. These data support the use of TTFields with GnP in unresectable LAPAC regardless of tumor location. Clinical trial information: NCT03377491 .
Systematic evaluation of machine learning optimization strategies for uveal melanoma detection using ultra-widefield photography.
9560 Background: Early differentiation of uveal melanoma (UM) from benign intraocular lesions remains clinically challenging, particularly in regions with limited access to specialized ocular oncologists. Diagnostic uncertainty can delay referral and treatment, highlighting a gap between current practice and the desired goal of timely, accurate early UM detection. Machine learning (ML) offers a promising approach to improving diagnostic support with early detection of UM. This study developed, trained, and optimized eight ML classification models for early UM detection using ultrawidefield fundus (UWF) imaging to evaluate methods to increase clinical utility and readiness of current models. Methods: 1,592 UWF images from 784 patients were retrospectively collected and labeled into four classes: UM (n = 364), choroidal nevus (n = 824), congenital hypertrophy of the retinal pigment epithelium (CHRPE, n = 102), and healthy controls (n = 302). Images with multiple lesions, significant media opacity, or poor tumor visualization were excluded. All models used a ResNet-50 architecture, and a binary UM/nevus classifier was developed as the control. Optimization strategies included 1) class expansion with healthy control and CHRPE images, 2) training dataset augmentation with multiple images per patient and post-treatment images, and 3) a region of interest (ROI) dilation preprocessing pipeline. A 70/15/15 train/validation/test split was used with patient-level stratification to prevent data leakage. The test set was held constant across all models to ensure fair comparison. Performance was assessed using per-class F1 score and area under the curve (AUC) with 5-fold cross-validation. Grad-CAM images were generated to aid in model interpretability. Results: The control model achieved an AUC of 0.94±0.01 and F1 scores of 0.90±0.01 for nevus and 0.80±0.03 for UM. ROI dilation achieved the greatest improvement (AUC 0.99±0.001; nevus F1 0.98±0.01; UM F1 0.95±0.01). Other optimization strategies showed variable impact, with multi-optos yielding the second-best performance (AUC 0.95±0.02; nevus F1 0.91±0.02; UM F1 0.83±0.04). A final model integrating the most effective strategies demonstrated robust performance (AUC 0.95±0.01; nevus F1 0.91±0.03; UM F1 0.82±0.06). Grad-CAM confirmed an appropriate lesion-centered focus. Conclusions: Our findings show that targeted ML optimizations can meaningfully improve performance for early UM detection. While ROI dilation preprocessing had the most performance gain, results varied across optimization pathways, indicating room for refinement. Further evaluation on external datasets is needed to assess generalizability, paving the way for end-stage clinical rollout.
Multi-modal molecular profiling of CDH1 and CLDN18.2 to identify predictive biomarkers for zolbetuximab efficacy in patients with advanced gastric adenocarcinoma: The ZOL-C3 study.
3148 Background: Zolbetuximab is a chimeric monoclonal antibody designed to bind to claudin 18.2 (CLDN18.2) and induce antibody-dependent cell-mediated cytotoxicity (ADCC). Although phase 3 trials established its efficacy in CLDN18.2-positive advanced gastric adenocarcinoma (GAC), robust predictive biomarkers beyond protein expression remain an unmet need. We investigated how epithelial-mesenchymal transition (EMT), mediated by CDH1 alterations and the tumor mutational burden (TMB), modulates the therapeutic target and affects zolbetuximab-mediated ADCC efficacy. Methods: This study employed a multi-layered integrative framework across four independent datasets to cross-validate genomic, transcriptomic, and phenotypic findings: (1) mechanistic characterization using the TCGA-STAD cohort (n=375); (2) the genomic screening of 2,483 patients with GAC in the Japanese national database (C-CAT), which identified 49 treated with zolbetuximab-containing regimens; (3) a spatial transcriptomic evaluation (10× Genomics Visium platform, GSE251950); and (4) immunohistochemical (IHC) validation of CLDN18.2 in 36 institutional GAC patients. The trial is registered at UMIN-CTR (UMIN000060223), a WHO ICTRP primary registry. Results: In the TCGA cohort, CDH1 low-expression was significantly associated with 3.2-fold lower CLDN18 mRNA levels than in the high-expression group (p <0.001). CDH1 expression was strictly regulated by promoter methylation (p <0.001), whereas CLDN18 down-regulation in CDH1-low tumors appeared to be driven by non-epigenetic EMT pathways. Spatial transcriptomics revealed a spatial correlation between CDH1 and CLDN18 mRNAs within the malignant epithelial compartment (p <0.001). The objective response rate within the zolbetuximab-treated C-CAT subset was 22.2%. TMB was markedly lower in responders than in non-responders (median 1.82 vs. 4.0 mut/Mb; p = 0.050). Furthermore, patients without CDH1 single nucleotide variants had higher ORR (p = 0.086). Institutional IHC validation revealed a superior response in cases that retained membranous CDH1 expression to that in cases with the loss of epithelial integrity (100% vs 50% ORR; p=0.088). Conclusions: These results demonstrate that CDH1/E-cadherin integrity and low TMB are critical co-determinants of CLDN18.2 expression and zolbetuximab efficacy. Maintenance of the epithelial phenotype appears to be essential for optimal zolbetuximab-mediated ADCC. The integration of comprehensive molecular profiling, including the CDH1 status and TMB, may significantly refine patient selection, with a shift from simple target positivity to a more biological enrichment strategy for zolbetuximab therapy.
Mutational signatures as potential predictors of HIPEC response and recurrence in ovarian cancer.
e17577 Background: While Hyperthermic intraperitoneal chemotherapy (HIPEC) has demonstrated overall survival benefit in advanced ovarian cancer (OC) patients, no predictive biomarkers exist for optimal patient selection. Methods: A cohort of 41 patients with primary or recurrent ovarian cancer underwent HIPEC with cisplatin 75 mg/m2 (60 minutes, 42°C at optimal cytoreductive surgery, CRS) at COH between 2012 and 2022 (NCT01970722). Tumor and normal samples or buffy coat were collected at time of HIPEC CRS and subjected to genomic DNA isolation. Whole exome sequencing (WES) genomic libraries were constructed. HIPEC response was defined by the following progression-free survival (PFS) cut-off values to distinguish good vs poor responders: 18 months in primary EOC patients (based on KGOG, CARCINO-HIPEC trials), and 12 months in recurrent EOC patients (based on MSK, CHIPOR HIPEC trials). Recurrence type was defined as peritoneal-only (only within the peritoneal cavity) vs. extraperitoneal recurrence (any type of recurrence which included extraperitoneal metastases). Mutational signatures for single and doublet-base substitutions (DBSs) and indels were analyzed using SigProfiler (Github, UC San Diego, CA) and fitted to the COSMIC v3.1 database, excluding noncontributory signatures. The Wilcoxon signed-rank test was used to compare cosmic mutational burden between good and poor responders, and peritoneal-only vs extraperitoneal recurrences. Results: A total of 36 EOC tumor samples were analyzed, after exclusion of 5 samples due to lack of exonic variants, or absent normal samples. 21 (58%) were primary EOC, and 15 (42%) were recurrent EOC. Eighteen (50%) patients were identified as good responders, and 18 (50%) as poor responders, with OS of NR (46.5, NR), and NR (27.2, NR), respectively (p=0.02). BRCA1/2 mutation was present in 11.2% of patients. Peritoneal-only recurrence occurred in 23 (64%) of patients, and extraperitoneal recurrence occurred in 13 (36%). Median PFS in the peritoneal-only group was 15.3 months (95%CI, 11.6, 19.5) vs 14.3 months (9.7, 20.8) in the extraperitoneal group. The most common gene mutation in the entire cohort was TP53 (47%). Top OncoPathways included TP53, RTK-RAS, NOTCH, and PI3K. The cosmic mutational signature SBS5 (unknown etiology) was enriched in good HIPEC responders compared to poor responders (p=0.017). Patients with peritoneal-only recurrence had higher SBS15 (DNA mismatch repair) mutational burden compared to patients with extraperitoneal recurrence (p=0.048). Conclusions: SBS5 cosmic signature in EOC tumors is associated with HIPEC response in ovarian cancer and could be validated as a prognostic biomarker in a larger validation cohort. DNA mismatch repair signature SBS15 is associated with peritoneal only recurrence after HIPEC in ovarian cancer, and its absent mutational burden may predict increased risk of extraperitoneal recurrence.
Temporal trends in intensive care unit outcomes for lung cancer patients: A 15-year analysis of the MIMIC-IV database.
e20693 Background: Lung cancer care has been transformed over the past decade. Targeted therapies and immunotherapy have extended median survival from roughly 10 months to over 38 months-a remarkable achievement. But here's the question no one has answered: when these patients become critically ill and need the ICU, are they faring any better than they were 15 years ago? Methods: We analyzed 1,689 lung cancer patients admitted to the ICU using the MIMIC-IV database (2008–2022), dividing the cohort into three treatment eras: 2008–2012 (n = 570), 2013–2017 (n = 503), and 2018–2022 (n = 616). We tracked hospital mortality, ICU mortality, why patients were being admitted, what interventions they received, and where they went after discharge. Results: Despite everything that's changed in lung cancer treatment, hospital mortality held steady at about 21% across all three eras (20.2%, 21.7%, and 20.9%; p = 0.834). ICU mortality told the same story: 12.6%, 12.3%, and 12.5% (p = 0.989). What did change was striking. The patients arriving in the ICU looked completely different. Metastatic disease rose from 51% to 61%. The reasons for admission shifted dramatically in respiratory failure dropped from 26% to just 1%, while sepsis climbed from essentially zero to 19%, and cardiac complications surged from 4% to 42%. We also saw less mechanical ventilation (39% down to 32%) and more hospice discharges (6% up to 11%). Conclusions: The oncologic revolution hasn't reached the ICU. Mortality stayed flat at 21% because what kills critically ill patients isn't cancer biology but it's acute organ failure. Meanwhile, the newer therapies brought their own problems: myocarditis, immune-related sepsis, and other complications that created entirely new pathways to the ICU. Add in the fact that we're now admitting patients with more advanced disease, and the survival gains were effectively cancelled out. These numbers matter for the conversations we have with patients and families. When a lung cancer patient becomes critically ill, the prognosis today is essentially what it was in 2008 and that reality should inform goals-of-care discussions.
Effects of genome-wide modifiers of response to next-generation FGFR inhibitors in solid tumors—An Indian multicentric study.
e23422 Background: FGFR alterations are established actionable drivers in solid tumors, with particular clinical relevance in hepatobiliary malignancies. The advent of selective and next-generation FGFR inhibitors has expanded therapeutic options; however, durable responses remain inconsistent. Emerging evidence suggests that co-occurring genomic alterations significantly influence sensitivity and resistance to FGFR-directed therapies. We report a large Indian multicentric genomic analysis evaluating FGFR alterations, actionability, and resistance mechanisms with direct relevance to contemporary FGFR inhibitors. Methods: Comprehensive genome-wide next-generation sequencing data from 791 patients with solid tumors across multiple Indian centers were analyzed. Alterations in FGFR1–4 were characterized by type, oncogenicity, and clinical actionability. Known and putative resistance mechanisms—including FGFR gatekeeper mutations and co-alterations in bypass signaling pathways—were systematically evaluated. Results: A total of 848 FGFR alterations were identified in 791 patients. FGFR1 (35.5%), FGFR3 (34.6%), and FGFR2 (23.7%) were most frequently altered. Missense variants predominated (56.5%), followed by amplifications (26.9%), deletions (9.3%), and fusions (2.7%). Only 4.2% of FGFR alterations met Level 1 or 2 evidence for clinical actionability, emphasizing the need for careful variant interpretation when considering FGFR inhibitor therapy. FGFR amplifications are also targetable through panTKI agents which result in variable therapeutic response. Clinically relevant resistance mechanisms were identified in 41.8% of patients. FGFR gatekeeper mutations—including FGFR2 N549K, FGFR3 V555M, and FGFR4 V550M were observed in about 2% cases and may impact sensitivity even to newer-generation FGFR inhibitors. Additionally, frequent co-alterations were detected in key oncogenic pathways, notably PIK3CA (15.5%), KRAS (13.4%), PTEN (6.8%), and NRAS (2.5%), suggesting parallel MAPK and PI3K pathway activation as major contributors to intrinsic or early acquired resistance. Our data indicate that both drug-agnostic (pathway bypass) and drug-specific (FGFR gatekeeper) mechanisms coexist at conspicuous frequencies. Conclusions: In the era of next-generation FGFR inhibitors, FGFR alterations alone are insufficient to guide precision therapy. This Indian multicentric pan cancer real-world dataset demonstrates that a substantial proportion of FGFR-altered tumors harbor concurrent genomic events with potential to attenuate therapeutic benefit. Genome-wide profiling should be considered standard in FGFR-driven solid tumors to refine patient selection, anticipate resistance, and inform rational combination or sequencing strategies in precision oncology practice.
Efficacy of immunotherapy-based regimens in Russian patients with <i>EGFR</i> -mutated metastatic NSCLC after EGFR TKI failure: A multi-institution real-world analysis.
e20628 Background: Patients with EGFR-mutated (EGFRm) metastatic non-small cell lung cancer (mNSCLC) generally derive limited benefit from immune checkpoint inhibitors (ICIs), yet outcomes may vary across clinical and molecular subgroups. We evaluated real-world outcomes of ICI-based therapy versus chemotherapy with/without bevacizumab after EGFR TKI progression in a Russian cohort. Methods: Multi-institution retrospective analysis (2019–2025) of EGFRm mNSCLC patients after EGFR TKI progression treated with subsequent systemic therapy. Regimens were categorized as chemotherapy ± bevacizumab (CT±B) or ICI-based therapy (ICI+CT±B). Progression Free Survival (PFS) was measured from start of the subsequent line to progression or death. Analyses were exploratory and unadjusted for confounding; subgroup analyses included EGFR genotype, T790M , PD-L1 expression, and ECOG. Results: A total of 258 patients were included; 70.6% were female. Median age was 62 years (range 28–87); 68% were never-smokers. EGFR variants included ex19del (59.2%), L858R (35.0%), and rare mutations (7.8%). T790M testing was performed in 24.6% of patients and was positive in 42.8% of those tested. Subsequent therapy consisted of CT±B in 68% and ICI-based therapy in 32%. Overall, mPFS favored ICI-based therapy versus CT±B (6.7 vs 4.7 months; p = 0.03) and versus CT+B (6.7 vs 5.5 months; p = 0.043). By genotype, mPFS favored ICI-based therapy in L858R (8.3 vs 5.2 months; p = 0.044) and showed a trend in rare mutations (6.8 vs 3.9 months; p = 0.064), but not in ex19del (5.9 vs 4.5 months; p = 0.27). By T790M status, no difference was observed in T790M+ disease (4.4 vs 4.2 months; p = 0.326), whereas mPFS was longer with ICI-based therapy in T790M− disease (7.0 vs 4.8 months; p = 0.02). The PFS benefit with ICI-based therapy was more statistically pronounced in PD-L1 ≥1% compared with PD-L1 = 0%, and did not vary by ECOG status. Conclusions: In our Russian real-world EGFRm mNSCLC cohort after EGFR TKI failure ICI-based regimens were associated with longer PFS compared with CT±B, with the most pronounced benefit in L858R and possibly rare EGFR mutations, limited or no benefit in ex19del and T790M+ disease, and stronger signal in PD-L1 ≥1%. Findings are exploratory and warrant validation in confounding-adjusted analyses and prospective studies.
Myelodysplasia-related mutation dynamics and outcomes in newly diagnosed (ND) acute myeloid leukemia (AML) treated with low-intensity therapy (LIT).
6531 Background: Outcomes of patients (pts) with ND AML with myelodysplasia related gene mutations (MRMs), while traditionally adverse, appear to be better with LIT regimens + venetoclax (Ven). The impact of MRMs as a minimal residual disease (MRD) marker warrants investigation. Methods: We retrospectively analyzed the impact of persistence/clearance of MRM at VAF <2% at best response of CR/CRi, in pts with LIT treated ND AML (with available genomic data), treated at our institution between 2017 – 2025 and had ≥1 MRM ( ASXL1, BCOR, EZH2, STAG2, SF3B1, SRSF2, U2AF1 , and/or ZRSR2 ) at diagnosis, at a VAF of ≥2%. We excluded pts with CBF-AML, APL, or treated secondary AML. MRM dynamics were tracked at baseline and best response (CR/CRi) and correlated with relapse free/overall survival (RFS/OS). Results: A total of 185 pts, median (med) age 69 yrs (range, 56-87 yrs), were included; 96% were ≥ 60 yrs. The common baseline MRM were SRSF2 (54%), ASXL1 (32%), STAG2 (16%) and BCOR (14%); 16% had >1 MRM. Overall, 17% had concurrent NPM1 mutation (mut), 9% FLT3 -ITD (AR>0.05), 22% RAS , 32% RUNX1 , 8% TP53 , and 32/183 pts (17%) had ELN2017 adverse cytogenetics (CTG). 171 pts (92%) were treated with LIT+Ven and 14 (8%) LIT without Ven; 105 (57%) received hypomethylating agents based, and the rest (43%) received cladribine + low-dose cytarabine based LIT. Overall, 148 (80%) achieved CR and 37 (20%) CRi. 117/168 (70%) pts with available data were MRD negative (-) by flow cytometry (FCM; <0.01%) at CR/CRi. At CR/CRi, 49 (26%) cleared their MRMs (MRM-) while 136 (74%) retained MRMs (MRM+); among pts with FCM MRD- CR/CRi, 33/117 (28%) were MRM- and 84/117 (72%) MRM+. Med cycle to CR/CRi was 1 (IQR 1-2). Pts with MRM- at CR/CRi had superior med RFS (24 [95% CI 9-18] vs. 14 mos [16-NR], p=0.008) and OS (54 [25-NR] vs. 21 mos [15-27], p=0.01) compared to MRM+ pts. Among 117 FCM MRD- pts, med RFS (NR vs. 14 mos [9-23], p=0.001) and med OS (NR vs. 19 mos [14-36], p=0.003) was still superior among MRM- pts (n=33) vs. MRM+ (n=84). 67 pts (36%) underwent a hematopoietic stem cell transplantation (HSCT) in CR1, 26/49 (53%) MRM- and 41/136 (30%) MRM+. Among FCM MRD- pts who did not undergo HSCT, 2-yr OS rate (86% vs. 48%, p=0.04) was better in MRM- pts (n=15) vs. MRM+ pts (n=58); among FCM MRD- pts who had HSCT, 2-yr OS rate trended to favor MRM- pts (n=18) (77% vs. 55%, p=.14) vs. MRM+ pts (n=26). On backward selected Cox MVA, clearance of MRMs at CR/CRi was independently associated with favorable OS (HR=0.54, 95% CI 0.31-0.95, p=0.03), along with BCOR, IDH2 mut, and HSCT, while RAS mut, FLT3 -ITD, and adverse CTG were unfavorable. Finally, among FCM-MRD- pts at CR/CRi, using the same Cox model, clearance of MRMs was independently favorable for OS (HR=0.42, 95%CI 0.20-0.93, p=0.03). Conclusions: In our analysis, the status of MRM clearance at CR/CRi in LIT treated AML with baseline MRM affected survival outcomes, including in pts FCM MRD- at CR/CRi.
Outcomes following CAR-T therapy in adults with relapsed/refractory B-ALL according to prior inotuzumab exposure.
e18626 Background: Chimeric antigen receptor T-cell (CAR-T) therapy represents a novel approach to relapsed and refractory B cell acute lymphoblastic leukemia (B-ALL). Its use is associated with toxicity events, namely cytokine release syndrome (CRS) and immune effector cell associated neurotoxicity syndrome (ICANS). Inotuzumab ozogamicin (InO) is a CD22 antibody-drug conjugate that was approved in 2017 for the treatment of R/R B-ALL. We report on the association between prior InO exposure and outcomes following CAR-T therapy. Methods: Retrospective cohort analysis was done through TriNetX, aggregating de-identified patient data from multiple healthcare organizations. Patient cohorts were constructed using clinical codes including all individuals >18 years of age with a diagnosis of B-ALL who subsequently received CAR-T. Cohort analysis was run to evaluate the differences in event free survival (EFS) and overall survival (OS) from 1 to 1825 days. CRS/ICANS were the primary endpoint, while mortality, proxy measures of toxicity (tocilizumab use, ICU admission, etc.), and acute/subacute hepatic disease represented secondary endpoints. Results: A total of 148 patients with and 618 without prior InO exposure were identified. After propensity score matching patients for demographic factors, age at CAR-T administration, sex, and prior chemotherapy/HSCT, 124 patients remained per cohort. The hazard of CRS/ICANS after CAR-T administration was significantly increased in the InO cohort (HR = 1.48, 95% CI 1.04-2.10). Prior InO use was associated with an increased hazard of CAR-T toxicity proxy events (HR 2.41, 95% CI 1.80-3.24) as well as increased cumulative incidence (RR 1.34, 95% CI 1.17-1.53). Hepatic toxicity or disease events occurred in 49.1% of InO patients compared to 35.4% in non-Ino patients (HR 1.65, 95% CI 1.11-2.44). InO exposure was also associated with significantly worse OS after CAR-T (HR = 1.97, 95% CI 1.27-3.04). Conclusions: InO exposure prior to CAR-T was associated with higher rates of CRS/ICANS, hepatic toxicity, and increased mortality in the analyzed B-ALL cohorts. Increased mortality may reflect underlying disease severity and cumulative treatment exposure in patients receiving prior InO. Inotuzumab use may identify a subgroup predisposed to a higher rate of complications that would benefit from additional supportive measures and monitoring post CAR-T. Comparison of post-CAR-T outcomes in InO vs InO-free patients. Outcome Inotuzumab (n=124) No Inotuzumab (n=124) Measure of Association CRS/ICANS 72 55 HR 1.48 (1.04, 2.10) CAR-T Event Proxies 113 84 HR 2.41 (1.80, 3.24) Hepatic Toxicity 61 44 HR 1.62 (1.11, 2.44) Death 55 33 HR 1.97 (1.27, 3.04)
Women’s perceptions and preferences regarding liquid biopsy for breast cancer screening.
e22548 Background: Liquid biopsy is a potential new noninvasive tool for cancer detection and may address barriers to current breast cancer screening methods. Little is known about preferences and perceptions regarding liquid biopsy among diverse populations of women. This study evaluated attitudes towards liquid biopsy for breast cancer screening that may influence its acceptance. Methods: An online survey was distributed between February 2024 and December 2025 to women aged 40–75 years without a history of invasive breast cancer enrolled in three cohorts: Mano a Mano (Mexican-American women in the Houston area), Project CHURCH (African-American women, also in Houston) and High Risk Breast Cohort (women at high breast cancer risk enrolled at MD Anderson Cancer Center). 520 participants completed a 29-item questionnaire that assessed screening preferences, barriers, and psychosocial factors. A subsample (n = 35) also completed a qualitative interview. Statistical analyses included descriptive statistics, Kruskal–Wallis, or Chi-square tests. Results: Among survey respondents, 57.8% indicated that they were extremely interested in liquid biopsy for breast cancer screening. Interest level differed across the three cohorts (p < 0.001) with the lowest proportion of extremely interested in the Mano a Mano (37%) and the highest in the High-Risk Breast cohort (62%). The most cited reason for adopting liquid biopsy was the potential to detect breast cancer earlier than current screening methods, rated as "very important" by 96% of participants in High-Risk Breast, 86% in Mano a Mano and 79% in Project CHURCH cohorts. The most cited concern was the test accuracy, rated as "very important" by 61% of participants in High-Risk Breast, 58% in Mano a Mano, and 58% in Project CHURCH cohorts. The proportion of participants who had never undergone breast cancer screening differed significantly across cohorts (p < 0.005), with Project CHURCH participants being the least likely to have had prior screening. Among those who had never undergone screening (n = 57), 80% indicated a willingness to pursue screening using liquid biopsy. Overall, 94% of participants interviewed had a positive attitude towards liquid biopsy screening. The major themes of concern elicited in qualitative interviews were its accuracy, cost, and relying on liquid biopsy as the only screening method. Conclusions: Liquid biopsy represents a promising strategy to increase breast cancer screening participation, particularly among under-screened populations. Improving accuracy, decreasing costs, and understanding how to integrate it with traditional screening methods will be needed to increase acceptance of liquid biopsy for breast cancer screening.
Time-restricted eating and metformin (TEAM) in invasive breast cancer or DCIS: A randomized, phase IIb, window-of-opportunity presurgical trial.
10524 Background: Preclinical studies indicate that metformin administered during fasting synergizes with feeding–fasting cycles to suppress tumor growth through activation of the PP2A–GSK3β–MCL1 pathway. The TEAM trial evaluated the feasibility, safety, biological, metabolic, and clinical effects of a short-term metabolic intervention combining prolonged nightly fasting and metformin in operable hormone receptor–positive breast cancer. Methods: TEAM is a randomized, phase IIb, presurgical window-of-opportunity trial. Women with hormone receptor-positive operable invasive breast cancer or DCIS were randomized 1:1 to either ≥16-hour nightly fasting plus metformin 750 mg bid and nutritional counseling (experimental [exp] arm) or to healthy WCRF lifestyle recommendations (control arm). Both groups wore a Continuous Glucose Monitor (CGM) per protocol data downloads. Treatment duration was 4–6 weeks before surgery. The primary endpoint is the absolute change in centrally assessed Ki67 between biopsy and surgery in invasive disease or DCIS. A co-primary endpoint was the difference in post-treatment Ki67 in cancer-adjacent DCIS. Analyses were intention-to-treat. Results: A total of 120 patients were randomized and completed the study (mean±SD, 32±8.6 days). As of December 31, 2025, 105 were evaluable for paired Ki67 analysis due to tissue availability. Since the complete results will be presented at the meeting, current data are descriptive with no inferential statistics and p-values, as per DSMB recommendation. Median baseline Ki67 was 15% in both arms. Median absolute Ki67 change was −3 percentage points (IQR −7 to 1) in the exp arm versus −2 (IQR −4 to 4) in the control arm. The median absolute change in invasive disease was -3 (-7 to 2) versus -1 (-3 to 4), and Ki67 reduction ≥3% occurred in 53% versus 32%, in the exp and control arm, respectively. Post-treatment Ki67 in cancer-adjacent DCIS was 3.5 (2-8.5) vs 4 (2-7). There were 3 pathological CR in the exp arm and 0 in the control arm during the presurgical window. Median weight decreased by 1.75 kg in the exp arm relative to the control arm. Mean tumor 18F-FDG PET SUV decreased by 22% versus 4% in a subgroup of 25 patients with tumors ≥ 15 mm. Reductions were observed in glucose, HOMA-IR, leptin, C-peptide, insulin and IGF-I. Changes in CIP2A and MCL1 were observed in patients with Ki67 reduction. Median adherence to ≥16-hour fasting was 100% (97.4-100); overall metformin adherence was 95% (85%-99%). No dose-limiting toxicities or grade ≥3 hypoglycemia occurred. Conclusions: These preliminary results show that short-term metabolic intervention was feasible, safe, highly compliant, and very inexpensive, and was associated with favorable antiproliferative and metabolic changes that warrant further investigation in breast cancer prevention and treatment. Full results will be presented at the meeting. Clinical trial information: NCT05023967 .
Risk of checkpoint inhibitor pneumonitis associated with anti-angiogenic agents in non–small cell lung cancer: An integrated analysis of meta-analysis of randomized controlled trials, FAERS, and real-world data.
e24163 Background: Immunotherapy with immune checkpoint inhibitors (ICIs) and antiangiogenic treatment represent the standard-of-cares for advanced non-small cell lung cancer (NSCLC). checkpoint inhibitor pneumonitis (CIP) is one of common and life-threatening immune-related adverse events. This study aimed to comprehensively evaluate the effect of anti-angiogenic agents on the risk of CIP in NSCLC using multi-level evidence. Methods: A triangulated research design integrating a meta-analysis of randomized controlled trials (RCTs), pharmacovigilance analysis, and a single-center real-world cohort study was conducted. RCTs published between January 1, 2005, and July 1, 2025, were identified from PubMed, Google Scholar, Scopus, and ClinicalTrials.gov for meta-analysis. Adverse event data were extracted from the FAERS for NSCLC patients treated with ICIs between January 1, 2015, and December 31, 2023, and multivariable logistic regression was applied to estimate adjusted odds ratios (ORs). A retrospective real-world cohort of advanced NSCLC patients treated between January 2019 and July 2025 was analyzed. CIP was defined and graded according to CTCAE version 5.0, in conjunction with radiologic assessment. Results: Meta-analysis including 7 eligible trials selected from 1,454 studies demonstrated that ICI combined with anti-angiogenic agents significantly reduced the risk of CIP compared with ICIs (OR: 0.72; 95% CI: 0.52-0.99; p=0.043). The FAERS data analysis showed that 2,800 of 31,755 NSCLC patients (8.8%) receiving ICIs developed CIP. Among 2,540 patients receiving combined ICI and anti-angiogenic therapy, 184 (7.2%) of CIP were reported. Subgroup analyses revealed a protective signal with anti-VEGF therapy for CIP risk (OR: 0.76; 95% CI: 0.64-0.91; p=0.003), whereas anti-VEGFR-2 agents and VEGFR tyrosine kinase inhibitors did not demonstrate a significant reduced risk of CIP. In a single-center cohort with 218 advanced NSCLC, CIP occurred in 7.4% of those receiving combined ICIs and anti-angiogenic agents vs. 24.0% of patients receiving ICIs (OR: 0.25; 95% CI: 0.09-0.67; p=0.005). However, combined ICIs and anti-angiogenic agents did not affect the incidence of ≥ grade 3 CIP (2.9% vs. 6.7%; OR: 0.42; 95% CI: 0.09-1.99; p=0.264). The median time to onset of CIP (167 vs. 143 days; p=0.936) and cycles of ICIs (6.0 vs. 5.0; p=0.873) was comparable in patients with combined ICIs and anti-angiogenic agents vs. those with ICIs. Conclusions: Consistent evidence from multiple data sources indicate that anti-angiogenic therapy particularly anti-VEGF agents are associated with a reduced risk of CIP in NSCLC receiving ICIs. These findings support risk-adapted combination strategies to mitigate immune-related pulmonary toxicity in NSCLC.
Prospective study of a novel portable scalp-cooling cap for prevention of chemotherapy-induced alopecia.
e24157 Background: Chemotherapy-induced alopecia (CIA) is one of the most distressing adverse events experienced by patients receiving systemic cancer therapy. Scalp cooling has emerged as a strategy to reduce CIA; however, its use in routine practice is limited by the high cost of the devices and the workload for medical staff. Aibou is a portable scalp-cooling system developed by a Japanese ice-pack company, in which patients bring their own ice packs and apply cooling using a specialized cap, without the need for hospitals to purchase cooling devices or staff to operate them. Methods: This prospective observational study enrolled patients with breast cancer at our institution who were scheduled to receive perioperative chemotherapy including anthracycline- and/or taxane-based regimens and used Aibou for scalp cooling between November 2024 and September 2025. The primary endpoint was the rate of successful hair preservation, assessed by two independent evaluators using standardized photographs; in cases of discordance, the worse (higher) Dean scale grade was adopted. Hair preservation success was defined as a Dean scale score of ≤2 at all assessment time points, assessed at baseline and 2–4 weeks after completion of each regimen. Secondary endpoints included scalp-cooling completion rate, adverse events, and patient satisfaction. Results: A total of 25 patients were enrolled, of whom 23 were included in the analysis after excluding two who discontinued chemotherapy due to adverse events unrelated to alopecia. The median age was 46 years (range, 35–65). Thirteen patients received sequential anthracycline–taxane regimens, and ten received taxane-based regimens alone. Based on independent evaluator assessments, successful hair preservation was achieved in 10 of 23 patients (43.4%), including 30.8% (4/13) in the sequential anthracycline–taxane cohort and 60.0% (6/10) in the taxane-only cohort. Scalp cooling with Aibou was completed in 22 of 23 patients (95.7%); one patient discontinued due to headache attributed to scalp cooling. No serious adverse events related to scalp cooling were observed. Eighteen of 23 patients (78.3%) reported being satisfied with Aibou, and 17 of 23 (73.9%) would recommend it to other patients. Conclusions: Aibou, a novel portable patient-operated scalp-cooling cap, showed favorable hair-preservation outcomes and acceptable tolerability. This device will help broaden the use of scalp cooling in routine oncology care.
A phase 2 study of telisotuzumab adizutecan (ABBV-400; Temab-A) in patients with advanced solid tumors harboring <i>MET</i> amplification.
TPS3157 Background: MET amplification is a rare alteration occurring at a rate of 0.06%–4.2%, in different solid tumor types across genomic databases. It is associated with poor prognosis and has limited treatment options. Currently, there are no therapies approved specifically to target MET -amplified tumors. Temab-A is an antibody-drug conjugate targeting c-Met and conjugated to a topoisomerase 1 inhibitor payload. In a first-in-human study (NCT05029882), Temab-A demonstrated encouraging activity (objective response rate 46%) and a manageable safety profile in MET -amplified solid tumors (Murciano-Goroff et al. ESMO 2025), supporting the continued investigation of Temab-A. The present phase 2 study evaluates Temab-A in adults and adolescents (12–17 years) with relapsed/refractory advanced solid tumors harboring MET amplification. Methods: This is an open-label, single-arm, global, multicenter phase 2 study (NCT07196644) enrolling approximately 125 patients (100 global; 25 China cohort) with locally advanced or metastatic solid tumors harboring MET amplification. Patients must have received ≥1 prior systemic therapy appropriate for their tumor type in the advanced/metastatic setting and have no satisfactory alternative treatment options. Patients must be ≥12 years old, have ECOG PS 0–1, measurable disease per RECIST v1.1 or RANO, and documented MET amplification by local next-generation sequencing or central FoundationOne CDx. Patients are treated with Temab-A 2.4 mg/kg IV every 3 weeks until disease progression, unacceptable toxicity, or other discontinuation criteria are met. Trial objectives are listed in the table. Enrollment began in November 2025; as of December 2025, 1 patient has enrolled. Clinical trial information: EU CT: 2024-518871-74-00. Objectives. Primary Evaluate efficacy (ORR per ICR, RECIST v1.1/RANO) and safety of Temab-A (AEs, vital signs, ECG, clinical laboratory testing) Secondary Further assess efficacy (DOR, PFS, OS, disease control) and characterize PK and immunogenicity (PK parameters, ADAs, nADAs) Exploratory Biomarker analyses (eg, c-Met expression) and PROs evaluation ADAs, antidrug antibodies; AEs, adverse events; DOR, duration of response; ECG, electrocardiogram; ICR, independent central review; nADAs, neutralizing antidrug antibodies; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; PK, pharmacokinetics; PROs, patient-reported outcomes; RANO, Response Assessment in Neuro-Oncology; RECIST v1.1, Response Evaluation Criteria in Solid Tumors version 1.1.
Clinical profile and treatment outcomes of lymphoma patients: A retrospective cohort study.
e19100 Background: Real-world data on lymphoma subtypes from low-income countries remain limited, particularly regarding treatment patterns and outcomes across diverse histologies. This study evaluates the clinical characteristics, treatment responses, and outcomes of patients with Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL) subtypes managed in routine clinical practice. Methods: This retrospective real-world study included 450 patients diagnosed with lymphoma between January 2019 to December 2025 at MNJ institute of oncology and regional cancer centre, India. Histological subtypes analyzed included Hodgkin lymphoma, non-Hodgkin lymphoma (diffuse large B-cell, follicular, mantle cell, anaplastic large cell lymphoma [ALCL], T-cell NHL), plasmablastic lymphoma, and Burkitt’s lymphoma. Diagnosis was established using histopathology and immunohistochemistry. Treatment regimens were selected based on institutional protocols and resource availability. Treatment response was assessed using standard clinical and radiological criteria. Results: A total of 427 patients included in the final analysis, with a median age of 40 years. 23 patients excluded. Hodgkin lymphoma (n = 127) constituted the largest proportion of cases, followed by non-Hodgkin lymphoma (n = 107), among which follicular lymphoma (n = 57) and diffuse large B-cell lymphoma (n = 55) were the most common subtypes. Aggressive lymphomas included Burkitt’s lymphoma (n = 20), plasmablastic lymphoma (n = 7), mantle cell lymphoma (n = 12), and T-cell non-Hodgkin lymphoma (n = 13), the majority of which presented with advanced-stage disease. The cohort shows a male predominance, with 274 males and 153 females. stage IV disease observed in 173 patients at diagnosis. At last follow-up, 272 patients were alive, while 71 patients had died due to disease progression or chemotherapy-related toxicity. Notably, 84 patients (19%) defaulted treatment, reflecting significant challenges related to treatment adherence in a resource-constrained setting. 20% of defaulted patients were confirmed alive, suggesting potential for recovery with appropriate strategies, commonly seen in Hodgkin’s lymphoma, Pediatric age group and who presented with poor initial response. Relapse Rate is 17%. Among relapse patients 60% are alive with effective salvage regimens are ICE, GDP, DHAP+/- Rituximab, 3% are dead, 37% defaulted. Overall survival till December 2025 is 70% [Hodgkin's (82%) > Follicular (79%) > DLBCL (65%)]. Advanced stage, older age, and treatment default were independent predictors of poorer survival. Conclusions: This retrospective analysis of 427 lymphoma patients provides comprehensive insights into treatment outcomes, relapse patterns, and survival in a real-world clinical setting. Outcome n (%) Overall Default rate 84/427 (19%) Alive 272/427 (63.7%) Dead 71/427 (16.6%) Response to treatment 276/427(64.6%) Relapse 72/427 (16.9%)
Effect of breast density reporting on supplemental MRI and ultrasound use.
10568 Background: Mammography has lower sensitivity in women with dense breasts due to masking, as both dense tissue and tumor tissue appears white. Supplemental breast imaging, such as MRI or ultrasound, can increase cancer detection among women with dense breasts, but its use is limited by screening capacity, cost, and potential overdiagnosis. The purpose of this study is to understand the impact of changing breast density on supplemental screening use in a longitudinal cohort of women undergoing screening mammography. Methods: Participants are from the Boston Mammography Cohort Study (BMCS), a cohort of women undergoing routine screening mammography within the Mass General Brigham health system in Boston, MA enrolled from 2008-2016, with follow up collected through 2022. Participants completed a questionnaire at enrollment which provided demographic, reproductive, lifestyle and behavioral factors. Electronic medical record review provided dates and results of all breast imaging received within the health system. We performed a longitudinal analysis using generalized estimating equations to evaluate the association between breast density and supplemental MRI and/or ultrasound. Women with a history of breast cancer and those without at least one mammogram with reported BIRADS breast density were excluded from the analysis. Results: Of the 2,696 women enrolled in BMCS, 87 were excluded for breast cancer prior to enrollment and 16 were excluded for lack of reported BIRADS breast density, leaving 2,593 in the analytic cohort. Of these women, 37.6% were aged 40-49 at the time of enrollment, 71.5% were non-Hispanic white, 46.5% had a body mass index (BMI) of 18.5-24.9, 51.9% were postmenopausal, and 78.0% had no first-degree relatives with breast cancer. Overall, 1.5% (n = 40) were diagnosed with breast cancer, 3.5% (n = 98) received a breast MRI, and 25.8% (n = 669) received a breast ultrasound. In unadjusted analyses, increased breast density was associated with higher rates of supplemental imaging. Compared with women who never had dense breasts, those who always had dense breasts had higher use of MRI (RR 3.7, p < 0.001), while ultrasound use was higher among women who sometimes (RR 1.5, p = 0.0002) and always (RR 1.7, p < 0.0001) had dense breasts. After adjusting for age, race, BMI, menopausal status, family history, and mammogram results, the association between breast density and supplemental imaging was attenuated. Women who sometimes had dense breasts remained more likely to undergo ultrasound and any supplemental imaging compared with women who never had dense breasts, but the associations were smaller and did not reach significant statistical significance. Abnormal mammogram results were the strongest predictor across all models. Conclusions: These findings suggest that breast density alone may be insufficient to guide supplemental screening guidelines and support more targeted, risk-informed approaches.
Liposomal irinotecan combined with oxaliplatin, 5-fluorouracil/leucovorin with or without camrelizumab as preoperative chemotherapy for borderline resectable pancreatic cancer: Preliminary results of a single-center phase II exploratory study.
e16465 Background: Borderline resectable pancreatic cancer (BRPC) lacks established standard preoperative chemotherapy regimens to improve resectability and long-term outcomes. This single-center phase II exploratory study evaluated the efficacy and safety of liposomal irinotecan + oxaliplatin + 5-fluorouracil/leucovorin (5-FU/LV) with/without camrelizumab as preoperative chemotherapy for BRPC(NCT06345300). Methods: This was a single-center, open-label, randomized phase II study. Eligible patients had histologically/cytologically confirmed BRPC, ECOG PS 0-1, and no prior systemic antineoplastic therapy. Patients were randomized to two cohorts: Cohort 1 (immunechemotherapy): liposomal irinotecan (56.5 mg/m², IV, Day 1) + oxaliplatin (85 mg/m², IV, Day 1) + 5-FU (2400 mg/m², IV, continuous infusion 46-48h, Day 1) + LV (200 mg/m², IV, Day 1) + camrelizumab (200 mg, IV, Day 1, q2w); Cohort 2 (chemotherapy): same regimen excluding camrelizumab. Both cohorts received 3 preoperative cycles (14 days/cycle). Efficacy was assessed per RECIST v1.1, safety per NCI-CTCAE v5.0. Primary endpoint: R0 resection rate; secondary endpoints: objective response rate (ORR), disease control rate (DCR), and safety profile. Results: As of Jan 17, 2026, 21 patients were enrolled in the intention-to-treat (ITT) population (Cohort 1: 12 [57.1%]; Cohort 2: 9 [42.9%]). Median age was 60.7 years (Cohort 1) and 60.25 years (Cohort 2). Efficacy was evaluable in 14 patients (Cohort 1: 8; Cohort 2: 6). Cohort 1: ORR 37.5% (3/8), DCR 62.5% (5/8); Cohort 2: ORR 33.3% (2/6), DCR 83.3% (5/6). Four patients (19.0% of ITT) underwent surgical exploration (Cohort 1: 1; Cohort 2: 3), all achieving R0 resection (100% R0 rate). Pathological re-staging: Cohort 1 (1 case: ypT1N0); Cohort 2 (1 ypT2N0, 1 ypT2N1, 1 ypT3N0). Safety was evaluable in all 21 ITT patients. Cohort 1: most common all-grade AEs were nausea (69.2%), diarrhea (50.0%), neutropenia (41.7%), constipation (33.3%); Grade ≥3 AEs included neutropenia (25.0%), malnutrition (16.7%), leukopenia (8.3%), hypoalbuminemia (8.3%). Cohort 2: most common all-grade AEs were nausea (44.4%), diarrhea (22.2%), leukopenia (22.2%); only Grade ≥3 diarrhea (11.1%) was severe. No unexpected serious AEs (SAEs) occurred in either cohort. Conclusions: This study initially confirms that liposomal irinotecan combined with oxaliplatin + 5-FU/LV ± camrelizumab as preoperative chemotherapy for BRPC effectively controls tumor progression and improves R0 resection rate with good safety, providing a potential preoperative chemotherapy option for BRPC patients. Clinical trial information: NCT06345300 .
Early detection of multiple myeloma and MGUS progression risk using artificial intelligence and routine laboratory data.
7519 Background: Multiple myeloma (MM) is a malignant plasma cell disorder that often presents with non-specific symptoms, resulting in delayed diagnosis and poorer outcomes. While artificial intelligence (AI) models using routine laboratory data have shown promise for MM detection, their ability to distinguish monoclonal gammopathy of undetermined significance (MGUS) from early MM and to identify patients at risk of progression remains limited. Methods: We conducted a retrospective study using routinely collected clinical and laboratory data from 89 patients evaluated at the Lincolnshire Hospital Group between 2003 and 2025. The cohort included confirmed MM cases, MGUS patients, and non-myeloma controls. Predictive features included hemoglobin, serum creatinine, calcium, albumin, immunoglobulin levels, paraprotein concentration, and derived laboratory indices. After rigorous data cleaning and preprocessing, eight supervised machine learning models were trained using a training–testing split. These included six machine learning algorithms (Lasso, gradient boosting, random forest, elastic-net, ridge regression, and support vector) and two baseline algorithms (Naïve (Mean) and Seasonal-Naïve). The training set included 69 patients (82%), and the test set included 20 patients (18%). Model performance was evaluated using precision, recall, F1 score, and regression evaluation metrics such as R² (R-squared) and MAE (Mean Absolute Error), with a specific focus on MGUS–MM differentiation. Results: The developed AI models demonstrated reliable performance in distinguishing MM from non-myeloma controls and showed improved discriminatory ability between MGUS and MM compared with standard laboratory threshold-based assessment. Specific laboratory feature patterns were associated with progression-consistent phenotypes, enabling risk stratification within the MGUS population. The best-performing model, Lasso Regression, achieved 0.01 MAE (months) and a nearly perfect score of 1.00 for R². This meant the best prediction done by the model had a 0.47 months error, i.e. 14 days difference from the actual transformation date. Gradient Boosting and Random Forest achieved 1.04 and 1.15 MAE, and 0.90 and 0.95 R², respectively. The results were implemented as a prototype clinical decision-support tool capable of generating rapid risk predictions using routinely available laboratory data. Conclusions: This study advances existing AI-based diagnostic approaches by addressing the clinically important challenge of early MM detection and MGUS risk stratification. AI-driven analysis of routine laboratory data may support earlier identification of patients at risk of progression, enabling improved monitoring, timely referral, and enhanced clinical decision-making.
PEG timing in head and neck cancer: Is there a sweet spot? A real world study.
e18116 Background: Malnutrition and dysphagia are common during chemoradiation for head and neck cancer, with 30–50% of patients malnourished at diagnosis and over 60% losing ≥5% body weight during treatment. Percutaneous endoscopic gastrostomy (PEG) tubes are widely used to support nutrition, yet the optimal timing of placement remains unclear. Prior studies have shown mixed results with no consistent impact on pneumonia, unplanned care, treatment completion, or survival. Current guidelines reflect this uncertainty and do not recommend an uniform approach to PEG timing. Methods: We conducted a retrospective cohort study using the TriNetX Network. Adults with oral cavity, oropharyngeal, or laryngeal cancers receiving definitive radiation therapy with or without cisplatin or carboplatin were included. Radiation start defined the index date. PEG placement was classified as timely if within 7 days of radiation initiation and untimely if after 7 days and up to 3 months. Outcomes within 90 days included aspiration pneumonia, IV hydration, acute kidney injury (AKI), and syncope or orthostatic events. Overall survival was assessed at 12 months. Propensity score matching was performed 1:1 using age, sex, cancer site grouping, gastrointestinal comorbidities, and other baseline characteristics, yielding two balanced cohorts of 267 patients. Results: Aspiration pneumonia occurred in 10.5% of patients with timely PEG placement and 11.6% with untimely placement (p=0.73). IV hydration was lower in the timely PEG cohort (44.9% vs 62.9%, p<0.001). AKI occurred in 15.3% of the timely cohort and 30.5% of the untimely cohort (RR 0.50, 95% CI 0.37–0.68). Syncope or orthostatic events occurred in 6.5% and 12.1%, respectively (RR 0.54, 95% CI 0.32–0.89). One-year overall survival was 77.1% with timely PEG and 82.1% with untimely placement (HR 1.34, p=0.57). Conclusions: This analysis shows that PEG timing acts as an important supportive care measure in treating patients with definitive intent concurrent chemoRT. Earlier PEG placement did not influence aspiration pneumonia or one year survival but was associated with fewer dehydration related complications, including IV fluid use, acute kidney injury, and syncope, suggesting improved tolerance of chemoradiation when nutritional access is established early. These findings move the PEG timing discussion toward objective outcomes that reflect day to day treatment tolerance. Early PEG placement should be considered for patients at higher risk of nutritional and volume compromise, while delayed placement appears reasonable for patients able to maintain intake. Assessment of treatment interruptions as a function of time from radiation initiation was not feasible in this dataset, though deeper evaluation of treatment continuity in relation to PEG timing represents an important area for future study. Residual confounding from unmeasured factors influencing PEG selection remains possible.
Effects of delayed sodium thiosulfate on cisplatin-induced ototoxicity in pediatric and adolescent patients with cancer: Results from the Japanese Children’s Cancer Group STS-J01 study.
10052 Background: Cisplatin is a cornerstone of therapy for pediatric, adolescent, and adult solid tumors but frequently causes irreversible, dose-dependent ototoxicity. Delayed administration of sodium thiosulfate (PEDMARK) has demonstrated otoprotection in Western trials. STS-J01 evaluated the efficacy, safety, and pharmacokinetics of delayed sodium thiosulfate anhydrous in Japanese pediatric and adolescent patients. Methods: STS-J01 was an open-label, single-arm phase II study enrolling patients aged 0–18 years with localized solid tumors receiving cisplatin-based chemotherapy. Sodium thiosulfate (12.8 g/m²) was administered intravenously 6 hours after completion of each cisplatin dose. The primary endpoint was incidence of hearing loss by American Speech-Language-Hearing Association (ASHA) criteria in patients aged ≥3 years, compared with the observational cohort of ACCL0431. Secondary endpoints included hearing loss by Brock classification, tumor response, safety, and pharmacokinetics. Results: Thirty-one patients were enrolled, including 25 evaluable patients in the primary cohort (Table). Hearing was preserved in the majority of patients, with 76% free of ototoxicity by ASHA criteria and 84% with no hearing loss by Brock classification, representing a significant reduction versus historical controls (relative risk 0.42; 95% CI 0.20–0.82; P=0.007). Antitumor efficacy was preserved, with an objective response rate of 95.8% and there was no evidence of treatment interference from sodium thiosulfate. Pharmacokinetic analyses demonstrated no clinically meaningful impact of sodium thiosulfate on cisplatin exposure. No serious adverse events were attributed to sodium thiosulfate transient electrolyte abnormalities were manageable and reversible. Conclusions: Delayed administration of sodium thiosulfate significantly reduced cisplatin-induced ototoxicity without compromising antitumor efficacy in Japanese pediatric and adolescent patients, confirming reproducibility of sodium thiosulfate otoprotection across populations. Clinical trial information: jRCT2061220018. Summary of STS-J01 trial. Age at diagnosis (years) 3–6 (N=4) 7–14 (N=17) 15-18 (N=4) Total (N=25) No Hearing Loss by ASHA 2/4(50%) 13/17 (76.4%) 4/4 (100%) 19/25 (76.0%) Disease Hepatoblastoma 1/2 (50%) - - 1/2 (50.0%) Germ Cell Tumors 1/1 (100%) 6/6 (100%) 1/1(100%) 8/8 (100%) Bone and Soft Tissue Tumors - 6/8 (75.0%) 3/3 (100%) 9/11 (81.8%) Medulloblastoma 0/1 (0%) 0/1 (0%) - 0/2 (0%) Others - 1/2 (50.0%) - 1/2 (50.0%) Chemotherapy Responder* 4/4 (100%) 16/16 (100%) 3/4 (75.0%) 23/24 (95.9%) *One case was unevaluated for treatment response.