Browse Articles

Observation on the application effect of biological polysaccharide in radiation-induced oral mucositis of head and neck tumors.

e13596 Background: To explore the effect of 2 different gargles on the prevention and treatment of radioactive oral mucositis in patients with head and neck cancer, and to provide a theoretical basis for clinical prevention and treatment of radioactive oral mucositis. Methods: A total of 200 patients with head and neck malignant tumors diagnosed via pathological examination and free from distant metastasis in our head and neck on cology departmert from August 2022 to November 2024 were selected as the research subjects,and according to the sequence of admission, they were allocated into an observation group (biopolysaccharide body therapy) and a control group (conventional fluid therapy)using a randomized number table. Both cohorts underwent radical intensity⁃modulated radiotherapy and were administered medications aimed at preventing mucositis throughout the radiotherapy period, concluding upon treatment completion. Using the American Radiation Oncology Collaboration Group's grading criteria for oral mucositis and the numeric rating scale for pain scoring,the extent of oral mucositis, oral pain, and xerostomia experienced by patients in both groups was assessed. Results: Both cohorts exhibited varying severities of radiation⁃induced oral mucositis, xerostomia, and oropharyngeal discomfort. In comparison to the control group, the severity of radiation⁃induced oral mucositis was mitigated and the onset of symptoms was delayed as well as the incidence of xerostomia was decreased in the experimental group,and no significant adverse drug reactions were reported among all enrolled patients, and these findings demonstrated statistical significance (P < 0.05). Conclusions: Both biopolysaccharide bodies andKangfuxin solution have a certain effect on the protection of radioactive oral mucositis, but the clinical treatment effect of biopolysaccharide bodies is better. There are no adverse reactions related to drugs, which is worthy of clinical application. Comparison of radiation-related oral toxicities between two groups. Outcome Control (n=100) Observation (n=100) Statistics P value Grade 2 oral mucositis, n (%) 53 (53.0) 35 (35.0) χ² = 16.17 ＜0.01 Grade 3 oral mucositis, n (%) 27 (27.0) 13 (13.0) χ² = 4.72 0.03 Xerostomia severity Distribution across grades Distribution across grades χ² = 42.590 0.008 Oral mucositis pain severity Distribution across grades Distribution across grades Z = 3.258 0.001 Time to grade III mucositis (days, mean ± SD) 24.2 ± 5.5 28.8 ± 6.4 t = −6.46 0.030 Time to grade IV mucositis (days, mean ± SD) 30.4 ± 3.6 38.0 ± 2.8 t = −16.47 0.219

Association of heparinoids exposure with cytokine release syndrome–related morbidity and mortality in patients receiving CAR-T or BITE therapies.

e14510 Background: Chimeric antigen receptor (CAR)-T and bispecific T-cell engaging (BITE) therapies mediate cytotoxicity primarily via interferon-gamma (IFN-γ). While IFN-γ is critical in anti-cancer immunity and infection control, excessive systemic levels drive inflammation. This resembles hemophagocytic lymphohistiocytosis (HLH) IFN-γ pathophysiology. Sulfated glycosaminoglycans (sGAGs) play a critical role in locally containing IFN-γ in animal models of viral infection. We hypothesized that soluble therapeutic sGAGs such as heparinoids may interfere with native sGAG, resulting in widespread systemic cytokine storm/cytokine release syndrome (CRS) leading to adverse outcomes. Methods: A retrospective review using the Global TriNetX Research Network (18 HCO with complete data) of patients receiving the first treatment of CAR-T/BITE or HLH (index events) from 09/2017-12/2025 was carried out and divided into two cohorts based on heparinoid exposure within 30 days. Propensity matching for age at index, sex, Black race, lactate dehydrogenase levels (a surrogate for disease burden), common comorbidities and concomitant medications (37 parameters) was done. Kaplan-Meier survival curves assessed overall survival (OS). The incidence of CRS, inflammatory markers and opportunistic infections was evaluated using odds ratios (OR) with 95% confidence intervals. Results: After cohort matching, 1,001 patients per group were included for comparison (Table 1). Heparinoid exposure resulted in significantly worse OS [log-rank p=0.0026, adjusted hazard ratio (aHR) = 1.3, proportionality P=0.11 suggesting well-balanced cohorts]. These patients also showed an increased risk of CRS, opportunistic infections and elevated mean erythrocyte sedimentation rate (ESR). In secondary exploratory analyses, comparing anticoagulation with heparinoids vs direct thrombin inhibitors (DOAC) (n = 194 each), a similar trend toward worse OS and CRS was observed. Interestingly, analyses restricted to patients receiving heparin for non-therapeutic indications, had similar adverse outcomes. Finally, patients with HLH (n= 3278 each) exposed to heparinoids had increased risk of death (aHR = 1.66). Conclusions: Our findings suggest that heparinoid, but not DOAC, exposure is associated with CRS and poor OS in patients receiving CAR-T/BITE, implicating the need for confirmatory studies to establish safer anticoagulation strategies in at-risk individuals. Patient’s outcome with/without heparinoid exposure. Outcome Heparinoid n=1001 (%) NO heparinoid n=1001 (%) aHR (p-value)/OR (CI) Death 318 (32) 231 (23) 1.3 (0.0026) CRS 251 (35) 207 (25) 1.6 (1.24,1.93) ESR, mean (SD) 24.7 (25) 30.99 (28) p = 0.015 Opportunistic infection 118 (15) 88 (11) 1.5 (1.08,1.95) Heparinoid n=194 (%) DOAC n=194 (%) Death 66 (34) 41 (21) 1.37 (0.1131) CRS 33 (24) 23 (17) 1.6 (0.86,2.83)

Efficacy and safety of donafenib combined with PD-1 inhibitors and transarterial chemoembolization (TACE) in unresectable hepatocellular carcinoma (uHCC): A retrospective study.

e16188 Background: Emerging evidence suggests that combining TACE with tyrosine kinase inhibitors (TKIs) and PD-1 inhibitors may improve outcomes in patients with uHCC. Donafenib has been approved in China as a standard first-line systemic therapy for uHCC, and preliminary data indicate that combination regimens incorporating donafenib are promising. This study aimed to evaluate the efficacy and safety of donafenib in combination with PD-1 inhibitors and TACE in patients with uHCC. Methods: This single-center, retrospective study included patients with uHCC who received donafenib plus PD-1 inhibitor and TACE at The First Affiliated Hospital of Bengbu Medical University between February 2022 and July 2025. Eligible patients had at least one measurable lesion per modified Response Evaluation Criteria in Solid Tumors (mRECIST). The primary endpoint was progression-free survival (PFS) assessed according to mRECIST. Secondary endpoints included objective response rate (ORR), disease control rate (DCR), overall survival (OS) and treatment-related adverse events (TRAEs). Results: A total of 48 patients with uHCC were enrolled. All patients received oral donafenib in combination with intravenous PD-1 inhibitors—camrelizumab, sintilimab, or tislelizumab—administered every 3 weeks (Q3W), along with TACE performed on an as-needed basis. The median age was 59 years (range, 47–75). 98.0% patients (47/48) had chronic hepatitis B virus (HBV) infection, and 43.8% (21/48) had baseline alpha-fetoprotein (AFP) levels ≥400 ng/mL. Thirty-one patients (64.6%) received the regimen as first-line therapy, and 17 (35.4%) as second-line therapy. As of the data cutoff (December 30, 2025), the median follow-up duration was 14.2 months (16.3 months in the first-line group; 11.7 months in the second-line group). Based on the mRECIST criteria, the best ORR was 72.9% overall (80.6% in first-line; 58.8% in second-line), and the DCR was 93.8% (93.5% in first-line ; 94.1% in second-line). The CR rate was 33.3%. Notably, the median time to response of 2.2 months in the first-line group indicated rapid therapeutic onset. Median PFS and OS were not reached. The estimated 6-month PFS rates were 92.0% (first-line) and 79.9% (second-line); the corresponding 12-month OS rates were 79.0% and 74.3%. TRAEs occurred in 33 patients (68.7%), with grade 3 events reported in 3 patients (6.3%). No grade 4 or 5 TRAEs were observed. Conclusions: The combination of donafenib, PD-1 Inhibitors,and TACE demonstrated robust antitumor activity and a favorable safety profile in patients with uHCC, both in first- and second-line settings. These encouraging results warrant validation in prospective, randomized controlled trials.

Comparative outcomes of first-line atezolizumab plus bevacizumab (A+B) and durvalumab plus tremelimumab (D+T) in hepatocellular carcinoma (HCC).

4163 Background: Standard first-line systemic therapy for hepatocellular carcinoma involves immune checkpoint inhibitors, including A+B and the STRIDE regimen consisting of D+T. However, there is limited data comparing outcomes of patients receiving these two regimens. We aimed to compare the survival and safety profiles of patients receiving A+B versus D+T as first-line systemic therapy for advanced HCC in a propensity score matched analysis. Methods: We performed a retrospective cohort study using data from TriNetX, a healthcare database of over 150 million patients in the United States. We identified adult patients with a history of HCC and included those who received A+B or D+T as first line treatment. Propensity score matching (PSM) was conducted, matching for patient demographics (age, sex, race), clinical history (alcohol use, ascites, esophageal varices, viral hepatitis), and laboratory parameters [platelets, albumin, bilirubin, INR, alpha-1-fetoprotein (AFP)]. Median overall survival (mOS) and relevant patient safety events were compared between those who received A+B or D+T. Results: We identified 2,762 patients who had HCC who were treated with A+B or D+T, of which 2,030 patients received A+B and 723 patients received D+T. Compared to patients in the A+B cohort, those who received D+T were more often older (68.5 vs 66.9 years, standardized mean difference (SMD) = 0.17), had lower platelet counts (186.4 vs 198.6, SMD = 0.11), had higher prevalence of ascites (33.1% vs 20.3%, SMD = 0.29), and alcohol use (23.8% vs 18.8%, SMD = 0.12). After PSM, 723 patients were included in survival analyses, with all variables adequately balanced except for albumin, which was higher in the A+B group (3.6 vs 3.5, SMD = 0.19). At 90 days after starting therapy, D+T was associated with higher incidence of immune-mediated colitis [5.3% vs 2.6%, odds ratio (OR) 2.1; 95% confidence interval (CI), 1.2-3.7], and a higher risk of hospitalization [35.4% vs 27.9%, OR 1.4; 95% CI, 1.1-1.8]. At 6 months, the risk of hypertension and proteinuria was higher in the A+B cohort [21.9% vs 13.7%, OR 0.57; 95% CI, 0.34-0.95], which persisted at 1 year [28.1% vs 16.5%, OR 0.51; 95% CI, 0.32-0.81]. There was no significant difference in mOS between those receiving A+B or D+T [18.3 months vs 22.8 months, hazard ratio (HR) 0.94; 95% CI, 0.79-1.1]. Conclusions: In patients with advanced HCC who received A+B or D+T as first-line systemic therapy, we found that A+B was associated with higher risk of hypertension and proteinuria, while D+T was associated with a higher risk of immune-mediated colitis and hospitalization. No significant difference in mOS was observed between the cohorts. These findings suggest that safety profiles, rather than survival differences, may be used to guide optimal first-line treatment regimens. Prospective studies are warranted to further assess differences in patient outcomes.

NAPISTAR 1-01: Results of phase 1 dose escalation of monotherapy with TUB-040, a novel NaPi2b-targeting exatecan ADC, in patients (pts) with platinum-resistant ovarian cancer (PROC).

5513 Background: NaPi2b is overexpressed in high-grade serous ovarian cancer and non-small cell lung cancer (NSCLC). TUB-040 is a DAR8 ADC targeting NaPi2b, comprising a humanized Fc-silenced IgG1 mAb conjugated to exatecan via a cleavable, cysteine-selective, stable, solubility-mediating P5 linker providing excellent stability and biophysical ADC properties. In preclinical studies, TUB-040 induces potent antigen-specific cytotoxicity and demonstrates strong bystander activity. Methods: NAPISTAR 1-01 is a phase 1/2a study in biomarker-unselected pts with PROC and NSCLC. TUB-040 was administered Q3W in dose escalation (range 0.5-5.3 mg/kg) to pts with PROC until progression or unacceptable toxicity. Results: As of Dec 1, 2025, 67 PROC pts received TUB-040 for a median of 10 cycles (1-25), median duration of exposure was 213 days (21-546), and 42 (63%) pts were ongoing. Median age: 62 years (34-81), ECOG PS 0-1, median prior lines: 4 (1-7), prior treatment included bevacizumab (84%), PARPi (76%), and mirvetuximab soravtansine (13%). Between dose levels 1.67-3.3mg/kg (N=46), the most common all-grade TEAEs were nausea (78%), fatigue (54%), neutropenia (43.5%), anemia (37%), constipation (34%), diarrhea (33%), vomiting (28%), alopecia (28%), and decreased appetite (26%). Gr≥3 heme TEAEs by dose are in Table 1. There were no fatal TEAEs or treatment discontinuations due to TEAEs. Three pts (6.5%) experienced Gr1 pneumonitis, which resolved. Between dose levels 1.67–3.3 mg/kg, 46 pts were evaluated for efficacy. The uORR was 63% [95% CI, 47.5-77.2] and cORR was 60.9% [95% CI, 45.3-75.1] (RECIST; v1.1), including two cCRs. Responses by dose are in Table 1. The majority of responses, 90% (26/29), were ongoing. Disease control rate was 96% [95% CI, 85.2-99.5]. Among CA-125 evaluable pts (N=42), CA-125 response occurred in 81% (34/42) [95% CI, 65.6-91.5]. cORRs were similar across subgroups, including: ≥4 prior lines of therapy (cORR=71%), prior exposures to bevacizumab (62%), mirvetuximab soravtansine (63%) or PARPi (67%). Five pts with stable disease remain on treatment and eligible for response. Conclusions: TUB-040 was well tolerated with promising clinical activity at low doses, offering a potential new treatment option with a differentiated, favorable benefit–risk profile and a wide therapeutic window in PROC pts. Dose optimization is currently ongoing. Clinical trial information: NCT06303505 . 1.67 mg/kg (N=10), n (%) 2.1 mg/kg (N=12), n (%) 2.50 mg/kg (N=12), n (%) 3.3 mg/kg (N=12), n (%) Total (N=46), n (%) Complete response 1 (10) 0 1 (8) 0 2 (4) Partial response 6 (60) 7 (58) 7 (58) 6 (50) 26 (57) Stable disease 2 (20) 5 (42) 3 (25) 6 (50) 16 (35) Progressive disease 1 (10) 0 1 (8) 0 2 (4) Overall responses (PR + CR) 7 (70) 7 (58) 8 (67) 6 (50) 28 (60.9) Gr≥3 neutropenia 0 2 (17) 4 (33) 6 (50) 12 (26) Gr≥3 anemia 0 0 1 (8) 5 (42) 6 (13) Gr≥3 thrombocytopenia 0 0 1 (8) 1 (8) 2 (4)

Use of multiplex digital profiling of DNA methylation heterogeneity to enable sensitive early detection of non–small cell lung cancer from low-volume liquid biopsies.

10557 Background: Early detection of non–small cell lung cancer (NSCLC) within liquid biopies remains limited by the extremely low abundance and heterogeneous nature of circulating tumor DNA (ctDNA) in plasma, particularly in patients with indeterminate pulmonary nodules. DNA methylation is an attractive early biomarker; however, conventional PCR- and sequencing-based assays have limited ability to sensitively and cost-effectively assess heterogeneous epiallelic methylation patterns at low variant fractions. We developed a digital microfluidic platform that enables multiplex, single-molecule profiling of DNA methylation heterogeneity to improve early NSCLC detection. Methods: We applied REM-DREAMing (Ratiometric-Encoded Multiplex Discrimination of Rare EpiAlleles by Melt), a digital high-resolution melt (dHRM) platform combining nanowell-based digital PCR with ratiometrically encoded, methylation-agnostic probes. Following bisulfite conversion, cfDNA was digitized into >40,000 nanoliter wells, enabling copy-by-copy assessment of methylation density across a five-gene biomarker panel previously associated with NSCLC. Analytical sensitivity and multiplex performance were compared against quantitative methylation-specific PCR (qMSP) and digital MSP (dMSP). Clinical feasibility was evaluated in plasma from patients with CT-detected indeterminate pulmonary nodules. Results: REM-DREAMing resolved heterogeneous methylation patterns at single-molecule resolution and detected methylated epialleles at fractions as low as 0.00005% in a high background of unmethylated DNA. In a cohort of 48 low-volume plasma samples, incorporating intermolecular methylation density distributions significantly improved classifier performance relative to binary methylation calls. The five-gene panel achieved 93% sensitivity at 90% specificity for early-stage NSCLC, with an AUC of 0.96. Compared with qMSP and dMSP, REM-DREAMing demonstrated superior sensitivity, multiplex scalability, and discrimination of partially methylated epialleles. Conclusions: Multiplex digital profiling of DNA methylation heterogeneity enables sensitive, low-cost detection of early-stage NSCLC from limited plasma input. By capturing epigenetic heterogeneity at the single-molecule level, this platform improves diagnostic performance in clinically challenging early detection settings and provides a scalable foundation for expanded biomarker panels. Analytical and clinical performance comparison. Method Heterogeneous Methylation Detection Detectable Copy Number Analytical Specificity NSCLC AUC qMSP No 10 ~0.01% 0.78–0.82 dMSP No 1 ~0.005% 0.85 Bulk MS-HRM Yes 10 ~0.1% 0.80 REM-DREAMing Yes (single-molecule) 1 0.00005% 0.96

A phase 2 study of neoadjuvant nivolumab + relatlimab versus nivolumab in patients with resectable high-risk basal cell carcinoma.

TPS9613 Background: The high response rate of neoadjuvant anti-PD-1 has changed the management of locally advanced cutaneous squamous cell carcinoma. However, little is known about the benefit of neoadjuvant immunotherapy for basal cell carcinoma (BCC). Anti-PD-1 is approved second line for advanced BCC where it has modest response rates compared to the first line setting. High LAG-3 expression within the BCC tumor microenvironment may function as a subdominant checkpoint to synergize with anti-PD-1. The current randomized prospective multicenter phase 2 trial evaluates the safety and efficacy of neoadjuvant nivolumab (anti-PD-1) + relatlimab (anti-LAG-3) versus nivolumab alone in resectable high-risk BCC. Methods: This multicenter study conducted across the University of California Melanoma and Skin Cancer Consortium will enroll up to 30 patients with resectable high-risk BCC. High-risk BCC is defined by size 2.0 cm or greater in the head and neck region or 4.0 cm or greater for the trunk and extremities. All patients must have surgically resectable disease that is at increased risk for cosmetic disfigurement, functional defects, poor oncologic control, or anticipated to require skin grafting or free flap reconstruction. Patients must also be checkpoint and hedgehog inhibitor naïve with ECOG performance status of <2. Key exclusion criteria include recent radiation therapy and any prior history of myocarditis. Patients will be randomized in 2:1 fashion to receive nivolumab 480 mg + relatlimab 160 mg or nivolumab 480 mg IV every 4 weeks for up to 4 doses. Patients may forgo surgery and continue beyond 4 doses for up to 1 year in cases of ongoing clinical benefit. Tumor tissue and peripheral blood will be interrogated for immune cell subsets and checkpoint expression before and after neoadjuvant treatment. The trial’s primary endpoint is the combined rate of pathological complete response, major pathologic response, and clinical response following neoadjuvant therapy with secondary endpoints including surgical de-escalation, safety, duration of response, and recurrence-free survival. This trial uses a Simon’s minimax two-stage design for the nivolumab + relatlimab cohort. Eleven patients will be enrolled in the first stage. If ≥2 patients respond to nivolumab + relatlimab, the study will continue enrollment for a total of 20 patients. This design yields a type I error rate of 0.1 and power of 0.8. This study was approved by UC San Diego’s Institutional Review Board; approval number is 810617. Trial Registration NCT06624475. Clinical trial information: NCT06624475 .

Predicting recurrence and survival in hormone receptor–positive (HR+)/HER2-negative (HER2–) early breast cancer (EBC) with machine learning models: Optimization of model complexity.

562 Background: A machine learning model was previously developed to predict distant recurrence (DR) in endocrine therapy–treated patients with HR+/HER2− EBC, with high accuracy achieved using 10 variables (Howard FM, et al. Clin Cancer Res . 2026). Here, optimal performance, based on the number and identity of variables, of recurrence models that include survival events (distant relapse–free survival [DRFS], overall survival [OS]) was determined. Methods: Retrospective data from patients with stage I-III HR+/HER2− EBC (diagnosed 1 January 2011 to 30 April 2024) were extracted from the Flatiron Health US electronic health record–derived database. Elastic net–penalized Cox proportional hazards–based machine learning models were developed with 2, 5, 10, 15, or 20 variables identified by gradient boosting to predict DR, DRFS, and OS; DRFS was defined according to STEEP criteria v2.0. Model performance was assessed using the C-index and Brier score (BS). Results: With modeling based on 7842 patients, DR prediction accuracy plateaued (C-index, 0.857; BS, 0.046) (Table) with 10 variables (N status, T status, tumor grade, Ki-67 score, age, time from diagnosis to surgery, percent estrogen receptor–positive cells, menopausal status [post-], Oncotype DX Recurrence Score, and percent progesterone receptor–positive cells). The DRFS (C-index, 0.740; BS, 0.089) and OS (C-index, 0.781; BS, 0.075) prediction models showed high accuracy with 5 to 10 variables, and accuracy continued to increase with up to 20 variables. In contrast to the DR prediction model, the DRFS and OS prediction models included functional status (ie, Eastern Cooperative Oncology Group performance status, Charlson Comorbidity Index) among the most impactful variables. Conclusions: DR model performance plateaued with 10 tumor-intrinsic variables. In contrast, optimal DRFS and OS prediction required expanding the model to 20 variables, including functional and comorbidity domains necessary to capture non-cancer competing risks critical for survival prediction in real-world populations, despite the modest statistical impact. The machine learning model for predicting DR, a reliable proxy for survival in HR+/HER2− EBC, is based on a lower number of risk factors that are commonly accessible in real-world clinics and has good accuracy. Model performance by outcome and number of variables. No. of variables DR (C-index a ) DR (BS b ) DRFS (C-index a ) DRFS (BS b ) OS (C-index a ) OS (BS b ) 2 0.812 0.044 0.667 0.093 0.611 0.081 5 0.849 0.046 0.718 0.092 0.770 0.075 10 0.857 0.046 0.736 0.089 0.774 0.075 15 0.858 0.046 0.737 0.089 0.776 0.076 20 0.851 0.047 0.740 0.089 0.781 0.075 a Concordance index; scores span from 0.5 to 1; 1 = perfect. b Brier score; scores span from 0 to 0.25; 0 = perfect.

Impact of remote therapeutic monitoring with patient-reported outcomes on hospitalization in real-world patients receiving therapy for metastatic solid tumors.

11005 Background: Remote therapeutic monitoring (RTM) using electronic patient-reported outcomes (ePROs) may help care teams respond earlier to patient needs and potentially avoid acute care events, but evidence from real-world practice is limited. This study evaluates the impact of using the Canopy ePRO-based RTM platform on acute care events. Methods: We studied 1,549 patients with metastatic solid tumor malignancies receiving systemic treatment at 5 community oncology sites between Jan 1, 2024 until July 20, 2025. Patients were invited to complete weekly ePRO symptom surveys. Symptoms that exceeded pre-specified severity thresholds triggered notifications to a dedicated triage nursing staff for evaluation and consideration of telephonic triage, urgent outpatient office evaluation or referral to the emergency department or hospital. Patients who enrolled and submitted ≥ 2 ePRO surveys within 45 days of first treatment were included in the RTM group, while those not enrolled or who did not submit 2 or more reports were in the control group. Hospitalization events were extracted from the Arkansas health information exchange. Inverse probability of treatment weighting was used to account for potential non-random selection of patients into the RTM group when performing statistical tests of difference, and achieved cohort balance in terms of age at index, time from first cancer diagnosis to treatment, site of cancer, sex, and race. Results: Over 18 months following first anticancer treatment through data cutoff date, 558 patients were in the RTM group compared to 991 patients in the control group. Among patients in the RTM group, 9.0% had a record of hospitalization compared to 13% in the control group, representing a 28% reduction (risk difference: -3.63, 95% CI -7.06 to -0.39, p: 0.032); 12% of patients in the RTM group had record of ED visit compared to 13% in the control group (risk difference: -1.77, 95% CI: -5.24 to 1.63, p: 0.310). The estimated cost savings per 1,000 patients/years utilizing RTM during treatment was $3,192,789 for hospitalizations and $93,330 for emergency department visits. Conclusions: RTM utilization with Canopy is associated with fewer hospitalizations and potentially fewer emergency visits compared to traditional outpatient management of patients with metastatic solid tumors, potentially reducing the cost of care. Prompt notification of patient symptoms resulting in preemptive treatment may help patients avoid acute care events. Unweighted clinical and demographic characteristics. Characteristic Enrolled & ActiveN = 558 Not EnrolledN = 991 Age at index 64 (55, 71) 70 (62, 78) Sex - Female 313 (56%) 515 (52%) Race/Ethnicity - White 484 (87%) 846 (85%) Race/Ethnicity - Hispanic 28 (5.0%) 58 (5.9%) Lung Cancer 131 (23%) 179 (18%) Breast cancer 98 (18%) 219 (22%) Colorectal cancer 88 (16%) 148 (15%)

Pathologic complete response to neoadjuvant immunochemotherapy in locally advanced head and neck squamous cell carcinoma: Characteristics of 20 responders.

e18011 Background: The use of neoadjuvant immunochemotherapy for locally advanced head and neck squamous cell carcinoma (HNSCC) is increasing. However, the characteristics of patients who achieve a pathologic complete response (pCR) remain poorly defined. This study aims to describe the clinical and pathological features of this deep-responding population. Methods: We retrospectively included patients with locally advanced oropharyngeal or hypopharyngeal carcinoma who were assessed as having a partial response (PR) after two cycles of neoadjuvant immunochemotherapy (between March 2023 and May 2025) and whose postoperative pathology confirmed pCR. Data collected included clinicopathological characteristics, p16 status for all oropharyngeal cancer patients, PD-L1 expression (TPS and CPS), and survival outcomes. The primary endpoint was to describe the characteristics of the pCR population. The secondary endpoint was event-free survival (EFS). Results: A total of 20 patients with pCR were included. There were 12 (60%) oropharyngeal and 8 (40%) hypopharyngeal carcinomas. All patients with oropharyngeal cancer were p16-positive. Among the 14 patients with known PD-L1 status, TPS was <1% in 4 (29%), 1-49% in 7 (50%), and ≥50% in 3 (21%). CPS was <1 in 4 patients (29%), 1-20 in 3 (21%), and ≥20 in 7 (50%). Three patients experienced grade 3 or higher adverse events, all of which were myelosuppression. No patient deaths occurred. All patients, after providing informed consent, opted for postoperative maintenance immunotherapy. Conclusions: Neoadjuvant immunochemotherapy can lead to pCR in a subset of patients with locally advanced HNSCC, and this population demonstrates favorable early outcomes. In this pCR cohort, all oropharyngeal cancer patients were p16-positive, and half of those with known PD-L1 status had a CPS ≥20. Although limited by sample size and short follow-up, these findings suggest that p16 positivity and high PD-L1 expression may be features associated with achieving deep pathological response, warranting further validation in prospective studies.

Lutetium-177-PSMA-617 ( ¹⁷⁷ Lu-PSMA-617) in neuroendocrine prostate cancer (NEPC).

e17046 Background: Lutetium-177–PSMA-617 (¹⁷⁷Lu-PSMA-617) improves survival in PSMA-positive metastatic castration-resistant prostate cancer (mCRPC). Although neuroendocrine prostate cancer (NEPC) is often associated with low PSMA expression, tumor heterogeneity exists, and some cases—particularly those with mixed adenocarcinoma–neuroendocrine features—can retain PSMA avidity on imaging. Clinical outcomes with ¹⁷⁷Lu-PSMA-617 in NEPC are poorly characterized. Methods: We performed a retrospective, single-institution analysis of patients with histologic or pathologic evidence of NEPC who demonstrated PSMA-avid disease on PET/CT and received ≥1 cycle of ¹⁷⁷Lu-PSMA-617 monotherapy. Outcomes included PSA decline ≥50% (PSA50), investigator-assessed radiographic response, time to progression (TTP; radiographic or clinical), and overall survival (OS). Time-to-event outcomes were estimated using Kaplan–Meier methods and reported as medians with two-sided 95% confidence interval (CI). Results: Ten patients initiated ¹⁷⁷Lu-PSMA-617 between July 2022 and October 2024. At time of data cutoff on 1/20/26, median follow-up was 38 months with most events occurring during the first year. Median age was 70 years (range: 64–78); 8 (80%) had ECOG performance status 0–1, and median baseline PSA was 16.7 ng/mL (range: < 0.02–335). Sites of metastasis included lymph node in 8 (80%), bone in 8 (80%), and liver in 5 (50%) pts. All patients had prior exposure to androgen receptor pathway inhibitor (ARPI) and taxane chemotherapy; 6 (60%) had received prior platinum-based therapy. Treatment-emergent NEPC was present in all cases, including adenocarcinoma with neuroendocrine differentiation (n = 7, 70%) and pure high-grade neuroendocrine or small cell carcinoma (n = 3, 30%). Tumor genomic profiling was available for all patients, with 7 (70%) harboring alterations in RB1, TP53, and/or PTEN. The median number of ¹⁷⁷Lu-PSMA-617 cycles was 3 (range: 1–6). Among 8 evaluable patients with detectable baseline PSA, 5 (63%) achieved PSA50. Best radiographic response by conventional or PSMA PET/CT imaging included 1 complete response CR, 1 partial response, 1 stable disease, 2 mixed responses, and 5 cases of progressive disease. Median TTP was 3 months (95% CI, < 1–9, 9/10 events) and median OS was 7 months (95% CI, 2–not reached, 7/10 events). Notably, one patient achieved an ongoing durable complete response lasting > 2 years following 177 Lu-PSMA-617 had completed neoadjuvant platinum-based chemotherapy and surgical resection with no residual NEPC identified. Conclusions: Selected patients with NEPC and PSMA-avid disease may derive clinical benefit from ¹⁷⁷Lu-PSMA-617, although presentation is heterogeneous and overall prognosis is poor. Benefit appears greatest in patients with limited neuroendocrine differentiation and prior sensitivity to NEPC-directed therapy. Prospective evaluation in larger cohorts is warranted.

Integrated NAL-IRI/5-FU/LV, benmelstobart, anlotinib with or without immunoradiotherapy for second-line metastatic pancreatic cancer: A phase II dual-cohort study.

e16374 Background: Metastatic pancreatic ductal adenocarcinoma (mPDAC) is an immune cold tumor characterized by stromal and vascular barriers that limit the efficacy of immunotherapy alone. This study evaluated an integrated regimen combining cytotoxic chemotherapy (NAL-IRI/5-FU/LV), anti-angiogenesis (anlotinib), and PD-L1 inhibition (benmelstobart), with or without stereotactic body radiation therapy (SBRT) for second-line treatment of mPDAC (NCT06662006). Methods: This investigator initiated, prospective, open label, non-randomized phase II study enrolled patients (aged 18–75, ECOG PS ≤2) with confirmed mPDAC who progressed after first line therapy. All patients received liposomal irinotecan 50 mg/m² IV (D1,15), leucovorin 400 mg/m² IV (D1,15), 5-FU 2.4 g/m² 46 h infusion (D1,15) in 28 day cycles, plus benmelstobart 1200 mg IV Q3W and anlotinib 12 mg PO (D1–14) Q3W. Patients were assigned to one of two cohorts based on metastatic patterns: Cohort A ( > 5 dispersed metastases) received systemic therapy only; Cohort B (≤5 lesions and/or clustered metastases) received systemic therapy combined with SBRT (24 Gy in 3 fractions within 3–7 days). The primary endpoint was objective response rate (ORR). Secondary endpoints included disease control rate (DCR), progression free survival (PFS), and overall survival (OS). Results: From March 2024 to December 2025, 46 patients were enrolled (Cohort A: n = 26; Cohort B: n = 20). With a median follow up of 5.9 months, 36 patients had at least one efficacy evaluation: 23 in Cohort A and 13 in Cohort B. The overall ORR was 25.0% (95% CI: 10.9–39.1%), with rates of 17.4% (95% CI: 1.9–32.9%) in Cohort A and 38.5% (95% CI: 12.0–64.9%) in Cohort B. The overall DCR was 69.4% (95% CI, 54.4–84.5%) and differed significantly between cohorts, with a higher DCR in Cohort B than in Cohort A (92.3% [95% CI, 77.8–100%] vs. 56.5% [95% CI, 36.3–76.8%]; P = 0.031). The mPFS was 4.7 mo (95% CI: 3.9–5.4) overall, with 4.3 mo (95% CI: 3.3–5.4) in Cohort A and 7.1 mo (95% CI: 4.0–10.2) in Cohort B. Correspondingly, the overall 6-month PFS rate was 35.5% (95% CI: 18.6–52.3%), with rates of 27.3% (95% CI: 8.7–45.9%) and 55.6% (95% CI: 23.1–88.0%) in Cohorts A and B. OS data were immature at cutoff. Grade ≥3 treatment related adverse events occurred in 30.4% of patients (14/46), with grade 4 events in 4.3% (2/46). Toxicities were predominantly hematologic, and no unexpected safety signals emerged. Conclusions: The combination of NAL-IRI/5-FU/LV, benmelstobart, and anlotinib showed manageable toxicity and promising efficacy in second-line mPDAC. The addition of SBRT in patients with limited or clustered metastases significantly improved disease control and provided sustained PFS benefit. These results support a precision approach integrating SBRT for oligometastatic-like disease in mPDAC, warranting further randomized validation. Clinical trial information: NCT06662006 .

Demographics and clinical characteristics of epithelioid leiomyosarcoma: A National Cancer Database analysis.

e23535 Background: Epithelioid leiomyosarcoma is a rare, aggressive variant of uterine leiomyosarcoma, defined by round-to-polygonal cells with eosinophilic cytoplasm. It demonstrates poorer survival than conventional leiomyosarcoma and limited response to chemotherapy and radiation. Due to its rarity, comprehensive epidemiological data remain limited. The demographics and clinical characteristics of patients with epithelioid leiomyosarcoma were analyzed using the National Cancer Database (NCDB). Methods: This retrospective cohort study queried the NCDB for patients with a histologically-confirmed diagnosis of epithelioid leiomyosarcoma between 2004-2020 (ICD-O-3 code 8891). Factors including age, sex, race, ethnicity, insurance status, facility type, treatment modalities, and survival outcomes were evaluated via descriptive statistics, regression analyses. Additional variables analyzed included treatment modalities, survival outcomes, and anatomical sites. Results: A total of 933 patients with epithelioid leiomyosarcoma were identified. Incidence showed a modest upward trend over the study period (R² = 0.36). Women were much more likely to be diagnosed (90.9%) compared to men, and the average age at diagnosis was 57 years. Most patients were non-Hispanic (88.5%), White (73.4%), lived in metropolitan areas exceeding one million residents (60.1%), and were privately insured (57.4%). The most common treatment facility type was an academic/research program (45.8%), followed by comprehensive community cancer programs (28.6%). Chemotherapy was the most frequently administered systemic treatment (41.2%), and surgery at the primary site was performed in 43.0% of cases. Mean tumor size was 109.94 mm, reflecting the aggressive nature of this malignancy. Survival analysis demonstrated two-year, five-year, and 10-year survival rates of 62.1%, 45.9%, and 35.6%, respectively, with a median survival of 90.5 months. Short-term mortality remained relatively low with 30-day and 90-day mortality rates of 1.5% and 3.9%, respectively. Conclusions: This is the first NCDB analysis of epithelioid leiomyosarcoma, providing novel characterization of socioeconomic and healthcare access factors in this patient population. These findings corroborate earlier reports that it predominantly affects women and frequently presents with large, aggressive tumors at the time of diagnosis. Patients are predominantly privately insured, reside in large metropolitan areas, and most commonly receive treatment at academic or research facilities. Despite relatively low short-term mortality rates, long-term survival remains poor, reflecting the aggressive nature of this malignancy. Future research is needed to better understand how demographic and socioeconomic factors influence diagnostic timing, treatment decisions, and survival outcomes in epithelioid leiomyosarcoma.

High ferroptosis score activity as a predictor of survival and rituximab response in diffuse large B-cell lymphoma.

e19073 Background: Ferroptosis is an iron-dependent form of cell death linked to redox imbalance, tumor immunity, and therapy sensitivity. In diffuse large B-cell lymphoma (DLBCL), its clinical relevance remains incompletely defined. We evaluated whether ferroptosis activity, quantified by a transcriptomic ferroptosis score (FerroScore), is associated with survival, treatment response, and underlying biological pathways in DLBCL. Methods: Gene expression and clinical data from GSE10846 (n=420) were analyzed. FerroScore was calculated using single-sample GSEA based on 20 canonical ferroptosis regulators (SLC7A11, GPX4, ACSL4, ALOX15, LPCAT3, NCOA4, AIFM2, GCLC, GCLM, FTH1, FTL, TFRC, HMOX1, NFE2L2, KEAP1, SCD, SAT1, PTGS2, CISD1, DPP4). Patients were dichotomized by the median FerroScore. Overall survival (OS) was assessed using Kaplan–Meier and Cox models. Associations with clinical features were evaluated, and pathway enrichment analyses were performed. Prognostic effects were examined separately in R-CHOP and CHOP treated patients. Results: In the GSE10846 cohort, patients with higher FerroScores showed significantly better OS (HR = 0.85, 95% CI 0.72–0.99, p= 0.04). When stratified by treatment, the association persisted among R-CHOP treated patients (p= 0.0004) but not in those treated with CHOP alone, suggesting that ferroptosis activity may enhance the efficacy of rituximab-containing regimens. FerroScore did not correlate significantly with LDH levels or ECOG status but tended to be higher in patients with advanced stage or extranodal involvement. Pathway level analysis revealed that tumors with high FerroScores were enriched for TNF-α/NF-κB signaling, apoptosis, and complement activation, indicating an immune-active, pro-inflammatory phenotype. Conversely, MYC target and KRAS downstream signaling were reduced, consistent with suppression of proliferative and metabolic stress pathways. Among individual genes, SLC7A11, CEBPB, and ERO1A were the most strongly correlated with FerroScore, reflecting a link between oxidative stress regulation and immune activation. Collectively, these findings suggest that ferroptosis-high DLBCL exhibits both inflammatory and stress adaptive molecular traits, which may contribute to improved therapy response and survival. Conclusions: FerroScore provides a transcriptomic measure of ferroptosis activity that captures clinically meaningful heterogeneity in DLBCL. High FerroScore is associated with better survival in R-CHOP treated patients and correlates with immune and apoptotic pathway activation. This study links ferroptosis biology to real-world treatment outcomes, identifying ferroptosis as a potential biomarker of therapy response and disease biology in DLBCL. Future validation needed to support its role in risk stratification and treatment personalization.

Interpreting Circulating Tumor DNA-Guided Risk Stratification After First-Line Therapy

Hematologic malignancies to define a distinct high-risk sepsis phenotype among hospitalized cancer patients.

e23245 Background: Sepsis is a leading cause of mortality among hospitalized cancer patients. Although cancer is broadly recognized as high risk, patients with hematologic malignancies are often treated as biologically uniform. We hypothesized that hematologic malignancies define a distinct high-risk sepsis phenotype and that mortality risk varies across subtypes. Methods: We performed a retrospective cohort study of adult hospitalized cancer patients with sepsis using a national inpatient database. Demographics, comorbidities, acute organ dysfunction, malignancy subtype, and hospital characteristics were extracted. Multivariable logistic regression identified factors independently associated with in-hospital mortality. Hematologic malignancies were first analyzed as a group relative to solid tumors. Subtype-specific analyses were performed for AML, ALL, CLL, CML, and MM, with odds of mortality calculated for each subtype compared with other hematologic malignancies. Models were adjusted a priori for age, sex, Charlson Comorbidity Index, cancer stage, septic shock, and hospital characteristics including elective status and region. Results: The cohort included 338,545 hospitalizations, with 62,830 deaths (18.6%). After adjustment, hematologic malignancy remained independently associated with increased in-hospital mortality (OR 1.21, 95% CI 1.15–1.28; p < 0.001), whereas solid tumors were not (OR 0.82, 95% CI 0.78–0.87; p < 0.001), supporting hematologic malignancy as a distinct high-risk sepsis phenotype. Acute organ failure was the dominant mortality driver, including acute respiratory failure (OR 5.29, 95% CI 5.07–5.52; p < 0.001) and dialysis-requiring acute kidney injury (OR 2.87, 95% CI 2.65–3.11; p < 0.001). Venous thromboembolism (OR 1.53, 95% CI 1.43–1.65; p < 0.001) and opioid use disorder (OR 1.36, 95% CI 1.14–1.64; p = 0.001) were also associated with mortality. Within hematologic malignancies, risk varied substantially. AML was associated with higher mortality (OR 1.82, 95% CI 1.75–1.91; p < 0.0001), while CLL showed a modest increase (OR 1.08, 95% CI 1.03–1.14; p = 0.003). In contrast, ALL (OR 0.76, 95% CI 0.70–0.83), MM (OR 0.81, 95% CI 0.78–0.85), and CML (OR 0.91, 95% CI 0.82–1.01) were associated with lower or neutral risk. These findings identify AML as the principal driver of excess sepsis mortality among hematologic cancers. Conclusions: Hematologic malignancies represent a distinct high-risk sepsis population, but they are not biologically uniform. AML defines a particularly high-risk sepsis phenotype, while ALL, MM, and CML show lower mortality risk. These findings support disease-specific risk stratification, early critical-care engagement, and tailored sepsis pathways, particularly for AML.

Quantifying time and financial toxicity of travel for intravesical chemotherapy: Rationale for in-home delivery of bladder cancer treatment.

e13516 Background: Approximately 63,000 adults are diagnosed with non–muscle-invasive bladder cancer (NMIBC) annually in the United States. Despite strong evidence of benefit, intravesical therapy is underutilized in routine practice for NMIBC. We sought to quantify the time and financial toxicity incurred by patients traveling for intravesical chemotherapy. Methods: In a retrospective study, all the patients with NMIBC who received outpatient intravesical treatment at a high-volume infusion center affiliated with the University of Minnesota between 2021 and 2024 were included. For each visit, total travel distance, driving time, and infusion duration were recorded. Indirect costs were estimated using two cost models incorporating mileage rates ($0.70 and $1.00 per mile) and lost productivity based on local median wages (using driving time and infusion duration per visit). Descriptive statistics were generated, and interquartile ranges (IQRs) were calculated for time to start the treatments, travel distance, and estimated per-visit cost. Results: A total of 114 patients with a median age of 78 years were included. 96/114 (84.2%) patients had high-risk NMIBC and 98/114 (85.9%) received intravesical Bacillus Calmette-Guérin as the intravesical agent. Overall, 68/114 (59.6%) waited more than 6 weeks and 24/114 (21.1%) waited more than 2 months to start the treatment. The median round-trip distance to the infusion center was 25 miles (IQR, 11.2–32.2). Across a typical six-cycle induction course, patients incurred $599 (IQR, $432–$674) in cumulative indirect costs with Model 1 and $647 (IQR, $452–$759) with Model 2, excluding medical copayments or lodging. (Table 1). Conclusions: A significant subset of patients experience delays in the start of intravesical treatment. Moreover, travel and time commitments for intravesical therapy contribute meaningfully to the financial and logistical toxicity of NMIBC care. Implementing strategies such as in-home intravesical treatment delivery could potentially mitigate these barriers and improve adherence to guideline-recommended therapy. Patient access/cost outcomes (N=114). Variable Value AUA risk category, n (%) (Intermediate risk / High risk) 18 (15.8%) / 96 (84.2%) Intravesical agent, n (%) (BCG / Gemcitabine–Docetaxel) 98 (85.9%) / 16 (14.1%) Time to start treatment ≥6 weeks, n (%) 68 (59.6%) Time to start treatment ≥2 months, n (%) 24 (21.1%) Round-trip distance (miles), median (IQR) 25 (11.2–32.2) Indirect cost per visit ($) – Model 1, median (IQR) 100 (72–112) Indirect cost per visit ($) – Model 2, median (IQR) 108 (75–126) Cumulative indirect cost over six-cycle ($) – Model 1, median (IQR) 599 (432–674) Cumulative indirect cost over six-cycle ($) – Model 2, median (IQR) 647 (452–759) AUA, American Urological Association; BCG, Bacillus Calmette-Guérin; IQR, Interquartile range.

Impact of insurance-related disparities on delay in treatment initiation among colorectal cancer patients in an underserved community teaching hospital.

e23382 Background: Timely treatment after colorectal cancer (CRC) diagnosis is critical, nationally therapy starts within 30–60 days. Patients in underserved communities, with public or no insurance, may experience delays. This study assessed insurance-related disparities in time to treatment initiation at a community teaching hospital. Methods: in a retrospective cohort study of CRC patients diagnosed in 2025,51 gastrointestinal malignancies identified, 21 were CRC, and 17 patients with available data were included. Patients were stratified into public/uninsured and private groups. Time to treatment initiation was recorded in days and categorized as ≤30, 30–60, and > 60 days. Continuous delays were compared using the Mann–Whitney U test, and delays > 60 days using Fisher’s exact test in GraphPad Prism 10 (p < 0.05). Results: Among 17 colorectal cancer patients, 10 were publicly insured/uninsured and 7 had private insurance. The cohort was 52.9% female, and 58.8% had advanced-stage disease (stage III–IV). Median time to treatment initiation was longer among publicly insured/uninsured patients compared with privately insured patients (72.5 vs 18.0 days; Mann–Whitney U = 16, p = 0.067). Delays exceeding 60 days occurred in 60% of publicly insured/uninsured patients versus 42.9% of privately insured patients (odds ratio [OR] 2.0, 95% CI 0.28–11.2; Fisher’s exact p = 0.64). Conclusions: Public insurance and lack of insurance were associated with a four-fold longer time to treatment initiation in colorectal cancer patients in Woonsocket ,RI community hospital, with publicly insured/uninsured patients exceeding national average. Although older age and advanced disease may contribute to delays due to diagnostic workup and care coordination, the consistent magnitude of delay observed suggests that insurance status remains an important independent factor .system-level interventions are needed to address insurance-related barriers and improve equity in oncologic care. Characteristics. number age at diagnosis sex cancer Rectal/Colon insurance medicare/private/uninsured date of diagnosis date of first treatment surgery chemort Tx type <30 30-60 >60days stage 1 83 m R M 5/30/25 8/26/25 Chem 88 days stage 4 2 66 f R M 10/1/25 10-Nov chem 40 stage 2 3 67 f C M 3/5/25 4/8/25 chemo 34 stage 4 4 77 m C M 9/5/25 11/21/25 Sur 77 stage 2 5 76 m C P 9/26/25 10/14/25 Chem 18 days stage3 6 85 f C P Nov-20 nov2020 Sur 7days stage2 7 83 m C M 6/4/25 8/27/25 rt 84 days stage 3 8 42 m R U 11/18/25 1/25/26 Chem 68 days stage 3 9 71 m R P 4/22/25 4/22/25 Sur 7 days stage1 10 69 m R M 4/30/25 7/20/25 cmo 90 days stage 4 11 83 f C M 7/23/25 8/27/25 Sur 35 days stage2 12 52 f C P 23-Sep 8-Dec Sur 76 days stage2 13 55 f C M 03/0802025 6/5/25 CMO 89 days stage 4 14 74 f C M 11/10/25 11/21/25 Sur 11 days stage 2 15 82 f C P 3/25/25 4/1/25 Sur 7 days stage2 16 70 m C P 5/21/25 8/11/25 Sur 82 days stage3 17 75 f R P 6/26/25 8/29/25 rt 64days stage 4

A phase II exploratory study of stratified therapy using toripalimab plus GP induction followed by concurrent immunoradiotherapy for locally advanced nasopharyngeal carcinoma.

e18070 Background: Immunotherapy is under extensive investigation for the treatment of locally advanced nasopharyngeal carcinoma (LANPC). The decreased tumor burden of induction immunotherapy may reduce the reliance on platinum-based chemotherapy and its associated toxicity during the concurrent phase. Therefore, this study explored response-adaptive therapy according to response of induction immunochemotherapy. Methods: This phase II study enrolled 30 patients with World Health Organization type II/III, stage III-IVA NPC. In the induction phase (Phase 1), patients received two cycles of toripalimab (240 mg, D1 Q3W) plus the GP regimen (gemcitabine 1000 mg/m² D1, 8, and cisplatin 80 mg/m² D1, Q3W), followed by efficacy assessment. Patients with disease progression (PD) proceeded directly to Phase 2. Other patients received two additional induction cycles. Prior to Phase 2, patients were stratified based on treatment response: 1)patients achieving partial response (PR) with ≥50% tumor reduction or a complete response (CR) received toripalimab concurrently with intensity-modulated radiotherapy (IMRT) for three cycles; 2) patients with PR <50%, stable disease (SD), or PD received concurrent IMRT with cisplatin and toripalimab. Results: Between October 2024 and December 2025, 30 patients were enrolled. One patient discontinued treatment after two cycles due to grade 3 immune-related keratitis. As of January 19th 2026, 29 patients completed therapy of Phase 1, 23 patients also completed the subsequent phase. All 29 patients completed four cycles induction therapy without PD. The overall objective response rate (ORR) was 89.66%, with 10 cases (34.48%) of CR, 16 cases (55.17%) of PR, and 3 cases (10.34%) of SD. Among PR patients, 12 (41.38% of total) achieved a tumor reduction of ≥50%. In safety analysis, 8 patients (26.67%) experienced grade 3-4 chemotherapy-related adverse events, most commonly anemia (10.00%), followed by nausea, vomiting, thrombocytopenia, and neutropenia (3.33% respectively). 1 patient (3.33%) developed a grade 3 immune-mediated adverse event (irAE) (keratitis). No grade 4 irAE was reported. Conclusions: Induction therapy with toripalimab combined with the GP regimen demonstrates a high tumor response rate and a manageable safety profile in patients with LANPC. These findings support the rationale for de-escalating platinum-based chemotherapy in subsequent concurrent treatment phases to potentially reduce toxicity. Clinical trial information: ChiCTR2400089302.

Association of AI-based quantitative B7-H3 evaluation in the tumor microenvironment with molecular subtypes in high-grade serous ovarian carcinoma.

5599 Background: High-grade serous ovarian carcinoma (HGSC) exhibits heterogeneous tumor microenvironments (TME) that correlate with clinical outcomes. We previously reported four molecular histopathological subtypes of HGSC based on gene expression profiles reflecting TME characteristics: Immune Reactive (IR), Mesenchymal (MT), Solid/Proliferative (SP), and Papillary/Glandular (PG). The IR subtype, characterized by tumor-infiltrating lymphocytes, demonstrates favorable prognosis (PMID:26993207, 30853361). We also identified lower B7-H3 expression in IR compared to non-IR subtypes (PMID:34799346). Here, we developed an AI-based B7-H3 IHC quantification method across TME compartments and investigated its association with molecular subtypes. Methods: Forty-five HGSC patients who underwent primary debulking surgery were enrolled. Molecular histopathological subtyping categorized cases into IR (n=16), MT (n=20), and Other (SP/PG, n=9). B7-H3 immunohistochemistry-stained whole slide images were analyzed using DeepPathFinder, an AI-based pathological image analysis tool that performs automated tissue segmentation and IHC marker quantification. B7-H3 expression was quantified across three tumor-associated compartments: tumor, peritumor (within 1000 μm of the tumor border), and necrosis. An integrated B7-H3 score was calculated as the mean density across these compartments. Survival analyses were performed using the Kaplan-Meier method and Cox proportional hazards regression. Results: The integrated B7-H3 score showed a stepwise increase across subtypes: IR (0.148) < MT (0.219) < Other (0.252) (Kruskal-Wallis p=0.091), and was significantly lower in IR compared to non-IR subtypes (0.148 vs 0.229, p=0.034). B7-H3 expression was consistently lower in IR across all compartments: tumor (IR: 0.171, MT: 0.205, Other: 0.257), peritumor (IR: 0.100, MT: 0.154, Other: 0.164), and necrosis (IR: 0.171, MT: 0.298, Other: 0.336; p=0.025). Survival analysis revealed significant prognostic differences among subtypes. Compared to IR, MT showed significantly worse outcomes (PFS: HR=8.35, 95% CI 3.00–23.24, p<0.0001; OS: HR=15.35, 95% CI 3.50–67.31, p=0.0003). The Other subtype also demonstrated inferior OS (HR=6.38, 95% CI 1.17–34.95, p=0.033). Both MT and Other (SP/PG) subtypes, which exhibited higher B7-H3 expression, were associated with worse prognosis. Conclusions: AI-based quantitative B7-H3 evaluation revealed subtype-specific expression patterns in the HGSC tumor microenvironment. B7-H3 expression was lowest in IR and higher in non-IR subtypes across all tumor-associated compartments. These findings suggest that patients with higher B7-H3-expressing subtypes may benefit from B7-H3-targeted therapy.