Browse Articles

Quantitative CT lung fibrosis and immune checkpoint inhibitor–related pneumonitis in NSCLC.

e20531 Background: Interstitial lung abnormalities (ILAs) are established risk factors for immune checkpoint inhibitor–related pneumonitis (ICI-P), but qualitative ILA assessment is subjective. We evaluated whether quantitative lung fibrosis (QLF) metrics from pretreatment computed tomography (CT) are associated with ICI-P in patients with non–small cell lung cancer (NSCLC) receiving immune checkpoint inhibitors (ICIs). Methods: We retrospectively analyzed 241 patients with metastatic NSCLC treated with ICIs. Pretreatment CT scans underwent quantitative analysis using commercial software (VIDA) to extract QLFs representing consolidation, ground-glass opacity (GGO), emphysema, reticulation, and honeycombing. QLFs were quantified by percentage, volume, and estimated mass at total lung, lobar (upper vs lower), and subregional (core [central] vs peel [peripheral]) levels. Associations between QLFs and ICI-P were evaluated using univariate logistic regression. Results: Regional QLFs demonstrated differential associations with ICI-P. Upper-lobe QLFs were not associated with pneumonitis (all p≥0.13). In contrast, lower-lobe mass-based fibrosis metrics were associated with increased ICI-P risk, including GGO mass (OR 1.005, 95% CI 1.001–1.010; p=0.030), quantitative ILD (QILD) mass (OR 1.004, 95% CI 1.000–1.008; p=0.031), and quantitative ILA (QILA) mass (OR 1.004, 95% CI 1.001–1.007; p=0.024). Total-lung QILA mass showed a trend toward association (OR 1.002, 95% CI 1.000–1.004; p=0.086). Subregional analysis demonstrated stronger associations in peripheral (peel) regions; peel consolidation volume was associated with ICI-P (OR 1.102, 95% CI 1.013–1.198; p=0.023), whereas no core metric reached significance. Conclusions: Quantitative fibrosis metrics from pretreatment CT, particularly lower-lobe and peripheral mass-based QLFs, are associated with immune checkpoint inhibitor–related pneumonitis in NSCLC. Incorporation of regional quantitative fibrosis measures into baseline imaging assessment may help identify patients at elevated pneumonitis risk and inform surveillance strategies. These findings support further multivariable validation of quantitative CT fibrosis metrics as imaging biomarkers for immunotherapy-related pneumonitis. Associations between quantitative lung fibrosis metrics and immune checkpoint inhibitor–related pneumonitis in NSCLC. Region Metric (Mass-based QLF) (gm) OR 95% CI p-value Lower lobe GGO Mass 1.005 1.001–1.010 0.030 Lower lobe QILD Mass 1.004 1.000–1.008 0.031 Lower lobe QILA Mass 1.004 1.001–1.007 0.024 Upper lobe No significant predictors — — ≥0.13 Total lung QILA Mass 1.002 1.000–1.004 0.086

Long-term survival results of perioperative chemoimmunotherapy with DCF and avelumab in locally advanced gastro-esophageal adenocarcinoma.

4096 Background: We have shown that avelumab, an anti-PD-L1 antibody, added to perioperative modified DCF chemotherapy (aDCF regimen), improves the pathologic complete response rate from 7% to 14% in locally advanced gastro-esophageal adenocarcinoma (GEA), in comparison to historical controls from our own institution. Little is known about the long-term survival results obtained with perioperative chemoimmunotherapy, a new standard of care. We are here reporting the 5-year survival results of our cohort of patients treated with aDCF. Methods: Single-arm phase II study of aDCF given every 2 weeks for 4 cycles before and after surgery with planned sample size of 50 operated patients in order to test hypothesis that pCR rate would improve from 7 to 20%. Main inclusion criteria were: histologically proven GEA, locally advanced disease (cT3-4 and/or N+), adequate organ function, WHO performance status 0-1. Main exclusion criteria were: histology other than adenocarcinoma, metastatic disease (M1), use of immunosuppressants, serious autoimmune disease, daily intake of more than 10 mg of prednisone. Pathological response (Tumour Regression Grade -TRG) was determined by the College of American Pathologists criteria: 0 = complete;1 = near complete/microscopic residual disease; 2 = moderate; 3 = poor/no response. MPR was defined as TRG 0 or 1. Data presented as median (range), KM determined survival. Patients with less than 5 years follow-up and who were disease free / alive were censored. Results: One of the 51 patients enrolled into the trial withdrew consent, leaving 50 patients in the survival analysis. Clinical disease burden was high (cT3 44/50, 88%, N+ 31/50, 62%) with 49/50 (98%) completing neoadjuvant aDCF. All patients proceeded to surgery with MPR found in 9/50 (18%) patients. The median follow-up for the cohort was 71.5 months (range 43-105) with an estimated mean DFS and OS of 62.5 months and 68.5 months corresponding to a 5-year DFS and OS of 55.3% and 61.1%. For patients with MPR, 5-year DFS and OS were 100%. There were no new safety signals. Conclusions: Perioperative aDCF resulted in long-term survival in the majority of patients treated on our trial, bringing more evidence that chemoimmunotherapy improves outcomes of patients with locally advanced GEA. Those outcomes seem enhanced among patients with MPR. Clinical trial information: NCT03288350 .

Hospital designation and survival disparities in young-onset gastrointestinal cancer.

1535 Background: Early-onset gastrointestinal (GI) cancers (<50 years) are increasing in incidence. We examined the impact of National Cancer Institute (NCI) designation and academic affiliation on mortality and care delivery for young-onset GI cancer patients in New York State. Methods: We performed a retrospective analysis using the Statewide Planning and Research Cooperative System (SPARCS) database from 2009 to 2024. Patients were stratified by admission type (emergency vs elective) and hospital site (NCI vs non-NCI; academic vs community). GI cancers included anal, biliary, colorectal, esophageal, gallbladder, liver, pancreatic, peritoneal, small intestine, and gastric cancers. Clinical characteristics were defined using the All Patient Refined grading system. Academic hospitals were classified via the Association of American Medical Colleges Database. Wilcoxon rank-sum tests compared continuous variables; Chi-squared or Fisher's exact tests for categorical variables. Linear regression was performed for continuous outcomes, and multinomial logistic regression for categorical outcomes, with significance at P ≤ 0.05. Results: A total of 29,753 young-onset GI cancer patients were included, of which 48.7% were emergency admissions. For emergency admissions, NCI sites treated patients with higher illness severity, and most had private insurance (P<0.001). There was no significant difference in length of stay (LOS) between NCI and non-NCI sites. Cost of treatment was significantly higher at NCI sites (P<0.001), which also performed more procedures (P = 0.002). NCI sites were associated with 36% lower odds of in-hospital mortality (OR=0.64 [0.55-0.75], P<0.001) despite higher illness severity. For elective admissions, in-hospital mortality was 0.5% at NCI sites vs 7.3% at non-NCI sites (OR=0.04 [0.03-0.05], P<0.001). For emergency admissions, academic sites treated patients with higher illness severity, and most patients had private insurance (P<0.001). Admission to an academic site was associated with increased LOS (β=0.03 [0.02-0.05], P<0.001). Cost of treatment was significantly higher at academic sites (P < 0.001), which performed more procedures (P <0.001). No significant difference in mortality was seen between academic and community sites. For elective admissions, in-hospital mortality was 0.5% at academic sites vs 5.8% at community sites (OR=0.08 [0.06-0.11], P<0.001). Across all hospitals and admission types, private insurance was independently associated with shorter length of stay (P<0.001). Conclusions: Treatment at NCI-designated centers for both emergent and elective admissions was associated with significantly lower in-hospital mortality among young-onset GI cancer patients despite higher illness severity. Insurance-based disparities highlight the need for interventions to ensure equitable access to high-quality cancer care for this vulnerable population.

SWIVEL: A randomized phase II trial of switching medications versus guideline-directed interventions for adjuvant aromatase inhibitor side effects in breast cancer patients.

TPS12171 Background: Approximately half of breast cancer patients do not complete the standard 5-year course of adjuvant hormone therapy due to side effects, a finding associated with increased recurrence rates and breast cancer-specific mortality. In clinical practice, side effects are commonly addressed by switching to an alternative hormone therapy; however, this strategy has not been prospectively evaluated. In prior retrospective work, adherence was significantly lower among patients who switched therapies compared with those who remained on the same therapy and received treatment for side effects (62% vs 91%, p=0.013); however, interpretation is limited by the retrospective design. Therefore, a prospective randomized clinical trial was initiated to address this evidence gap by directly comparing medication switching with guideline-directed symptom management. Methods: SWIVEL is an ongoing phase II randomized trial comparing medication switching with guideline-directed interventions for the management of aromatase inhibitor side effects. Eligible patients are postmenopausal women or men with stage I–III ER-positive/HER2-negative breast cancer who are planning to initiate aromatase inhibitor monotherapy in the adjuvant setting. Upon enrollment, participants will be screened every 4 weeks with a validated single-item questionnaire (FACIT/GP5). Those who respond that they are ‘quite a bit’ or ‘very much’ bothered by side effects of treatment will be randomized to either (1) switch to a different hormone therapy, which may include an alternative aromatase inhibitor or tamoxifen, or (2) initiate an evidence-based intervention for side effects in accordance with NCCN guidelines and patient-provider preference. Participants may crossover between strategies at any time. The primary outcome is change in patient-reported symptom burden at 3 months following randomization or at crossover, whichever occurs first, measured by the FACT-ES survey. Secondary outcomes include patient-reported symptom burden at 6 months assessed in an intention-to-treat analysis using FACT-ES and PROMIS measures, and medication adherence assessed at 3, 6, 12, and 24 months using a multi-modal approach incorporating validated patient-reported measures (VOILS), pharmacy records, and urine testing for drug metabolites. Additional outcomes include patient-reported quality of life (FACT-G) and sexual function (FSFI), factors associated with adherence including access to care among rural communities, and evaluation of a novel screening tool to identify patients at higher risk for non-adherence. The study is open at Dartmouth Health and has enrolled 7 of 200 planned participants (NCT# 07071038). Clinical trial information: RCT07071038 .

Clear cell–type squamous cell carcinoma: An NCDB analysis.

e21549 Background: Clear cell–type squamous cell carcinoma (ccSCC) is a rare, aggressive histologic variant of squamous cell carcinoma characterized by clear cytoplasmic features resulting from glycogen accumulation. Immunohistochemical evaluation is essential to distinguish ccSCC from other clear cell neoplasms. Given its rarity, ccSCC is mainly documented in small case series, though available reports highlight recurrence, metastatic potential, and limited therapeutic guidance, underscoring the challenges in management. Population-level data are needed to better define its presentation. To address this, the National Cancer Database (NCDB) was analyzed to determine the demographic profile of patients diagnosed with ccSCC. Methods: A retrospective cohort analysis of the 2004–2020 NCDB identified 476 patients with histologically confirmed ccSCC (ICD-O-3 code 8084). Demographic and clinical variables (age, sex, race, ethnicity, primary site, stage, urban/rural residence, income, insurance status, facility type, treatment, and survival) were analyzed using descriptive statistics, and incidence trends were assessed via regression analysis. Results: This study identified 476 patients with ccSCC diagnosed between 2004 and 2020. The mean age at diagnosis was 67.0 years (SD = 12.3, range = 27–90 years), and the cohort was predominantly female (55.3%), White (82.6%), and non-Hispanic (91.0%). The upper lobe of the lung was the most common primary site (45.2%), and 44.5% presented with Stage I disease. Socioeconomic patterns demonstrated that 80.3% of individuals resided in metropolitan areas and 33.3% lived in the highest income quartile (≥$74,063). Medicare coverage was most common (61.6%), and treatment was largely concentrated in academic/research (40.4%) and comprehensive community cancer programs (34.9%). Tumors measured an average of 46 mm. Surgical intervention was performed in 72.3% of patients, yielding negative margins in 66.0%; radiation and chemotherapy were administered in 32.4% and 34.2% of cases, respectively. Long-term outcomes were poor, with overall survival of 62.7% at two years, 43.4% at five years, and 26.0% at ten years, and a mean survival of 73.3 months. Conclusions: This NCDB analysis provides one of the first large, population-level assessments of ccSCC and reveals several clinically meaningful patterns. Although nearly half of patients presented with Stage I disease, long-term survival remained poor, indicating that early clinical stage may not mitigate the aggressive behavior of this variant. Tumors most often originated in the upper lobe of the lung and occurred predominantly in older, Medicare-insured individuals living in metropolitan areas. Surgical management was common, yet margin-negative resection did not translate into durable survival outcomes. These findings highlight the need for better prognostic tools and more effective treatment strategies for this rare malignancy.

Patient characteristics and real-world utilization of darolutamide in advanced prostate cancer among a national cohort of veterans.

e23426 Background: Darolutamide (DARO) has demonstrated efficacy and a tolerable safety profile in prostate cancer (PC) clinical trials spanning nonmetastatic castration-resistant PC (nmCRPC) and metastatic castration-sensitive prostate cancer (mCSPC), resulting in FDA approval. Real-world evidence of DARO utilization, however, remains limited. Here we summarize patient characteristics and utilization patterns of DARO across the PC spectrum among a national cohort of United States Veterans treated in the Department of Veterans Affairs (VA). Methods: We conducted a retrospective, observational study among Veterans with advanced PC diagnosed from January 2019 through December 2024. This cohort was identified from the Veterans Health Administration’s Corporate Data Warehouse. Patients were indexed into the study at first initiation of DARO. Advanced PC was defined as diagnosis with mCSPC, nmCRPC, or metastatic castration-resistant PC (mCRPC). Analyses of these disease states were operationalized through analysis of structured data (e.g., prostate-specific antigen values) and unstructured data (e.g., clinical notes). Veterans were included in this analysis if treatment with DARO was initiated after diagnosis of advanced PC. The follow-up period extended for a minimum of 2 months after index or until patient’s date of death, the final date of activity in database, or the end of the study period. Use of androgen deprivation therapy (ADT) and/or docetaxel alongside DARO was examined. Results: 1,729 Veterans with advanced prostate cancer and who had DARO ± ADT during the study period were identified. These Veterans were 57.7% White, 27.1% Black, 5.3% Hispanic and 9.9% Other or Unknown. 75.6% resided in urban locations with a median age of 75 yrs. At the time of treatment initiation, 783 (45.3%) were diagnosed with mCSPC, 301 (17.4%) were diagnosed with nmCRPC, and 645 (37.3%) were diagnosed with mCRPC. No notable differences in patient characteristics were observed between these clinical subgroups. Among the veterans treated with DARO ±ADT, 198 (11.5%) Veterans received DOCI therapy concomitantly and of this, 153 (77.3%) had mCSPC. Conclusions: Our findings suggest that DARO is increasingly utilized in advanced PC. Our results suggest that VA physicians are prescribing DARO in mCSPC and mCRPC as off-label indication, potentially reflecting confidence in its efficacy and favorable safety profile.

Geographic and demographic differences in genomic testing and outcomes in pancreatic cancer.

e22629 Background: National Comprehensive Cancer Network (NCCN) guidelines recommend universal tumor genomic profiling for patients with pancreatic ductal adenocarcinoma (PDAC) to identify actionable alterations and guide treatment. Real-world implementation of next-generation sequencing (NGS) may vary by geographic and demographic factors, potentially contributing to disparities in treatment delivery and outcomes. Methods: We conducted a retrospective cohort study of patients with PDAC treated at a single academic center serving the Washington, DC–Maryland–Virginia (DMV) region. Clinical, demographic, treatment, and outcome data were abstracted from the medical record. Residential zip codes were categorized as Washington, DC or suburban Maryland/Virginia. Receipt and timing of tumor genomic profiling, treatment intensity, and overall survival (OS) were evaluated descriptively by geographic location and patient demographics. Results: Twenty-two patients with PDAC were identified. Median age was 69 years, 55% were male, and 64% were Black. Sixteen patients resided in Washington, DC and six in suburban Maryland or Virginia. Tumor genomic profiling was performed in 93% of DC residents and 100% of suburban residents. Suburban residents were more likely to receive multi-agent chemotherapy and undergo curative-intent resection. Median OS was approximately 14 months among DC residents and 23 months among suburban residents. Receipt and timing of tumor genomic profiling varied by geographic and demographic subgroups, including race, age, and treatment intensity. Common genomic alterations, including KRAS and TP53 , did not differ across demographic groups. Conclusions: In this real-world cohort, implementation of tumor genomic profiling and treatment intensity for PDAC varied by geographic location and patient demographics and was associated with differences in survival outcomes. These findings suggest that geographic factors, which may reflect underlying differences in access to care, influence delivery of guideline-concordant precision oncology and highlight the need for strategies to promote equitable delivery of tumor genomic profiling. Demographics, treatments, and outcomes by geographic location. Washington, DC (n=16) MD/VA Suburbs (n=6) Median age, years 67 74 Male sex, n (%) 8 (50) 4 (67) Black, n (%) 9 (56) 5 (83) Tumor genomic profiling performed, n (%) 15 (93) 6 (100) Received multi-agent chemotherapy, n (%) 13 (81) 6 (100) Curative-intent resection, n (%) 4 (25) 4 (67) Median OS a , months ~14 ~23 Common genomic alterations KRAS, TP53 KRAS, TP53 a. Overall survival.

Efficacy and dosimetry of yttrium-90 radiation segmentectomy for unresectable hepatocellular carcinoma: A retrospective study in China.

e14555 Background: Radiation segmentectomy (RS) with yttrium-90 ( 90 Y) microspheres has evolved as a targeted ablative approach for hepatocellular carcinoma (HCC). However, real-world evidence on its efficacy and dosimetric correlates remains limited in Chinese patients. This study assesses clinical outcomes and dosimetry of Radiation segmentectomy in unresectable HCC. Methods: We retrospectively analyzed patients with hepatocellular carcinoma (HCC) who received 90 Y - RS at our center between June 2022 and October 2025. The primary endpoint was objective response rate (ORR). Secondary endpoints included disease control rate (DCR), downstaging rate, and safety. Tumor response was assessed using mRECIST criteria, and adverse events were graded according to CTCAE v5.0. Results: 50 patients were included and were followed more than 3 months. The baseline characteristics of the patients are shown in the Table. The median age was 55 years (IQR, 50-59 years) and 84% of patients were male. Hepatitis B infection was common (84%),. The median diameter of solitary lesions was 82.2 mm（IQR 57.5-103.9 mm), median maximum tumor diameter was 60.86 mm (IQR 43.7-94.5 mm), with 78% having tumors > 5 cm and 56% having ≥3 lesions. BCLC staging distribution was A (12%), B (24%), and C (64%). Portal vein tumor thrombus was observed in 54% of patients, and 36% had extrahepatic lymph node metastases. The median 90 Y dose was 2 GBq (IQR, 1.3–3.2 GBq), delivering a median radiation dose of 243 Gy (IQR, 200–300 Gy) to the target tissue. Most patients (72%) received concomitant systemic therapy. Treatment responses according to mRECIST criteria showed an objective response rate (ORR) of 74% and disease control rate (DCR) of 82%, with complete response (CR) rate of 52.0% and partial response (PR) rate of 22%. Notably, 58% of patients achieved downstaging, with 11 patients (22%) undergoing potentially curative procedures (5 resections, 6 liver transplants). Treatment was well tolerated, with no grade ≥3 adverse events reported. Additionally, a tumor dose threshold of 215.5 Gy was identified as a predictor of objective response (AUC = 0.69). Conclusions: In this retrospective study, Yttrium-90 radiation segmentectomy demonstrated favorable safety and efficacy in Chinese patients，achieving an impressive objective response rate (ORR) of and a substantial downstaging rate. A tumor dose threshold of 215.5 Gy was identified as a predictor of treatment response. Patient Characteristics (N=50) Median IQR 25, 75 or n (%) Age, years 55 49.5 - 58.5 Diameter of solitary lesions 82.2 57.5 - 103.9 Max tumor diameter, mm 60.9 43.7 - 94.5 Number of tumors > 3 28 (56%) PVTT 27 (54%) AFP, ng/mL 286.4 16.3 - 1001.2 Treatment Responses (mRECIST) Objective Response Rate 74% Disease control rate 82%

Multifaceted graphene-based materials for comprehensive environmental sustainability

Inhibiting the integrated stress response restores cognition

Multifunctional Buffer Layer for Bolstering the Stability and Photovoltaic Performance of Perovskite Solar Cells

ABSTRACT Despite the impressive power conversion efficiency (PCE) of perovskite solar cells (PSCs), their long‐term operational stability remains compromised by endogenous ion migration and interfacial recombination. Herein, we report a robust strategy by introducing a novel multifunctional cathode buffer layer based on 4,4′‐((1,10‐Phenanthroline‐3,8‐diyl)bis(ethyne‐2,1‐diyl))dianiline (BAE‐Phen), which exhibits excellent thermal stability. Theoretical simulations and experimental characterizations reveal that BAE‐Phen operates through synergistic mechanisms: its phenanthroline core strongly coordinates with metal ions to decelerate detrimental electrode corrosion, while its extended π‐conjugated backbone enhances π–π stacking with the [6,6]‐phenyl‐C 61 ‐butyric acid methyl ester (PCBM) electron transport layer, facilitating efficient charge transfer. Consequently, the optimized BAE‐Phen‐based devices achieve a champion PCE of 27.07% (certified 26.85%). Notably, unencapsulated devices retained 90.5% of their initial PCE after 2000 h of thermal aging at 85°C. Furthermore, encapsulated devices maintain nearly 100% of their initial performance after 2200 h of continuous maximum power point tracking under 1‐sun illumination, demonstrating exceptional thermal and operational stability. This work presents a strategic interface engineering approach using a multifunctional molecular buffer, providing pivotal insights into the synergistic optimization of charge transmission and ionic to electronic stability for next‐generation photovoltaics.

Conditional survival patterns and individualized prognostic prediction in malignant peritoneal mesothelioma

Living after lymphoma: A retrospective study of survivorship outcomes from a tertiary cancer centre in rural India.

e24130 Background: With improving survival in lymphoma, long-term treatment-related complications have emerged as a major determinant of quality of life. However, comprehensive survivorship data from India remain scarce. This study evaluates metabolic, endocrine, bone, cardiovascular, and psychosocial outcomes among adult lymphoma survivors at a tertiary cancer centre in rural India. Methods: This retrospective observational study included adult Hodgkin and non-Hodgkin lymphoma survivors treated between 2012 and 2022, who completed curative-intent therapy and remained disease-free for ≥3 years. Survivorship evaluations performed within the last 2 years were reviewed. Data on demographics, disease characteristics, treatment exposures, metabolic parameters, endocrine function, bone health, cardiovascular events, secondary malignancies, and psychosocial outcomes were analysed using descriptive statistics. Results: A total of 100 survivors were included (62 males, 38 females), with a median age of 58.5 years (range 24–81). Diagnoses included diffuse large B-cell lymphoma (45%), follicular lymphoma (17%), Hodgkin lymphoma (20%), and others (18%). Most patients had advanced disease (stage III–IV, 59%). Treatments included chemotherapy in 97% (R-CHOP 52, ABVD 19, BR 11, others 15), radiotherapy in 20%, and autologous stem cell transplant in 9%. Metabolic complications were common: obesity in 40.4%, hypertension in 31%, dyslipidemia in 56%, and diabetes mellitus in 31.1%. Mean BMI was 24.3±5.1 kg/m². Mean total cholesterol was 210±35 mg/dL, LDL 130±36 mg/dL, triglycerides 175±86 mg/dL, and HbA1c 6.38±1.1%. Hypothyroidism was detected in 8.4%. Vitamin D insufficiency or deficiency was observed in 76.5%. Bone health assessment revealed osteopenia in 37% and osteoporosis in 23%. Two patients developed new-onset left ventricular dysfunction, three experienced cerebrovascular events, and two developed premature cataracts before the age of 50. Two patients developed tuberculosis post-treatment. Second primary malignancies occurred in two patients (DLBCL→AML; ALCL→ Hodgkin lymphoma). Psychosocial impact was notable, with five patients remaining unmarried and one experiencing marital disruption attributed to fear of recurrence and social stigma. Conclusions: Lymphoma survivors in rural India experience a substantial burden of metabolic, skeletal, cardiovascular, and psychosocial complications. These findings highlight the urgent need for structured, multidisciplinary survivorship care models tailored to resource-limited settings to improve long-term outcomes and quality of life.

Vimseltinib vs. pexidartinib: An indirect comparison of response rate in the treatment of tenosynovial giant cell tumor.

e23567 Background: Tenosynovial giant cell tumor (TGCT) is a locally aggressive but rare neoplasm of the synovium, tendon sheaths, and bursae frequently affecting major joints with a mean age of diagnosis at 35 to 50 years. Surgical resection is the mainstay of treatment. While generally not life-threatening, inadequate treatment is associated with a significant decrease in quality of life. TGCT also overexpresses colony-stimulating factor I (CSF 1) which has been successfully targeted with an inhibitor, Pexidartinib, which has been the only systemic therapy available for TGCT not amenable to surgical resection. Vimseltinib is an overall switch control-tyrosine kinase inhibitor which does not have the associated hepatotoxicity of pexidartinib. No head-to-head clinical trial between these agents has been performed; this indirect treatment comparison seeks to establish their relative efficacy, by comparing rates of radiological response by RECIST criteria in their major trials. Methods: Data pertaining to the ENLIVEN (pexidartinib vs. placebo) and MOTION (vimseltinib vs. placebo) clinical trials was analyzed; overall response rates were extracted along with 95% confidence intervals. ENLIVEN was a phase 3 multinational randomized double-blind placebo controlled clinical trial with 1:1 randomization and MOTION was a phase 3 randomized double-blind placebo controlled clinical trial with 2:1 randomization. The placebo arm was used as the common comparator. Baseline patient population characteristics were compared along with treatment emergent adverse events. The relative overall response was compared using the Bucher method for indirect treatment comparison. Upper and lower limits of the 95% confidence intervals were also calculated. Results: A hazard ratio of 1.03 was obtained using indirect treatment comparison, indicating that these agents have similar overall response rates. The upper and lower limits of the confidence intervals were 0.66 and 1.59 which makes the comparison statistically insignificant. Inclusion of baseline patient characteristics were also modeled and did not significantly alter the outcome. Simulation studies of patient compliance and disease progression due to discontinuation/suspension secondary to adverse drug related reactions tends to favor vimseltinib. Conclusions: Vimseltinib appears to be as effective as pexidartinib in producing similar response rates as documented by radiological response by RECIST criteria. However, given that vimseltinib appears to have a better toxicity profile than pexidartinib, it may offer a better treatment option to patients with TGCT as both agents have been shown to have similar response rates.

Preclinical efficacy of the anti-EBV agent brincidofovir (BCV) in EBV-associated gastric carcinoma (EBVaGC).

4028 Background: Immune checkpoint inhibitor (ICI) plus chemotherapy is an approved therapy for EBVaGC whose tumors express PD-L1 (CPS ≥1). However, some patients are resistant to ICI or exhibit myelosuppression, making them unable to tolerate chemotherapy. BCV, an antiviral drug with no myelotoxicity, is effective against EBV (EC50: 30 nM), and has demonstrated anti-tumor activity against head and neck cancer and EBV-associated lymphoma. Recently, we found that SNU-719, an EBVaGC cell line, is sensitive to BCV in vitro (IC50: 1.95 µM). To seek conventional chemo-free regimens for EBVaGC, we have studied the preclinical efficacy of BCV monotherapy and combination therapy with ICI against EBVaGC. Methods: BCV was tested for its ability to induce dsDNA breaks in SNU-719 cells by γ-H2A.X immunostaining. To assess synergy of BCV with ICI, we analyzed BCV-induced immunogenic cell death (ICD) in SNU-719 cells using calreticulin and HMGB-1 assays. Additionally, we assessed BCV-induced expression of PD-L1 using qPCR and FACS. Since BCV induced both ICD and PD-L1 expression in vitro , we then evaluated the in vivo anti-tumor activity of BCV with or without ICI in humanized NOG mice bearing SNU-719 xenografts. Specifically, mice received intraperitoneal administration of vehicle, BCV, pembrolizumab (Pembro), or Pembro+BCV. Subcutaneous tumor growth was monitored for each group. After euthanasia, the SNU-719 xenografts were also used to measure the mRNA expression levels of EBV genes including EBNA1 . Results: BCV induced both dsDNA breaks and ICD in SNU-719 cells, similar to the positive control bleomycin. It also induced PD-L1 mRNA and protein expression. Subsequent in vivo studies showed that, while BCV or Pembro monotherapy showed partial efficacy, the combination therapy of BCV and Pembro was effective in all mice. Notably, treatment with BCV significantly reduced the expression levels of EBNA1 mRNA ( p < 0.05) in the tumor tissue of SNU-719 xenografts. Conclusions: EBV is an oncovirus and its gene product EBNA1 is an oncoprotein for EBVaGC. This study showed a significant reduction of EBNA1 mRNA expression by BCV. Therefore, in addition to its own anti-tumor effects, BCV may contribute significantly to the suppression of EBVaGC tumorigenesis via its anti-EBV activity. From a clinical perspective, BCV was effective against EBVaGC xenografts when combined with ICI. This less toxic regimen may be beneficial for EBVaGC patients, especially those with myelosuppression.

INB-200: Phase I study and characterization of gene-modified autologous gamma delta (γδ) T cells in newly diagnosed glioblastoma multiforme (GBM).

2077 Background: Cell therapy for glioblastoma (GBM) holds promise but previous studies have not demonstrated durable responses or improved progression free survival (PFS). We provide updated findings from our assessment of any patients who completed treatment with drug-resistant immunotherapy (DRI) in newly diagnosed GBM. Methods: DRI combines the polyclonal upregulation of stress-antigens with concurrent local delivery of MGMT-modified γδ T cells, engineered to resist temozolomide (TMZ) driven lymphodepletion. The standard-of-care (SOC) chemotherapy TMZ induces transient upregulation of tumor-associated NKG2D-L stress antigens enabling the functional γδ T cells to identify and eradicate residual cancer cells during the period of maximum tumor vulnerability. Patients received 1 x 10 7 DRI cells/dose into the resection cavity with 150 mg/m 2 of IV TMZ on Day 1 of each maintenance cycle.17 patients received treatment, including 3 who received only a single dose (Cohort 1). 14 patients received repeated doses (3 to 6), (median age 64 (range: 33-75); 93% IDH-WT, 50% MGMT-unmethylated, 57% subtotal resection, KPS 80). 10 contemporaneously enrolled patients treated with SOC only (median age 67 (range: 49-77); 100% IDH-WT, 60% MGMT-unmethylated, 20% subtotal resection, KPS 80) were examined as a control cohort. Results: No DLTs, CRS, ICANS or treatment related deaths were reported. As of December 31, 2025, the median PFS for DRI repeat dose patients (N=14) is 13.0 months (m), vs. SOC control (N=10) of 6.6m (+97%) demonstrating prolonged disease control in newly diagnosed GBM. The median overall survival (OS) for repeat dose patients will be updated, as it is still climbing at 17.2m+ vs. SOC control of 13.2m. Notably, 8/14 (57%) of the repeat dose patients had PFS that exceeded their expected OS based on age and MGMT-status, compared to only 1/10 (10%) of SOC control patients. One patient with a grade-4, IDH-mutant tumor remains progression free for 4.6 years with no additional maintenance therapies. Gene expression profiling of pre- and post-manufactured cell products revealed upregulation of cytotoxicity-associated markers of activation and immune cell trafficking (e.g., IFN-g, GZMB, PRF1 ). Patient correlate analysis showed that CRS associated serum cytokines IL-6, IL-1B , CCL2 , and TNF-a were not increased over baseline during DRI + maintenance-phase TMZ. Peripheral γδ and total T cell levels, however, improved incrementally with each dose frequency cohort, suggesting an influence on systemic immune reconstitution through local DRI delivery. T cell persistence was demonstrated in recurrent GBM biopsies from patients that revealed significant γδ T cell infiltration and necrosis well beyond the final dose of TMZ and DRI γδ T cells. Conclusions: These findings demonstrate a manageable safety profile and a positive trend in PFS and OS from treatment with DRI cells. Clinical trial information: NCT04165941 .

ctDNA MRD combined with CODEX2 to identify high-risk colorectal cancer with potential sensitivity to immunotherapy.

e15664 Background: Circulating tumour DNA (ctDNA)-based minimal residual disease (MRD) detection identifies colorectal cancer (CRC) patients at high risk of relapse after curative surgery, but lacks associated therapeutic stratification. We developed CODEX2, a validated histological H-score quantifying CDX2 silencing, and hypothesised that integrating MRD status with tumour-intrinsic biology could refine post-surgical risk stratification and identify biologically actionable subsets. Methods: Forty-six resected CRC patients with whole-exome sequencing-based ctDNA MRD assessment 6 weeks post-surgery were analysed (21 MRD-, 25 MRD+). CDX2 expression was evaluated in matched FFPE primary tumours using the CODEX2 H-score and classified as high or low. Multivariable Cox models included MRD, CODEX2, MSI status and clinicopathological variables, with Akaike information criterion (AIC) for model selection. Transcriptomic analyses assessed immune pathways and Consensus Molecular Subtypes (CMS). Results: Eight patients were MRD+/CODEX2-low, six of whom were microsatellite stable (MSS). These MSS tumours displayed a CMS1-like, immune-inflamed transcriptomic profile enriched for interferon-γ signalling despite microsatellite stability. Combined stratification identified MRD+/CODEX2-low patients as the highest-risk group for relapse (HR 47.0; p = 0.00004), followed by MRD+/CODEX2-high cases (HR 5.8; p = 0.016), compared with MRD-/CODEX2-high patients. In multivariable analysis, both MRD positivity (HR 2.8; p = 0.01) and CODEX2-low status (HR 2.1; p = 0.04) independently predicted shorter relapse-free survival, while MSI status was not retained. The combined MRD+CODEX2 model outperformed MRD alone (ΔAIC -12.4). Conclusions: Integrating ctDNA-based MRD detection with CODEX2 refines post-surgical risk stratification in CRC. MRD-positive patients with CODEX2-low MSS tumours represent a biologically distinct, immune-inflamed, high-risk subgroup with MSI-like features and potential susceptibility to immunotherapeutic strategies.

Symptom monitoring program linked to electronic referrals (SyMPLER): A single-arm pilot feasibility study of remote symptom monitoring with palliative care self-referral.

TPS12164 Background: Symptoms often fluctuate during an individual’s cancer trajectory, and this can impact patients’ health-related quality of life (HRQOL). Two evidence-based approaches—symptom monitoring programs (SMPs) and specialty palliative care (PC)—have independently been shown to improve HRQOL through identification and treatment of cancer-related symptoms, especially among patients with thoracic cancers (i.e. lung cancer). However, PC remains underutilized when patients report uncontrolled symptoms via SMPs. Therefore, we developed SyMPLER, a digital SMP with an integrated option for PC self-referral. The primary aim of this phase 1 clinical trial is to determine the feasibility of SyMPLER among patients newly diagnosed with a thoracic cancer. If feasible for remote symptom reporting, SyMPLER may overcome logistical and attitudinal barriers towards provider-driven PC referral processes for cancer symptom management. Methods: SyMPLER is a prospective single-arm feasibility study of a novel digital health intervention for remote symptom reporting with an optional PC self-referral for cancer symptom management (NCT06396598). SyMPLER includes 5 main components: (1) a brief education about PC, (2) weekly text reminders to encourage symptom logging, (3) on-demand symptom reporting using the Edmonton Symptom Assessment System on a mobile health application interface, (4) an active choice option for PC self-referral, and (5) the ability to request a callback from oncology clinicians to discuss symptom concerns. Per the IRB-approved protocol, adults ( > 18 years) with any stage thoracic malignancy, without a prior PC referral, are consented within the first 12 weeks of outpatient oncology care. The primary outcome is frequency of symptom reporting with a feasibility benchmark set at 50% of participants logging symptoms at least monthly over a 6-month intervention period. PC referrals (secondary outcome) and exploratory outcomes, including telephone encounters with the oncology team, HRQOL (Function Assessment of Cancer Therapy – Lung; FACT-L), and patient satisfaction with medical care (FAMCARE-P13) are measured every 3 months for up to 12 months post-enrollment. Study enrollment commenced in February 2024. The target enrollment of 94 evaluable participants was achieved in August 2025. Data collection is ongoing for 23 patients still active on study. Clinical trial information: NCT06396598 .

The efficacy and limitations of artificial intelligence (AI) in radiologic tumor assessment.

e20002 Background: Numerous Artificial Intelligence (AI) models for evaluating radiologic images are emerging, yet limited data exist to demonstrate the efficacy of these tools, particularly within oncology. Notably, these tools often require assistance from radiologists to verify their accuracy, and thus further research is needed to help characterize the current efficacy and limitations of AI in radiologic tumor assessment. This study evaluated ChatGPT-5.2 in thoracic tumor identification and measurement in order to enhance our understanding of the current efficacy and limitations of AI in oncologic radiology assessments. Methods: 61 publicly available CT scans with known thoracic tumors from the Lung CT Diagnosis collection (The Cancer Imaging Archive) were analyzed, with 1 dominant tumor per scan. A single radiologist identified tumors using a representative slice and measured maximal anteroposterior (AP), transverse (TV), and craniocaudal (CC) diameters, with tumor volume calculated using the ellipsoid formula. We sequentially prompted ChatGPT-5.2 to identify tumors on axial and coronal images and to generate corresponding measurements using DICOM metadata. Radiologist and AI measurements were compared using paired t-tests. Results: ChatGPT showed no significant differences from radiologist measurements for axial anteroposterior (26.2 vs 25.2 mm, p=0.176) or transverse diameters (26.9 vs 26.2 mm, p=0.397) across 61 tumor assessments. In contrast, craniocaudal diameters (34.6 vs 24.4 mm, p<0.001) and derived tumor volumes (21.2 vs 14.4 mL, p<0.001) were significantly overestimated. Average percent differences were modest for AP (7.3%) and TV (5.0%) but markedly higher for CC (51.5%) and volume (79.0%). Percent differences for volume measurements ranged from -100% to 745.6% different from the radiologist assessment. The median number of prompts required for correct tumor identification by ChatGPT was 3 for both axial and coronal images. Conclusions: Although ChatGPT-5.2 is not a radiology-specific AI model, this tool utilizes similar approaches to radiology-specific models and can be trained to assess images. While the average percentage differences between ChatGPT and radiologist-assessed images appeared similar, we found wide variation in assessment percentage from image to image. Findings highlight that although AI models may provide a mechanism for aggregating and analyzing large volumes of data, further work is needed to identify how best to incorporate AI tools in radiologic tumor assessment. Comparison of radiologist and ChatGPT-5.2 tumor measurements (n = 61). Measurement Radiologist Mean ChatGPT Mean p -value (paired t -test) Mean % Difference % Difference Range AP (mm) 25.2 26.2 0.176 7.3% -38.9 to 75% TV (mm) 26.2 26.9 0.397 5.0% -65.4 to 120% CC (mm) 24.4 34.6 <0.001 51.5% -100 to 380% Tumor Volume (mL) 14.4 21.2 <0.001 79.0% -100 to 745.6%

Accessibility of study contact information in ClinicalTrials.gov: A comprehensive analysis.

1534 Background: While the official purpose of ClinicalTrials.gov is to satisfy regulatory requirements from the U.S. Food and Drug Administration meant to safeguard human subjects, the database also serves as the most comprehensive record of ongoing clinical trials for patient and provider use. Previous studies have focused on compliance with results reporting and completeness of information within restricted subsets of the database. Taking the perspective of a potential participant or their provider, the purpose of this study was to assess the accessibility of ClinicalTrials.gov contact information across actively recruiting oncology and non-oncology studies. Methods: This study included all actively recruiting interventional studies listed in the ClinicalTrials.gov database as of June 11, 2025. We assessed the completeness of location and contact information by multiple factors including oncology versus non-oncology study, sponsor type, trial phase, study start date, and date of last posted update. Within oncology studies, we further analyzed information availability by disease subsite. Results: We identified 47,703 actively recruiting interventional trials, among which 12,400 (26.0%) pertained to oncology. Within these oncology studies, 3178 (25.6%) had an industry sponsor, 9729 (78.5%) had a study phase listed (among which 59.4% were Phase 2 studies), and 4311 (34.8%) involved an FDA-regulated drug. Across both oncology and non-oncology studies, 2678 (5.6%) were missing any central contact information. While all trials had at least one study location listed, 3492 (7.3%) did not include a location-specific contact name and 9994 (21.0%) had neither a location-specific contact phone nor email address (24.1% for oncology, 19.8% for non-oncology). Within oncology trials, 50.9% of studies with an industry sponsor lacked any location-specific contact phone or email, compared to 11.3% for NIH sponsors and 14.9% for other (including academic) sponsors. Completeness of location-specific contact information was similar across trial phases. By oncology disease subsite, thoracic trials (29.8%) or those including multiple disease subsites (33.8%) most often lacked location-specific contact phone and email, whereas head and neck (15.4%) and neuro-oncology (16.3%) trials were least likely to do so. Conclusions: Location remains a primary means by which many potential trial participants filter clinical trial opportunities. Missing location-specific contact information, particularly among trials with industry sponsors, is an area of potential improvement that could be addressed by increased attention or more stringent registration requirements. While emerging strategies to improve clinical trial accessibility include artificial intelligence-powered platforms, these are only as good as the primary data on which they are built.