Browse Articles
Discover research articles across all indexed journals
Validation of a generative AI-based clinical decision support system for surgical extent in thyroid cancer in Brazil: Potential aid to de-escalating treatment to preserve function.
1604 Background: The decision between thyroid lobectomy (TL) and total thyroidectomy (TT) for differentiated thyroid carcinoma involves balancing oncological control with morbidity. While NCCN guidelines support TL for low-risk cases, clinical practice often favors TT due to risk aversion, exposing patients to lifelong levothyroxine dependence. We validated "Korina," a clinical decision support system based on the Gemini Large Language Model developed at Instituto do Câncer do Ceará (ICC), Brazil, to assess its ability to optimize surgical decision-making and identify candidates suitable for organ-preserving surgery. Methods: We validated the AI model using a retrospective thyroid cohort (N = 100). The median age was 47 years (IQR 38–55), 87% were female, and histology was predominantly papillary thyroid carcinoma (99%) with 1% follicular. The median reported nodule size was approximately 1.7 cm (IQR 1.2–2.4). We (three independent medical evaluators, one head neck specialist, blinded) compared the surgical recommendations of the AI and a senior head and neck surgeon against NCCN guidelines (ground truth). The primary outcome was the accuracy of the indicated surgical extent (TL vs. TT). Performance metrics included AUC, Accuracy, Precision, and Recall (Sensitivity) for identifying TL candidates. Results: In the validation analysis, the AI model demonstrated superior concordance with guidelines compared to standard clinical assessment. The AI achieved an Accuracy of 0.894 and an AUC of 0.895, significantly outperforming the clinician (Accuracy 0.532; AUC 0.542). Critical analysis revealed that the clinician had a high rate of unnecessary TT recommendations, with a Recall of only 0.083 (8.3%) for the TL class. In contrast, the AI model achieved a Recall of 0.833 (83.3%) for TL. Conclusions: The Gemini-based system (the first in Brazil to the best of our knowledge) outperformed standard clinical judgment in adhering to NCCN guidelines. The clinician demonstrated a bias toward aggressive surgery, missing over 90% of eligible lobectomy cases. Implementing this AI tool potentially could safely increase the rate of partial thyroidectomies, directly benefiting patients by preserving thyroid function and eliminating the lifelong burden of hormone replacement therapy for a significant proportion of cases. Performance metrics of AI vs. clinician (reference: NCCN). Model Precision Recall (Sensitivity) F1-Score Accuracy AUC AI (Gemini) 0.952 0.833 0.889 0.894 0.895 Clinician 1.000 0.083 0.154 0.532 0.542
Evaluation of the cytostatic properties of berberine derivatives on a cell culture of acute T-cell lymphoblastic leukemia (Jurkat).
e19114 Background: T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematological malignancy that accounts for 25% of all adult cases of T-ALL. Currently, the frequency of disease recurrence and resistance to therapy remains quite high, so there is an urgent need to improve therapy by testing new compounds that may potentially have antitumor activity. Promising compounds include berberine, whose high antitumor activity we previously confirmed on glioblastomas, and its derivatives. The aim of the study was to evaluate the cytotoxic effect of berberine derivatives on Jurkat cell culture in an in vitro experiment. Methods: Jurkat cell culture cells were cultured in a complete RPMI1640 nutrient medium without phenolic red (Gibco, USA) with the addition of 10% FBS (Hyclone, USA), 1% glutamine (Biolot, Russia). A total of 11 compounds were tested: berberine and its derivatives ZbRNN, Rub, NAPh, Rub HCl, 12NO2, EPOX, Mag-2, Mag-1, AkR(1) Fuz, 21A. Dose-response curves were constructed for each compound, measuring the level of metabolically active cells by colorimetric method using tetrazolium salt XTT (2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide). The exposure time was 72 hours. Viability was determined as the ratio of the number of living cells in the experimental wells to the control wells, expressed as a percentage. Based on the data obtained, dose-response curves were constructed and IC50 values were determined using the dcr package of the R programming language. Results: Compounds MAG-1, MAG-2, 21A, ZbRNN, Rub, 12NO2, EPOX, and NAPh demonstrated high activity (IC50 below 1000 nmol/L) against the Jurkat cell line. The IC50 value for the studied substances was: 236.4±33.5 nmol/L, 31.8±9.1 nmol/L, 126.3±27.2 nmol/L, 103.8±13.3 nmol/L, 136.4±18.7 nmol/L, 308.4±39.3 nmol/L, 455.7±54.1 nmol/L and 789.5±26.6 nmol/l, respectively (p<0.05). For compounds Rub HCl and AkR(1) Fuz, the IC50 value exceeded 1000 nmol/l and amounted to 1756.2±99.1 nmol/l and 1435.7±112.8 nmol/L (p<0.05). For berberine, the IC50 value was 447.6±46.1 nmol/L. Conclusions: The results obtained indicate the pronounced cytostatic properties of a number of berberine derivatives, of particular interest are compounds that are close to and superior to berberine in cytostatic properties.
Short-course (3-month) enzalutamide monotherapy in biochemically recurrent prostate cancer (BCR): Preliminary prostate specific membrane antigen tumor volume (PSMA-TV) data.
e17109 Background: Enzalutamide (enza) monotherapy is a treatment option for BCR. Although EMBARK allowed dose interruption after 9 months of enza monotherapy, a previous study (n = 38) at the National Cancer Institute (NCI) evaluated two 3-month intermittent courses of enza without androgen deprivation (ADT). The results indicated that PSA control (disease below PSA baseline) was a median of 308 days with 3 months of enza for course 1 and 273 days for course 2, emphasizing a therapeutic approach focused on treatment free survival. No patients (pts) had progression on CT or bone scan after PSA recovery to baseline PSA after 3 months of enza for up to 2 intermittent courses. (Madan, RA et al., JITC, 2021). Although prostate specific membrane antigen (PSMA) scans are increasingly being used for response assessment in prostate cancer with presumed informative potential, it remains unknown how most therapies including enza monotherapy will impact PSMA-TV. PSMA-TV has been proposed as key PSMA readout (e.g. RECIP 1.0) but has rarely been prospectively evaluated in therapeutic trials to date. Methods: NCT06096870 is a clinical trial (n = 50) at the NCI randomizing BCR pts to either enza monotherapy or enza + PDS01ADC (IL-12 immunocytokine). Pts must have negative CT, bone scan, T > 100, and PSMA+ BCR. All pts are treated with enza for 3 months without ADT. PSMA scans are done at baseline and after 3 months of enza. Upon PSA recovery to baseline, all pts are eligible for another course of 3 months of enza if CT and bone scan remain negative. This analysis evaluates the first group of pts who had enza monotherapy only with paired PSMA before/after enza. Baseline PSMA-TV was compared with PSA response data after 3 months of enza. Results: Among the first 13 enza monotherapy pts, medians were age = 71 years, PSA = 22.5 ng/ml, PSA doubling time = 3.5 months, PSMA-TV = 11.1 milliliters. Median PSA response was -98% (-81% to -100%) for a duration of 259 days (including only 84 days of enza). The median PSMA-TV response was -71.7% (+13.2% to -100%). Of note, among 4 pts with PSA = 0 ng/ml after enza, PSMA-TV = 0 in 2 pts, PSMA-TV = -86% in a 3 rd pt but PSMA-TV = +2.8% in the 4 th pt. Furthermore, the PSMA TV increase in 4 th pt did not predict poor clinical response and that pt even had longer PSA control than a pt with a 100% PSA decline and complete response on PSMA. Conclusions: This is the first data presented evaluating PSMA-TV changes after enza in BCR without ADT. Despite this small data set, lack of concordance of PSA and PSMA responses should highlight a need to collect more data to better understand the clinical utility of PSMA scans in defining therapeutic response in BCR. It further indicates the need for caution in the use of PSMA imaging presumptively to define therapeutic benefit or lack thereof in clinical practice. Clinical trial information: NCT06096870 .
Critical survivorship challenges among breast cancer survivors in a low- and middle-income country: A tertiary cancer center experience from Southern India.
e24136 Background: With improving breast cancer survival, long-term survivorship care needs are increasing steadily. Data from low- and middle-income countries (LMICs) remain limited despite accounting for a substantial proportion of global breast cancer survivors. Understanding context-specific survivorship challenges is essential for effective survivorship care delivery models. This study aimed to identify interim data that is prevalent and potentially modifiable for survivorship challenges among breast cancer survivors at a tertiary cancer centre in Southern India. Methods: This cross-sectional observational study included 56 non-metastatic breast cancer survivors who had completed definitive treatment at least three months before assessment. Survivorship needs were evaluated using a culturally adapted, bilingual, validated 24-item questionnaire assessing physical, emotional, cognitive, sexual, lifestyle, and financial domains on a 5-point frequency scale. Descriptive analyses were performed. Results: The 56 survivors were evaluated at a median follow-up of 31.5 (range 9–117) months from diagnosis, with a median age of 40 (range 27–72) years. Most underwent modified radical mastectomy (89%) and received adjuvant chemotherapy (86%). At assessment, 82% were receiving adjuvant endocrine therapy (median 23 mths). Survivorship challenges were predominantly related to physical recovery and lifestyle disruption. Persistent fatigue and reduced physical activity were reported by 43% of survivors, while musculoskeletal pain and sleep disturbance each affected 41%, clustering as post-treatment physical deconditioning. Financial toxicity was prominent, with 71% reporting financial concerns at least sometimes. Domain-wise analysis showed the burden in financial toxicity and lifestyle impact among 89.2% of survivors, followed by emotional well-being and fear of recurrence (17.6%), cognitive concerns (14.3%), physical symptoms (12.5%), and fatigue or sleep disturbance (10.7%). Sexual and reproductive health concerns were rarely reported, suggesting sociocultural under-reporting. Conclusions: Breast cancer survivors in this LMIC cohort demonstrate a distinct survivorship burden profile dominated by financial toxicity, physical deconditioning and lifestyle disruption with relatively low prevalence of sexual life problems. These findings underscore a mismatch between Western models and survivor priorities in resource-constrained settings. At the institutional level, these results will be used to refine survivorship-focused clinical workflows addressing the most distressing domains during routine oncology follow-up.
Prevalence and financial burden of WHO-defined disability among childhood cancer survivors: A report from the St. Jude Lifetime Cohort study (SJLIFE).
12114 Background: Survivors of childhood cancer are at high risk for chronic health conditions that may result in disability. However, the prevalence and financial burden of disability, defined using the World Health Organization International Classification of Functioning, Disability and Health (WHO ICF), have not been described in this population. Methods: Disability-related items ascertained using clinical and patient-report data from SJLIFE were classified using standardized WHO ICF childhood cancer components, including impairments in body function or structure (strength, neurocognition, vision, hearing, speech, balance, neuropathy, ataxia, hemiparesis, tone, bowel and bladder, flexibility, gait, posture, pain, fatigue), activity limitations (motor function, communication, personal care), and participation restrictions (physical activity, education, independent living, employment, quality of life). Aggregate disability severity scores (ADSS) were calculated by summing item-level indicators (0 = none/mild; 1 = moderate, severe, or life-threatening disability) across 55 items spanning 3 components (38 body impairments, 5 activity limitations, and 12 participation restrictions). Disability was classified using either age- and sex-specific z-scores derived from 815 community controls (0 = ≥ −1.5; 1 = < −1.5) or NCI Common Terminology Criteria for Adverse Events (CTCAE) grades (0 = < 2; 1 = ≥ 2). Logistic regression evaluated associations between disability and financial burden (cancer-related financial impact), adjusting for sex, race, age at evaluation, and cancer treatment exposures (chemotherapy, cranial radiation, amputation). Results: Among 3,549 survivors (mean age at evaluation 33.6 years; range 18.0–68.9; 52.1% male; 80.7% White non-Hispanic), the most common diagnoses were acute lymphoblastic leukemia (31.9%), central nervous system tumors (14.5%), and Hodgkin lymphoma (10.5%). The point (most recent visit) prevalence (95% CI) of ≥1 item rated as moderate or greater disability was 87.4% (86.3–88.5). Component-specific prevalence was 74.6% (73.2–76.0) for body impairments, 32.9% (31.3–34.4) for activity limitations, and 63.3% (61.7–64.9) for participation restrictions; all were higher than controls (p < 0.001). Higher overall ADSS was associated with greater financial hardship (OR = 2.97; 95% CI 2.05–4.30) with similar associations across all components, including body impairments (OR = 2.67; 95% CI 2.09–3.41), activity limitations (OR = 2.28; 95% CI 1.91–2.72), and participation restrictions (OR = 2.61; 95% CI 2.13–3.21). Conclusions: WHO-defined disability is highly prevalent among adult survivors of childhood cancer and is strongly associated with financial burden. These findings highlight the need for integrated survivorship care that addresses both medical and socioeconomic risk factors.
Assessing cervical cancer screening awareness and community-identified barriers to prevention in Peoria, Tazewell, and Woodford Counties, Illinois.
e22509 Background: Cervical cancer is largely preventable through routine screening and human papillomavirus (HPV) vaccination; however, gaps in prevention and early detection persist in many U.S. communities despite national declines in incidence and mortality. Peoria, Tazewell, and Woodford Counties in Illinois have a higher cervical cancer incidence than the national average (11.6 vs 7.5 per 100,000 women), reflecting missed opportunities for risk reduction and early detection. This study assessed cervical cancer prevention behaviors, knowledge gaps, and community-identified barriers to care to inform targeted, community-engaged prevention strategies. Methods: We conducted a cross-sectional community needs assessment among 180 adults recruited through healthcare facilities, community outreach programs, and digital platforms. Analyses were restricted to age-eligible participants for cervical cancer screening (ages 21–65; N = 163). Self-administered surveys assessed screening history, HPV vaccination status, knowledge of cervical cancer risk factors and screening guidelines, perceived personal risk, access to screening and vaccination resources, barriers to care, and preferences for prevention education. Descriptive statistics were used to characterize prevention gaps and intervention opportunities. Results: Among age-eligible participants, only 52.8% reported receiving cervical cancer screening, while 37.4% reported never being screened; 11.7% of previously screened participants were overdue, with more than five years since their last screening. While 75% of participants were aware of cervical cancer screening tests and 67.8% were aware of the HPV vaccine, only 38.7% reported HPV vaccination, and 19.6% were unsure of their vaccination status. Knowledge gaps were evident, with 21.3% believing screening was only necessary if symptoms occurred and 31.7% unsure of recommended screening intervals. Risk perception was low, with 62.6% not believing they were at risk for cervical cancer and 23.9% unsure. Although most participants reported knowing where to obtain screening or vaccination, barriers to prevention included transportation (16.7%), cost (15.0%), lack of time (13.3%), and fear of results (10%). 71.7% endorsed the need for additional prevention education and preferred in-person community-based or healthcare provider–led formats. Conclusions: There are substantial gaps in screening uptake, HPV vaccination, and risk perception in this high-burden region. These findings highlight actionable opportunities for community-engaged, prevention-focused interventions aimed at improving early detection and reducing preventable cervical cancer risk. This research was made possible with funding provided by The Community Health Advocacy initiative through University of Illinois Chicago and OSF HealthCare.
Phase II study of utidelone plus fruquintinib for the treatment of platinum-resistant recurrent ovarian cancer (FRUTD trial).
5579 Background: Patients (pts) with platinum-resistant ovarian cancer (PROC) have a poor prognosis, and traditional cytotoxic drugs have limitations in efficacy and safety. Therefore, we are exploring the combination of utidelone capsule (UTD2), a novel oral microtubule inhibitor, with fruquintinib capsule (F), a tyrosine kinase inhibitor targeting VEGFR 1,2,3, for the treatment of platinum-resistant recurrent ovarian cancer. Methods: This study is an open-label and Simon two-stage phase II clinical trial, aiming to enroll 35 pts with platinum-resistant recurrent epithelial ovarian cancer, fallopian tube cancer, or primary peritoneal cancer. All pts were high-grade serous ovarian cancer, with prior 1-5 lines of systemic therapy, and no more than 3 lines of subsequent therapy after platinum-resistant recurrence. Subjects received F in combination with UTD2 in 21-day cycles. The dose of F was 5 mg once daily, taken orally from day 1 to day 14 of each cycle. The dose of UTD2 was 60 mg/m²/day once daily, taken orally from day 1 to day 5 of each cycle. Primary endpoint was objective response rate (ORR) per RECIST v1.1. In Stage 1, 14 pts received per protocol treatment, and the study would proceed to Stage 2 (enrolling additional 21 patients) if ≥2 objective responses (CR/PR) were observed in Stage 1. Secondary endpoints included investigator-assessed progression-free survival (PFS), disease control rate (DCR), duration of response (DoR), overall survival (OS) and Safety. Results: From March 18, 2025 to December 22, 2025, 19 pts were enrolled. Median age was 59 years (range, 44-68) with an ECOG PS score of 1. Median prior lines systemic therapy was 3 (range, 1-5) with 1, 2, 3, 4 and 5 lines for 4 pts (21.1%), 2 pts (10.5%), 6 pts (31.6%), 5 pts (26.3%), and 2 pts (10.5%), respectively. Median number of chemotherapy lines after platinum resistance was 1 (range, 0-3), with 0, 1, 2, or 3 lines of chemotherapy for 6 pts (31.6%), 6 pts (31.6%), 5 pts (26.3%), and 2 pts (10.5%), respectively. 14 pts were evaluated for efficacy with an outcome of 1 complete response, 8 partial responses and 5 stable diseases. ORR was 64.3% (95% CI: 38.8-83.7) and DCR was 100%. Median PFS was 7 months (95% CI: 2.8-7.2) and median OS has not reached. The Grade 3 treatment-related AE (TRAE) included neutropenia (n=2), mucositis oral (n=1), palmar-plantar erythrodysesthesia syndrome (n=1), skin ulceration (n=1), diarrhea (n=1), pain (n=1) and neurotoxicity (n=1). There were no ≥Grade 4 TRAE. 1 patient discontinued UTD2+F due to TRAEs. Conclusions: Utidelone plus Fruquintinib demonstrated encouraging efficacy for the treatment of PROC with a tolerable safety profile. This study is still actively ongoing, and further data will be provided at time of presentation. Clinical trial information: NCT06973421 .
Phase I/II trial of anti-CLDN18.2/PD-L1 recombinant humanized bispecific antibody Q-1802 plus XELOX in treatment-naive CLDN18.2-positive advanced GC/GEJ.
4036 Background: First-line therapy for HER2-negative advanced GC/GEJ remains a huge unmet clinical need. Claudin 18.2 (CLDN18.2), a tight junction protein, is specifically expressed in normal gastric mucosa but aberrantly overexpressed in various cancers, making it a high-value therapeutic target. Q-1802 is a first-in-class humanized bispecific antibody targeting both CLDN18.2 and PD-L1. Q-1802 can bind CLDN18.2 on tumor cells and kill them through Fc mediating antibody-dependent cellular cytotoxicity (ADCC) and phagocytosis (ADCP). Meanwhile, it can activate immune system to eliminate tumor cells by blocking PD-1/PD-L1 binding. Methods: Qure-1802-201 was a multicenter, open-label, nonrandomized phase Ⅰ/Ⅱ trial. Eligible pts: CLDN18.2-positive (≥40% tumor cells with 2+/3+ membranous staining), HER2-negative, treatment-naïve, histologically confirmed unresectable locally advanced/metastatic GC/GEJ. Pts received Q-1802 (10 mg/kg or 20 mg/kg Q2W) plus standard XELOX (oxaliplatin IV d1; capecitabine PO d1-14 Q3W). Primary endpoints: DLT/MTD (Phase Ⅰ) and ORR per RECIST v1.1 (Phase Ⅱ). Secondary endpoints: efficacy, safety, pharmacokinetics, pharmacodynamics, immunogenicity. Results: As of Dec 11, 2025, 62 pts were enrolled (45 in 10 mg/kg, 17 in 20 mg/kg). No DLT observed; MTD not reached. All pts had TEAEs and Q-1802-related TEAEs; the latter mainly included nausea (75.8%), AST increase (61.3%), nausea (59.7%), ALT increase (51.6%), fatigue (50.0%). Incidences of ≥3 Grade TEAEs and Q-1802-related ≥3 Grade TEAEs: 74.2% and 43.5%, respectively. Most common Q-1802-related ≥3 Grade TEAE: thrombocytopenia (8.1%), followed by neutropenia (6.5%), anemia, WBC decrease, hypokalemia (4.8% each). Rate of Q-1802 permanent discontinuation due to TEAEs: 6.5%; no treatment-related death. ORR and median progression-free survival (mPFS) in 60 efficacy-evaluable pts were 70.0% (42/60; 95% CI: 56.8%–81.2%) and 11.3 months (95% CI: 8.0–16.8). A correlated trend was observed between efficacy and CLDN18.2 expression. In the 10 mg/kg cohort, ORR and mPFS were 73.0% (27/37; 95% CI: 55.9%–86.2%) and 11.3 months (95% CI: 7.1–16.8) in pts with CLDN18.2 high expression (2+/3+≥70%, N = 37), which were improved to 81.8% (9/11; 95% CI: 48.2%–97.7%) and 12.2 months (95% CI: 2.7–NE) when CLDN18.2 high expression companied with PD-L1 CPS≥5 (N = 11). DCR was nearly complete (non-response in 1 of 60 pts). ORR in the 20 mg/kg cohort (70.6%, 12/17; 95% CI: 44.0%–89.7%) was similar to 10 mg/kg (69.8%, 30/43; 95% CI: 53.9%–82.8%.). Overall survival (OS) in all pts and PFS in the 20 mg/kg cohort are under follow-up. Conclusions: Q-1802 plus XELOX has manageable safety and promising antitumor activity as first-line therapy for CLDN18.2-positive/HER2-negative advanced GC/GEJ, supporting a phase III trial (Q-1802 10 mg/kg as recommended dose). Clinical trial information: NCT05964543 .
Are socio-economic status indicators barriers for enrollment in phase 1 clinical trials?
1652 Background: Socio-economic Status (SES) indicators create barriers to access Phase 1 clinical trial (Ph1) enrollment, particularly for underrepresented minorities (URM). However, it remains unclear whether they directly affect enrollment after screening, or represent upstream barriers. We evaluated whether SES indicators impacted the enrollment rate in Ph1. Methods: All patients were screened for Ph1 at Hackensack University Medical Center (2014-2025). Patient ZIP codes were linked to the U.S. Census data (ACS 2018-2022) to derive neighborhood-level SES indicators: median household income (MHI), educational attainment (Bachelor), poverty rate (PR), unemployment rate (UR), and median home value (MHV). SES indicators were compared to national and state census benchmarks. The primary outcome was enrollment rate by SES indicators. We performed univariate analyses (Mann-Whitney U tests, chi-square), stratified analyses by race/ethnicity and gender, temporal trend analysis, and multivariable logistic regression. Results: A total of 1,316 patients were screened for 140 Ph1. Demographics were non-Hispanic White (NHW) 66.2%, Hispanic 19.0%, African American/Black (AA) 7.8%, Asian 6.5%. No SES variable predicted enrollment: median income (enrolled $62,646 vs not-enrolled $61,820, p=0.450), education (38.3% vs 37.9%, p=0.598), poverty (12.9% vs 12.3%, p=0.416), unemployment (6.1% vs 6.0%, p=0.422), home value ($277,461 vs $275,697, p=0.605). SES quartile analysis showed no enrollment gradient (Q1=61.7%, Q2=58.1%, Q3=60.5%, Q4=63.8%; p=0.491). Findings were consistent across all racial/ethnic groups (all p>0.05), see Table 1, gender (all p>0.05), age groups <65 and ≥65 years (all p>0.05), and time periods 2014-2019 vs 2020-2025. When compared to their respective racial/ethnic groups U.S. Census averages, it demonstrated atypical SES profiles: Hispanic patients had higher education (36.6% vs 20.6%), AA showed higher SES across all indicators, and Asians showed lower SES, suggesting regional selection bias in who reaches Ph1 screening. Conclusions: When patients reach Ph1 screening, SES indicators do not predict enrollment, demonstrating that equitable enrollment practices can be achieved at screening. Although there were no visible SES-related barriers at the enrollment point, disparities exist among groups likely suggesting referral patterns prior to screening. Future research should examine how to improve access to Ph1 in URM and upstream inequities. SES analysis stratified by race/ethnicity. Race SES Variable p-value NHW MHI 0.4476 NHW Education 0.4585 NHW PR 0.7483 NHW UR 0.6957 NHW MHV 0.4982 Hispanic MHI 0.8270 Hispanic Education 0.4104 Hispanic PR 0.0646 Hispanic UR 0.2401 Hispanic MHV 0.9677 AA MHI 0.9569 AA Education 0.7355 AA PR 0.9112 AA UR 0.3962 AA MHV 0.3076 Asian MHI 0.2037 Asian Education 0.3200 Asian PR 0.2067 Asian UR 0.0867 Asian MHV 0.1677
Consolidation PD-(L)1 inhibitor after definitive chemoradiotherapy for unresectable stage III NSCLC: Systematic review and meta-analysis of randomized controlled trials.
e20054 Background: Consolidation immune checkpoint inhibition after definitive chemoradiotherapy (CRT) improves outcomes in unresectable stage III non-small cell lung cancer (NSCLC). With newer randomized evidence, we evaluated the efficacy and safety of consolidation PD-(L)1 inhibitors versus placebo after definitive CRT. Methods: PubMed, Embase, Cochrane Library, and ClinicalTrials.gov were searched for randomized controlled trials enrolling adults with unresectable stage III NSCLC without progression after definitive CRT and randomized to consolidation PD-(L)1 inhibitor versus placebo. PFS and OS were pooled as hazard ratios (HRs) and discontinuation due to adverse events (AEs) as risk ratios (RRs) using DerSimonian-Laird random-effects; heterogeneity was assessed with I². Treatment-related deaths were summarized descriptively. Results: Three RCTs (N = 1,501) were included (PACIFIC, GEMSTONE-301, PACIFIC-5). Consolidation PD-(L)1 inhibition improved PFS (pooled HR 0.63, 95% CI 0.53-0.77; I² = 40.5%). OS favored PD-(L)1 inhibition (pooled HR 0.69, 95% CI 0.52-0.93) but was interim/immature in two trials with substantial heterogeneity (I² = 63%). Discontinuation due to AEs was higher with PD-(L)1 inhibition (pooled RR 1.72, 95% CI 1.23-2.40; I² = 0%). Treatment-related deaths were rare and reported in two trials (8/526 vs 0/260; GEMSTONE-301: 4/255 vs 0/126; PACIFIC-5: 4/271 vs 0/134). Conclusions: Consolidation PD-(L)1 inhibition after definitive CRT significantly improves PFS in unresectable stage III NSCLC, with an OS signal that requires longer follow-up. Discontinuation due to AEs is increased. Longer follow-up is needed to clarify long-term survival outcomes and infrequent safety events. Sugemalimab is approved for NSCLC indications in China and the UK but is not FDA-approved in the US. Trial-level efficacy (HRs) and AE-discontinuation (counts and RRs). Footnote: Time-to-event outcomes pooled as HRs. OS was interim/immature in GEMSTONE-301 and PACIFIC-5. AE-discontinuation uses reported safety-set denominators (PACIFIC-5 efficacy mITT differs from safety set). Treatment-related deaths were not pooled. RCT Agent N (Tx/Placebo) PFS HR (95% CI) OS HR (95% CI) Discontinuation due to AEs (Tx vs placebo) RR (95% CI) PACIFIC Durvalumab 476/237 0.55 (0.45-0.68) 0.72 (0.59-0.89) 73/475 vs 23/234 1.56 (1.01-2.43) GEMSTONE-301 Sugemalimab 255/126 0.64 (0.48-0.85) 0.44 (0.27-0.73) 29/255 vs 6/126 2.39 (1.02-5.60) PACIFIC-5 Durvalumab 252/129 (mITT); 271/134 (safety) 0.75 (0.58-0.99) 0.87 (0.66-1.17) 39/271 vs 11/134 1.75 (0.93-3.31) Pooled (DL random-effects) - - 0.63 (0.53-0.77) 0.69 (0.52-0.93) - 1.72 (1.23–2.40); I²=0%
Progression-free survival outcomes of first-line maintenance therapy in ovarian cancer stratified by <i>BRCA</i> and homologous recombination status: A multi-center real-world analysis.
e17596 Background: Poly (ADP-ribose) polymerase (PARP) inhibitors are widely used in ovarian cancer; however, real-world data on survival outcomes by BRCA and homologous recombination (HR) status are limited. We assessed progression-free survival (PFS) across different maintenance agents stratified by BRCA mutation and HR status. Methods: A retrospective analysis was conducted using medical records of patients diagnosed with epithelial ovarian cancer between January 2019 and February 2025 at four tertiary hospitals in South Korea. Patients were categorized into three groups: BRCA -mutated, BRCA wild-type with HR-deficiency (HRd), and BRCA wild-type with HR-proficiency (HRp). In the BRCA -mutated group, PFS was evaluated among patients who received first-line maintenance therapy with niraparib, olaparib, or olaparib plus bevacizumab. In the HRd group, patients treated with bevacizumab, niraparib, or olaparib plus bevacizumab were analyzed, while in the HRp group, patients who received bevacizumab alone or niraparib as first-line maintenance therapy were included. Kaplan–Meier analyses were performed separately within each group to assess the impact of maintenance strategies on PFS. Results: A total of 397 patients were included ( BRCA -mutated, n = 200; HRd, n =126; HRp, n =71). Overall, initial PFS comparisons between maintenance and no-maintenance group were conducted within each subgroup. Maintenance therapy significantly improved PFS in both the BRCA -mutated group (median PFS, 58.8 vs. 23.5 months; p < 0.0001) and the HRd group (26.9 vs. 14.1 months; p < 0.0001), while a nonsignificant trend was observed in the HRp group (16.5 vs. 12.1 months; p = 0.109). Subgroup analyses were subsequently performed to evaluate PFS according to the agent of maintenance therapy. In the BRCA -mutated group, no statistically significant difference in PFS was observed among patients treated with niraparib, olaparib, or olaparib plus bevacizumab (51.9 months vs. not reached vs. not reached; p =0.756). In the HRd group, niraparib and olaparib plus bevacizumab were associated with significantly longer PFS compared with Bevacizumab monotherapy (33.7 and 33.1 vs. 15.9 months; p <0.001), while no significant difference in PFS was observed between niraparib and olaparib plus bevacizumab ( p =0.433). In the HRp group, no significant differences in PFS were observed among patients treated with bevacizumab, niraparib, or no maintenance therapy (18.7 vs. 15.6 vs. 12.1 months; p =0.263). Conclusions: In this real-world analysis, PARP inhibitor maintenance significantly improved PFS in patients with BRCA-mutated and HR-deficient ovarian cancer, with consistent benefit observed across different agents and regimens. In contrast, maintenance therapy provided limited benefit in patients with HR-proficient ovarian cancer.
CheckMate-9DW 4-year follow-up: Overall survival by depth of response and outcomes in long-term survivors.
4104 Background: Nivolumab plus ipilimumab (NIVO + IPI) was globally approved as a first-line (1L) treatment for unresectable hepatocellular carcinoma (HCC) based on the phase 3 CheckMate 9DW trial (NCT04039607). We report 4-year follow-up results and overall survival (OS) by depth of response (DpR). Methods: Methods were reported previously (Yau T. Lancet 2025). Briefly, adults with previously untreated unresectable HCC were randomized 1:1 to receive NIVO + IPI (then NIVO for ≤ 2 years) or lenvatinib/sorafenib (LEN/SOR). OS analyses were done by best percentage change in tumor burden from baseline. We also present exploratory analyses on long-term survivors (survival ≥ 4 years after randomization). Results: A total of 668 patients (pts) were randomized to NIVO + IPI (n = 335) or LEN/SOR (n = 333); among 325 pts treated with LEN/SOR, 275 (85%) received LEN. At a median (range) follow-up of 52.5 (44.0–66.1) mo, NIVO + IPI continued to show OS benefit (95% CI) vs LEN/SOR (HR 0.78 [0.65–0.93]) with higher objective response rate (ORR; 95% CI) by blinded independent central review (BICR; 36% [31–42] vs 13% [10–17]) and more durable responses. Pts with deeper tumor reduction had numerically improved median OS on both NIVO + IPI and LEN/SOR; the trend was more pronounced with NIVO + IPI specifically in pts with tumor shrinkage ≥ 30% (Table). More pts treated with NIVO + IPI (n = 62) were long-term survivors vs pts treated with LEN/SOR (n = 33). In long-term survivors, ORR by BICR was higher with NIVO + IPI vs LEN/SOR (76% vs 24%), with higher rates of complete response (CR; 21% vs 6%; Table); median duration of response was not reached in either group. Among long-term survivors, median (range) treatment-free interval was 23.3 (0.1–56.6) mo with NIVO + IPI and 0.5 (0.1–54.0) mo with LEN/SOR; 26 of 62 (42%) and 2 of 33 (6%) patients, respectively, were off study treatment and remained free from subsequent treatment. Baseline characteristics and incidence of TRAEs in long-term survivors were consistent with all randomized pts. Conclusions: 1L NIVO + IPI continued to show sustained efficacy benefit vs LEN/SOR in unresectable HCC with no new safety concerns at the 4-year follow-up. These analyses suggest deeper responses with NIVO + IPI may be associated with improved survival outcomes. Long-term survivors treated with NIVO + IPI had higher ORR and CR rates vs LEN/SOR, and were more likely to remain free of subsequent treatment. These results reinforce NIVO + IPI as a 1L treatment for unresectable HCC. Clinical trial information: NCT04039607 . DpR, a % NIVO + IPI(n = 290) b LEN/SOR(n = 286) b Median OS (95% CI), mo Median OS (95% CI), mo −100 to ≤−60 NR (NE) NR (28.3–NE) −60 to ≤−30 31.1 (19.9–40.4) 20.9 (15.1–30.9) −30 to 20 17.0 (14.8–22.0) 20.5 (17.1–22.9) ≥20 14.3 (6.0–17.5) 7.8 (4.8–19.3) Best overall response in long-term survivors a NIVO + IPI (n = 62) LEN/SOR (n = 33) ORR (95% CI), % 76 (63–86) 24 (11–42) CR, % 21 6 a By BICR. b Response evaluable pts. NE, not estimable; NR, not reached.
Longitudinal changes in coping among early phase cancer clinical trial (EPCT) participants.
11035 Background: EPCT participants use varying strategies to cope with uncertainty related to their cancer, treatment, and prognosis. However, little is known about how coping strategies change over time among EPCT participants and how longitudinal changes correlate with patient-reported outcomes (PROs) and clinical outcomes. Methods: We prospectively enrolled adults with cancer participating in EPCTs at Massachusetts General Hospital from 4/2021-1/2023. Participants completed monthly PROs assessing coping strategies (Brief COPE), symptoms (Edmonton Symptom Assessment System [ESAS]), quality of life (QOL; Functional Assessment of Cancer Therapy General), hope (Herth Hope Index), and financial wellbeing (COST tool). We used regression models to assess associations of baseline (B/L) PROs with changes in coping scores over time (B/L to month 1 [M1] and B/L to month 2 [M2]). We also explored how changes in coping predicted clinical outcomes (time on trial [ToT], overall survival [OS]). Results: We enrolled 195 of 251 eligible patients (78% enrollment), and 188 completed the B/L surveys (96% response, median age=63 [range: 32-89], 56% female, most common cancer types: gastrointestinal [34%] and breast [21%]). Higher B/L QOL predicted decreased behavioral disengagement coping at M1 (B=-0.01, p=.037) & M2 (B=-0.02, p=.001) and self-blame at M2 (B=-0.02, p=.038), as well as increased emotional support (B=0.22, p=.001) and religion (B=0.02, p=.024) coping at M1. Higher B/L ESAS symptoms predicted decreased use of emotional support coping at M1 (B=-0.01, p=.048). Higher B/L ESAS physical symptoms predicted decreased use of emotional support at M1 (B=-0.02, p=.046) & M2 (B=-0.03, p=.016). Higher B/L psychological symptoms predicted increased use of self-blame at M2 (B=0.08, p=.007). Higher B/L financial wellbeing predicted increased emotional support at M1 (B=0.02, p=.025). Higher B/L hope predicted increased positive reframing at M1 (B=0.07, p=.010) and religion coping at M2 (B=0.05, p=.038) as well as decreased behavioral disengagement at M1 (B=-0.04, p=.001) & M2 (B=-0.06, p=.001). For clinical outcomes, increased acceptance (HR=0.85, p=.028) and religion (HR=0.85, p=.016) coping at M1 and increased positive reframing (HR=0.84, p=.010) at M2 predicted longer ToT. Increased use of behavioral disengagement at M1 predicted shorter ToT (HR=1.42, p=.002). Increased use of positive reframing (HR=0.83, p=.009) and self-blame (HR=0.78, p=.012) coping at M1 predicted better OS; increased use of behavioral disengagement (HR=1.41, p=.007) predicted worse OS. Conclusions: In this longitudinal cohort study, EPCT participants’ B/L PROs correlated with changes in their coping over time. We also found longitudinal changes in coping predicted ToT and OS. These findings highlight opportunities to enhance care delivery and outcomes for EPCT participants by addressing their PROs and coping behavior over time.
Neoadjuvant versus adjuvant chemotherapy in cardiac tumor resection: A propensity-matched analysis of perioperative and long-term outcomes.
11555 Background: Optimal timing of systemic therapy for primary cardiac malignancies is undefined. We compared outcomes after surgical resection when chemotherapy was administered Neoadjuvant (≤4 months before surgery) compared to Adjuvant (≤4 months after). Methods: We conducted a retrospective cohort study using the TriNetX database of patients undergoing cardiac tumor resection who received chemotherapy either Neoadjuvant (≤4 months before surgery) or Adjuvant (≤4 months after). Patients were matched 1:1 using propensity scores based on demographics, comorbidities, and treatment variables, yielding 220 patients per group. Primary outcomes were all-cause mortality, infection, and transfusion within 30 days of surgery. Secondary outcomes included tumor recurrence at 1 year (90–365 days), all-cause mortality, and new-onset heart failure during 1–3 years of follow-up. Time-to-event analyses were performed using Kaplan–Meier methods with log-rank testing and hazard ratios (HRs); absolute and relative risks are reported as risk differences (RD) and risk ratios (RR). Results: After 1:1 propensity-score matching, 220 patients were included in each group. In the perioperative period (1–30 days), all-cause mortality occurred in 18/217 patients (8.3%) in the Neoadjuvant chemotherapy group compared with 30/213 (14.1%) in the Adjuvant group (RD −0.058, 95% CI −0.117 to 0.002; p=0.057), with a corresponding HR of 0.57 (95% CI 0.32–1.02; p=0.056). Infectious complications were lower with Neoadjuvant chemotherapy (12/220, 5.5%) than with Adjuvant chemotherapy (21/220, 9.5%) (RD −0.041, 95% CI −0.090 to 0.008; p=0.103; HR 0.54, 95% CI 0.27–1.10; p=0.090), while transfusion requirements were similar between groups (14.1% vs 12.3%; p=0.57). At 1 year (90–365 days), malignant cardiac tumor recurrence occurred in 27.7% of Neoadjuvant patients versus 21.4% of Adjuvant patients (RD +0.064, 95% CI −0.017 to 0.144; p=0.12; HR 1.16, 95% CI 0.80–1.70). During longer-term follow-up (1–3 years), all-cause mortality was comparable between groups (44.7% vs 47.4%), with longer median survival observed in the Neoadjuvant group (791 vs 516 days; HR 0.81, 95% CI 0.62–1.07; p=0.14). New-onset heart failure occurred at similar rates in both groups (p=0.98). Conclusions: In this cohort study of cardiac tumor resections, Neoadjuvant chemotherapy showed lower 30-day mortality and infection than Adjuvant chemotherapy, without significant differences in transfusion, 1-year recurrence, long-term mortality, or new-onset HF. These hypothesis generating findings suggest neoadjuvant chemotherapy is not associated with worse early or late outcomes and warrant confirmation in prospective, multi-center studies with histology-specific stratification and standardized systemic regimens.
MELODY: A prospective non-interventional multicenter cohort study to evaluate different imaging-guided methods for localization of malignant breast lesions (EUBREAST-4/iBRA-NET/AGO-B-062; NCT05559411).
TPS652 Background: In the last decades, the proportion of breast cancer patients receiving breast-conserving surgery has increased, reaching 70-80% in developed countries. In case of non-palpable lesions, surgical excision requires some form of breast localization. While wire-guided localization has long been considered gold standard, it carries several limitations, including logistical difficulties, the potential for displacement and patient discomfort, and re-excision rates reaching 21% (in DCIS up to 30%). Other techniques (radioactive seed or radio-occult lesion localization, intraoperative ultrasound, magnetic, radiofrequency, and radar localization) have been developed with the aim of overcoming these disadvantages. However, comparative data on the rates of successful lesion removal, negative margins, and re-operations are limited. In most studies, the patient perspective, addressing e.g. discomfort and pain, has not been evaluated. The aim of MELODY (MEthods for LOcalization of Different types of breast lesions) is to evaluate different imaging-guided localization methods with regard to oncological safety, patient-reported outcomes, surgeon and radiologist satisfaction and economic impact. Methods: The EUBREAST and the iBRA-NET have initiated the MELODY study to assess breast localization techniques and devices from several perspectives (NCT05559411, http://eubreast.org/melody). MELODY is a prospective intergroup cohort study which enrolls female and male patients planned for breast-conserving surgery with imaging-guided localization for invasive breast cancer or DCIS. Multiple or bilateral lesions and neoadjuvant chemotherapy are allowed. Primary outcomes are: 1) Intended target lesion and/or marker removal, independent of margin status on final histopathology, and 2) Negative resection margin rates at first surgery. Secondary outcomes are, among others: rates of second surgery and secondary mastectomy, Resection Ratio (defined as actual resection volume divided by the calculated optimum specimen volume), duration of surgery, marker dislocation rates, rates of marker placement or localization failure, patient-reported outcomes, rates of “lost markers”, radiologist and surgeon satisfaction, and health economic evaluation of the different techniques. Enrollment started in January 2023. Up to 26 January 2025, 9727 patients from 27 countries were enrolled in the study. The study is expected to complete patient enrollment in year 2026. The study will be conducted in 30 countries and is supported by the Oncoplastic Breast Consortium (OPBC), AWOgyn, AGO-B, SENATURK, the American Society of Breast Surgeons (ASBS) and the Korean Breast Cancer Study Group (KBCSG). Clinical trial information: NCT05559411 .
TrioMBM: A multicenter, phase II trial of relatlimab, nivolumab, and ipilimumab in patients with asymptomatic and symptomatic melanoma brain metastases.
TPS9604 Background: Melanoma brain metastases (MBM) are a leading cause of morbidity and mortality for patients (pts) with advanced melanoma. Modern systemic therapies including immune checkpoint blockade (ICB) are insufficient at controlling MBM, particularly in patients with who are symptomatic and/or require corticosteroids. The CheckMate 204 study established nivolumab (NIVO) 1 mg/kg + ipilimumab (IPI) 3mg/kg followed by maintenance NIVO as the standard treatment (tx) for pts with MBM based on intracranial IC) objective response rate (ORR) of 55% and 6-month progression-free survival (PFS) of 64% in the asymptomatic cohort ( Tawbi et al. NEJM 2018 ). In symptomatic pts, intracranial benefit rate (iCBR) was only 22% and median PFS was 1.2 mos ( Tawbi et al. Neuro Oncol 2021 ). Triplet tx with the addition of LAG-3 inhibition with relatlimab (RELA) is now being explored as a mechanistically-driven means to improve efficacy of ICB. Arm 2B of RELATIVITY-048 (NIVO 480 mg Q4W + RELA 160 mg Q4W + IPI 1 mg/kg Q8W) reported promising efficacy (ORR 59%, 3y PFS 52%) and 2 of 3 (67%) pts with MBM had confirmed responses ( Ascierto et al. ASCO 2024 ) further justifying the need to explore this combination in this population. We hypothesize that triplet ICB will be more effective than NIVO+IPI for the tx of symptomatic and asymptomatic MBM. Methods: In this multicenter, investigator initiated, phase II study evaluating triplet ICB (NCT06712927) pts with MBM will be assigned to one of two cohorts: Cohort A (asymptomatic, n=40) and Cohort B (symptomatic, n=20) defined as having symptoms attributable to brain metastases and/or requiring steroids >10 mg prednisone/day or equivalent. All pts will receive NIVO 480mg + RELA 160mg IV Q4W (FDA approved dose) plus IPI 1mg/kg IV Q8W. Pts must have ≥1 untreated intracranial lesion (5-40mm) and no prior ICB for unresectable stage III/IV melanoma (>6 mo since neoadjuvant or adjuvant ICB). Primary endpoint is iCBR (CR+PR+SD ≥6 mo) determined by modified RECIST 1.1; secondary endpoints include safety, ORR, duration of response, and PFS. Extensive correlative analyses of peripheral blood mononuclear cells, plasma, tumors, and cerebrospinal fluid are planned. The study is currently enrolling at Stanford Cancer Center with multisite expansion underway. Clinical trial information: NCT06712927 .
Machine learning modeling for predicting 60-day prognosis in newly diagnosed advanced solid tumors.
1622 Background: Accurate mid-term prognosis estimation in patients newly diagnosed with advanced, non-operable or metastatic cancer is critical for evaluation and treatment prioritization and palliative care planning. A 60-day time horizon is clinically meaningful, as major oncologic assessments typically occur at 2–3-month intervals. Existing prognostic tools have shown limited clinical adoption. Methods: We conducted a retrospective analysis of patients evaluated at the Sheba Medical Center Oncology Rapid Diagnosis Clinic between 2017 and 2025. Eligible patients had metastatic or locally advanced, non-operable solid tumors. A raw dataset was generated using the MDClone platform (overall 60-day mortality ~18%). A case-enriched training cohort (n=524; 39.3% 60-day mortality) and an independent test cohort (n=343; 19% 60-day mortality) were randomly selected. Seventy-four features were included, encompassing demographics, laboratory values, ECOG performance status, and clinical variables extracted from physician notes using a large language model (LLM). LLM extraction accuracy was validated against expert manual review in 32 randomly selected cases (accuracy 0.91–1.0). Logistic Regression, Random Forest, and XGBoost models were trained with grid-search hyperparameter optimization to predict 60-day survival (y=1) versus death (y=0). Performance was evaluated on the independent test set. Wilson score intervals were used for proportion confidence intervals, and Hanley–McNeil method for AUC confidence intervals. Feature importance for the top-performing model was assessed using SHAP. Results: Performance of different ML models is summarized in table 1. Random Forest demonstrated the best overall performance. All models showed high positive predictive value for survival but limited ability to predict short-term mortality. The most influential Random Forest features included INR, CRP, albumin, ECOG status, total protein, LDH, emergency-related hospitalizations, age, neutrophil count, and history of cardiovascular disease. Conclusions: A Random Forest–based model accurately identifies patients likely to survive 60 days, supporting timely evaluation and initiation of oncologic treatment. Prediction of short-term mortality remains limited, suggesting intrinsic unpredictability of mid-term outcomes or the need for additional clinical or biological features. Metric (95%CI) Random Forest Logistic Regression XGBoost ROC AUC 0.84 (0.79-0.88) 0.81 (0.76-0.86) 0.80 (0.75-0.85) Sensitivity 0.89 (0.85-0.92) 0.82 (0.77-0.86) 0.80 (0.75-0.84) Specificity 0.62 (0.49-0.72) 0.62 (0.49-0.72) 0.63 (0.51-0.74) PPV 0.91 (0.87-0.94) 0.90 (0.86-0.93) 0.90 (0.86-0.93) NPV 0.57 (0.45-0.68) 0.45 (0.35-0.55) 0.43 (0.33-0.53) Accuracy 0.84 (0.80-0.87) 0.78 (0.74-0.82) 0.77 (0.72-0.81)
Artificial intelligence predicted natural killer cell fitness to guide daratumumab and transplant strategies in newly diagnosed multiple myeloma.
7520 Background: Daratumumab (Dara) requires CD16-expressing natural killer (NK) cells for anti-myeloma activity. The use of bone marrow transplant (ASCT) results in prolonged immunosuppression, including lower levels of NK-cells. We evaluated if artificial intelligence (AI) predicted CD16 from routine hematoxylin and eosin (H&E) stained bone marrow biopsy slides could be used to guide Dara and transplant treatment strategies. Methods: We analyzed 212 newly diagnosed patients from the HealthTree registry treated with VRd (n=135; median follow-up 24.5 months (mo)) and D-VRd (n=77; 16.6 mo). H&E slides were processed using GigaTIME, a foundation model via zero-shot inference. Patients were dichotomized at median predicted CD16 and stratified by ASCT status. Subgroups: transplant-eligible (TE: age<65, ASCT; n=53), deferred (TD: age<70, no ASCT; n=34), and ineligible (TI: age≥70; n=17). Primary endpoint: time to next treatment (TTNT). Multivariate Cox regression adjusted for age, cytogenetic risk tested interaction between CD16, treatment and ASCT. Results: D-VRd achieved superior outcomes vs VRd (23.4% event rate vs 74.8%, p<0.001) despite lower ASCT use among D-VRd treated patients (50.6% vs 71.1%, p=0.005). In the VRd group, AI-predicted CD16 was highly prognostic in non-transplanted patients: High-CD16 achieved median TTNT 33.1 vs 5.1 mo for Low-CD16 (p=0.0053). ASCT rescued Low-CD16 patients (median 5.1 to 26.7 mo, p<0.001). Conversely, in D-VRd, CD16 did not stratify outcomes (p=0.48), indicating that Dara can rescue low-immune-fitness disease. Specifically, Dara rescued low CD16 non-transplanted patients from progression (86.8% 18-mo event-free survival for D-VRd versus 28.6% 18-mo event-free survival for VRd; p<0.001). High CD16 patients on D-VRd without ASCT achieved 86.3% event-free survival, non-inferior to ASCT (82.2%, p=0.21), supporting transplant deferral. The CD16×treatment×ASCT interaction was significant (HR 0.06, p=0.045). Among transplant-deferred patients, D-VRd improved 18-mo event-free rates vs VRd (70.5% vs 40%, p=0.029), with the largest benefit in Low CD16 (p=0.016). TI patients on D-VRd achieved 100% event-free at 18 mo vs 38.2% on VRd (p=0.011). Conclusions: We validated that AI-quantified CD16 from routine H&E guides Dara and transplant strategies in NDMM (n=212). The CD16×treatment×ASCT interaction (HR=0.06, p=0.045) confirms that CD16 predicts treatment benefit, not just prognosis. In VRd, low CD16 identifies a high-risk group (median TTNT 5.1 mo) needing intensification. D-VRd rescues low-immune-fitness patients (18-mo event-free survival: 86.8% vs 28.6% with VRd, p<0.001) and enables safe transplant deferral in high-fitness patients (82.2% with ASCT vs 86.3% without, p=0.21). This novel strategy enables precision medicine using standard H&E slides, with immediate clinical implications.
Transplant-associated thrombotic microangiopathy after hematopoietic stem cell transplantation in adults: Incidence and inpatient outcomes in the U.S. National Inpatient Sample, 2016-2022.
e18570 Background: Transplant-associated thrombotic microangiopathy (TA-TMA) is one of the severe and fatal under recognised complications after hematopoietic stem cell transplantation (HSCT). Contemporary population-level estimates of incidence and inpatient outcomes are limited. Methods: We performed a retrospective cohort study using the National Inpatient Sample (NIS), 2016-2022. Adult hospitalizations with HSCT were identified using ICD 10 codes. TA-TMA was defined by ICD 10 codes during the index hospitalization. Survey weights generated national estimates. We compared demographics, length of stay (LOS), and total hospital charges (TOTCHG) by TA-TMA status using survey adjusted methods. Multivariable survey weighted logistic regression assessed associations between TA-TMA and inpatient mortality and complications, adjusting for demographic and clinical covariates. Results: An estimated 344,905 adult HSCT hospitalizations were identified nationally, with TA-TMA occurring in 0.3%. TA-TMA patients were younger (mean 52.2 vs 58.4 years) and more often female (54.3% vs 42.4%; p=0.0005); race distribution differed (p=0.0167), with higher proportions of Black and Hispanic patients. TA-TMA was associated with substantially greater resource utilization (LOS 19.7 vs 7.5 days; TOTCHG $424,329 vs $115,770) and higher unadjusted inpatient mortality (12.5% vs 4.7%; p<0.001). Adjusted analyses showed TA-TMA was independently associated with increased inpatient mortality (aOR 2.98, 95% CI 1.94–4.58) and complications, including acute kidney injury (aOR 6.60, 95% CI 4.92–8.86), renal failure (2.27, 1.65–3.12), dialysis (3.23, 2.00–5.21), respiratory failure (2.33, 1.75–3.11), liver injury (2.06, 1.32–3.21), acute GVHD (6.60, 3.74–11.64), and chronic GVHD (3.78, 2.46–5.80). Conclusions: In a national cohort of adult HSCT hospitalizations, TA-TMA was uncommon but carried substantial excess mortality, multi organ complications, and markedly higher LOS and hospital charge. These findings underscore the need for heightened vigilance, early recognition, and standardized risk mitigation strategies for TA-TMA in HSCT recipients, such as risk stratification, aggressive management of triggers, and risk based therapeutic interventions. Outcome/ Metric No TA-TMA TA-TMA Adjusted OR (95% CI) Inpatient Mortality 4.7% 12.5% 2.98 (1.94-4.58) Length of Stay, Days (mean) 7.5 19.7 Acute Kidney Injury 24.7% 65.4% 6.60 (4.92-8.86) Renal Failure 23.5% 35.1% 2.27 (1.65-3.12) Dialysis 3.3% 9.6% 3.23 (2.00-5.21) Respiratory Failure 15.8% 27.9% 2.33 (1.75-3.11) Liver Injury 6.2% 12.0% 2.06 (1.32-3.21) Acute GVHD 1.0% 7.2% 6.60 (3.74-11.64) Chronic GVHD 3.9% 14.9% 3.78 (2.46-5.80) Percentages are survey weighted. Adjusted odds ratios (aORs) from survey weighted multivariable logistic regression.
Incidence and risk factors for pneumonitis in patients with urothelial carcinoma treated with enfortumab vedotin plus pembrolizumab: A retrospective cohort study.
e16600 Background: Enfortumab vedotin plus pembrolizumab (EV+P) is first line standard of care for locally advanced or metastatic urothelial carcinoma (LA/mUC). Both agents carry pneumonitis risk, yet real-world incidence and predictors of pneumonitis with this combination remain poorly characterized. Methods: We conducted a retrospective cohort study of patients receiving EV+P for LA/mUC at MD Anderson Cancer Center. Pneumonitis was adjudicated by a pulmonologist and graded per CTCAE v5.0. Given the high risk for mortality from other causes, cumulative incidence was estimated using Fine-Gray competing risk analysis with death as competing event. Risk factors were evaluated using univariable subdistribution hazard regression. Results: Among 264 patients (median age 72 years; 76% male; 67% metastatic; median follow-up 10.3 months), pneumonitis developed in 17 (6.4%; 95% CI, 4.1-10.1%): Grade 1-2, 9 (53%); Grade ≥3, 8 (47%). Median time to onset was 69 days (IQR, 48-82), with 82% of events occurring within 90 days. During follow-up, 70 deaths occurred; the 90-day cumulative incidence was 5.4% (95% CI, 2.5-8.3%). One patient (6%) required ICU admission; none required mechanical ventilation. Corticosteroids were administered in 7 (41%). In competing risk regression, baseline interstitial lung abnormalities (ILA; sHR 2.99; 95% CI, 0.84-10.66; P = 0.09), prior ICI therapy (sHR 2.46; 95% CI, 0.81-7.44; P = 0.11), and COPD (sHR 2.36; 95% CI, 0.80-6.96; P = 0.12) showed the strongest signals, though statistical significance was not reached (Table). Pneumonitis incidence was higher in patients with ILA (17.6% vs 5.7% without), and 12.9% with prior ICI vs 5.6% without. In time-dependent Cox analysis, pneumonitis was associated with a trend toward increased mortality (HR 1.67; 95% CI, 0.76–3.69; P = 0.20), though this did not reach statistical significance. Conclusions: Pneumonitis occurred in 6.4% of patients receiving EV+P; however, 47% of events were Grade ≥3 and most occurred within 90 days. Baseline ILA, prior ICI therapy, and COPD showed higher incidence, representing hypothesis-generating signals for risk stratification. Close pulmonary monitoring during early treatment cycles may be warranted in patients with these features. Risk factors for pneumonitis (fine-gray subdistribution hazard model). Variable sHR (95% CI) P-value Age (per 10 years) 0.88 (0.66-1.17) 0.39 Male sex 0.57 (0.21-1.55) 0.27 Ever smoker 1.18 (0.44-3.17) 0.75 Interstitial lung abnormalities 2.99 (0.84-10.66) 0.09 Prior ICI therapy 2.46 (0.81-7.44) 0.11 COPD 2.36 (0.80-6.96) 0.12 Prior therapy lines (per line) 1.16 (0.99-1.35) 0.07 ECOG ≥2 0.63 (0.15-2.73) 0.54 Abbreviations: CI, confidence interval; COPD, chronic obstructive pulmonary disease; ECOG, Eastern Cooperative Oncology Group; ICI, immune checkpoint inhibitor; sHR, subdistribution hazard ratio.