Browse Articles
Discover research articles across all indexed journals
Application of locally developed referral criteria for palliative care in metastatic hormone receptor–positive breast cancer patients: A non-randomized controlled trial.
e24080 Background: Persons with advanced cancer are often referred to palliative care in the late stages of their disease. This trial aimed to evaluate the impact of palliative care consultation based on locally developed referral criteria, compared to usual care among patients with metastatic hormone-positive breast cancer. Methods: This nonrandomized controlled trial used a nonequivalent pre–post control-group design. A total of 107 women with stage IV hormone-positive, HER2-negative breast cancer were included at the Outpatient Breast Cancer Department of University Hospital Centre Zagreb, Croatia. The intervention consisted of consultation with a hospital supportive and palliative care team guided by locally developed referral criteria, while the control group received usual care. Symptom burden was assessed using the Edmonton Symptom Assessment System–revised (ESAS-r). Quality of life using was assessed using the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 15 Palliative (EORTC QLQ-C15-PAL). Assessments were conducted at three time points in both the intervention and control groups. Results: At baseline, the intervention group (N = 52) reported higher symptom burden (Mann–Whitney U = 1097; p = 0.04) and lower quality of life (Mann–Whitney U = 884; p < 0.001) compared with controls (N = 55). Similar differences were observed at 3 months. At 6 months, no significant differences were found between the intervention and control groups in symptom burden (Mann–Whitney U = 1271; p = 0.23) or quality of life (Mann–Whitney U = 1174; p = 0.11). Within-group analyses showed no significant change in symptom burden over six months in either group. Quality of life declined in the intervention group between baseline and 6 months (Wilcoxon signed-rank test = −1.960; p = 0.050), while no significant change was observed in the control group. Conclusions: This study indicates that patients referred to palliative care based on locally developed criteria had significantly higher symptom burden and lower quality of life at baseline, reflecting greater clinical complexity. Over six months, symptom burden remained stable in both groups, and between-group differences were no longer observed. These findings suggest that referral criteria successfully identified patients with greater needs and that early palliative care integration was associated with symptom stability over time, despite poorer baseline status. Clinical trial information: https://osf.io/vp5jr/overview .
Dynamic changes in immune checkpoint and cell populations following lymphodepletion and CAR-T therapy in hematologic malignancies.
7542 Background: Chimeric antigen receptor T-cell (CAR-T) therapy has significantly improved outcomes in relapsed and refractory hematologic malignancies; however, durable remissions remain limited in patients with high-risk features, including early progression to ≥2 extranodal sites, high tumor burden, and elevated serum lactate dehydrogenase. Proposed mechanisms of resistance include CAR-T cell exhaustion, impaired expansion and cytotoxicity, antigen loss or downregulation, and inadequate tumor infiltration. In addition, immunosuppressive interactions among myeloid-derived suppressor cells (MDSCs), regulatory T cells, tumor-associated macrophages, and inhibitory immune checkpoints such as PD-1, PD-L1, LAG-3, TIM-3, and VISTA may further limit CAR-T efficacy. Lymphodepleting chemotherapy alters immune homeostasis by affecting T-, B-, and NK-cell populations, modulating co-stimulatory signaling, reducing regulatory populations, and reshaping cytokine networks, influencing both response and toxicity. Methods: We conducted a pilot study (n = 10) to characterize peripheral immune alterations in patients receiving lymphodepleting chemotherapy followed by CAR-T therapy. Serial peripheral blood samples were collected at predefined time points before and after lymphodepletion and CAR-T infusion (D -5, D0 and D28+/-2). Flow cytometry was used to quantify immune cell subsets and evaluate immune checkpoint expression, including PD-1, PD-L1, LAG-3, and VISTA. Results: Lymphodepleting chemotherapy resulted in a marked reduction in T cells, B cells, and MDSCs. By D28, most immune populations recovered to pre-lymphodepletion levels; however, granulocytic MDSCs remained relatively suppressed. Preliminary molecular analyses demonstrated trends in both VISTA and LAG-3 expressions on T cells in multiple myeloma patients following lymphodepletion, without significant changes in overall circulating immune cell frequencies. Among patients who experienced disease recurrence (n = 3), a trend toward decreased CD45⁺CD3⁺ and CD45⁺CD4⁺ T-cell frequencies on D28 compared with D0 was observed relative to patients achieving complete remission (n = 7). Conclusions: These preliminary data suggest that lymphodepleting chemotherapy and CAR-T therapy induce dynamic peripheral immune remodeling, including sustained suppression of select myeloid populations and differential immune checkpoint expression. Interpretation is limited by the small sample size and pilot nature of the study. Nonetheless, observed trends in post-infusion T-cell dynamics associated with disease recurrence support further investigation. Larger, longitudinal studies are needed to validate these findings, elucidate mechanisms of resistance, and identify immune biomarkers predictive of response, toxicity, and durable remission following CAR-T therapy.
Bytes to blood: Artificial intelligence in leukemia management—A 2025 update.
e18557 Background: Artificial intelligence (AI) has been proposed as a tool to aid in the diagnosis, treatment, and monitoring of leukemias given their genetic complexity, subtype heterogeneity, array of treatments, and need for relapse detection. AI has several potential applications in the management of leukemia. First, it can be used to detect leukemia. Using AI to detect subtle nuances in lab values can ensure these deadly cancers are never missed and that complications, such as disseminated intravascular coagulation, are recognized quickly. Second, AI can be used to risk-stratify patients and personalize treatments. Leukemias are among the most genetically complex cancers and, while they have well-characterized risk profiles, tailoring treatments remains a challenge despite our advances. And third, AI can be used for surveillance ( e.g. , minimal/measurable residual disease [MRD] testing) and relapse management. There is a role for AI in predicting risk of relapse and determining when treatment is indicated for a patient who is MRD + , as not all leukemic cells detected can cause relapse. Using AI to tailor treatment plans for patients who ultimately do experience relapse is an especially intriguing potential role for AI. The objective of this narrative review is to determine the current state of AI utilization in the management of leukemia, specifically in diagnosis, risk stratification, treatment planning, and relapse detection. Methods: PubMed was indexed for articles published between January 2022 and November 2025 using the search term "artificial intelligence" AND "leukemia" together. Initial search identified 355 full-text articles. Upon critical appraisal by the authors, 55 articles were deemed relevant to the scope of our work and included in this review. Generative AI was not used in any aspect of the production of this manuscript. The authors have no financial disclosures. Results: The use of AI in the management of leukemia is largely limited to diagnosis, specifically via interpretation of routine clinical data ( e.g. , blood counts, blood smears, metabolic labs) and ancillary testing ( e.g. , flow cytometry, FISH, sequencing). The use of AI in leukemia risk stratification, treatment planning, and relapse detection is largely unexplored. Conclusions: AI has extraordinary potential to improve several aspects of leukemia management. However, it is currently underutilized and primarily focused on initial diagnosis. The real power of AI lies in optimizing treatment intensity to reduce short- & long-term toxicity and the likelihood of relapse, thereby improving patient outcomes. A. P. Runde and S. M. Koo contributed equally.
Barriers and enablers to pediatric palliative care integration in low- and middle-income countries: A systematic review and meta-synthesis.
10055 Background: Globally around 21 million children require pediatric palliative care yearly of whom more than 97-98% of the children in greatest need live in low- and middle- income countries. Nevertheless, presently, only around 5-10% of children and their families in those situations receive palliative care while the development of pediatric-specific palliative care frameworks and services remains very much less developed than adult palliative care in many health systems. Methods: A systematic search was performed in PubMed database, Embase database and Global Health for studies published between 2015 and 2025 using PRISMA guidelines. Inclusion criteria: Original research on the models of PPC, access to opioids or educational interventions in the context of LMICs. Articles were screened for inclusion by two independent reviewers and were evaluated for quality using the MMAT. Data have been synthesized with the help of a thematic framework. Results: Out of the 427 identified records, 48 studies in 32 countries met the inclusion criteria. Important barriers identified were restrictive legislation, and 68% of the studies mentioned "opiophobia" and legal barriers to access of pediatric morphine, median morphine consumption in LMICs less than 1% of high income countries. While workforce deficit only of 12% of LMIC children's or pediatric doctors (pediatricians) reported some type of formal training in PPC during residency. Also, financial toxicity: End-of-life care financial costs for families in LMIC often outweigh cost up to, or more than, 50% of their yearly income. On the other hand, successful enablers were task-shifting to community health workers and integration of PPC into existing oncology protocols. Quantitative synthesis demonstrated the use of integrated hospital in-home models to decrease emergency room visits by 35% in the last month of life. Conclusions: Large disparities still exist in PPC provision in LMICs, which is related to regulatory and educational gaps. But some localized "hub-and-spoke" models provide an outline of how to integrate in a sustainable fashion. For ASCO's global mission, these findings indicate that "Zero Pain" initiatives require pediatric specific legislative reform and mandatory oncology palliative training to address the gap in quality of survival.
Multicenter phase 2 study of hydroxychloroquine (HCQ) and chlorphenesin carbamate (CPC) in combination with mFOLFIRINOX in patients with advanced pancreatic adenocarcinoma (PDAC).
4189 Background: PDAC is one of the most lethal malignancies, with most patients (pts) presenting with advanced disease. Although combination chemotherapy, including modified FOLFIRINOX (mFOLFIRINOX) has improved survival, overall prognosis remains poor. Our preclinical studies suggest that HCQ and CPC synergistically enhance the antitumor activity of mFOLFIRINOX by inhibiting autophagy-dependent and AKT-mediated epithelial-mesenchymal transition pathways. Methods: This multicenter, single-arm phase 2 study included pts with locally advanced unresectable or metastatic PDAC in the first-line setting from three academic institutions in Korea. HCQ (200 mg) and CPC (250 mg) were administered orally twice daily in combination with mFOLFIRINOX for up to 24 cycles or until disease progression or unacceptable toxicity occurred. The primary endpoint was safety profile. Secondary endpoints included objective response rate (ORR), progression-free survival (PFS), overall survival (OS), cumulative incidence of new distant metastases, and health-related quality of life (HRQOL). Results: Between December 2021 and February 2023, 45 pts were screened and a total of 40 patients received study treatment. The median age was 61 years, and 23 (57.5%) were male; 73% (n = 29) and 27% (n = 11) had metastatic and locally advanced disease, respectively. The median number of cycles patients received was 13 (range, 1-24). Dose reductions for HCQ and CPC occurred in four (10%) pts each, with median > 90% of treatment compliance for both agents. Safety profiles are mostly consistent with known adverse events (AEs) with mFOLFIRINOX; grade 3-4 neutropenia (40%), nausea (13%), anemia (10%) were most common severe AEs. In the efficacy analysis set (n = 39), the ORR was 39% (CR or PR; n = 15), and the disease control rate was 92.3% (CR, PR or SD; n = 36). Curative-intent conversion surgery was performed in five pts (13%). With a median follow-up of 33.3 months, the median PFS was 7.4 months (95% CI, 6.2–13.3); in pts with objective response, median PFS was 13.6 months (95% CI, 7.4–33.8). The median OS was 14.8 months (95% CI, 9.6–27.9); median 20.5 months (95% CI, 14.8–NE) for locally advanced disease and median 13.9 months (95% CI, 9.2-NE) for metastatic disease. The cumulative incidence of new distant metastases was 38.8% at 12 months. HRQOL indicators did not deteriorate during the first 6 months of study treatment. Conclusions: HCQ and CPC in combination of mFOLFIRINOX were well tolerated and showed clinically promising efficacy outcomes in pts with advanced PDAC. Biomarker analysis using proteomics is ongoing. Clinical trial information: NCT05083780 .
Development of a novel blood-based translational PCR screen for monoclonal B-cell lymphocytosis.
e19033 Background: Over 10 million Americans have monoclonal B-cell lymphocytosis (MBL), a precursor to CLL, with age-associated increase in frequency, from ~5% at 45 years to >50% in nonagenarians. Despite significant associations with other malignancies and infectious disease hospitalizations, the natural history of MBL and CLL, and its causal link with allied conditions remains poorly defined. Our health system-wide screening based on routine CBC results identified 112 (10.5%) MBL cases among patients with persistent absolute lymphocytosis (PAL). However, flow cytometry diagnostic for monoclonal B cell expansion were only tested on 1070 out of 5747 PAL patients due to workflow constraints and lack of exposure to the clinical significance of MBL among non-oncology practitioners. Hence, we developed a cost-effective and high-throughput IGHV-D-J PCR screen to identify PAL cases with monoclonal B cells to streamline patients for definitive flow cytometry testing. Methods: Whole blood PAL samples were subjected to RBC lysis, RNA extraction, and reverse transcribed. Adapted from our published unique molecular identifier (UMI)-based methodology, linear amplification of the IGHV regions were performed with UMI-tagged primers to ensure amplification accuracy. Next, the NGS library was generated by semi-nested PCR using IGHV and IGHC primers and sequenced on the Illumina Miseq platform with depth of 40x based on estimated lymphocyte counts. Raw reads were processed and analyzed using a custom workflow built with MiXCR. Results: As research-based clonality testing is challenging for a system-wide screening test, we optimized the PCR strategies using varying cell counts expected for blood volume of CBC tests. No differences in the generation of IGVH-D-J libraries and average product sizes (~600bp) were observed between whole blood and FACS-sorted samples. To determine the correlation of monoclonal B cell percentage and UMI-based detection of unique IGVH-D-J rearrangements, healthy white blood cell were mixed with 0.5 - 12.5% of CLL B cells of known clonotype and clonality data by flow cytometry and our PCR screen were compared. This indicated a significant correlation between CLL-specific IGHV-D-J rearrangements and the percentage of spiked CD5 + CD19 + CLL cells determined by flow cytometry (R 2 = 1.0; P <0.001). Most importantly, we successfully identified the CLL-specific clones with sufficient sensitivity for all MBL counts. Conclusions: Despite reported associations of MBL with hematological malignancies and other conditions, large scale prospective studies remain challenging. To this end, we have developed an efficient semi-nested PCR test that identifies expanded monoclonal B cells using residual whole blood after routine CBC testing. With the catchment area of Northwell Health, this screening test will provide a crucial stepping stone to better understanding MBL as a clinical entity.
Fovinaciclib versus placebo combined with fulvestrant for advanced breast cancer after endocrine therapy: A randomized, double-blind phase 3 trial.
1101 Background: Approximately 70% breast cancer (BC) patients have hormone receptor (HR)-positive and human epidermal growth factor receptor 2 (HER2)-negative disease. Fovinaciclib is a novel, potent, oral selective cyclin-dependent kinase 4 and 6 (CDK4/6) inhibitor. A multicenter, randomized, double-blind, placebo-controlled phase 3 trial was conducted to compare fovinaciclib versus placebo plus fulvestrant in patients with advanced BC that relapsed or progressed on previous endocrine therapy. Methods: Eligible patients were randomly assigned (1:1) to receive 200 mg fovinaciclib (fovinaciclib arm) or placebo (placebo arm) orally once daily on days 1–21 in 28-day cycles. All patients also received intramuscular fulvestrant 500 mg (cycle 1: days 1 and 15; subsequent cycles: day 1). Premenopausal or perimenopausal patients also received subcutaneous goserelin 3.6 mg (day 1). The primary endpoint was blinded independent central review (BICR)-assessed progression-free survival (PFS). Secondary endpoints included other efficacy endpoints and safety. Exploratory endpoints included patient-reported outcomes. Results: A total of 319 patients were randomized to the fovinaciclib ( n = 160) or placebo arm ( n = 159). Baseline characteristics were balanced. Interim analysis at 9-month median follow-up (data cutoff date May 11, 2023) showed significant PFS benefit from adding fovinaciclib (p = 0.0001). Sustained PFS benefit of greater extent (median 14.0 months versus 5.6 months, hazard ratio 0.48, 95% confidence interval 0.36–0.65) was observed at updated analysis at 23-month follow-up (data cutoff date August 30, 2024). Trends of the PFS benefit were observed in all subgroups. Consistent improvements were observed in investigator-assessed PFS and objective response rate, disease control rate, clinical benefit rate, duration of response, and time to progression according to BICR or investigator's assessments. At updated analysis, all patients in the fovinaciclib arm and 148 (93.7%) in the placebo arm reported at least 1 treatment-emergent adverse event (TEAE). The most common TEAEs were hematological toxicities. TEAEs leading to study drug discontinuation were reported in 9 (5.6%) and 2 (1.3%) patients in the respective arms. No grade ≥3 venous thromboembolism or interstitial lung disease was reported. The two arms showed similar overall longitudinal profiles of global health status, function domains and symptom domains assessed with the European Organisation for Research and Treatment of Cancer Core Quality of Life questionnaire. Conclusions: Adding fovinaciclib to fulvestrant conferred statistically significant and clinically meaningful PFS benefit and consistent improvements in other survival outcomes, along with manageable safety. These findings support fovinaciclib as a later-line option. Clinical trial information: NCT05438810 .
Safety and efficacy of ifebemtinib combined with garsorasib in KRAS G12C mutant colorectal cancer from a randomized part of a phase Ib/II study.
3591 Background: Studies (Yaeger R, et al.; Fakih, MG, et al.) have shown when RASi is combined with optimal therapeutic partner(s), clinical efficacy and benefit can be maximized. Ifebemtinib (ifebe) is a highly potent and selective oral inhibitor of focal adhesion kinase (FAK) that exhibits synergistic effect with RASi in both preclinical and clinical setting. Garsorasib (D-1553) is a novel KRAS G12Ci approved in China for treatment of KRAS G12C mutant NSCLC. We previously reported a promising objective response rate (ORR) of 44.4% in patients (pts) with KRAS G12C mutant CRC receiving ifebe + D-1553, with the durability of efficacy yet to be confirmed. Herein, we present the updated results with median follow-up of (14.4 months) from a randomized cohort of phase Ⅰb/Ⅱ study, comparing ifebe + D-1553 vs D-1553 alone in KRAS G12C mutant CRC, to elucidate the relative therapeutic contribution of ifebe and durability of response. Methods: Metastatic KRAS G12C mutant CRC pts with at least 1 prior line of systemic anticancer therapy were enrolled in a randomized part and randomized 1:1 to ifebe (100mg QD) + D-1553 (600mg BID) or D-1553 (600mg BID) alone. Results: As of January 4, 2026, 36 previously treated metastatic CRC pts were enrolled, with 18 pts in ifebe + D-1553 group and 18 pts in D-1553 group (≥3 prior lines of treatment: 33.3% vs 27.8%; liver metastasis: 55.6% vs 72.2%). The median treatment duration was 8.5 months vs 3.6 months and 4 pts (3 [16.7%] vs 1 [5.6%]) were still on treatment. The hazard ratio (HR) for progression free survival (PFS) and overall survival (OS) was 0.47 and 0.44, respectively, demonstrating the durability of efficacy in pts receiving ifebe + D-1553 (Table). The safety profiles of ifebe + D-1553 in CRC pts are comparable to each single agent. No ifebe or D-1553 related death or AEs leading to drug withdrawal were reported. Conclusions: The combination of ifebe and D-1553, a dual-oral regimen, exhibits favorable safety profiles and better efficacy in treating KRAS G12C mutant CRC, with mOS exceeding 16 months. Clinical findings from this randomized study have shown robustness of improvement in overall ORR, PFS, DOR and OS benefit with this combination regimen, indicating that adding ifebe to a RASi can further improve clinical outcome to CRC patients with RAS G12C mutation. Clinical trial information: NCT06166836 . Efficacy. ifebe + D-1553(N=18) D-1553(N=18) Objective response rate (ORR), % 44.4 (21.5, 69.2) 16.7 (3.6, 41.4) Disease control rate (DCR), % 100.0 (81.5, 100.0) 77.8 (52.4, 93.6) Median duration of response (mDoR), months 12.1 (1.41, NE) 6.8 (6.3, NE) Median PFS, months 8.4 (3.0, 11.9) 4.0 (2.0, 5.6) HR for PFS (95% CI) 0.47 (0.23, 0.98) Median OS, months 16.4 (13.1, NE) 7.8 (5.3, NE) HR for OS (95% CI) 0.44 (0.17, 1.15)
Post-COVID changes to healthcare utilization, expenditures, and care prioritization among US cancer survivors.
e23131 Background: The COVID-19 pandemic disrupted cancer care delivery, leading to adaptations. We examined cost burden (CB), medical debt (MD), total healthcare expenditures (THE), out-of-pocket (OOP) costs, and forgone/delayed care (FoDC) among US cancer survivors (CanS) following COVID. Methods: Using a nationally representative Medical Expenditure Panel Survey (MEPS), we identified adult CanS in pre-COVID (PRE) (2018 & 2019) and post-COVID (POS) (2022 & 2023) eras. Outcomes included CB (OOP costs >10% of family income, or >5% if income <200% federal poverty level), MD (inability to pay medical bills or doing so over months), and FoDC (inability to afford or delay care or prescriptions). Survey-weighted chi-square and Wilcoxon rank-sum tests compared variables. Expenditures were inflation adjusted. Multivariable (MVA) Quasi-Poisson regression adjusting for age, sex, race, insurance, poverty, spouse status and COVID era identified predictors of THE and OOP costs. All analyses used survey design and were weighted. P < 0.05 was significant. Results: A total of 3,880 PRE (weighted (w): 20,309,197) and 3,346 POS (w: 21,766,685) CanS were identified. Prevalence of CB in PRE and POS CanS was 11%. Reported MD was lower POS among all-comers (18% vs 22%; p<0.01), and across all races, but more pronounced in Hispanic (18% vs 27%) and Black individuals (20% vs 29%). Patients with only public insurance experienced less MD POS (17% vs 23%) as did those never married (16% vs 26%). In adjusted regression, Younger CanS (18-45) (adjusted relative risk (aRR) 0.3, ref = 65+) and Hispanic (aRR 0.5) and Black (aRR 0.6) CanS (ref = white) were less likely to experience CB, while both privately insured (aRR 2) and private + Medicare (aRR 1.3) experienced more CB compared to public only insurance (all p <0.01). In MVA, THE was higher POS by $4,000 per CanS (p<0.01) while OOP costs remained unchanged. Female CanS (aRR 1.4), younger (18-45) (aRR 2.4) and middle-aged (45-64) CanS (aRR 2) compared to those 65+, CanS without a spouse at home (aRR 1.4), CanS experiencing CB (aRR 1.6), and poor CanS (aRR 1.78) were more likely to report FoDC while POS CanS (aRR 0.7) were less likely to report FoDC. Conclusions: This is the first nationwide study to examine changes to CanS priorities and healthcare spending after the COVID pandemic. POS CanS experienced higher THE, while risk of CB and OOP costs remained unchanged, suggesting that the decrease in MD, likely due to pandemic-era financial relief policies and improved access to cancer care with telehealth and other Cancer Moonshot initiatives, might have led to the decreased rates of FoDC among CanS. Despite population-level improvements, younger and middle-aged CanS and female CanS remain at risk of forgoing care when forced to prioritize. These findings underscore the need for targeted interventions to address persistent disparities among vulnerable CanS.
Hepatic arterial infusion chemotherapy plus lenvatinib and PD-1 inhibitors as first-line treatment for hepatocellular carcinoma with high tumor burden and portal vein tumor thrombus: A target trial emulation study (CHANCE 2416).
4132 Background: Patients with advanced hepatocellular carcinoma (HCC) characterized by high tumor burden (beyond up-to-seven criteria) and portal vein tumor thrombus (PVTT) have a median overall survival (OS) of lower than 4 months without intervention. Global guidelines rely on Phase III trials that often under-represented this high-risk subgroup, creating a therapeutic gap. While systemic combinations are standard, they may lack the rapid cytoreductive potency required to prevent liver failure in this subgroup. We emulated a target trial to evaluate if adding hepatic arterial infusion chemotherapy (HAIC) to lenvatinib and PD-1 inhibitors (H+L+P) improves outcomes compared to lenvatinib and PD-1 inhibitors (L+P) alone. Methods: Using multicenter data from 22 centers in the Chinese Liver Cancer Clinical Study Alliance (CHANCE) registry, we compared first-line H+L+P (n = 127) vs. L+P (n = 117) in patients with advanced HCC (BCLC stage C without extrahepatic spread) and PVTT (Vp1-4). To minimize selection and immortal time biases, we employed stabilized inverse probability of treatment weighting (sIPTW) and a cloning-censoring-weighting framework. The primary endpoint was OS; secondary endpoints included progression-free survival (PFS), site-specific PFS, objective response rate (ORR), and safety. Results: Baseline covariates were well-balanced after weighting (SMD < 0.1). After sIPTW adjustment, H+L+P demonstrated a significant survival benefit: median OS was 25.6 months (95% CI: 22.0–33.7) vs. 15.9 months (95% CI: 12.6–21.3) for L+P (adjusted HR, 0.60; 95% CI: 0.43–0.82; p = 0.0015). Crucially, a favorable survival trend was preserved even in the high-risk Vp3-4 subgroup (HR 0.77). Median PFS (RECIST v1.1) was 13.0 vs. 6.8 months (adjusted HR, 0.58; p = 0.0002). H+L+P significantly delayed progression in intrahepatic lesions (median 14.8 vs. 7.8 months; p = 0.0006) and PVTT (median 24.9 vs. 15.5 months; p = 0.0183). Intrahepatic ORR (mRECIST) was 87.4% for H+L+P vs. 28.9% for L+P (p < 0.0001), with a PVTT response rate of 81.6% vs. 22.1% (p < 0.0001). Grade ≥3 adverse events (AEs) occurred in 37.0% of the H+L+P group vs. 9.4% in the L+P group. Common Grade ≥3 AEs in the H+L+P arm included abdominal pain (20.5%) and leukopenia (5.5%); no treatment-related deaths occurred in the H+L+P group. Conclusions: In this target trial emulation, adding HAIC to lenvatinib and PD-1 inhibitors significantly improved OS and PFS in patients with high-risk HCC and PVTT. The survival advantage was driven by rapid and profound locoregional control, particularly of the tumor thrombus, with a manageable safety profile. Clinical trial information: NCT06631326 .
Interim analysis results from Duravelo-2: Zelenectide pevedotin (zele; BT8009) in patients (pts) with previously treated locally advanced/metastatic urothelial carcinoma (la/mUC).
4566 Background: Zele is a highly selective Bicycle Drug Conjugate (BDC) targeting Nectin-4, a protein overexpressed in la/mUC. There remains an unmet need in la/mUC for safer and more tolerable treatments. Here, we report an interim analysis (IA) for zele monotherapy dosage selection from the Phase 2/3 Duravelo-2 study (NCT06225596/BT8009-230) in previously treated pts with la/mUC. Methods: Adults with la/mUC were enrolled into 2 cohorts: previously untreated pts eligible for platinum-based chemotherapy (Cohort [Co]1) or pts with ≥1 prior systemic therapy (Co2). IA was conducted to determine the optimized dosage of zele + pembrolizomab (Co1) or zele monotherapy (Co2). Herein, we report IA results from Co2. Pts in Co2 were randomized 1:1 to zele 5 mg/m 2 on Days [D]1/8/15 or zele 6 mg/m 2 on D1/8 on a 21-D cycle. Dosage optimization included safety, efficacy, and pharmacokinetic data in pharmacometrics and utility score analyses to quantify benefit-risk. Results: IA was performed at 27 weeks of follow-up (N=30 each zele dose group). Median zele treatment (tx) duration was 4.80 months. Confirmed ORRs by BICR among randomized, dosed pts with measurable disease at baseline were 28% (8/29; 2 complete responses [CR], 6 partial responses [PR]; 95% CI 12.7–47.2) and 30% (8/27; 3 CR, 5 PR; 95% CI 13.8–50.2) with zele 5 mg/m 2 and 6 mg/m 2 , respectively. Zele-related adverse events (AEs) were reported in 93% (47% Gr ≥3) of pts at 5 mg/m 2 and 93% (52% Gr ≥3) at 6 mg/m 2 . Gr ≥3 zele-related AEs for 6 mg/m 2 (≥5% pts) included neutrophil count decreased and neutropenia (10% each) and anemia, asthenia, and ALT increase (7% each). No zele-related Gr 5 AEs were reported at either dosage. Zele-related AEs of clinical interest (AECIs) for 6 mg/m 2 are summarized in the Table. Notably, no zele-related severe skin reactions of any Gr were reported at either dosage. Zele dose reductions and discontinuations due to zele-related AEs occurred in 24% and 0%, respectively, for 6 mg/m 2 . At Week 27, a third of pts remained on tx. Conclusions: Zele monotherapy at 5 mg/m 2 on D1/8/15 and 6 mg/m 2 on D1/8 on a 21-D cycle demonstrate encouraging response rates and safety profiles with the potential to differentiate from ADCs in previously treated pts. The 6 mg/m 2 D1/8 regimen provides a more favorable benefit-risk profile to meet the need for safer and more tolerable treatments, with enhanced potential for combinability and convenience, leading to fewer discontinuations. Clinical trial information: NCT06225596 . Zele-related AECIs in pts with previously treated la/mUC treated with zele 6 mg/m 2 on D1/8 (N=30). AECI, n (%) Any Grade Grade ≥3 Peripheral neuropathy 11 (38) 1 (3) Skin reactions 8 (28) 0 Eye disorders 3 (10) 0 Hyperglycemia 1 (3) 1 (3)
Mechanism-driven comparative safety signals of capivasertib versus CDK4/6 inhibitors: A FAERS pharmacovigilance study.
e23444 Background: Capivasertib, a first-in-class AKT inhibitor approved for hormone receptor-positive/human epidermal growth factor receptor 2-negative (HR+/HER2–) advanced breast cancer, exhibits a toxicity profile distinct from cyclin-dependent kinase 4/6 (CDK4/6) inhibitors. Post-marketing pharmacovigilance offers an opportunity to characterize real-world safety patterns beyond clinical trials. We compared adverse event (AE) reporting profiles of capivasertib and CDK4/6 inhibitors using the FDA Adverse Event Reporting System (FAERS). Methods: FAERS reports from 2011-2026 were queried for capivasertib and CDK4/6 inhibitors (palbociclib, ribociclib, abemaciclib). AEs of interest were grouped according to the Medical Dictionary for Regulatory Activities (MedDRA). Disproportionality analyses were performed using reporting odds ratios (RORs) with 95% confidence intervals (CIs) calculated from 2×2 contingency tables, with statistical significance evaluated using chi-square testing (1 df). Results: A total of 2,112 capivasertib reports and 137,831 CDK4/6 inhibitor reports were identified. Capivasertib demonstrated striking enrichment of metabolic toxicities consistent with AKT-pathway inhibition, including hyperglycemia (ROR 62.23; 95% CI 51.53–75.14; p < 1×10⁻¹⁰) and diabetic ketoacidosis (ROR 120.54; 95% CI 79.06–183.77; p < 1×10⁻¹⁰). Dermatologic immune-mediated events were also significantly over-reported, including rash (ROR 6.07; 95% CI 5.36–6.88; p < 1×10⁻¹⁰) and erythema multiforme (ROR 64.40; 95% CI 43.90–94.49; p < 1×10⁻¹⁰). Gastrointestinal AEs demonstrated a heterogeneous pattern, with increased diarrhea reporting (ROR 1.79; 95% CI 1.59–2.01; p < 0.001) but reduced nausea reporting (ROR 0.55; 95% CI 0.45–0.66; p < 0.001). In contrast, capivasertib was associated with markedly lower hematologic toxicity compared with CDK4/6 inhibitors, including neutropenia (ROR 0.04), leukopenia (ROR 0.02), and thrombocytopenia (ROR 0.20), all p < 0.001. A mortality signal was observed for capivasertib (ROR 1.76; 95% CI 1.56-2.00; p < 0.001), though interpretation is limited by confounding from disease severity, prior therapy exposure, and inherent FAERS reporting bias. Conclusions: Capivasertib exhibited a distinct toxicity signature characterized by pronounced metabolic and dermatologic adverse events and substantially lower hematologic toxicity compared with CDK4/6 inhibitors. These findings align with the drug’s mechanism of AKT inhibition and underscore the importance of proactive metabolic monitoring and early dermatologic management in clinical practice. Prospective and comparative real-world studies are warranted to contextualize the observed mortality signal and refine patient selection.
Phase 3, multicenter, randomized, controlled open-label study: Efficacy and safety of M701 (EpCAM×CD3 bispecific antibody) intraperitoneal infusion in advanced epithelial solid tumor with malignant ascites vs paracentesis.
104 Background: Malignant ascites (MA) is a severe complication of advanced epithelial cancers, imposing a heavy symptom burden, compromising systemic therapy efficacy, and correlating with poor survival and impaired quality of life. While only one pharmacologic therapy is approved; guideline-recommended paracentesis provides merely transient symptomatic relief. Consequently, a significant unmet medical need persists for effective therapies that durably control MA. Methods: Patients with malignant ascites (MA) secondary to advanced gastric, colorectal, or ovarian cancer after failure of standard anticancer therapy were enrolled. Patients were randomized to Arm T, receiving therapeutic paracentesis plus intraperitoneal M701, or Arm C, receiving paracentesis alone. All patients continued background systemic treatment. The primary endpoint was puncture-free survival (PuFS), defined as the time from Day 18 to the subsequent paracentesis or death. Secondary endpoints included overall survival (OS), time to next paracentesis (TTNP), patient-reported outcomes (PROs) evaluated using the EORTC QLQ-C30 questionnaire and a Likert scale, and safety profiles between the two arms. Results: As of January 15, 2026, 312 pts were enrolled (Arm T, n = 206; Arm C, n = 106). Median PuFS was significantly longer in Arm T than in Arm C (87.59 vs 49.96 days; HR = 0.57, 95% CI: 0.42–0.78; p = 0.0003). TTNP was also markedly improved with M701 (186.7 vs 55.1 days; HR = 0.42, 95% CI: 0.29–0.60, p < 0.0001). Overall survival was comparable between arms (p = 0.8407, HR = 0.97, 95% CI: 0.74–1.28). Patient-reported outcomes showed a longer median time to global health status deterioration in Arm T (72.0 vs 39.0 days; HR = 0.73, p = 0.049). A statistically significant between-group difference was observed in the change from baseline in total Likert score at Day 116 (Visit 12) (mean difference −2.45; p = 0.0130), indicating sustained symptom relief with M701. Grade ≥3 treatment-emergent adverse events (TEAEs) occurred in 70.39% of patients in Arm T and 51.89% in Arm C. Serious adverse events (SAEs) occurred in 48.54% and 33.02% of patients in the two arms, respectively. Only one grade 1 cytokine release syndrome (CRS) event was reported in Arm T, and the incidence of CRS-like symptoms (e.g., pyrexia and dyspnea) was very low. Conclusions: Intraperitoneal M701 was well tolerated when combined with systemic therapy. In patients with epithelial cancer-associated MA, M701 significantly prolonged puncture-free survival with comparable overall survival. These encouraging findings further support the clinical development of M701 and its potential therapeutic role in the management of malignant ascites. Clinical trial information: NCT06432296 .
Real-world association between the androgen receptor gene polymorphisms (CAG repeats) and aggressiveness of prostate cancer at diagnosis in the Medstar Health Network.
e17143 Background: Clinical manifestations of prostate cancer (PC) range from indolent to aggressive disease. Androgen receptor (AR) signaling is central to PC biology, and the AR gene contains a polymorphic CAG trinucleotide repeat in exon 1. Shorter repeat lengths ( < 20) have been associated with increased PC susceptibility; however, their relationship with disease aggressiveness and clinical outcomes remains unclear. Established indicators of aggressiveness include prostate-specific antigen (PSA), clinical stage, and Gleason score (GS), while time to progression and overall survival reflect long-term prognosis. Clarifying the association between AR CAG repeat length and these outcomes may improve understanding of PC biology and risk stratification at diagnosis. Methods: Men with biopsy-proven prostate cancer were included in this retrospective analysis within the MedStar Health Network from 2016–2024. Next-generation sequencing of tissue biopsy samples was used to determine androgen receptor (AR) CAG repeat length in exon 1. Patients were categorized as having short ( < 20 repeats) or long (≥20 repeats) CAG repeat length. PSA level, Gleason score, and clinical stage at diagnosis were obtained from medical records. Ordinal logistic regression was used to assess associations between CAG repeat length ( > 20 vs ≤20) and Gleason risk category or clinical stage at diagnosis, while linear regression was used to assess the association with PSA level. Time-to-event outcomes were evaluated using Cox proportional hazards regression, and Kaplan–Meier survival curves were compared using the log-rank test. Results: A total of 128 men were included in the cohort, of whom 43 (33.6%) had short AR CAG repeats and 85 (66.4%) had long repeats. Using ordinal logistic regression, AR CAG repeat length ( > 20 vs ≤20) was not significantly associated with Gleason score (p = 0.62) or clinical stage at diagnosis (p = 0.17). Linear regression showed no significant association between PSA levels and CAG repeat category (p = 0.47). In Cox proportional hazards analyses, CAG repeat length > 20 was not associated with all-cause mortality compared with ≤20 repeats (HR, 0.91; 95% CI, 0.43-1.95; p = 0.814) or with time to disease progression (HR, 1.13; 95% CI, 0.65-1.98, p = 0.663). Conclusions: In this retrospective cohort, AR CAG repeat length was not significantly associated with established markers of tumor aggressiveness at diagnosis or with clinical outcomes, including progression and overall survival. Despite limited sample size and event rates, this study demonstrates the feasibility of incorporating somatic AR CAG repeat analysis into clinical datasets. Larger, prospective studies are warranted to better define the potential prognostic role of AR CAG repeat length and to explore its utility in risk stratification and personalized management of PC.
Hybrid AI framework for automated prostate cancer disease state ascertainment from real-world electronic health records.
e17003 Background: Accurate ascertainment of prostate cancer (PCa) disease states—biochemical recurrence (BCR), castration-sensitive prostate cancer (CSPC), and castration-resistant prostate cancer (CRPC)—is critical for automated clinical trial enrollment, quality improvement initiatives, and real-world evidence generation. Essential information for disease state ascertainment resides in unstructured EHR data, requiring labor-intensive manual abstraction that is inconsistent and not scalable. We developed an automated hybrid AI framework using large language models to identify PCa disease states from real-world EHR data. Methods: This retrospective study included patients with histopathologically confirmed PCa (2017-2022) and ≥3 years follow-up. Sequential EHR data—including PSA and testosterone levels, clinical notes, pathology and radiology reports—were retrieved across temporal points. BCR was identified using rule-based algorithms (post-radical prostatectomy [RP]: PSA ≥0.2 ng/mL ×2; post-radiation therapy [RT]: nadir + 2 ng/mL ×2). CRPC was determined using androgen deprivation therapy (ADT) timing, PSA kinetics, testosterone < 50 ng/dL, and radiographic progression per PCWG/EAU-ASCO criteria. GPT-4o with structured parametrized prompts extracted metastatic status and radiographic progression. Independent clinicians performed manual annotation for validation. The framework was developed on 10% of data with evaluation on 90%. External validation used a held-out cohort of clinical trial patients. Results: Among 200 PCa patients (median age 66 years [IQR 60-71]; 91% White, 93% non-Hispanic), the framework demonstrated high accuracy: PCa diagnosis (92%), RP dates (94%), RT dates (93%), and ADT initiation (92%). For disease state identification, BCR detection achieved 90% accuracy, castration status 93%, and metastatic state (M0/M1) 87%. External validation on trial-enrolled patients confirmed robust performance: 26/31 (84%) CRPC trial patients and 26/27 (96%) CSPC trial patients were correctly classified, validating the framework’s reliability for clinical trial matching. Conclusions: This hybrid LLM-based framework accurately identifies PCa disease states using real-world EHR data with minimal manual intervention. The system offers a scalable, reproducible approach for automated disease state identification, enabling streamlined clinical trial matching and enhanced real-world data curation in oncology.
First 1000 patients report of global practice patterns and clinical outcomes in the Collaborative Ocular Melanoma Natural History (OMNi) real-world database (NCT04588662).
9595 Background: OMNi (NCT04588662) is an ambispective natural history database developed to provide contemporary real-world data of UM. The overall objectives of OMNi are to characterize regional/international UM management practice patterns and associated clinical outcomes in an effort to inform best practice recommendations and provide real world evidence. Here we present characteristics, demographics and practice patterns in the first 1000 patients enrolled globally. Methods: OMNi utilizes the Pulse Infoframe Healthie platform, enabling the longitudinal structured collection of data mapped to Observational Medical Outcomes Partnership. Results: Median age at diagnosis was 60 (15-94), with 47% of patients were female. Enrolment was from Australia (309), UK (93), USA (121) and Canada (477). When known, family history of UM was present in 2% (15/733). Recorded primary treatment modalities included plaque brachytherapy 44%, Enucleation 18%, Stereotactic radiotherapy 9%, Proton beam 3%, other radiotherapy 1%, other/ unknown 23%. Within the cohort, 30 patients were diagnosed with de novo metastatic disease and 343 patients had metastatic disease at any time. Of those, metastatic location was 166 extrahepatic and hepatic, 154 hepatic only, 18 extrahepatic only. HLA-A2 status was *0201 in 172/ 297(58%). Systemic treatments were administered to 176 patients first line– Tebentafusp (71 fist line); Ipi/Nivo (58); Ipi (7); PD-1 inh (15). The total number of patients receiving systemic treatments were Tebentafusp 108; Ipi/Nivo 90; Ipi 22; PD-1 (82). In total, 133 patients received liver directed therapy. At last follow up, where available 29% of patients had died from UM (179/611). Data to be presented will include genetic profiling summaries. Conclusions: The OMNi dataset can serve and aid in interpretation of clinical trial outcomes in the real-world, and accelerate the development of diagnostics and therapeutics to provide comparative data.
The quantum scalpel: Unveiling new dimensions in cancer therapy.
e15000 Background: The Cell Memory Disc (CMD) model describes each human cell as a quantum wave–particle system composed of visible (vCMD), invisible (iCMD), and empty (eCMD) components embedded within a multidimensional superposition of states. This framework suggests that malignant transformation, metastasis, and treatment resistance may arise from quantum-level discontinuities, tunneling events, and pre-treatment state transitions not captured by classical models. Understanding CMD kinetics under therapeutic stress may provide new insights into cancer recurrence. Methods: A theoretical modeling approach was used to characterize CMD behavior during exposure to high-intensity therapeutic stress. The model integrates CMD wave–particle duality, coherence–decoherence transitions, Higgs-singlet–mediated information transfer, and transient quantum gateways within CMD space-time. Simulated interactions between stress energy and CMD states were evaluated to describe transitions among robustness (R-phase), fragility (F-phase), and mimicry (Mi-phase). Results: Modeling predicts that severe stress can induce transient quantum gateways enabling Higgs singlets to access uncollapsed CMD states, facilitating pre-emptive state transitions before treatment takes effect. This mechanism may allow cancer cells to adopt stress-resistant configurations, contributing to recurrence. The model further suggests that vCMD discontinuity and iCMD particle mobility may enable metastasis without classical cell migration. Simulations indicate that conventional therapies preferentially eliminate R-phase cells while allowing F-phase cells to survive, enter Mi-phase, and later re-collapse into new proliferative states, consistent with clinical patterns of relapse. Conclusions: The CMD framework provides a structured, testable model for interpreting cancer behavior through quantum-level mechanisms, including state superposition, tunneling, and pre-treatment adaptation. These findings preserve the full conceptual structure of the CMD theory while generating hypotheses regarding recurrence, metastasis, and resistance. Future experimental work is needed to evaluate CMD-based therapeutic strategies such as CMD masslessness, CMD vacuumlessness, wave-interference modulation, and CMD orchestration.
Clinical characteristics and outcomes of germline and sporadic NF1-associated gastrointestinal stromal tumors: Insights from a real-world patient registry.
11538 Background: Neurofibromatosis type 1 (NF1)-associated gastrointestinal stromal tumor (GIST) is a rare, distinct molecular subtype, typically lacking KIT or PDGFRA mutations. Standard KIT-targeted therapies show limited efficacy, highlighting the need for alternative systemic approaches. Real-world data on the clinical features, treatment patterns, and outcomes of NF1-associated GIST are limited. Methods: The Life Raft Group (est. 2000) Registry is an international, observational registry. Patients with GIST and documented NF1 mutations without another primary oncogenic alteration were included. Demographic, tumor, and treatment characteristics were summarized; group differences were assessed using Fisher’s exact test, and disease-free survival (DFS) and overall survival (OS) were estimated by Kaplan–Meier methods. Results: A total of 38 patients were identified (1.3% of the total registry, n = 2990) and stratified as germline NF1 (n = 18) and sporadic NF1 (n = 20). Median age at diagnosis was 57 years; 53% (n = 20) of patients were female, with a higher proportion of germline NF1 than sporadic NF1 (60%, n = 12 vs 40%, n = 8). Conversely, among males, sporadic NF1 was more common than germline NF1 (67%, n = 12 vs 33%, n = 6). Tumors predominantly arose in the small bowel (79%). Germline NF1 patients exhibited significantly higher rates of multifocal disease compared with sporadic NF1 patients (85% vs 15%, p = 0.0006). Most patients presented with localized disease (84%, n = 32), with similar proportions between germline and sporadic NF1 (78%, n = 14 vs 90%, n = 18). All localized patients underwent surgical resection, achieving R0 margins in 72% of cases. Among patients who developed advanced disease (de novo metastatic or recurrent disease; n = 18), 67% received systemic therapy including MEK inhibitors in 22% (n = 4). One patient achieved durable disease control on selumetinib for nearly six years and remains on treatment. Median follow-up was 37.6 months (2.5–185). In localized patients, 5-year DFS was 65% (95% CI: 39–100) for germline and 46% (95% CI: 25–84) for sporadic NF1; median DFS was 55.4 months in sporadic NF1 and not reached in germline NF1. 10-year OS was 78% (95% CI: 62–98) overall, 88% (95% CI: 67–100) for germline, and 83% (95% CI: 64–100) for sporadic localized patients. Survival estimates in metastatic patients were highly variable due to small numbers. Conclusions: In this real-world registry analysis, NF1-associated GIST shows small bowel predominance and frequent multifocality, particularly in germline NF1. Most patients achieve disease control with surgery. MEK inhibitors may demonstrate potential activity in advanced NF1 GIST. These findings highlight phenotypic differences between germline and sporadic NF1-associated GIST and support further evaluation of MEK-targeted therapy in patients with metastatic disease.
Assessment of gedatolisib (geda) pharmacokinetics (PK), drug-drug interactions (DDI), and exposure-safety and exposure-efficacy relationships.
e13069 Background: Gedatolisib is a highly potent, multitarget PI3K/AKT/mTOR (PAM) inhibitor that targets all class I PI3K isoforms and both mTOR complexes, mTORC1 and mTORC2. Safety and efficacy have been demonstrated in adults with HR+/HER2- advanced breast cancer (ABC) after progression on/after CDK4/6 and aromatase inhibition. Because i n vitro studies demonstrate that geda inhibits CYP2C8, BCRP, and MATE1/2-K and is a substrate for P-gp, we conducted a two-part DDI study in healthy volunteers. We also performed exposure-response analyses using pooled PK, safety, and efficacy data from clinical trials, including the phase 3 VIKTORIA-1 study. Methods: Adults aged 18-55 years were eligible for PK/DDI study. In Part 1, single oral doses of repaglinide (0.5 mg; CYP2C8), rosuvastatin (5 mg; BCRP) and metformin (500 mg; MATE1/2-K) without geda, and after washout, with geda (180 mg) were administered. In Part 2, two single doses of geda (45 mg), without itraconazole (P-gp inhibitor), and after washout, with multiple itraconazole doses (200 mg/daily) were given. Serial PK samples were collected. Safety and tolerability were assessed based on AEs and clinical findings. For exposure analyses, geda concentrations from phase 3 (180 mg 3w on/1w off), FiH dose-escalation (10 to 319 mg qw) and Phase 1b (180 mg qw or 3w on/1w off) were pooled & analyzed by population PK methods to derive predicted estimates of C max , C min and cumulative Cycle 1 AUC. Clinical data from Phase 1b & Phase 3 were used for exposure-efficacy analysis. Safety endpoints were Gr ≥3 stomatitis/mucositis, any grade AE and Gr ≥3 AE causing dose reduction/withdrawal, and hyperglycemia, myelosuppression, GI AEs, rash and infusion reactions. Results: With concomitant geda + repaglinide, C max and AUC inf were 43.7% and 4.7% lower, respectively); geda + rosuvastatin, C max and AUC inf were 36.4% and 44.4% higher; geda + metformin, C max and AUC inf were 45.5% and 67.7% higher. With concomitant geda + itraconazole, C max and AUC inf were 17.3% and 1.4% lower. Geda was generally well tolerated, with TEAEs of mild-to-moderate severity. In exposure analyses, tumor size change and PFS stratified by quartiles of Cycle 1 geda PK parameters did not show clear or consistent relationship to exposure. In pts with stable or progressive disease or partial response, best overall response did not appear related to exposure. Logistic regression analysis of geda PK parameters did not show statistically significant and/or clinically relevant relationships between exposure and safety endpoints. Conclusions: Dose modification of CYP2C8, BCRP and MATE1/2-K substrates is not required when co-administered with geda, and geda dose modification is not required with P-gp inhibitors. Exposure-efficacy and exposure-safety assessments support the currently recommended geda dose of 180 mg administered in a 3w on/1w off regimen. Clinical trial information: NCT05501886 ; NCT00940498 ; NCT02684032 .
Mid-treatment circulating tumor HPV DNA to guide early radiotherapy cessation in HPV-associated oropharyngeal cancer.
e18054 Background: Circulating tumor HPV DNA (ctHPVDNA) is an emerging biomarker in HPV-associated oropharyngeal squamous cell carcinoma (HPV+OPC) with established prognostic and surveillance value. However, limited data exist on its role in treatment adaptation during radiotherapy (RT). We evaluated whether mid-treatment ctHPVDNA clearance could identify patients who safely discontinue RT early due to severe toxicity without compromising oncologic outcomes. Methods: We conducted a single-institution retrospective cohort study of patients with HPV+OPC treated with definitive RT or chemoradiotherapy. Plasma ctHPVDNA testing was performed at week 4 of RT. Patients experiencing grade ≥3 acute toxicity who discontinued RT after approximately 60 Gy were compared with patients who completed standard 70 Gy based on ctHPVDNA status. One-year progression-free survival (PFS) was estimated using the Kaplan–Meier method and compared using log-rank testing. Results: Thirty-two patients with HPV+OPC were identified. Twenty-four (75%) experienced grade ≥3 acute toxicity, and 18 underwent mid-treatment ctHPVDNA testing. Ten patients achieved ctHPVDNA clearance at week 4 and discontinued RT at approximately 6 weeks (Group A), while eight patients remained ctHPVDNA-positive and completed 7 weeks of RT (Group B). Baseline characteristics, including age, stage, and smoking history, were similar between groups. At a median follow-up of 18 months, no locoregional or distant recurrences were observed in Group A. Group B experienced two disease recurrences, including one distant failure, and one disease-related death. One-year PFS was 100% in Group A versus 72% in Group B (p=0.08). Conclusions: Mid-treatment ctHPVDNA clearance was associated with excellent short-term disease control despite early cessation of RT due to toxicity. These findings suggest ctHPVDNA may serve as a clinically useful biomarker to guide adaptive radiotherapy de-escalation strategies in HPV+OPC. Prospective validation is warranted.