Browse Articles
Discover research articles across all indexed journals
Non-blood liquid biopsy in nipple discharge for decoding tumor-derived CNV signatures for breast cancer detection and prognostication.
e12578 Background: Nipple discharge is a prevalent symptom affecting about 80% of women during reproductive years, yet only less than 5% of such cases are malignant. However, accurate non-invasive diagnosis in these women remains a longstanding clinical challenge. Here, we evaluated somatic copy number variations (CNVs) in nipple discharge DNA (ndDNA) as diagnostic biomarkers and explored their cellular origins. Methods: Breast cancer-specific CNV features were first derived from The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA; 1,086 tumors and 942 controls). A diagnostic model constructed from these features was then tested in two in-house cohorts: nipple discharge (39 malignant, 18 benign) and plasma cell-free DNA (cfDNA; 149 malignant, 34 benign). To trace the cellular origin of nipple discharge CNVs, we performed single-cell RNA-sequencing on matched tumor tissues (n=5) and integrated the resulting inferCNV profiles with nipple discharge CNV profiles at single-cell level. Results: We identified 20 recurrent somatic CNVs (12 amplifications and 8 deletions). The diagnostic model demonstrated robust diagnostic performance in four external validation datasets. In the nipple discharge cohort, it achieved 74.4% sensitivity and 94.4% specificity for malignancy. In contrast, no significant CNV feature was detected in the plasma cfDNA cohort. Paired analysis revealed high concordance between nipple discharge CNVs and inferCNV. CNV release fidelity varied among epithelial subclusters. High-fidelity subsets displayed greater CNV burden, active estrogen signaling, and stronger immune interactions, and were associated with better prognosis. Conclusions: Our findings provide the first definitive evidence that nipple discharge is a tumor-proximal, highly specific non-blood liquid biopsy for breast cancer—surpassing plasma cfDNA by preserving tumor genomic signatures with minimal dilution and background noise. Beyond diagnostics, single-cell tracing deciphers luminal epithelial subcluster-selective ndDNA shedding, illuminating biomarker release mechanisms and advancing liquid biopsy paradigms across cancers.
PROgress Tracker Breast Cancer Registry: Reporting worry of illness from a longitudinal peer-led, national patient-reported outcomes (PRO) registry.
11112 Background: PROgress Tracker Breast Cancer Registry is a national, ethics approved non-interventional patient-reported outcome measure (PROMs) registry, using a peer-to-peer model directed by patient advocacy group Breast Cancer Canada. Here, initial results report breast cancer survivors nature of cancer illness worry and intensity based on age, stage and molecular subtype at diagnosis. Methods: PROgress Tracker was launched in 2023, with enrollment over 10 years of Canadians with Stage 0-IV breast cancer. Participants self-refer and complete a series of validated PROMs (demographic, socioeconomic, clinical, global wellbeing) via a digital platform, capturing current status on a quarterly basis for up to 10 years. Real-world analysis assisted by AI ( January 2026; Claude [Opus 4.5], manufacturer: Anthropic; data extraction, statistical analysis with means, percentages, patient-level aggregation, stratification by clinical variables, longitudinal pattern analysis and visualization; all outputs validated against source data ) of FACT-B PROM data capture and analysis methodology was used. Results: To date, 823 participants have enrolled with 186 participating for ≥2 years. From participation start, longitudinal worry is reported using 3 factors from FACT-B PROM; hereditary risk of family members (HR), if stress impacts illness (SI), and if participant’s condition will worsen (CW). Highest degree of reported worry scores was in HR 40.4% (n = 74), SI 31.7% (n = 59) and CW at 17.4% (n = 32). Worry intensity varied by age, stage and molecular subtype. Participants diagnosed < 50 years reported higher levels in all 3 factors (HR 51% mean 2.53; SI 46.9% mean 2.24; CW 24.5% mean 1.78) to those at > 50 years (HR 37.1% mean 1.92; SI 25.9% mean 1.63; CW 14.3% mean 1.32). Stage IV participants show 66.7% high worry (CW and HR) compared to stage I at 8.5% (CW). Molecular subpopulations show higher worry in HR- / HER2- (TNBC) n = 34 (HR 48.5% mean 2.24; SI 38.2% mean 2.00; CW 17.6% mean 1.59) compared to HR+ / HER2- n = 81 (HR 37.5% mean 2.04; SI 35.4% mean 1.90; CW 17.3% mean 1.41) and HR+ / HER2+ n = 13 (HR 30.8% mean 2.08; SI 23.1% mean 1.77; CW 7.7% mean 1.38). Non-linear trajectory over time shows 24% (n = 183) experienced high worry at registry entry, declining at 12 months to 15.8% with a secondary peak at 18 months in 21.7% of participants. Conclusions: Patient-reported outcomes show the burden of worry by breast cancer survivors for hereditary risk to family and individuals’ illness. These findings offer directional insights for psychosocial screening, care resources and tailored education throughout survivorship.
Phase 1/2 study of the novel Werner helicase inhibitor EIK1005 as monotherapy and in combination with pembrolizumab in patients with advanced solid tumors, including MSI-H or dMMR tumors.
TPS3170 Background: Cancers with deficient mismatch repair (dMMR) have high microsatellite instability (MSI-H) and are reported to be sensitive to inhibition of the Werner syndrome helicase (WRN) via synthetic lethality. WRN inhibition in MSI-H cells leads to loss of genomic integrity and ultimately cell death while not affecting microsatellite stable (MSS) and healthy cells. Standard of care for patients (pts) with MSI-H tumors includes surgery and immunotherapy (checkpoint inhibitors [CPIs] like pembrolizumab [pembro]). While these are viable treatment options, depending on the indication, ~30 to 50% do not respond to first-line immune CPIs, or acquire resistance. A critical unmet medical need exists for new therapies for such patients (pts). EIK1005 is a novel, potent, high-affinity and selective WRN inhibitor (WRNi) with the potential for targeted antitumor activity in MSI-H or dMMR solid tumors. In vitro studies demonstrated sustained target engagement leading to cell death in MSI-H cells while sparing MSS and healthy cells. In a Phase 1 single ascending dose study in healthy participants, EIK1005 was well tolerated at the doses evaluated. Methods: EIK1005-002 (NCT#07262619) is a multicenter, multi-part, Phase 1/2 study of EIK1005 monotherapy and combination with pembro, in participants with advanced solid tumors including MSI-H and dMMR tumors. The study has three parts; Part 1A (monotherapy dose escalation [DE]): EIK1005 will be evaluated at ~ three dose levels (DL) in participants without alternative therapeutic options (n~40). Preference is given to pts with MSI-H or dMMR cancer for whom prior CPI therapy has failed, or with MSS cancer who progressed after ≥1 regimen of platinum, alkylating or topoisomerase chemotherapy as evidence suggests MMR alterations and MSI induction in previously MSS cancer. Part 1B (combination DE): participants with MSI-H or dMMR tumors (n~40) will receive EIK1005 + pembro (Q3W). Part 2 (dose optimization [DO]): participants (n~80) with MSI-H and dMMR tumors and slowly progressive disease are randomized 1:1 to receive EIK1005 at one of two doses from Part 1A. Key eligibility criteria: participants ≥18 years, life expectancy ≥3 months, unresectable and/or metastatic solid tumor, measurable disease at baseline (RECIST v1.1), progression after or intolerance to ≥ 1 advanced-setting standard treatment regimen (Part 1A), has MSI-H or dMMR tumor (Parts 1B and 2) and slowly progressive disease (Part 2). Primary objectives: safety and tolerability of EIK1005 monotherapy and + pembro to determine the MTD (Part 1) and to inform DO (Part 2). Secondary objectives: preliminary antitumor activity, characterize plasma PK profile of EIK1005 monotherapy and + pembro. This study began in December 2025. Clinical trial information: NCT07262619 .
The effects of Mediterranean diet, physical activity, and vitamin D on breast cancer recurrence and cardiometabolic health: A multicenter randomized trial.
511 Background: Low quality diets and physical inactivity are two modifiable lifestyle factors that together with vitamin D deficiency have been associated with unfavorable breast cancer (BC) outcomes and cardiometabolic health. We therefore tested the effect of a lifestyle program on BC recurrence and cardiometabolic risk factors on a background of vitamin D sufficiency. Methods: Women aged 30-74 years were enrolled in a phase III trial (NCT02786875) within 12 months of their primary BC diagnosis (stages I-III) in 7 oncologic centres in Italy. Consenting participants were randomized to one of two lifestyle treatments lasting 33 months: 1) positive control (n = 249), standard advice on a Mediterranean diet (MedDiet) and avoidance of sedentary behavior; 2) intervention (n = 243), high intensity advice on low-glycemic index MedDiet and 30-minute additional daily brisk walking. Oral vitamin D3 was administered throughout. Data were collected quarterly on health status, anthropometry, 7-day food records, daily steps, serum 25(OH)D concentrations and metabolic syndrome parameters. The primary outcome was BC recurrence (intention to treat analysis) at 33 months, with cardiometabolic risk factors as secondary outcomes. Hazard ratios (HR) and 95% confidence intervals (CI) for BC recurrence were computed using Cox proportional hazards models adjusted for recruitment center. One-year changes in treatment variables and cardiometabolic risk factors were tested by paired Student’s t-test within each treatment arm and by analysis of covariance between treatment arms, adjusted for baseline values. Results: Approximately 58% of patients were aged ≥50 years, 94% were postmenopausal, 86% had stages I-IIB and 75% hormone-positive BC. A total of 37 recurrences were observed with no differences between arms (HR = 0.94, 95%CI: 0.49-1.79). Independently of randomization arm, women with hormone-positive BC who were more adherent to the intervention program (41%), showed 76% lower risk of BC recurrence (HR = 0.24, 95% CI: 0.06-0.95). Weight loss of 3 kg, representing a 4.2% drop of initial body weight, was significantly greater in the test arm compared to 1.7 kg or 2.4% in the control arm (p = 0.0034). This translated in reductions of body mass index from 27.5±5.7 to 26.3±5.4 kg/m 2 in the intervention arm and from 27.2±5.1 to 26.6±5.1 kg/m 2 in controls (p = 0.0032). The presence of metabolic syndrome dropped by 65% in the intervention arm and 34% in controls (p = 0.008). The probability of metabolic syndrome remission was 2.5-fold higher in the intervention arm. Conclusions: A safe and low-cost lifestyle program with low-glycemic index MedDiet, daily brisk walking and oral vitamin D supplementation resulted in weight loss, reduced metabolic syndrome prevalence and was significantly associated with lower recurrences in women with hormone-positive BC. Clinical trial information: NCT02786875 .
Efficacy of velzatinib (IDRX-42) in patients with advanced/metastatic GIST by line of therapy and circulating tumor DNA response in the phase 1/1b StrateGIST 1 trial.
11520 Background: Most gastrointestinal stromal tumors (GIST) are driven by mutations in KIT or PDGFRA . Resistance to standard tyrosine kinase inhibitor (TKI) therapy is largely driven by acquisition of secondary mutations in KIT exons 13/14/17/18. Velzatinib (subject to USAN approval) is a broad-spectrum TKI designed to inhibit most common primary KIT mutations (exons 9 and 11) as well as secondary mutations. We present updated circulating tumor DNA (ctDNA) results from StrateGIST 1, a phase 1/1b study evaluating the efficacy and safety of velzatinib across primary and secondary mutations in patients (pts) with advanced/metastatic GIST. Methods: The study design of StrateGIST 1 has been previously described (Schöffski et al. J Clin Oncol 2024). Clinical activity was assessed by mRECIST v1.1. Blood samples for ctDNA analyses (Guardant360 assay) were collected for all pts before and serially during treatment. Data analysis was performed in R and hazard ratios were calculated using Cox proportional hazards regression. Results: As of September 2025, 256 pts received velzatinib across all lines of therapy. Velzatinib showed clinical activity against both primary (Exon 9 ORR: 44% [95% CI: 29.1, 60.1]; Exon 11 ORR: 19% [95% CI: 12.7, 27.6]) and secondary (Exon 13 ORR: 17% [95% CI: 8.6, 27.9]; Exon 14 ORR: 20% [95% CI: 2.5, 55.6]; Exon 17 ORR: 23% [95% CI: 14.6, 33.2]) mutations. Clinical activity is supported by ctDNA analysis (n=180 pts); >99% reduction in ctDNA variant allele frequency (VAF) was observed across a broad spectrum of individual mutations (Exon 9: 80%; Exon 11: 67%; Exon 13: 80%; Exon 14: 80%; Exon 17: 83%). We also evaluated the predictive value of different exon mutations in KIT detected at baseline via ctDNA. Focusing on pts with primary mutations only, KIT exon 9 mutations had 3-fold lower risk of disease progression compared to exon 11 mutations (HR=3.06). No significant difference was observed between major secondary mutations (exon 13 vs exon 17). Additionally, 46 pts without detectable KIT/PDGFRA mutations at baseline showed a lower progression risk (HR=0.41) compared to 180 pts with detectable KIT/PDGFRA mutations, underscoring the prognostic and/or predictive value of ctDNA analysis, even in low-shedding indications such as GIST. Conclusions: Velzatinib has broad activity against clinically relevant KIT mutations in pts with advanced and/or metastatic GIST and can substantially reduce ctDNA levels across a broad spectrum of clinically meaningful KIT mutation profiles. Baseline exon-level KIT mutation status and ctDNA detectability provide potentially useful prognostic and predictive data, justifying further study to discern the relationship between clinical benefit and molecular response. This study (NCT05489237) is funded by GSK. Clinical trial information: NCT05489237 .
Oncologic safety of autologous fat grafting as part of local–regional management of breast cancer.
e12623 Background: Autologous fat grafting (AFG) is increasingly incorporated into breast reconstruction after mastectomy or breast-conserving surgery to improve contour and symmetry. Despite its widespread use, concerns remain regarding its potential impact on local and regional oncologic control, particularly in the setting of prior multimodal breast cancer treatment. We evaluated local and regional oncologic outcomes following AFG after surgical treatment of breast cancer. Methods: A retrospective single-center cohort study was performed including patients who underwent autologous fat grafting after surgical treatment for breast cancer between June 2023 and December 2024. Primary endpoints were local and regional oncologic events after the first AFG procedure. Secondary endpoints included distant progression and procedure-related complications. Associations between oncologic outcomes, nodal status, and number of AFG procedures were analyzed. Statistical analysis was conducted using StatTech v4.11.1; categorical variables were compared using χ² or Fisher’s exact test. A two-sided p value < 0.05 was considered statistically significant. Results: Fifty-eight patients were included (median age 46.5 years). Mastectomy-based procedures were performed in 84.3% of cases, including skin- and nipple-sparing techniques with immediate reconstruction; 13.7% underwent mastectomy without lymph node dissection, and 1.7% breast-conserving surgery. Most patients had early-stage disease (DCIS–T2), with node-negative status in 65.5%. A single AFG procedure was performed in 87.9% of patients; 12.1% underwent two or more sessions. Median follow-up was 19 months. No local or regional recurrences were observed during 2024. In 2025, oncologic events were documented in 3.4% of patients and were not temporally or statistically associated with AFG exposure (p > 0.05). Nodal status was associated with subsequent oncologic events (p = 0.008). Procedure-related complications (oleogranuloma or liponecrosis) were significantly associated with the number of AFG sessions (p < 0.001). Conclusions: In this single-center cohort, autologous fat grafting performed after breast cancer surgery was not associated with impaired local or regional oncologic control during short-term follow-up. Nodal status remained the primary factor associated with oncologic events, while an increased number of AFG procedures was linked to higher complication rates. These findings support the oncologic safety of AFG as part of breast reconstruction after surgical treatment of breast cancer, while highlighting the need for longer follow-up.
Trends and disparities in cervical cancer and heart failure mortality: A 21-year retrospective study of the US population using the CDC database.
e17524 Background: Cervical cancer remains one of the leading causes of female cancer mortality, while heart failure drives chronic disease burden. Their coexistence complicates outcomes through shared risk factors, cardiotoxicity, and mortality. This study examined cervical cancer mortality trends in patients with heart failure, assessed disparities, and aimed to inform integrated management strategies. Methods: We analyzed death certificates from CDC WONDER Database (1999–2020), using ICD-10 C53 (cervical cancer) and I50 (heart failure). Female mortality was stratified by race and urbanization. Joinpoint Regression Program V5.4.0 calculated Annual Percent Change (APC) and Average Annual Percent Change (AAPC). Results: A total of 5,534 deaths were identified. Age-adjusted mortality rates (AAMR) declined overall, but trends varied. Among females, mortality fell from 0.1102 per 100,000 in 1999 to 0.076 in 2020 (AAPC -1.28%). Three phases emerged: a modest decline (1999-2006, APC -1.45%), a rapid decline (2006-2009, APC -17.6%), and a concerning reversal (2009-2020, APC +3.27%). The latter increase may reflect reduced screening, declining vaccination rates, or COVID-19 disruptions.By urbanization, large central metropolitan areas showed a steady decline (APC -4.13%). By race, White women demonstrated a consistent reduction (APC -2.61%), indicating sustained benefits of screening and vaccination. Conclusions: Cervical cancer mortality in heart failure patients declined from 1999-2020, reflecting prevention and healthcare advances. However, disparities persist, with notable increases among women since 2009. These findings highlight successes but stress the urgent need for vaccination and screening.
Prognostic associations of the G8 geriatric screening tool in the EORTC Brain Tumor Group trials 1608 and 1709 in patients with newly diagnosed glioblastoma.
e13777 Background: The G8 screening assessment tool has been validated to assess frailty in older patients with cancer. A score of 14 or less has commonly been considered to identify patients at risk of frailty. The G8 score has been implemented for all European Organisation for Research and treatment of Cancer (EORTC) clinical trials that included patients 70 years or older since 2017. Here we assessed the prognostic implications of the G8 score in older patients with newly diagnosed glioblastoma who were enrolled in EORTC clinical trials. Methods: Data from 99 patients aged 70 or older enrolled in the 1608 STEAM trial on the multikinase inhibitor zotiraciclib (NCT 03224104) (n = 23) or the 1709 MIRAGE trial on the proteasome inhibitor marizomib (NCT 03345095) (n = 76) were analyzed. We defined a G8 low group (<14) and a G8 high group (15-17). Results: The median age was 73 years, 21 patients (48%) were women and 23 patients (52%) were men. The Karnofsky performance status was 90 or more in 48 patients (52%). Steroids were taken at baseline by 55 patients (59%). The mean G8 at baseline was 14. Forty-four patients were assigned to the G8 low group whereas 49 patients were assigned to the G8 high group. Patients with a Karnofsky performance status of 90 or 100 were more often in the G8 high group (HR 5.44, 95% CI 2.17-14.46). Patients in the G8 high group had higher means for global health status scale and functioning scores at baseline than patients in the G8 low group. Treatment delivery measured by the relative dose intensity was better in patients with a high G8 than in patients with a low G8 score (HR 1.22, (95% CI 0.91-5.96). Patients in the G8 low group tended to experience more toxicity than patients in the G8 high group. Patients in the G8 high group tended to have a longer progression-free (HR 0.66, 95% CI 0.42-1.03, p = 0.066) and overall survival (HR 0.67, 95% CI 0.42-1.07, p = 0.096). A statistically significant effect was however noted when using a pre-specified adjustment model showing longer progression-free survival and overall survival in patients in the G8 high group. Conclusions: The G8 score may be a powerful tool for prognostic assessment and for patient stratification in old and frail patients with glioblastoma. It may help to identify patients in need of early involvement of a palliative care team.
Comparison of different induction chemotherapy regimens in pediatric and adolescent patients with locally advanced nasopharyngeal carcinoma.
e18062 Background: Induction chemotherapy (IC) is a first-line treatment for locally advanced nasopharyngeal carcinoma (LA-NPC). However, data on the efficacy and safety of different IC regimens in pediatric patients remain scarce, as this group is often excluded from clinical trials. Methods: This retrospective study analyzed pediatric NPC patients (age ≤ 21 years) who received IC between 2006 and 2022. To balance confounding factors, inverse probability of treatment weighting (IPTW) was applied. Results: Among 493 included patients, the median follow-up was 69 months. Before adjustment, the taxane-plus-platinum (TP) group showed significantly shorter distant metastasis–free survival (DMFS) than the non-TP group, who received platinum plus 5-fluorouracil or capecitabine (TPF/C), platinum plus 5-fluorouracil (PF) or gemcitabine plus platinum (GP), (83.7% vs. 90.9%; P=0.03). Similar results were observed after IPTW adjustment. Conversely, the non-TP regimen was associated with a higher incidence of grade 3-4 toxicities (36.1% vs. 28.3%; P=0.004). Subgroup analysis revealed that TP-based IC was associated with worse DMFS in patients aged ≤14 years, those with EBV-DNA ≤4000 copies/mL, and those with N3 stage disease. Conclusions: For pediatric LA-NPC, a non-TP IC regimen was associated with superior PFS and DMFS compared to a TP regimen, despite a higher rate of severe toxicities. The TP regimen may be a suboptimal choice, particularly for patients under 14 or those with high-risk features.
Utility of deep learning–derived PSMA PET biomarkers for metastatic prostate cancer.
e17002 Background: PSMA PET is increasingly used in staging and management of metastatic prostate cancer. Although qualitative interpretation and SUV-based heuristics of specific regions have shown prognostic value, we currently lack quantitative PET biomarkers that can aid in risk stratification at key therapeutic decision points (e.g., taxanes, ARPI, radioligand therapy). End-to-end deep learning applied directly to whole-body PSMA PET volumes enables pixel-level analysis without manual lesion segmentation, offering a practical pathway toward deployable imaging biomarkers. The purpose of this study was to investigate the utility of deep learning derived PSMA PET prognostic biomarkers in patients with metastatic prostate cancer. Methods: In this retrospective multi-institutional study, 486 PSMA PET (Ga-68 and F-18) studies for patients with stage IV metastatic prostate cancer were identified. Studies from patients diagnosed or initially treated at the host institution were split 80:20 into train/validation (344:86 studies, respectively); cases diagnosed or treated elsewhere comprised an external test set (32 studies). Whole-body PSMA PET images were aligned, resampled, and normalized. A pretrained ConvNeXtV2 deep learning model was fine-tuned to predict 3-year post-scan mortality and performance was assessed using area under the receive operating characteristic curve (AUC). Kaplan–Meier survival curves were generated to compare overall survival (OS) between patients with high vs low predicted mortality within 3 years following PSMA PET, and clinical variables were compared between groups to assess biologic correlates to the imaging biomarker. Results: Median OS across the entire dataset was 2.1 years (range 0–18). The model achieved an accuracy and AUC of 0.885 and 0.774 in the internal validation set, and 0.969 and 0.973 in the external testing set for predicting survival in 3 years after the PSMA PET scan, respectively (Table). A prediction of high mortality was significantly associated with a greater initial PSA (132 vs 86, p = 0.05), higher Gleason score (p = < 0.001), presenting with a more advanced stage at diagnosis (p = 0.004), and lower OS (2.0 vs 3.6 yrs, log-rank p = < 0.001) compared to the low predicted mortality group. Conclusions: We developed and externally validated prognostic PSMA PET biomarkers for metastatic prostate cancer using deep learning. PSMA PET biomarkers represent an objective scalable method for risk stratification that can aid in personalizing treatment. Future work will focus on deriving PSMA PET biomarkers to predict response to specific therapies. Accuracy Sensitivity Specificity PPV NPV F1-Score ROC-AUC Validation 0.885 0.625 0.911 0.417 0.960 0.500 0.774 External Test 0.969 1.000 0.964 0.800 1.000 0.889 0.973 PPV = Positive Predictive Value; NPV = Negative Predictive Value.
Multi-cohort validation of a semaphorin prognostic signature to identify an immune- excluded subtype and therapeutic sensitivity in lung adenocarcinoma.
e20016 Background: Semaphorins, classically implicated in axonal guidance, have increasingly recognized roles in tumor progression and immune regulation. However, their integrated prognostic significance and biological relevance in lung adenocarcinoma (LUAD) remain incompletely characterized. Methods: A retrospective multi-cohort study used TCGA-LUAD RNA-sequencing data (N = 503) for training and 12 independent GEO LUAD cohorts (total N = 1,820) for external validation. Differential expression analysis of 20 semaphorin genes was performed between short- and long-progression-free survival (PFS) subgroups (median cutoff: 438 days). Six prognostic candidates (SEMA3A, SEMA4A, SEMA6B, SEMA4B, SEMA3F, SEMA7A; P < 0.05) were incorporated into a multivariable Cox model with ridge regularization (λ = 0.1) to mitigate multicollinearity. A risk score was derived from penalized regression coefficients, with patients stratified by median score. External validation used multivariable Cox models adjusted for age, sex, and stage, with random-effects meta-analysis (DerSimonian–Laird). Functional enrichment (GSEA; GO and KEGG), tumor microenvironment profiling (ESTIMATE, xCell, ssGSEA), immune checkpoint analysis, and drug sensitivity prediction (oncoPredict/CTRP2) were performed. Results: The 6-gene signature demonstrated robust prognostic stratification. In TCGA, high-risk patients exhibited inferior overall survival (OS; HR = 1.43, 95% CI 1.17–1.74, P < 0.001) and PFS (HR = 1.27, 95% CI 1.10–1.48, P = 0.001). Meta-analysis across external cohorts confirmed consistent performance for OS (12 cohorts, N = 1,820; pooled HR = 1.41, 95% CI 1.26–1.58, P < 0.001) and recurrence (6 cohorts, N = 995; HR = 1.35, 95% CI 1.09–1.67, P = 0.006). High-risk tumors displayed a pro-proliferative, immune-excluded phenotype, with enrichment of mitotic spindle checkpoint regulation, collagen fibril organization, and p53 signaling. Immunologically, these tumors showed marked depletion of CD8⁺ T cells, B cells, and myeloid dendritic cells, with significantly lower Immune and Stromal Scores (all P < 0.0001). Despite this immune-excluded state, high-risk tumors selectively overexpressed CD276 (B7-H3; P < 0.0001) and CD274 (PD-L1; P < 0.01). Drug sensitivity modeling predicted increased susceptibility to taxane-based chemotherapy (docetaxel and paclitaxel; P < 0.0001) and reduced sensitivity to carboplatin (P < 0.01). Conclusions: We identified a semaphorin-based signature consistently associated with survival, immune contexture, and therapeutic response patterns in LUAD, supported by large-scale external validation. These findings highlight the potential role of semaphorin signaling as a clinically informative axis for risk stratification and treatment optimization.
First-in-human in vivo CAR T-cell generation using a CD8-targeted lipid nanoparticle platform in relapsed/refractory B-cell lymphoma.
7014 Background: R/R B-cell lymphomas remain challenging despite advances in ex vivo CAR T-cell therapies, which are limited by complex logistics, manufacturing delays, and cost. In vivo CAR T-cell generation using targeted delivery offers a paradigm shift. We report the first-in-human study of a CD8-targeted lipid nanoparticle (CD8-tLNP) platform to generate functional CD19 CAR T cells in vivo. Methods: The CD8-tLNP was constructed by conjugating an anti-CD8 VHH to the surface of LNP encapsulating mRNA coding a CD19-directed chimeric antigen receptor, Safety, pharmacodynamics, and tolerability were first evaluated in non-human primates following intravenous administration. Subsequently, 4 patients with R/R CD19⁺ B-cell lymphoma received four intravenous doses of CD19-CD8-tLNPs. Peripheral blood was collected at serial time points to assess CAR⁺ CD8⁺ T cells, B-cell depletion, cytokine levels, and clinical laboratory parameters. Antitumor activity was evaluated by PET-CT imaging, and adverse events were graded according to standard criteria. Results: The platform showed favorable safety and PD in monkeys. To date, seven patients with relapsed or refractory CD19⁺ B-cell lymphoma have been treated. Among the evaluable patients, two patients achieved PR with decreased tumor lesions, one maintained SD by day 28; four patient remains pending efficacy evaluation. Following i.v. administration, CAR gene and surface expression in peripheral blood CD8⁺ T cells peaked within 4–6 hours and declined thereafter, consistent with transient in vivo CAR T-cell generation. This was accompanied by rapid and near-complete peripheral B-cell clearance, with complete depletion (<1 B cell/μL) observed between 12 hours and day 3. Three booster doses administered at 3-day intervals sustained complete B-cell depletion and CAR T-cell activity. The treatment was generally well tolerated. No cytokine release syndrome of grade ≥2 was observed. Transient increases in CRP and ferritin were observed and were clinically manageable, without significant elevations in liver enzymes or prolonged cytopenias. Two patients showed naïve B-cell reconstitution by day 32 or 47. Conclusions: This first-in-human study demonstrates the feasibility and early clinical activity of in vivo CAR T-cell generation using a CD8-targeted LNP platform in patients with relapsed or refractory B-cell lymphoma. The approach enables rapid, transient CAR T-cell expression, repeatable dosing, and a favorable safety profile without lymphocyte clearance and complex manufacturing in ex vivo CAR T-cell. These findings provide clinical proof of concept for a scalable, off-the-shelf CAR T-cell therapy and support further clinical evaluation in hematologic malignancies. Clinical trial information: 2025-02-02.
Neoadjuvant/adjuvant serplulimab vs. placebo combined with chemotherapy for PD-L1–positive gastric cancer: A randomized, double-blind, multicenter phase 3 study.
4009 Background: The addition of immune checkpoint inhibitors (ICIs) to chemotherapy (chemo) in the neoadjuvant/adjuvant setting for gastric cancer (GC) has yielded mixed efficacy. Target population and treatment regimen remain to be optimized. This study compared the efficacy of serplulimab (anti-PD-1 antibody) plus chemo versus placebo plus chemo as neoadjuvant therapy and serplulimab versus chemo as adjuvant therapy for resectable GC. Methods: Patients with PD-L1 positive (PD-L1 CPS ≥ 5), histologically confirmed, untreated, resectable gastric or gastroesophageal junction (GEJ) adenocarcinoma were randomized 1:1 to receive 3 cycles of neoadjuvant intravenous (iv) serplulimab (4.5 mg/kg) or placebo, plus chemo (iv oxaliplatin, 130 mg/m 2 and oral S-1, 40-60 mg based on body surface area), followed by post-surgery adjuvant serplulimab monotherapy (up to 17 cycles) or adjuvant chemo (5 cycles), Q3W. The primary endpoint was investigator (INV)-assessed event-free survival (EFS). Efficacy superiority was first evaluated and established in patients with PD-L1 CPS ≥ 10 followed by that in the PD-L1 CPS ≥ 5 population. Results: Between November 26, 2019 and April 19, 2024, 588 patients were enrolled across 57 sites, with 292 randomized to the serplulimab group and 296 to the placebo group. As of the data cutoff date of August 19, 2025 with a median follow-up duration of 35.9 months, INV-assessed median EFS was significantly longer in the serplulimab group (n = 193) than in the placebo group (n = 217) in the PD-L1 CPS ≥ 10 population (not reached [NR], 95% CI 43.0–not estimable [NE] vs. 42.0 months, 95% CI 20.5–NE; HR 0.65, 95% CI 0.47–0.90; p = 0.0082), and in the PD-L1 CPS ≥ 5 population (NR, 95% CI 37.9–NE vs. 35.9 months, 95% CI 21.3–52.0; HR 0.73, 95% CI 0.56–0.94; p = 0.0152), meeting the protocol-specified superior criteria. Blinded independent central review-assessed median EFS was consistent with that of the investigator. Pathological complete response rate was higher in the serplulimab group than in the placebo group (21.6% vs 6.4%; odds ratio 3.95). Median OS was immature in both groups. Grade ≥ 3 treatment-related adverse events (TRAEs) were reported in 136 (46.6%) and 172 (58.5%) patients in the serplulimab and placebo groups, respectively. Discontinuation of any component of the study regimen due to TRAEs occurred in 19 (6.5%) and 31 (10.5%) patients in the respective groups. Conclusions: Neoadjuvant serplulimab plus chemo, followed by adjuvant serplulimab monotherapy, significantly prolonged EFS in patients with PD-L1-positive, resectable GC or GEJ adenocarcinoma. Improvements in other efficacy endpoints were also observed along with a favorable safety profile compared to neoadjuvant/adjuvant chemo. This world-first chemotherapy-free regimen (adjuvant) represents a promising treatment option for this indication. Clinical trial information: NCT04139135 .
Household costs of breast cancer and coping mechanisms in Kenya: A case of Kenyatta National Hospital.
11051 Background: In 2020, cancer was the third leading cause of mortality in Kenya, with breast cancer the most prevalent cancer and the second leading cause of cancer-related deaths. Patients and the health system faced a substantial economic burden due to breast cancer, which requires long-term medical attention as a chronic condition. Research has identified cost as a primary obstacle to accessing breast cancer healthcare services, particularly in low- and middle-income countries (LMICs). Objectives: The objective of this study was to examine the household costs of breast cancer among patients receiving treatment at Kenyatta National Hospital, Kenya’s national referral hospital. Specifically focusing on cost incurred, drivers of cost and cost coping mechanisms. Methods: Cross-sectional descriptive research design was utilized to quantify the cost of breast cancer among 103 breast cancer patients at the Kenyatta National Hospital in 2024. Data were collected via a semi-structured questionnaire on the direct medical costs, direct non-medical costs, and indirect costs incurred, as well as cost coping mechanisms. To assess the drivers of cost, multivariable linear regression model was used to determine associations between the independent variables and cost outcomes. Results: The mean total costs of breast cancer were estimated to be USD 14,255 (KES 2,138,330). Direct medical costs (USD 8,916/KES 1,337,475) were the largest component, followed by direct non-medical costs (USD 1,917/KES 287,625) and indirect costs (USD 3,421/KES 513,229). Targeted therapy was the largest cost component within direct medical costs, representing 45.37% of direct medical costs. Luminal B and Stage IV breast cancer incurred the highest total costs. Patients primarily relied on public health insurance and personal income and savings to cater for the costs. Despite all participants using the National Health Insurance Fund (NHIF), now Social Health Authority (SHA), it only covered 25-50% of the direct medical costs. Out-of-pocket expenditures were significant, with 97% of respondents using personal income and savings, 29% liquidating assets, and 22% taking loans. Conclusions: Breast cancer patients and their households incur significant costs in Kenya. This underscores the need for urgent health financing reforms that ensure broadened insurance coverage, and social protection systems that ensure Kenyans have access to comprehensive cancer care and prevent cancer-related financial toxicity and impoverishment.
Real-world safety of tarlatamab in patients receiving recent or concurrent CNS radiation.
e20126 Background: Tarlatamab, a delta-like ligand 3 (DLL3)-directed bispecific T-cell engager, has revolutionized the treatment of small cell lung cancer (SCLC) since FDA approval in 2024. However, data are limited regarding the safety of tarlatamab with recent or concurrent CNS radiation therapy (RT) for brain metastases and how recent CNS RT may impact risk of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). We describe CNS RT exposure relative to tarlatamab initiation and characterize associated toxicities. Methods: We conducted a retrospective chart review of patients with platinum-refractory SCLC and high-grade neuroendocrine carcinomas treated with tarlatamab between June 1, 2024, and June 1, 2025, at the University of Chicago Medical Center. Clinical data including demographics, performance status, disease characteristics, CNS RT history, and ICANS/CRS events were collected. Results: Twenty-four patients received tarlatamab during the study period (median age 64 years; 46% male; 75% White and 21% Black, median 2 lines of prior therapy [range 1-4]). Seven patients had brain metastases prior to tarlatamab initiation (four treated and three untreated with CNS RT). Overall, 13 of 24 patients (54%) developed CRS (n=5 grade 2 events), and 5 of 24 patients (21%) developed ICANS (n=1 grade 2 event). 3 of 7 patients (43%) with brain metastases developed CRS (n=1 grade 2 event) and 2 of 7 patients (29%) developed ICANS (n=1 grade 2). Eleven patients (46%) underwent CNS RT within 6 months of tarlatamab initiation (interval range: −6.0 to +5.5 months), including 4 patients who received CNS RT prior to tarlatamab initiation. Among patients who received CNS RT prior to tarlatamab, CRS occurred in 1 of 4 patients (25%) and ICANS in 2 of 4 patients (50%). No cases of radionecrosis were observed in patients treated with CNS RT during tarlatamab therapy. Conclusions: In this retrospective cohort, neither the presence of brain metastases nor treatment with CNS RT within six months of tarlatamab initiation was associated with clearly increased rates of CRS or ICANS. These real-world findings support the cautious use of tarlatamab in patients requiring CNS RT, with close monitoring. Larger studies are needed to further define the impact of RT timing on immune-related toxicities. Subgroup of patients receiving CNS RT within 6 months of tarlatamab treatment with CRS/ICANS events. Patient No. Duration of Tarlatamab Therapy (months) RT–Tarlatamab Interval (months) CRS Grade ICANS Grade 1 6.0 +2.1 months 1 none 2 4.0 +5.5 months 2 none 3 1.0 −2.0 months none none 4 5.7 −1.9 months 1 none 5 10.8 −4.0 months none 1 6 10.5 +0.7 months 2 none 7 1.9 +1.4 months none none 8* 9.5 +4.2 months 2 none 9 6.0 +0.5 months 1 none 10** 4.6 −6.0 months none 2 11 13.1 +2.5 months 2 none Majority of patients with SCLC. *Atypical carcinoid of the lung, transformed to SCLC. **Neuroendocrine tumor of the lung.
Association of genomic and clinical factors with outcomes with front-line chemoimmunotherapy in patients with advanced biliary tract cancer.
4135 Background: Anti-PD1/L1 therapy plus gemcitabine/cisplatin is standard first-line therapy in advanced biliary tract cancer (BTC), yet outcomes remain heterogeneous and prognostic biomarkers are poorly defined. We aimed to identify genomic and clinical factors associated with outcomes in advanced BTC patients (pts) treated with first-line gemcitabine, cisplatin and durvalumab (GCD). Methods: All pts with advanced BTC treated with first-line GCD at Memorial Sloan Kettering who underwent pre-treatment tumor genomic profiling (NCT01775072) were included. The primary objective was to study the association between genomic and clinical features and investigator-assessed progression free survival (PFS); evaluation of overall survival (OS) was a secondary objective. Cox regression analyses evaluated associations with PFS/OS. A multivariable Cox PFS model used variables selected by LASSO-based stability selection and was stratified by tumor subtype. Results: 236 pts were eligible for analysis. Tumor subtypes included intrahepatic cholangiocarcinoma (CCA) (56%), gallbladder cancer (23%), extrahepatic CCA (16%), and BTCs not otherwise specified (5%). Common genomic alterations were TP53 (36%), ARID1A (24%), CDKN2A (23%), KRAS (17%), IDH1 (13%), SMAD4 (12%), MTAP (11%) and ERBB2 (7.6%). With a median follow up of 23 months (mo), median PFS for the cohort was 9.1 mo (95% CI 7.7–11.0). OS at 12 and 24 mo was 68% (95% CI: 62–74) and 45% (95% CI, 38–53), respectively. On univariable analysis, KRAS and FGFR2 alterations were associated with shorter PFS (Table). KRAS alterations were also associated with an increased risk for all-cause mortality on univariable analysis (HR 1.81, 95% CI: 1.17–2.81, p=0.008). Median PFS and OS for the KRAS -altered group were 5.6 (95% CI: 3.6-7.8) and 11 (95% CI: 7.8-42) mo respectively, compared to 11 (95% CI: 8.1-12) and 21 (95% CI: 19-28) mo for KRAS -WT patients. On multivariable analysis, KRAS alterations and liver metastases were independently associated with shorter PFS; higher baseline albumin was protective (Table). Conclusions: Among pts with advanced BTC treated with front-line GCD, KRAS alterations were independently associated with poor PFS. These findings highlight the heterogeneity of benefit from GCD and support development of biomarker-informed strategies to refine patient selection and incorporation of KRAS-targeted strategies to improve outcomes in patients with KRAS -mutant BTCs. Factors associated with PFS with GCD. Analysis (95% CI), p Factor Univariable HR Multivariable HR Alteration KRAS 2.17 (1.48–3.19), <0.001 2.06 (1.36–3.12), <0.001 FGFR2 1.85 (1.02–3.34), 0.042 - Baseline CA19-9 1.14 (1.05–1.23), 0.002 - Metastases Liver 1.78 (1.25–2.54), 0.001 2.06 (1.41–3.01), <0.001 Bone 1.9 (1.05–3.43), 0.034 - Baseline albumin (per 1-unit increase) 0.43 (0.29–0.64), <0.001 0.44 (0.30–0.66), <0.001
Outcomes of endovascular embolization in cancer-associated gastrointestinal bleeding: A nationwide inpatient analysis.
e16340 Background: Advances in endovascular techniques have improved the technical success of embolization for gastrointestinal bleeding (GIB). However, embolization carries a risk of ischemic complications, particularly in colonic territories. This study aimed to evaluate the utilization patterns and clinical outcomes of endovascular embolization in gastrointestinal cancer-associated GIB. Methods: We performed a retrospective analysis of adult hospitalizations with cancer-associated GIB using the National Inpatient Sample database from 2016 to 2022. Cancer types included stomach, duodenum, pancreas, gallbladder, small intestine, and colon. Variables included age, sex, hospital size, severity (significant bleeding, invasive ventilation, blood transfusion, sepsis, peptic ulcer disease), and do-not-resuscitate status. Patients who underwent embolization were matched to non-embolized patients using propensity score matching with a logistic regression model. The covariates were used for nearest-neighbor matching using a caliper width of 0.01, and balance was assessed using standardized mean differences and variance ratios. In the matched cohort, crude mortality was assessed using χ² and mortality risk using logistic regression. Results: A total of gastrointestinal bleeding admissions was N = 47,085 stomach; 6,210 duodenum; 55,435 pancreas; 7,180 gallbladder; 8,400 small intestine; 106,165 colon, with endovascular embolization performed in 1.5%, 3.6%, 3.3%, 2.2%, 3.2%, and < 1% of cases, respectively. Embolized patients were similar or younger than non-embolized patients (stomach: 65 vs 65; duodenum: 66 vs 69; pancreas: 67 vs 66; gallbladder: 63 vs 68; small intestine: 66 vs 67; colon: 70 vs 69 years). Crude mortality did not differ between embolized and non-embolized patients with stomach (9.9% vs 9.7%), duodenal (11.1% vs 7.7%), pancreatic (12% vs 13.6%), gallbladder (6.5% vs 13.7%), or small intestinal cancers (11.3% vs 7.9%) (all p > 0.05), while higher crude mortality was observed in embolized patients with colon cancer (12.7% vs 6.7%, p < 0.05). After adjustment, embolization was associated with lower in-hospital mortality only in pancreatic cancer (aOR 0.5, p < 0.05). No significant associations were observed for stomach (aOR 0.9), gallbladder (aOR 0.04), or colon cancers (aOR 1.4) (all p > 0.05). Adjusted models for duodenal and small intestinal cancers were unstable due to complete separation from severe illness markers. Conclusions: Endovascular embolization was associated with significantly lower mortality in pancreatic cancer, while no benefit was observed for stomach, gall bladder, or colon cancers. Prospective studies are needed to define the optimal timing and modality of intervention, and hospital-level factors influencing mortality in these patients.
A phase 2 trial of androgen deprivation therapy interruption in patients responding exceptionally to androgen receptor pathway inhibitor in metastatic hormone-sensitive prostate cancer (A-DREAM / Alliance A032101).
5004 Background: Patients (pts) with metastatic hormone-sensitive prostate cancer (mHSPC) treated continuously with testosterone suppression (TS) plus an androgen receptor pathway inhibitor (ARPI) can experience cumulative toxicity. We conducted this single-arm phase 2 trial (NCT05241860) to test the hypothesis that pts who achieve favorable response to TS + ARPI can have prolonged treatment-free interval with testosterone recovery after treatment interruption (TI). Here we report on the primary endpoint. Methods: Eligible pts had mHSPC by conventional imaging with prostate-specific antigen (PSA) ≥ 5 ng/ml and testosterone ≥ 150 ng/dl prior to starting TS+ARPI, with PSA < 0.2 ng/ml and testosterone < 50 ng/dl at the time of enrollment after having received TS for 540-750 days and ARPI for ≥ 360 days. Prior local therapy, radiation to metastases and docetaxel were permitted. After enrollment, pts discontinued TS and ARPI and were followed with PSA and testosterone levels every 3 months (mo), CT/MRI and bone scan at least every 6 mo, and FACT-P questionnaire for patient-reported outcomes (PROs) every 6 mo. Treatment was resumed for PSA ≥ 5 ng/ml, radiographic change (progressive disease [PD] per RECIST 1.1 on CT/MRI or unconfirmed PD per PCWG3 on bone scan), or prostate cancer-related symptoms. The primary endpoint was proportion of men remaining treatment-free 18 mo after TI with eugonadal testosterone ( > 150 ng/dl). Target enrollment was 75 pts to differentiate 18-mo treatment-free rates of 0.30 (H 0 ) and 0.45 (H a ). Results: Of 79 pts enrolled between 07/2022 and 03/2024, 78 were eligible and underwent TI. Among these 78, median PSA prior to starting TS+ARPI was 19 (range 5-6759), 27 (35%) had high volume disease, 31 (40%) never received local therapy, and 55 (71%) did not receive radiation to metastases. By 18 mo after TI, 67% (52/78) recovered testosterone and 58% (45/78) remained treatment-free; 41% (32/78) remained treatment-free with testosterone recovery (80% CI 33.5-48.9%, one-sided p 0.0249). At median follow-up of 21.2 mo, 35% (27/78) resumed initial TS+ARPI after meeting re-initiation criteria (of whom 1 required treatment switch for PD 9 mo later); 5% (4/78) resumed prior to meeting criteria; 9% (7/78) pursued alternative therapy instead of resuming TS+ARPI per protocol; 5% (4/78) withdrew; 1 died of myocardial infarction prior to resuming treatment. 4 pts died, 1 of prostate cancer. Conclusions: The primary objective was achieved with 41% of favorable responders to TS+ARPI remaining treatment-free with testosterone recovery at 18 mo after TI. Analyses of biomarkers and PROs are in progress, and pts will be followed for long-term outcomes. Support: U10CA180821, U10CA180882, UG1CA189823, https://acknowledgments.alliancefound.org, Veracyte Inc. Clinical trial information: NCT05241860 .
Impact of PD-L1 and HPV biomarkers on immune checkpoint inhibitor efficacy in locally advanced HNSCC: A systematic review of randomized trials.
e18050 Background: The role of immune checkpoint inhibitors (ICIs) in locally advanced head and neck squamous cell carcinoma (LA HNSCC) remains uncertain, with randomized trials showing inconsistent results in heterogeneous populations. We conducted a systematic review of randomized trials evaluating ICI-based strategies in LA HNSCC, with outcomes stratified by PD-L1 expression, HPV/p16 status, and cisplatin eligibility to identify patients most likely to benefit from ICIs. Methods: MEDLINE and EMBASE databases were systematically searched up to January 10, 2026. Randomized controlled trials (RCTs) evaluating ICIs in patients with LA HNSCC were included. The primary outcome was progression-free survival (PFS). A inverse variance method was used to calculate the estimated pooled hazard ratio (HR) for PFS with 95% confidence interval (CI). Heterogeneity was assessed with Cochran’s Q test. Random effects model was applied. Results: A total of 3,605 patients from 7 randomized trials (5 phase III, 1 phase II/III, and 1 phase II) were included. These studies evaluated ICI-based strategies versus standard therapy in LA HNSCC, including ICI with radiotherapy (RT) in cisplatin-ineligible patients (NRG-HN004, and GORTEC 2015 PembroRad), ICI added to cisplatin-based chemoradiotherapy (KEYNOTE-412, and JAVELIN Head and Neck 100), perioperative pembrolizumab or postoperative nivolumab with standard of care treatment (KEYNOTE-689, and NIVOPOSTOP GORTEC 2018), and atezolizumab maintenance following definitive therapy (IMvoke010). In the overall population, no significant difference in PFS/EFS/DFS was observed between ICI and standard therapy (HR 0.90; 95% CI: 0.77–1.06; p=0.20). However in subgroup analyses stratified by PD-L1 expression, patients with PD-L1 positive tumors demonstrated improved PFS with ICIs compared with control (HR 0.78; 95% CI: 0.67–0.91; p<0.0001). In contrast, PD-L1 negative tumors demonstrated inferior PFS in the ICI arm (HR 1.31; 95% CI: 1.02–1.68; p=0.03). No significant differences in PFS were observed based on HPV or p16 status. A subset analysis of cisplatin-eligible LA HNSCC trials evaluating the addition of ICIs to standard therapy exhibited a similar pattern. ICI use in PD-L1 positive patients demonstrated significantly improved PFS (HR 0.76; 95% CI: 0.63–0.92; p<0.0001) while PD-L1 negative patients demonstrated decreased PFS (HR 1.28; 95% CI: 0.99–1.66; p=0.06). In cisplatin ineligible populations ICI regimens did not improve PFS compared with cetuximab plus RT. Conclusions: This study showed that although there was no significant difference in PFS/EFS/DFS in the overall population, the PDL-1 positive subgroup experienced significantly improved PFS with ICIs compared with control while the PDL-1 negative subgroup demonstrated inferior PFS in the ICI arm; these results were mirrored in the cisplatin-eligible subgroup.
Pre-treatment circulating tumour DNA (ctDNA) to predict survival and aid decision-making in head and neck cancer patients: 5-year outcome analysis.
e18017 Background: HPV-associated (HPV+) head and neck squamous cell carcinoma (HNSCC) has a favourable prognosis compared to HPV-independent disease (HPV-), forming the basis for de-escalation trials. However, 25% of HPV+ patients experience recurrence, with no biomarker to identify this high-risk subset. In Stage IV disease (HPV+ and HPV-), deciding which patients should have radical treatment remains challenging, particularly in borderline curative cases where morbidity is balanced against chance of cure. This study investigated whether pre-treatment somatic variants could identify high-risk patients within favourable subgroups, and predict suitability for curative treatment in advanced stage disease. Methods: This prospective observational study recruited 41 HNSCC patients undergoing curative surgery or chemoradiotherapy. HPV status was determined by immunohistochemistry/in situ hybridisation. Blood samples were taken pre-treatment for ctDNA extraction. Somatic mutations were detected using a next-generation sequencing panel targeting 17 genes. Libraries were sequenced on Illumina MiSeq or NextSeq platforms. Pre-treatment somatic variants were profiled using ClinVar/COSMIC databases and classified as benign, variants of unknown significance or pathogenic. Kaplan-Meier survival analysis and Cox regression were performed using Python 3.13 with univariate analysis to identify variables associated with disease-free survival (DFS) at 5 years. Log rank tests compared survival and Firth’s penalisation was used in complete separation. Results: Thirty-five patients remained in the study, and 25 were HPV+. Five-year DFS was 66% with disease-specific survival of 77%. Thirty-one somatic variants were identified across 13 genes. HPV- patients had worse DFS vs HPV+ patients (HR 3.9, 95CI:1.3-11.2, p=0.0124). Within HPV+ subgroup, those with pre-treatment circulating variants had a worse 5-year DFS compared to HPV+ with no variants (HR 9.7, 95CI:1-80), p=0.04). Within HPV-, pre-treatment pathogenic variants resulted in worse DFS compared to HPV- and non-pathogenic variants (HR 8.2, 95CI:0.95-70, p=0.055). Notably, TP53 variants were associated with relapse ≤28months. Stage IV disease had worse DFS vs Stage I-III (HR 5.3, 95CI:2-15, p=0.002). Furthermore, Stage IV with a circulating variant had worse DFS than Stage IV with no variants (log rank p=0.028). Stage I-III with circulating variants had worse DFS than those without (HR 10.2, 95CI:1-85, p=0.032). Conclusions: Pre-treatment somatic variants are a potential prognostication tool and with further validation, could be used in decision-making. Within HPV+ and disease-stage subgroups, they identified patients with worse DFS who may be suitable for additional monitoring, treatment escalation or palliation. Funding: ORACLE Cancer Trust, Oxford NIHR Biomedical Research Centre.