Browse Articles
Discover research articles across all indexed journals
Investigating the molecular characteristics underlying desmoid tumor clinical trajectory.
11583 Background: Desmoid tumors are locally aggressive mesenchymal neoplasms that primarily affect adolescents and young adults. They demonstrate a range of clinical behavior from spontaneous regression to refractory growth leading to morbidity and sometimes mortality. However, there are no prognostic biomarkers for patients with desmoid tumors and we have a limited understanding of the biology behind the range of clinical behavior. Thus, we aimed to characterize the clinical course of patients with banked desmoid tumors at our institution from across the age spectrum, and correlate clinical behavior with genetic alterations and gene expression signatures. Methods: We identified 64 adult patients through our adult sarcoma biobank and 5 patients through our pediatric biobank with desmoid fibromatosis and banked samples. Retrospective chart review was conducted including patient characteristics, disease features, treatment course, and outcomes. Genetic analysis was completed on 61 patients (43 targeted exome sequencing, OncoPanel). To characterize gene expression, we performed single nucleus RNA sequencing (snRNAseq) on 15 samples from 12 patients, with concurrent CTNNB1 genotyping to accurately identify tumor cells. Primary analyses included cell annotation of clusters and differential gene expression between progressing and stable/regressing tumors. Results: A total of 69 patients were included in this study. Median age at diagnosis was 35 years (range 2–83), with 68% female. Median tumor size at diagnosis was 5.2cm (range 0.7–25.0). Tumor location included abdominal wall (41%), mesenteric (25%), trunk/breast (14%), extremity (16%), and head/neck (5%). At time of tissue collection 39% of tumors were progressing, 23% were stable, and 10% were regressing (28% unknown). Among 61 sequenced tumors, 72% and 21% harbored CTNNB1 and APC mutations, respectively, and 7% showed no mutations in either gene. snRNAseq resulted in a dataset of over 180,000 high-quality nuclei. Cell composition differences were identified across the spectrum of disease states. Concurrent single-nucleus CTNNB1 genotyping revealed expression differences in the tumor cell compartment between progressing and non-progressing tumors, including collagen/extracellular matrix proteins and mesenchymal transcription factors. Conclusions: In this cohort of adult and pediatric desmoid tumors, integration of clinical annotation and molecular characteristics identified gene expression programs associated with tumor progression. These findings provide new insights into the biology underlying variable clinical trajectory in desmoid tumors, implicating altered collagen/extracellular matrix and mesenchymal transcription factor programs. Ongoing bulk RNA sequencing and immunohistochemical analyses are being performed to assess the potential of these signatures as predictive biomarkers and therapeutic targets.
The nodal nexus: Assessing the lymph node dynamics and outcomes in node-positive carcinoma of the penis.
e17030 Background: Although early-stage, organ-confined disease has favorable outcomes, lymph node metastasis is the strongest prognostic factor. Nearly one-third of patients present with nodal involvement, and 20–25% of clinically node-negative cases harbor occult metastases. Precise risk stratification in node-positive disease is therefore critical to optimize treatment and improve survival. Methods: This retrospective study analyzed medical records of 1200 patients treated for carcinoma of the penis at the Cancer Institute (WIA) between 1990 and 2022. Of these, 354 patients were pathologically node-positive and included in the final analysis. Demographics, tumor characteristics, nodal status, treatment modalities, and outcomes were recorded. Descriptive statistics summarized baseline data. Kaplan-Meier methods were used for survival analysis. Cox regression was employed to identify factors associated with overall survival (OS) and disease-free survival (DFS). Results: The cohort comprised 354 node-positive patients with a mean age of 52.7 years; 24.3% were smokers and 28.2% had comorbidities, mainly diabetes (10.5%) and hypertension (5.1%). Grade 2 tumors predominated (54.5%), with lymphovascular and perineural invasion seen in 7.6% and 3.2%, respectively. Nearly half were pN3 (48.6%), and extranodal extension (ENE) was present in 35.9%. Penile-preserving surgery was performed in 77.4%, with IIBD, PLND, and SIBD in 79.9%, 87.6%, and 88.4%, respectively. Adjuvant chemotherapy and PORT were administered in 20.1% and 5.4%. Recurrence occurred in 30.5%, predominantly nodal (12.7%). Five-year overall survival was ~60%, with a median OS of 100.3 months; median DFS was not reached, with 25% recurring within 13.8 months. Poor OS was associated with recurrence (HR 3.79), pN3 (HR 1.89), and pN2 disease (HR 1.75), while adjuvant chemotherapy was protective (HR 0.50). High LNR, ENE, bilateral nodal disease, and pelvic metastases correlated with inferior outcomes. Conclusions: Groin dissection is a key intervention in managing node-positive penile cancer. Nodal stage and recurrence are the strongest predictors of survival. Multimodality approaches, including early lymphadenectomy and adjuvant therapy, improve outcomes in high-risk patients. Despite available prognostic markers, most data remain retrospective and institution-specific. There is a need for multicenter prospective trials and individualized, multidisciplinary care strategies for this challenging subgroup.
Progestin-only implant and breast cancer risk in adolescents and premenopausal women: A systematic review and meta-analysis.
10594 Background: Progestin-only long-acting reversible contraception via subdermal etonogestrel implants (Implanon/Nexplanon) is widely used. However, the breast cancer (BC) risk signal potentially attributable to implant exposure remains under-characterized, limiting evidence-based, shared contraceptive counseling. Methods: We systematically searched PubMed/MEDLINE, Embase, and the Cochrane Library and screened reference lists for observational cohort and case-control studies comparing etonogestrel/progestin-only implant exposure versus non-use in adolescents and reproductive-aged/premenopausal women (as defined by each study) without prior BC. Studies restricted to hereditary breast cancer predisposition cohorts (BRCA1/2 pathogenic variant carriers) were excluded. The primary endpoint was incident BC (invasive with/without in situ). Maximally adjusted estimates were extracted and harmonized to odds ratios (ORs). Risk of bias was assessed using ROBINS-E. Pooled estimates were generated using inverse-variance random-effects meta-analysis. Heterogeneity was quantified with I2 and Tau2. Results: Five studies met inclusion criteria, including three nationwide cohort studies from Denmark/Sweden and two nested case–control studies from the UK and Australia. Across cohort studies, implant exposure contributed 1,659,309 person-years of follow-up with 510 incident BC events. The largest case-control dataset included 67,470 BC cases. Study-specific adjusted ORs ranged from 0.93 to 1.24. In pooled analysis, implant exposure was associated with an increased risk of incident BC (OR 1.23. 95% CI 1.19–1.27, P < 0.0001), with no between-study heterogeneity (I2 = 0%, Tau2 = 0). Conclusions: In large, contemporary population-based observational studies excluding hereditary breast cancer predisposition cohorts, use of etonogestrel/progestin-only subdermal implants is consistently associated with a modestly increased risk of incident breast cancer (~23%) compared with non-use. While the absolute risk increase is likely small in younger women, the consistent signal across study designs supports its relevance for individualized risk–benefit discussions and shared contraceptive counseling in routine clinical practice. Study (Year) Design / Country Population Implant Type Odds Ratio (95% CI) Mørch et al. (2017) Nationwide cohort, Denmark Women 15-49 y Progestin-only 0.93 (0.48-1.80) Fitzpatrick et al. (2023) Case-control, UK Women 16-39 y Progestin-only 1.22 (0.93-1.60) Hadizadeh et al. (2025) Register-based cohort, Sweden Women 13-49 y Etonogestrel 1.22 (1.11-1.35) Hultstrand et al. (2022) Nationwide cohort, Sweden Women 15-34 y Etonogestrel 1.22 (0.98-1.52) Tuesley et al. (2025) Case-control, Australia Women 20-55 y Etonogestrel 1.24 (1.17-1.32)
National practice patterns of adjuvant breast hypofractionated radiation therapy completion in early-stage breast cancer.
e13500 Background: The seminal nationwide examination of treatment completion between standard fractionation radiation therapy (SFRT; typically 5-6 weeks) and hypofractionated radiation therapy (HFRT; typically 3-4 weeks) assessed both treatments from 2004–2015, revealing that HFRT yielded significantly superior completion rates. (PMID 32693964) We sought to assess whether these trends have persisted in wake of the COVID19 pandemic and the increased adoption of ultrahypofractionated (one-week) radiation therapy. Methods: Using the National Cancer Database, we identified women with early stage, node-negative, invasive breast cancer treated with lumpectomy followed by whole-breast RT from 2018 through 2022. We defined SFRT as a dose of 1.8–2.0 Gray (Gy) per fraction, HFRT as 2.66–2.70 Gy/fraction, and ultrahypofractionated radiation therapy (URT) as 5.2-6.0 Gy/fraction. Completion of RT was defined as receiving a total dose of at least 46 Gy for SFRT, 40 Gy for HFRT, and 26 Gy for URT. Multivariable logistic regression models were used to evaluate predictors including fractionation type and distance from treatment facility. Results: A total of 34,043 patients were identified, with 95.9% completing RT. A majority of patients underwent HFRT (81.7%) with 97.4% completing their radiation courses. 11.0% of patients underwent SFRT while 7.3% underwent URT. HFRT utilization was 79.42% in 2018 and 78.27% in 2022. Despite similar rates of HFRT use, there was a significant increase in URT in that time (0.26% 2018 to 17.52% 2022; p < 0.0001) Age (>60 vs <=60) was associated with URT adoption (p<0.0001), with those over 60 more frequently adopting URT. Distance from treatment center was also associated with fractionation regimen (p trend <0.0001), with people living farther away more frequently adopting URT. Conclusions: From 2018–2022, use of URT increased, while utilization of SFRT decreased, reinforcing the ongoing shift towards shorter-course treatment in this patient population. HFRT regimens continue to demonstrate improved completion rates compared with SFRT, with rates of URT increasing commensurate with SFRT decrease in early-stage breast cancer RT. Further assessment of the impact of the COVID19 pandemic on these trends is ongoing.
Heterogeneity of stereotactic radiosurgery quality across the community.
e14049 Background: Stereotactic radiosurgery (SRS) is a ubiquitous technique for the treatment of brain metastases that demands highly localized and precise dose delivery. Standardized SRS audits: testing the delivery accuracy of an SRS treatment to a standardized anthropomorphic phantom, is routinely performed by the Radiation Quality Assurance Lab, including for participants in NCI clinical trials (as IROC). As these audits are performed according to each institution’s clinical processes, we analyzed SRS audit data to characterize variability across the community. Methods: We evaluated 930 irradiations (673 institutions, 2013–2023), stratifying for delivery system (GammaKnife, CyberKnife, TomoTherapy, and LINACs) and beam collimator. Radiotherapy quality metrics were examined (Conformity (CI), Gradient Index (GI), and V12Gy), as was the radionecrosis risk (NTCP) estimated using HyTEC model. Statistical analysis utilized ANOVA for subgroup differences and independent t-tests to compare pass vs. fail cohorts. Results: The overall audit pass rate, based on criteria established by the NCI’s clinical trial network, was 87%. Delivery platform strongly influenced the quality of the treatment (Table 1), and dedicated platforms typically showed distinct strengths. Gammaknife achieved the best dose fall-off and lowest V12Gy, while Cyberknife demonstrated superior overall conformity. Standard LINAC delivery yielded slightly higher normal tissue exposure, while the TomoTherapy platform exhibited the poorest conformity (0.67) and gradients (7.37). Crucially, and unexpectedly, suboptimal plan quality (i.e., worse designed dose distributions) predicted phantom delivery failure (i.e., poorer delivery accuracy). Institutions that struggled to deliver the plan dose accurately were the same as those that designed poorer plans (irradiating significantly larger volumes of normal brain (p<0.001). This association suggest a potential double-whammy for patients that may impact patient outcomes and even clinical trial results. Conclusions: SRS plan quality is intrinsically linked to delivery platform and hardware selection, and showed substantial variability across the community. Additionally, delivery accuracy was also heterogeneous, with 13% of institutions failing to meet accepted NCI standards. The quality of the treatment plan and the quality of the treatment delivery were significantly correlated, highlighting the need for ongoing evaluation of SRS. Category Pass (%) Conformity Gradient V12Gy[cc] Technique GK 62 (98%) 0.78 ± 0.08 2.7 ± 0.3 12.0 ± 1.7 CK 71 (90%) 0.86 ± 0.08 3.7 ± 0.8 14.1 ± 3.5 Elekta 148 (82%) 0.73 ± 0.11 3.7 ± 1.2 17.0 ± 7.3 Varian 513 (88%) 0.78 ± 0.11 3.2 ± 0.8 14.6 ± 10.5 Collimator LD 329 (89%) 0.76 ± 0.12 3.5 ± 0.9 16.4 ± 12.9 HD 195 (88%) 0.80 ± 0.08 3.2 ± 0.8 13.6 ± 4.7 Cone 101 (76%) 0.74 ± 0.11 3.3 ± 1.2 14.9 ± 7.1 Fail* Total: 118 (13%) 0.74 ± 0.13 3.7 ± 1.9 18.8 ± 17.3 Pass* Total: 812 (87%) 0.78 ± 0.11 3.4 ± 1.0 14.8 ± 7.9
Th1/Th2 immune polarization and cancer of unknown primary (CUP) risk in autoimmune disease: A global healthcare cohort study.
e14550 Background: Cancer of unknown primary (CUP) represents 2–5% of global malignancies and remains among the top 10 cancers despite advances in imaging, molecular diagnostics, and immunotherapy. Prior studies show increased CUP risk in autoimmune diseases, but the influence of inflammatory patterns is unclear. Emerging data suggest Type 2 inflammation promotes tumorigenesis through IL-4/IL-13–driven proliferation, Bcl-2/Bcl-xL–mediated apoptosis resistance, and M2/MDSC-mediated immunosuppression, whereas Type 1 inflammation enhances IFN-γ responses, M1 activation, and direct tumor cytotoxicity. We hypothesized that Th1 versus Th2 inflammatory signatures differentially affect CUP development. Methods: Using the TriNetX global research network (2010-2024), we identified adults (≥18 years) with a total of 36 autoimmune diseases stratified by predominant inflammatory pattern: Type 1 (rheumatoid arthritis, multiple sclerosis, Crohn's disease), Type 2 or mixed (Sjögren syndrome, SLE, dermatomyositis, ulcerative colitis), and vasculitis. Propensity score matching (1:1) generated matched controls without autoimmune disease. Cox proportional hazards models estimated hazard ratios (HR) and odds ratio (OD) adjusting for age, sex, race, smoking, and HIV/AIDS. Diagnosis of ADHD was included as internal control. Results: After matching, 4,334,934 patients were included (median follow-up 3.3 years). CUP occurred in 2,406 patients with autoimmune disease versus 1,101 controls (HR 2.19, 95% CI 2.04–2.35, p < 0.0001). Sixteen autoimmune diseases were associated with increased CUP risk. Stratification showed higher risk in Th2-dominant or mixed diseases (HR 2.50, 95% CI 2.43–2.57) compared with Th1-dominant diseases (HR 2.04, 95% CI 1.94–2.15), supporting a role for chronic Type 2 inflammation in creating an immunosuppressive tumor microenvironment. Addison’s disease showed the highest individual risk (HR 3.68, 95% CI 1.98–6.76), possibly reflecting chronic steroid exposure. Patients with multiple autoimmune diseases had markedly higher risk (HR 2.89, 95% CI 2.53–3.30) than those with a single disease (HR 1.88, 95% CI 1.74–2.03), suggesting cumulative immune dysregulation. Biomarker and treatment comparisons showed no significant differences in MSI-H or HER2 positivity. Likewise, there is no difference for use of platinum chemotherapy (OR 1.12, 95% CI 0.65–1.92) or HER2-targeted therapy (OR 0.82, 95% CI 0.50–1.36). CUP patients with certain autoimmune diseases were more likely to receive immune checkpoint inhibitors (OR 1.25, 95% CI 1.06–1.46). Conclusions: This large real-world study shows that Type 2 inflammatory autoimmune diseases confer higher CUP risk than Type 1 diseases, and that multiple autoimmune diseases further amplify risk, highlighting immune dysregulation as a key pathogenic driver.
Immune context as a modulator of MSI-high and TMB-high as predictors of immune checkpoint inhibitor response in breast cancer.
e17638 Background: Microsatellite instability–high (MSI-H) and tumor mutational burden–high (TMB-High) are tissue-agnostic biomarkers for immune checkpoint inhibitor (ICI) response. Their predictive value in breast cancer (BC) remains unclear due to low prevalence and heterogeneous immune contexts. We investigated the immune and transcriptomic features of MSI-H and TMB-High BC, with comparison to colorectal cancer (CRC) and uterine endometrial carcinoma (UCEC), where MSI-H/TMB-High are established predictors of ICI benefit. Methods: We analyzed TCGA BC (n = 402), CRC (n=355), and UCEC (n=173) cohorts, with an independent UAE BC cohort (n=85). Immune cell composition was inferred using xCell 2.0. Differential gene expression (DEG), gene set enrichment analysis (GSEA), and protein–protein interaction (PPI) network analyses were performed for TMB-High versus TMB-Low samples. Per-sample immune profiling was conducted for MSI-H BC cases treated with ICI. Results: MSI-H BC was rare (TCGA 0.5%; UAE 2.35%) and overlapped with TMB-High. In CRC and UCEC, MSI-H vs MSS comparisons revealed lymphoid-dominant immune-hot profiles, characterized by enrichment of CD8⁺ and CD4⁺ T cells, PD-1–expressing CD8⁺ T cells, NK cells, and B/plasma cells, consistent with established responsiveness to ICI. In TMB-High vs TMB-Low comparisons, CRC demonstrated coordinated immune activation supported by both immune cell enrichment and upregulation of immune-related differentially expressed genes, including pathways related to NK cell–mediated cytotoxicity and interferon signaling. In contrast, UCEC TMB-High tumors showed strong immune cell enrichment. In contrast, BC exhibited a distinct immune context. In TCGA BC, TMB-High samples showed significant immune differences relative to TMB-Low, but these reflected a mixed immune phenotype with concurrent activation of lymphoid, myeloid and regulatory populations. Comparison between TCGA MSI-H and MSS samples did not reveal significant difference in immune cell proportions. In the UAE BC cohort, TMB-High samples did not show enrichment of lymphoid immune populations, and MSI-H vs MSS comparisons revealed limited differences, with increased macrophage proportions observed in MSI-H cases. Per-sample analysis of MSI-H BC treated with ICI demonstrated that the responder exhibited coordinated adaptive immune infiltration resembling canonical MSI-H CRC/UCEC immune profiles, whereas the non-responder lacked such lymphoid engagement. Conclusions: MSI-H and TMB-High are tissue-agnostic biomarkers of ICI response. Canonical MSI-H tumor types such as CRC and UCEC exhibit lymphoid-dominant, immune-hot phenotypes that most reliably translate into clinical benefit. In contrast, MSI-H BC display a mixed immune context, often constrained by myeloid and stromal components, limiting the predictive utility of MSI-H and TMB-High status alone.
Predictive markers of response and toxicity to tarlatamab in patients with extensive stage small cell lung cancer (ES-SCLC): A multi-institutional real-world analysis.
8097 Background: Tarlatamab (tarla) is a novel immunotherapeutic agent for the treatment of relapsed ES-SCLC associated with favorable response and survival outcomes. While tarla’s unique side effect profile of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) is well described, factors influencing response or toxicity have yet to be fully established. Methods: We performed a retrospective, multivariate analysis of patients with ES-SCLC receiving tarla at three NCI-designated comprehensive cancer centers from 7/2024-10/2025. CRS and ICANS were reported using ASTCT consensus guidelines. Dysgeusia was graded using CTCAE v5.0. Overall response rate (ORR) was determined using clinical interpretation of radiographic assessments from patients with available post-treatment imaging. Time to event analyses were performed. Cox Proportional Hazards regression analyses assessed the association between patient characteristics and PFS and OS outcomes. Logistic and ordinal regression analyses evaluated the predictive value of variables contributing to the presence and severity, respectively, of CRS and ICANS. Results: 115 patients received at least 1 dose of tarla, with median age of 67 years, 55% female and ECOG PS ranging from 0-3 (median of 1). Rates of observed CRS, ICANS, and dysgeusia were 46%, 28%, and 47%, respectively. ORR was 44% and 6-month PFS/OS rates were 30.4% (95% CI: 22.5-41.1%)/54.9% (95% CI: 45.4-66.4%). Excluding patients on tarla <30 days, presence of dysgeusia was associated with a lower rate of progression (HR 0.44, p = 0.00691) and death (HR 0.34, p = 0.0145). Neither CRS nor ICANS were associated with treatment response. Both sum of liver metastasis diameter >10 cm (OR 4.4, p = 0.02) and elevated baseline LDH (p = 0.01) were significantly associated with increased likelihood of CRS, while female sex was a protective factor (OR 0.33, p = 0.01). Increased ECOG PS (OR 2.3, p = 0.02), sum of brain metastasis diameter >2 cm (OR 3.6, p = 0.02), development of CRS (OR 2.8, p = 0.04), elevated baseline LDH (p = 0.04) and elevated baseline ferritin (p =0.004) were all associated with increased risk of developing ICANS. Conclusions: In our multicenter cohort of patients with ES-SCLC, tarlatamab continued to demonstrate anticancer activity with similar ORR and 6 month PFS compared to existing DeLLphi-304 data. Importantly, the presence of dysgeusia was found to be a marker of treatment response associated with improved PFS and OS. CRS and ICANS incidence was associated with several baseline patient and disease characteristics. Collectively, these preliminary results identify several predictive biomarkers of toxicity that can be utilized in clinical practice, though validated models are needed to better stratify low and high-risk populations.
Leveraging single-cell trajectory analysis and machine learning to build a robust prognostic classifier for triple-negative breast cancer to pinpoint CCDC171 as a master regulator.
1136 Background: Triple-negative breast cancer (TNBC) is limited by high heterogeneity and a lack of robust biomarkers for risk-stratified management. While chemotherapy is standard, traditional bulk signatures fail to capture high-resolution cellular transitions driving recurrence. This study integrates scRNA-seq with multi-algorithmic machine learning to identify key regulatory programs. We developed a trajectory-informed prognostic classifier to improve individual risk assessment and uncover potential therapeutic targets in the TNBC tumor microenvironment (TME). Methods: We performed single-cell RNA sequencing (scRNA-seq) on 24 TNBC samples to map the cellular landscape and applied pseudotime trajectory analysis to identify dynamic gene programs associated with cell-fate decisions. Prognostic genes derived from trajectory-informed differential expression were integrated with bulk RNA-seq data from TCGA-TNBC and validated in two independent cohorts (GSE58812, GSE135565). A robust prognostic model was constructed using an extensive machine-learning framework combining forward stepwise Cox regression and gradient boosting (StepCox[forward] + GBM). Model performance was evaluated using Harrell’s C-index and Kaplan–Meier analysis, and predictive utility was assessed in neoadjuvant treatment cohorts. Virtual gene knockout experiments were used to analyze the function of CCDC171. Results: We developed an 18-gene prognostic signature that robustly stratified TNBC patients into high- and low-risk groups across multiple cohorts (TCGA: HR = 55.12, p = 2.9×10 -12 ; GSE58812: HR = 3.96, p = 6.3×10 -4 ; GSE135565: HR = 7.73, p = 4.0×10 -3 ). The model outperformed existing prognostic signatures and generalized to non-TNBC breast cancer subtypes. Four genes (CRISP3, PDZK1IP1, CCDC171, IGFL1) were consistently retained across top-performing algorithms. CCDC171 emerged as a central regulator whose expression correlated with poor survival and coordinated a network involving complement activation (C1QB), metabolic reprogramming (CA8), and mitochondrial stress response (MTRNR2L8). In silico knockout of CCDC171 downregulated these effectors and suppressed oncogenic pathways including PI3K–Akt signaling, ECM interaction, and platinum resistance. Conclusions: Our study presents a TNBC prognostic model grounded in TME cellular dynamics and machine learning, with superior predictive accuracy and cross-subtype applicability. CCDC171 is identified as a novel master regulator of a multi-effector network driving tumor aggressiveness, offering a potential therapeutic target. This integrative framework supports risk-stratified management and personalized therapy in breast cancer.
Impact of <i>KRAS</i> co-mutations on the efficacy of durvalumab in patients with locally advanced non–small cell lung cancer (LA-NSCLC) treated with concurrent chemoradiotherapy (ChRT).
8550 Background: Durvalumab improves survival after platinum-based ChRT in LA-NSCLC. Although immunotherapy is effective in KRAS-mutant tumors across treatment lines, co-mutations are associated with poorer prognosis and treatment resistance. Data efficacy in this context remains limited. Methods: We conducted a retrospective, single-institution analysis of patients with locally advanced NSCLC to evaluate the impact of KRAS mutations on progression-free survival (PFS) and overall survival (OS) after concurrent chemoradiotherapy (ChRT) and durvalumab, accounting for key clinical factors and co-mutations (TP53, STK11, KEAP1). Patients with other actionable genomic alterations (EGFR, ALK, ROS1, BRAF, MET, HER2, RET, FGFR) were excluded. PFS was defined from the first durvalumab dose to progression or death and estimated using the Kaplan–Meier method with comparisons by log-rank test. Results: A total of 290 patients were included (113 KRAS-mutant, 177 non-KRAS [non-AGA]). Median age was similar (69.7 vs 68.5 years), and baseline characteristics were comparable, except that KRAS-mutant patients were more often female(65% vs 38%, p < 0.001) and enriched for adenocarcinoma histology (90% vs 42%; squamous 2.7% vs 49%; p = 0.009) and PD-L1 expression (negative: 34% vs 48%; 1–49%: 25% vs 28%; ≥50%: 41% vs 34%). No significant differences in PFS were observed between KRAS and non-KRAS patients (median PFS: 22.9 months [95% CI, 16.0–29.2] vs 19.6 months [95% CI, 13.4–35.2]; log-rank p = 0.087), nor in overall survival (OS) (median OS: 44.5 months [95% CI, 34.4–NR] vs 40.9 months [95% CI, 30.0–59.4]; log-rank p = 0.40). In contrast, PFS differed significantly across KRAS-mutant subgroups: isolated KRAS mutations (median PFS 23.9 months [95% CI, 17.4–30.0]), KRAS with STK11 or KEAP1 co-mutations (8.3 months [95% CI, 3.8–20.8]), and concurrent KRAS/STK11/KEAP1 mutations (3.0 months [95% CI, 2.7–NR]; log-rank p < 0.0001). OS showed a similar numerical trend without statistical significance (49.4 months [95% CI, 37.4–NR], 27.3 months [95% CI, 18.4–NR], and 10.0 months [95% CI, 6.7–NR], respectively; log-rank p = 0.08). Conclusions: While KRAS mutations alone were not associated with inferior outcomes with durvalumab, concurrent STK11 and KEAP1 co-mutations were associated with markedly worse clinical outcomes.
Effects of baseline neurological disorders on composite gastrointestinal toxicity risk in cancers of the lip, oral cavity, and pharynx, and in breast cancer through the gut-brain axis: A propensity-matched cohort study.
e18127 Background: Neurological disorders (NDs) such as depression and anxiety may influence the gut-brain axis, potentially exacerbating gastrointestinal adverse events during cancer treatment. However, their impact on adverse events (AEs) in cancers of the lip, oral cavity, and pharynx (CLOP) and breast cancer (BC) remains under explored. We evaluated the association between baseline NDs and the incidence of gastrointestinal AEs within 180 days following diagnosis in patients with CLOP or breast cancer. Methods: This study leveraged retrospective cohort TriNetX federated health research network database. Adults aged 18 years or older diagnosed with CLOP or BC were divided into cohorts with and without baseline NDs. Propensity score matching (1:1) was applied for demographics, race, ethnicity, body mass index, and ECOG performance status. Primary outcomes were the risk of gastrointestinal AEs (diarrhea, mucositis, nausea/vomiting, digestive system diseases (DD), fluid/electrolyte imbalance (FE), dry mouth (DM), GI mucositis (GIM), and a composite measure (CM) of all AEs) assessed via risk ratios (RR), hazard ratios (HR) from Kaplan-Meier survival analysis, and number of instances, excluding patients with prior outcomes. Results: For CLOP (64,558 matched patients per cohort) with NDs had higher risks--for diarrhea, RR was 2.39 (95% CI, 2.23-2.56) and HR 2.28 (95% CI, 2.12-2.45); mucositis, RR 1.44 (95% CI, 1.38-1.50) and HR 1.38 (95% CI, 1.32-1.44); nausea and vomiting, RR 1.74 (95% CI, 1.67-1.81) and HR 1.68 (95% CI, 1.60-1.75); DD, RR 1.24 (95% CI, 1.19-1.27) and HR 1.19 (95% CI, 1.16-1.23); FE, RR 1.52 (95% CI, 1.46-1.58) and HR 1.44 (95% CI, 1.38-1.50); DM, RR 1.56 (95% CI, 1.48-1.66) and HR 1.48 (95% CI, 1.39-1.58); GI mucositis, RR 2.01 (95% CI, 1.75-2.29) and HR 1.89 (95% CI, 1.65-2.17); CM, RR 1.56 (95% CI, 1.52-1.61) and HR 1.54 (95% CI, 1.49-1.59). For BC (315,342 matched patients per cohort), similar patterns emerged: DD, RR 1.78 (95% CI, 1.75-1.81) and HR 1.72 (95% CI, 1.69-1.75); diarrhea, RR 2.48 (95% CI, 2.39-2.57) and HR 2.35 (95% CI, 2.26-2.44); mucositis, RR 2.39 (95% CI, 2.25-2.53) and HR 2.25 (95% CI, 2.12-2.38); nausea and vomiting, RR 1.98 (95% CI, 1.93-2.01) and HR 1.89 (95% CI, 1.84-1.94); FE, RR 2.01 (95% CI, 1.95-2.07) and HR 1.89 (95% CI, 1.84-1.95); GIM, RR 2.31 (95% CI, 2.17-2.48) and HR 2.18 (95% CI, 2.04-2.33); DM, RR 3.38 (95% CI, 2.97-3.88) and HR 3.18 (95% CI, 2.78-3.62); CM, RR 1.98 (95% CI, 1.94-2.02) and HR 1.89 (95% CI, 1.86-1.94). Conclusions: Baseline ND are associated with significantly higher risks of gastrointestinal AEs in the 6 months following CLOP or BC diagnosis, suggesting a role for the gut-brain axis in treatment toxicities. Screening for neurological conditions may inform personalized supportive care strategies.
Association of SPOP mutation with survival outcomes following <sup>177</sup> Lu-PSMA-617 radioligand therapy in metastatic castration-resistant prostate cancer (mCRPC).
e17055 Background: 177 Lu-PSMA-617 ( 177 Lu) radioligand therapy (RLT) is a standard treatment for mCRPC after progression on androgen receptor (AR) pathway inhibitors, with or without prior taxane chemotherapy. Clinical outcomes remain heterogeneous, and predictive genomic biomarkers are lacking. SPOP-mutant prostate cancers exhibit relative genomic stability and preserved AR signaling, which may confer differential sensitivity to PSMA-targeted RLT. We aimed to evaluate the association between SPOP mutations status and treatment outcomes following 177 Lu RLT. Methods: We performed a retrospective analysis of mCRPC patients treated with 177 Lu at a single academic center (2023-2025). Clinical and genomic data were extracted from the EMR. Genomic profiling focused on DNA damage repair (DDR) genes, tumor suppressor genes (TSGs: TP53, RB1, PTEN), and SPOP mutation status. Overall survival (OS) and radiographic progression-free survival (rPFS) were assessed using Kaplan-Meier and Cox proportional hazards models. Results: Among 43 evaluable patients (median age 67 years; 69% Gleason 8-10, 84% received prior taxane therapy), 6 (13.9%) were SPOP-mutant and 37 were wild-type (WT). SPOP mutation was associated with a trend toward improved outcomes, with hazard ratio (HR) 0.20 for OS (p = 0.1) and 0.47 for rPFS (p = 0.1). Identified SPOP mutations clustered within the substrate-binding MATH (Meprin and TRAF Homology) domain, most commonly at known hotspot residues Y87 (n = 3), F102 (n = 1), and F133 (n = 1). Median OS was not reached in SPOP-mutant patients compared with 13.6 months in SPOP-WT patients (log-rank p = 0.09), while median rPFS was 9.0 vs. 5.1 months, respectively (log-rank p = 0.1). SPOP mutation was not significantly associated with baseline clinical, metastatic, genomic, or treatment-related characteristics. In exploratory Cox models with limited adjustment for key clinical variables, the association between SPOP mutation and improved OS (HR range 0.16-0.22) and rPFS (HR range 0.38-0.50) remained directionally consistent. Conclusions: In this real-world cohort of mCRPC patients, SPOP mutation was associated with a consistent trend toward improved survival following 177 Lu therapy. Despite limited numbers of SPOP-mutant cases, these findings suggest that SPOP-mutant mCRPC may represent a biologically distinct subgroup with enhanced sensitivity to PSMA-targeted RLT that is not readily explained by baseline disease burden or treatment history, supporting further investigation in larger, prospective biomarker-driven studies.
Exploring the genotoxic and tumorigenic potential of artificial food dyes and implications for colorectal cancer risk in socioeconomically disadvantaged communities.
e22580 Background: The artificial food dyes Blue No. 1, Blue No. 2, and Red 40 are common in processed foods and have been increasingly scrutinized for their potential biological effects. Experimental evidence links these dyes to DNA damage, immune dysregulation, and colonic inflammation, raising concerns about their possible role in cancer-related pathways despite regulated safety thresholds. Exposure is disproportionately higher in socioeconomically disadvantaged communities due to reliance on ultra-processed foods, food deserts, and constraints within nutrition assistance programs like SNAP. As early-onset colorectal cancer incidence rises, these populations face uneven burdens. Chronic dietary exposure to synthetic dyes may influence colorectal tumorigenesis through inflammatory signaling and PANX1-mediated immune responses. This study synthesizes current evidence on artificial food dyes and nutritional inequity to evaluate their potential contribution to EOCRC risk in vulnerable groups. Methods: A narrative literature review was conducted using peer-reviewed studies and regulatory reports from PubMed, NIH-affiliated databases, major scientific journals, and federal agencies, with findings synthesized qualitatively to assess biological pathways and EOCRC disparities related to Blue No. 1, Blue No. 2, and Red 40. Results: Across peer-reviewed studies, Blue No. 1, Blue No. 2, and Red 40 were associated with biological processes relevant to colorectal carcinogenesis. Blue No. 1 showed genotoxic effects in human lymphocytes, including DNA strand breaks, an early event in colorectal tumorigenesis. Blue No. 2 exposure was linked to tumor formation under certain experimental conditions and caused reproductive toxicity and tissue disruption in animal models, suggesting dysregulated cellular repair pathways. Pannexin-1 (PANX1) hemichannel activity promoted colonic inflammation, pathogenic CD4⁺ T-cell responses, and tumor progression, with elevated PANX1 expression associated with worse survival in colon cancer patients. Although Blue No. 1 inhibits PANX1 activity, its broader genotoxic effects suggest multiple, potentially opposing biological interactions. Red 40 consistently exacerbated colitis and immune-mediated intestinal inflammation through microbiome-dependent pathways and pro-inflammatory cytokine signaling. Population-level studies showed that low-income and SNAP-participating populations consume more ultra-processed foods and experience worse EOCRC outcomes, highlighting structural exposure risks. Conclusions: Synthetic food dyes may contribute to EOCRC through genotoxic, inflammatory, and immune-signaling pathways. Structural inequities increase exposure in low-income communities, highlighting the need for human based studies to clarify risks and guide policy.
First-in-human phase I/II study of a highly selective FGFR2 inhibitor 3HP-2827 in patients with advanced solid tumors harboring <i>FGFR2</i> alterations.
3114 Background: Molecular profiling of cholangiocarcinoma (CCA) is highly recommended to guide access to targeted therapies such as FGFR2 inhibitors. 3HP-2827 is a highly selective and potent FGFR2 inhibitor targeting FGFR2 alterations. Here we report the findings of 3HP-2827 in advanced solid tumors harboring FGFR2 alterations. Methods: This is a phase I/II study to evaluate the safety, tolerability, PK, and preliminary efficacy of 3HP-2827 in pts with advanced solid tumors harboring FGFR2 alterations. Pts failed to standard therapy with/without prior pan-FGFR inhibitor (FGFRi) are eligible for enrollment. Bayesian Optimal Interval (BOIN) and back filling design were employed in the dose-escalation stage of phase I, followed with the expansion stage. Treatment-related adverse events (TRAEs), PK, and anti-tumor activity (RECIST v1.1) were assessed. Results: As of Jan 8, 2026, a total of 52 pts (46 with CCA, 6 with others) were enrolled from 12 centers, received 3HP-2827 at doses of 60-180 mg once daily (QD), including 42 with FGFR2 fusion/rearrangement (f/r), 6 with FGFR2 mutation. 19 pts (36.5%) had prior FGFRi. No dose-limiting toxicities (DLTs) were observed at the dose levels of 60 mg, 120 mg, 180 mg QD. 3HP-2827 had favorable PK with doses≥120 mg QD providing FGFR2 occupancy＞90%. The most common TRAEs were FGFR2 on-target toxicities including dry mouth (69.2%, G3 1.9%), nail toxicity (67.3%, G3 3.8%), stomatitis (50%, G3 3.8%), dry eyes (26.9%, G3 1.9%), PPE (19.2%, G3 3.8%). 12 pts (23.1%) experienced at least one G3/G4 TRAE and no G5 TRAEs occurred. 7 pts (13.5%) experienced at least one serious adverse event, 4 (7.7%) of them were considered related to 3HP-2827. No discontinued treatment or death due to TRAEs. Anti-tumor activities were observed from 60 mg QD. Among 15 CCA pts (FGFR2 f/r, FGFRi-naive) who had at least one post-treatment tumor assessment, objective response rate (ORR) was 80% (95% CI: 51.9, 95.7), and disease control rate (DCR) was 100% (95% CI: 78.2, 100). The ORR and DCR in 15 CCA pts (FGFR2 f/r, FGFRi-refractory) was 26.7% (95% CI: 7.8, 55.1) and 86.7% (95% CI: 59.5, 98.3). The ORR and DCR in 4 advanced solid tumor pts (FGFR2 mutation, FGFRi-naive) was 75% (95% CI: 19.4, 99.4) and 100% (95% CI: 39.8, 100). Conclusions: 3HP-2827 was found to be safe and tolerable, with encouraging anti-tumor activity observed in pts with FGFR2 alterations. Further clinical efficacy will be explored in the expansion stage and phase II stage. Clinical trial information: NCT06378593 .
Artificial intelligence (AI) foundation model as a predictor of efficacy of next-generation checkpoint inhibition with botensilimab (BOT) + balstilimab (BAL) in solid tumors using pretreatment H&E images.
2535 Background: Predicting immunotherapy response is challenging in treatment-refractory/resistant (R/R) tumors where heterogeneity limits conventional biomarker utility. BOT (Fc-enhanced anti–CTLA-4) augments T-cell priming, depletes Tregs, and activates antigen-presenting cells. BOT+BAL (anti–PD-1) has shown activity across “cold” and R/R tumors including PD-L1–low and tumor mutational burden–low disease; thus, non-conventional predictive biomarkers are needed. A self-supervised AI foundation model applied to routine pretreatment H&E images was used to infer spatial transcriptomics and identify features associated with therapeutic benefit from BOT+BAL in microsatellite-stable colorectal cancer (MSS CRC), sarcoma, and ovarian cancer. Methods: A self-supervised AI foundation model was trained to infer spatial transcriptome expression from H&E images using purpose-built multimodal training data from thousands of human tumors profiled with multimodal assays. Pretreatment H&E images (not used for training) from 121 BOT+BAL–treated patients (pts) with R/R MSS metastatic CRC (included 20 pts with liver metastases), sarcoma, and ovarian cancer from the phase 1b C-800-01 trial (NCT03860272) were analyzed. Pt-specific embeddings were used to fit cross-validated classifiers predicting BOT+BAL clinical benefit (defined as complete or partial response [CR/PR] or stable disease [SD]). Results: Fitted logistic regression classifiers separated responders and non-responders in all tumor types, as measured by area under the receiver operating characteristic curve (AUROC; ranges from 0.5 [chance] to 1.0 [perfect separation]). Cross-validated AUROC was 0.61 in MSS CRC, 0.67 in sarcoma, and 0.77 in ovarian cancer (table; shows all findings). The model-recommended population is predicted to have a higher rate of clinical benefit, as measured by cross-validated precision. Beyond this binary analysis, the concordance index (C-index; measures accuracy and ranges from 0.5 [chance] to 1.0 [perfect]) for predicting overall survival (OS; via Cox proportional hazards model) was >0.5 in all tumor types, and highest in ovarian cancer. Conclusions: A self-supervised AI foundation model applied to routine pretreatment H&E images predicted BOT+BAL responses in MSS CRC, sarcoma, and ovarian cancer. These findings support AI-derived, H&E–based biomarker strategies for BOT+BAL and warrant prospective validation. Sample characteristics and model predictions. MSS CRC Sarcoma Ovarian Sampled population a No. of pt samples 67 30 24 Pts with CR/PR or SD 65% 57% 54% Modeled population AUROC 0.61 0.67 0.77 OS prediction, C-index 0.58 0.69 0.78 Model recommended population Population size 43% 50% 46% Estimated pts with CR/PR or SD 76% 67% 73% a Data cutoff: Mar 13, 2025; analyses ongoing.
Everolimus and aromatase inhibitors for advanced and recurrent low-grade serous ovarian carcinoma.
5569 Background: Treatment options for advanced and recurrent low grade serous ovarian carcinoma (LGSOC) are limited. Mutations in the PI3K/AKT/mTOR pathway have been described in LGSOC, but the use of everolimus and aromatase inhibitors (AI) has not been studied. Our primary objective was to determine the disease control rate (DCR) of everolimus (used off-label) and an AI in patients with LGSOC. Secondary objectives included progression free survival (PFS) and overall survival (OS). Methods: Diagnosis codes and medication lists were used to identify all patients with LGSOC from a single tertiary-care institution who were treated with everolimus and an AI between 2000- 2025. Data were extracted from the electronic medical record including demographics, staging, treatment, CA-125, imaging studies, reported side effects, recurrence, and survival data. Response was defined with RECIST 1.1 when measurable disease was present, otherwise clinical response was used. CA-125 was used independently as an additional measure of response. Results: A total of eight patients were identified, six with stage IIIC disease and two with stage IVB disease. Patients had between 0 and 4 debulking surgeries. Prior to treatment with everolimus and an AI, patients had 1-6 prior lines of therapy. All patients had received prior endocrine therapy, and five received chemotherapy. All patients were treated with everolimus 10mg and an AI (6 with letrozole and 2 with anastrozole). Patients remained on this regimen between 2 and 20 months (median = 15 months). One patient continues treatment today after 21 months of treatment. By RECIST criteria, two patients had a partial response, and four patients had stable disease. One patient has had clinically stable disease (soft tissue thickening and a stable CA-125) for 21 months and continues treatment. When comparing initial CA-125 to nadir while on treatment, CA-125 was stable in three patients and decreased 29%-64% (median = 50.5%) in four patients. One patient did not have CA-125 available in the EMR. Overall, the DCR was 87.5% (7 of 8 patients). Median PFS was 16.8 months, and median OS was 26 months. Of the seven patients who discontinued treatment, reasons for discontinuation included progression (n=3), death from disease (n=2), and inability to tolerate (n=2, one with grade 3 pneumonitis). Conclusions: In a group of eight patients with advanced LGSOC, everolimus and an AI demonstrated a DCR of 87.5%, a median PFS of 16.8 months, and a median OS of 26 months. While larger retrospective studies or prospective studies are needed, this regimen should be considered for advanced, recurrent LGSOC.
Longitudinal plasma and CSF ctDNA profiling as used to define therapeutic landmarks and compartmentalized clonal evolution in rhabdomyosarcoma.
11517 Background: Current risk stratification in rhabdomyosarcoma (RMS) relies on static pretreatment variables, which fail to capture dynamic tumor evolution and molecular residual disease under therapeutic pressure. Although liquid biopsy offers a non-invasive solution, clinical implementation has been constrained by a reliance on retrospective datasets and the lack of prospectively validated therapeutic landmarks or compartmentalized monitoring strategies. Methods: We conducted the largest multi-center, prospective liquid biopsy study to date, enrolling 143 treatment-naïve RMS patients (Aug 2020–Feb 2025; 91.6% intermediate/high-risk). We utilized ultra-deep (>5,000×) sequencing (506-gene panel) on matched tumor, longitudinal plasma (n=292), and cerebrospinal fluid (CSF) from patients with parameningeal or central nervous system (CNS) metastatic disease. Critically, paired tissue and plasma were collected at disease progression (PD) to map evolutionary trajectories. Genomic alterations were correlated with progression-free survival (PFS) and overall survival (OS) using univariate and multivariate Cox regression. Results: Baseline profiling identified MYOD1 mutation as the strongest independent predictor of inferior PFS (HR 6.52, P=0.003), while MYCN amplification defined a novel "ultra-high-risk" fusion-positive subset (OS, P<0.001). Baseline ctDNA positivity (66.7%) correlated with advanced disease (P<0.001). Longitudinally, persistent ctDNA at Cycle 5 Day 1 (C5D1) conferred a 12-fold increased progression risk (HR=12.2, P=0.004), establishing C5D1 as the definitive prognostic landmark. Patients with positive longitudinal ctDNA harbored a nearly 6-fold increased risk of progression (HR = 5.97, 95% CI 2.25–15.85, P < 0.001) and a 12-fold increased risk of death (HR = 12.4, 95% CI 1.42–108.37, P = 0.023) compared to those with sustained negativity. Analysis of paired samples at relapse revealed subclonal replacement driven by therapeutic pressure, enriching for acquired resistance alterations in TP53 and DNA damage response (DDR) pathways undetectable by static assays. Additionally, paired CSF-plasma profiling in parameningeal/CNS disease identified the CNS as an evolutionary "sanctuary site" harboring distinct, compartment-specific drivers (e.g., TP53, HRAS) divergent from plasma. Conclusions: This prospective study establishes ctDNA as a critical biomarker for RMS, defining MYOD1, MYCN, and TP53 as key genomic risk factors and validating C5D1 as the optimal therapeutic landmark. Deep longitudinal profiling uncovers subclonal replacement and compartmentalized CNS evolution as drivers of treatment failure, advocating for comprehensive, compartment-specific monitoring in next-generation risk-adapted trials. Clinical trial information: NCT05778955 .
Impact of advanced practice provider–led new consults in community oncology: Expanding access and volume in benign hematology.
1531 Background: Community oncology practices face rising patient volumes and workforce shortages, leading to delays in access to care. Although Advanced Practice Providers (APPs) are increasingly utilized to expand capacity, the structure and impact of APP-led new consult models, particularly in benign hematology, remains unclear. This analysis evaluated the operational and clinical impact of APP-led benign hematology consult programs to determine whether this model improves timely access and supports high-quality, patient-centered care. Methods: A retrospective, multi-site analysis was conducted in four community oncology practices within The US Oncology Network that implemented APP-led new consult models in benign hematology. Key performance indicators included total new consult volume, time to first appointment (T2F), percentage of APP visits, and APP engagement. Data were obtained from internal scheduling platforms, productivity dashboards, and operational reports. Pre- and post- implementation comparisons assess changes in access and throughput. Results: Across all sites, APP-led consults significantly enhanced access and capacity. In a four-year period at Practice A, median T2F decreased by 50% (24 to 14 days; 36 to 18 days across two regions). During the same timeframe, monthly new consults increased from 208 to 1,045, and total new patient volume rose 403% (2,495 to 12,544). APP participation expanded from 93 to 162 clinicians. In Practice B, APP-led consults increased from 55 (2024) to 406 (2025; 638% YOY growth), with daily new visits rising from 0 to 3 per day and APP share of total practice visits increasing from 0% to 23%. The model generated 295 physician hours saved and reduced T2F for new cancer patients from 8 to 4 days. Physicians and patients expressed confidence and satisfaction with APP new visit care. Practice C grew APP-led consults from 171 (2024) to 235 (2025; 37% YOY growth), increasing APP share of total practice visits from 5% to 23%. Practice D expanded from 2 to 7 visits/day (September 2024 to Oct 2025) with consults rising from 81 to 1,221 (1,407% YOY). APP visit share grew from 6% to 59%, and T2F decreased from 15 days to 7 days. Conclusions: APP-led new consult models in benign hematology improved access, operational efficiency, and growth within community oncology settings. By optimizing APP utilization, the model addressed workforce shortages, accelerated time to care, and enabled early triage of patients. Broad APP engagement and sustained growth across diverse settings demonstrates its scalability and impact. Together, these findings support the integration of APP-led new consults as a practical and transformative approach to addressing critical gaps in oncologic care delivery. Ongoing prospective evaluation will further define the model's impact and inform broader adoption of this approach and best practices across the oncology landscape.
Safety and efficacy of ifebemtinib (IN10018) combined with garsorasib (D-1553) in KRAS <sup>G12C</sup> mutant solid tumors from a phase Ib/II study: 2-year follow-up from single-arm of non-small-cell lung cancer (NSCLC).
3113 Background: Studies have shown that combining RAS inhibitors (RASi) with appropriate partners may enhance efficacy. Ifebemtinib (ifebe) is a potent and selective oral focal adhesion kinase (FAK) inhibitor demonstrating synergy with RASi. D-1553 is a KRAS G12C inhibitor approved in China for KRAS G12C –mutant NSCLC. We previously reported encouraging antitumor activity with ifebe plus D-1553. Here we present updated 2-year efficacy and safety results, including exploratory TP53 subgroup analyses. Methods: This is a phase Ib/II, open-label study to evaluate the safety, tolerability, pharmacokinetics, and antitumor activity of D-1553 in combination with ifebe in patients with locally advanced or metastatic NSCLC harboring KRAS G12C mutations regardless of PD-L1 expression. Patients received ifebe 100 mg once daily (QD) plus D-1553 600 mg twice daily (BID). Results: As of January 4, 2026, 33 treatment-naïve NSCLC patients (81.8% stage IV) were enrolled; median follow-up was 24.9 months (Range: 1.05, 31.77). The objective response rate (ORR) was 90.3% (95% CI: 74.2, 98.0) and disease control rate (DCR) was 96.8% (95% CI: 83.3, 99.9). The median progression free survival (PFS) was 22.14 months (95% CI: 9.63, NE) with the 12- and 24-month PFS rates were 67.9% and 40.0%, respectively. The median duration of response (DOR) was 19.38 months (95% CI: 12.85, NE). The 24-month overall survival (OS) rate was 69.7%, and mOS was not reached (95% CI: 27.76, NE). TP53 wild-type patients (n=17) showed more favorable outcomes, with higher ORR (94.1% vs. 81.8%), unreached mPFS (vs. 14.2 months), and unreached mOS (vs. 17.7 months) compared with TP53-mutant patients (n=13). The safety profile of ifebe plus D-1553 was generally comparable to that of each single agent. SAEs were reported in 33.3% of patients, with 18.2% experiencing ifebe-related SAEs. Most TRAEs were grade 1–2; grade ≥3 TRAEs occurred in 24.2% of patients, including one grade 4 event. No treatment-related deaths or discontinuations were reported. Conclusions: Ifebe plus D-1553 exhibited robust, durable antitumor activity and a favorable safety profile in treatment-naïve KRAS G12C -mutant NSCLC patients regardless of PD-L1 expression. Notably, TP53-wild-type KRAS G12C tumors typically show poor responsiveness and short durability to ICI therapies, likely due to reduced PD-L1 expression and lower tumor mutational burden. The superior outcomes in TP53 wild-type patients suggest this combination may address an unmet need in this distinct population. Further studies are needed to validate these findings and define the regimen’s optimal role in the evolving treatment landscape. Clinical trial information: NCT06166836; NCT05379946 .
Reduced treatment adherence and cancer mortality in patients with depression.
11040 Background: Depression affects up to 20% of cancer patients and is associated with poorer cancer outcomes. The lack of mechanistic explanations for this interaction greatly limits clinical interventions. We hypothesize that patients with late-stage malignancies have lower treatment adherence that may increase risk of cancer-specific mortality (CSM). Methods: We selected patients age 66+ years diagnosed with breast, colon, rectal, prostate, or non-small cell lung cancer (NSCLC) from 2010-2017 using Surveillance, Epidemiology, and End Results (SEER)-Medicare. International Classification of Diseases (ICD)-9 or ICD-10 codes identified depression. Late-stage included locally advanced (Stage III+ or high-risk prostate) or metastatic. Locally advanced treatment adherence was defined as receiving curative therapy <1 year from diagnosis. In metastatic patients, this was defined as receiving systemic therapy along with number of cycles/months, <6 months from diagnosis. Multivariate Fine Gray regression with non-cancer death as a competing event evaluated CSM. Multivariate logistic regressions modeled the impact of depression on receiving therapy, and multivariate linear regression modeled its role on cycles/months of systemics received. Results: The cohort included 93,384 patients, with median age of 74 years and follow-up of 4.2 years. In total, 21,690 patients (23.2%) had depression, highest in NSCLC (27.1%) and lowest in prostate (13.4%). Depression was associated with 6-36% increased risk (p<0.01) of CSM (Table 1). Locally advanced patients with depression had 15-32% lower odds (p<0.01) of receiving curative treatment. Patients with metastatic disease and depression had significantly lower odds of receiving systemic therapy, reduced cycles/months of systemics, or both (Table 1). Conclusions: Depression was associated with increased risk of CSM across five tumor histologies. Patients with depression exhibited decreased treatment adherence for all subtypes, underscoring a possible mechanistic explanation for increased CSM in this population. Multivariate regressions for depression effect. Endpoint CSM:SHR 95% CI, p Received treatment:Locally advancedOR 95% CI, p Received treatment:MetastaticOR 95% CI, p Therapy cycles/months:Metastatic,B Coefficient, p Histology Breast 1.11 [1.02-1.24], <0.01 0.78 [0.65-0.92], <0.01 0.85 [0.71-1.03], 0.10 -0.52, <0.01 Colon 1.15 [1.10-1.20], <0.01 0.71 [0.63-0.79], <0.01 0.67 [0.59-0.75], <0.01 -0.63, <0.01 Rectal 1.13 [1.02-1.25], <0.01 0.68 [0.53-0.87], <0.01 0.67 [0.52-0.88], <0.01 -0.24, 0.33 Prostate 1.36 [1.25-1.48], <0.01 0.85 [0.78-0.93], <0.01 0.94 [0.69-1.31], 0.71 -0.22, 0.02 NSCLC 1.06 [1.04-1.09], <0.01 0.75 [0.67-0.84], <0.01 0.94 [0.88-0.999], 0.045 -0.21, <0.01 SHR: Subdistribution hazard radio; OR: odds ratio; CI: confidence interval.