Browse Articles
Discover research articles across all indexed journals
Phase II open-label trial of neoadjuvant immunotherapy in combination with CAPOX for resectable non-metastatic proficient mismatch repair (pMMR) colon cancer: NICER study.
TPS2691 Background: Colorectal cancer (CRC) is the third most common malignancy worldwide. Immune checkpoint inhibition has transformed outcomes in deficient mismatch repair (dMMR/MSI-H) CRC but has demonstrated limited efficacy in metastatic proficient mismatch repair (pMMR) tumors, likely related to low tumor mutational burden and limited immune infiltration. Emerging neoadjuvant studies suggest that earlier exposure to immunotherapy may enhance antitumor immunity and induce pathologic responses in pMMR colon cancer. Neoadjuvant chemotherapy may further augment immunogenicity through tumor debulking, antigen release, and modulation of the tumor immune microenvironment. Combination chemo-immunotherapy has improved outcomes across multiple solid tumors and may overcome resistance to immune checkpoint blockade. Early neoadjuvant immunotherapy studies in pMMR colon cancer have demonstrated encouraging pathologic responses without compromising surgical outcomes. The NICER study evaluates the safety, feasibility, and biological activity of neoadjuvant atezolizumab plus CAPOX in resectable pMMR colon cancer. Correlative studies include longitudinal circulating tumor DNA (ctDNA) analysis and paired tissue assessment to characterize immune microenvironmental changes following treatment. Methods: This is a phase II, open-label study evaluating neoadjuvant atezolizumab in combination with capecitabine and oxaliplatin (CAPOX) in patients with resectable, non-metastatic pMMR colon adenocarcinoma. The planned accrual is 28 patients. Participants receive four 3-week cycles of neoadjuvant atezolizumab plus CAPOX, followed by standard-of-care surgical resection. Adjuvant chemotherapy is permitted per investigator discretion for high-risk patients. Key eligibility criteria include pMMR colon adenocarcinoma with tumor ≥12 cm from the anal verge and at least one high-risk feature (elevated CEA, low lymphocyte-to-monocyte ratio, poor differentiation, lymphovascular or perineural invasion, CT-defined T3–T4 disease ≥4 cm, or regional lymphadenopathy). Exclusion criteria include metastatic disease, autoimmune disorders, recent malignancy, or synchronous colorectal primaries. The primary endpoint is tumor regression grade (TRG) using the modified Ryan scoring system. The regimen will be considered promising if the post-therapy TRG 1 rate is ≥15%, compared with an expected rate of 0% with upfront surgery. Secondary endpoints include pathologic complete response, R0 resection rate, lymph node yield, safety, surgical timing, recurrence outcomes, and quality of life. As of January 2026, 7 of 28 planned patients have been enrolled since accrual began in July 2025. ClinicalTrials.gov ID NCT05870800. Clinical trial information: NCT05870800 .
Association of baseline protein–energy malnutrition with mortality and healthcare utilization in patients with hepatocellular carcinoma treated with VEGF-based systemic therapy.
e16166 Background: Patients with hepatocellular carcinoma (HCC) frequently have advanced liver disease and frailty at the time of systemic therapy initiation. Protein–energy malnutrition (PEM) is common in this population; however, whether PEM reflects acute illness at treatment initiation or is associated with persistent vulnerability after therapy begins remains unclear. We evaluated the association between baseline PEM and early clinical outcomes after excluding deaths and events within the first 30 days of VEGF-based therapy. Methods: We conducted a retrospective cohort study using the TriNetX Global Collaborative Network. Adults with HCC treated with VEGF-based systemic therapies (bevacizumab, sorafenib, lenvatinib, regorafenib, cabozantinib, or ramucirumab) were identified, excluding patients with prior liver transplantation. Baseline PEM was defined by ICD-10 codes for malnutrition, cachexia, or adult failure to thrive within 6 months prior to therapy initiation. Deaths and outcomes within the first 30 days after VEGF initiation were excluded to minimize bias from acute illness at treatment start. Baseline race and ethnicity differences were addressed through propensity score matching. Propensity score matching (1:1) was used to balance demographics, liver disease severity (including cirrhosis, ascites, portal hypertension, and varices), comorbidities, metastatic burden, and prior liver-directed therapies. Outcomes from days 30–90 included overall survival (OS), hospitalization, ICU admission, and gastrointestinal (GI) bleeding. Results: After propensity score matching, 957 patients with baseline PEM were compared with 957 patients without PEM, with well-balanced baseline characteristics. Between days 30 and 90 after VEGF initiation, patients with PEM had worse overall survival (80.6% vs 86.0%; hazard ratio [HR] 1.44, 95% CI 1.13–1.84; p = 0.003). PEM was also associated with higher hospitalization rates (14.3% vs 9.9%; risk ratio [RR] 1.44; p = 0.003). ICU admission rates were similar (4.0% vs 3.6%; p = 0.63), and GI bleeding did not differ between groups (7.1% vs 7.1%; RR 1.00, 95% CI 0.72–1.38). Conclusions: After excluding early deaths and events within the first 30 days of treatment, baseline protein–energy malnutrition remained associated with worse short-term survival and increased hospitalization among patients with HCC treated with VEGF-based systemic therapy. These findings suggest that PEM is not solely a marker of acute illness at treatment initiation but reflects persistent vulnerability during early therapy. Routine nutritional assessment and early supportive interventions may improve outcomes in this population.
QUINTESSENTIAL-2: A phase 3 study of arlocabtagene autoleucel versus standard of care in adult patients with relapsed and refractory multiple myeloma (RRMM) exposed to lenalidomide.
TPS7577 Background: Despite advances in multiple myeloma (MM) treatment, most patients (pts) will relapse, highlighting the need for new drug classes in RRMM. Further, with the extensive use of lenalidomide (LEN), an immunomodulatory drug (IMiD), in frontline and maintenance therapies, LEN-refractoriness has become increasingly common and poses an additional challenge as the disease is less likely to respond to subsequent treatment. G protein-coupled receptor class C group 5 member D (GPRC5D) is a promising therapeutic target for MM as the receptor is highly expressed on malignant plasma cells; it has little to no expression on non-plasma cell immune populations and limited expression on other tissues. Arlocabtagene autoleucel (arlo-cel) is a GPRC5D-directed autologous CAR T cell therapy that has demonstrated safety and efficacy in pts with RRMM in a first-in-human phase 1 study. ORR was 94% and 91% in pts with 1-3 and ≥3 prior LOT (pLOT), respectively, after a single arlo-cel infusion (150×10 6 CAR T-cells); ORR was 92% in pts with ≥3 pLOT following a 75×10 6 CAR T cell infusion. These phase 1 study outcomes support further clinical development of arlo-cel. Methods: QUINTESSENTIAL-2 (NCT06615479) is a randomized, open-label, multicenter, phase 3 confirmatory study comparing the efficacy and safety of arlo-cel vs SOC in adults with RRMM and prior LEN. Key inclusion criteria were age ≥18 years, confirmed diagnosis of MM per IMWG criteria, 1-3 pLOT (may include a proteasome inhibitor, IMiD, anti-CD38 antibody, and BCMA-targeted therapy), be exposed to LEN (≥2 consecutive cycles, unless PD was the best response to LEN-containing treatment or if there was LEN intolerance or unacceptable toxicity), have measurable disease, and ECOG PS of 0 or 1. Pts who received prior GPRC5D-targeted therapy are excluded. Eligible pts will be randomized 1:1 across 2 arms. Arm A includes leukapheresis within 3-4 days of randomization, mandatory bridging therapy ≤6 days of randomization with DPd (daratumumab, pomalidomide, dexamethasone) or Kd (carfilzomib, dexamethasone) per investigator choice and lymphodepleting chemotherapy prior to a single arlo-cel infusion. Arm B includes SOC of DPd or Kd per investigator choice, dosed per labeling, until PD. Primary endpoints are PFS and minimal residual disease (MRD) negativity in CR. Key secondary endpoints include OS and ORR. Other endpoints include safety, MRD-negative status, CR rate, time to response, duration of response, pharmacokinetics, and pt-reported outcomes. Pts will be followed for ≤5 years after the last pt is randomized, and with a long-term follow-up study (≤15 years post-infusion) for pts receiving arlo-cel. Anticipated enrollment is 440 pts at ~125 sites globally. The first pt was enrolled in March 2025. © American Society of Hematology (2025). Reused with permission. Clinical trial information: NCT06615479 .
Prevalence of post-traumatic stress disorder and associated factors among adult sarcoma patients.
11557 Background: A cancer diagnosis and treatment can be extremely stressful and even traumatic for sarcoma patients. However, research on prevalence and risk factors for developing post-traumatic stress disorder (PTSD) symptoms among sarcoma patients has not been well studied. Methods: This study used survey data from participants recruited between 4/2022 and 12/2025 for the “Cohort to Augment the Understanding of Sarcoma Survivorship Across the Lifespan” (CAUSAL) study. PTSD symptoms were assessed via the 22-item Impact of Event Scale-Revised (IES-R) at cohort enrollment, and a total IES-R score of ³33 indicates PTSD symptomatology. Associations of demographics and clinical characteristics with PTSD symptoms were evaluated using logistic regression models. The association between PTSD symptoms and health-related quality of life (HRQOL), measured via T-scores for the seven core domains of the PROMIS-57 scale, was then assessed using multivariable linear regression models adjusted for potential confounders. Results: Among 1,112 adult sarcoma patients (509 on active treatment and 603 survivors), there are 605 females and 507 males, aged between 18 and 84 (mean age: 54.8). 18.1% of sarcoma patients/survivors reported experiencing PTSD symptoms. Female (20.5%) and younger (15.9-24.3%) sarcoma patients were significantly more likely to experience PTSD symptoms, whereas those with higher household incomes were significantly less likely to experience PTSD symptoms (11.6%-17.6%). Patients who were 5+ years post-diagnosis were less likely to experience PTSD symptoms with adjusted odds ratios (aOR) and 95% confidence intervals (CIs) of 0.56 (0.34, 0.89), compared to those who were < 18 months post-diagnosis. Regular exercise (aOR [95 CI] = 0.59 [0.42, 0.83]) and having emotional support (aOR [95 CI] = 0.63 [0.41, 0.97] for medium and 0.26 [0.15, 0.42] for high support) were inversely associated with PTSD prevalence. PTSD symptoms were significantly associated with a decreased T-score of physical function and social roles and activities with ß(95% CI) of -4.2 (-5.7, -2.7) and -6.8 (-8.3, -5.2), respectively and increased scores for anxiety (ß = 8.9; 95% CI: 7.6, 10.0), depression (ß = 9.4; 95% CI: 8.2, 11.0), fatigue (ß = 7.0; 95% CI: 5.5, 8.5), sleep disturbance (ß = 7.9; 95% CI: 6.6, 9.3), and pain interference (ß = 6.7; 95% CI: 5.2, 8.2). Conclusions: PTSD symptoms were reported by an average 18% sarcoma patients and survivors, particularly among female, younger patients, non-exercisers, and those with lower income and low emotional support. PTSD was associated with poor HRQOL. Long-term survival, regular exercise, and increased emotional support were inversely associated with the prevalence of PTSD symptoms. Psychological support and exercise programs should be offered during active cancer treatment and follow-up surveillance to reduce PTSD symptoms in sarcoma patients and survivors.
Immune checkpoint inhibitor adoption, timing, and survival in an urban lung cancer cohort: A landmark analysis.
e20697 Background: Although immune checkpoint inhibitors (ICIs) improve survival in randomized trials, real-world use is shaped by clinical selection, complicating retrospective survival analyses. We evaluated predictors of ICI receipt and used landmark analyses to assess whether timing of ICI initiation independently impacts survival in a diverse urban lung cancer population. Methods: We conducted a retrospective cohort study with 211 lung cancer patients diagnosed between 2010 and 2025 at a tertiary academic center. Predictors of ICI receipt were assessed using multivariable logistic regression. Overall survival (OS) was evaluated using multivariable Cox proportional hazards models with ICI exposure treated as a time-dependent covariate. Among ICI recipients, exploratory landmark analyses at 90 and 120 days were performed to examine associations between timing of ICI initiation and OS. Results: Seventy-two patients (34.1%) received ICI therapy. Advanced disease stage was the strongest predictor of ICI receipt with Stage IV patients significantly more likely to receive ICIs (OR 7.11, 95% CI 2.86–20.1, p<0.001). Trends toward higher ICI use were observed among older patients (OR 1.03 per year, p=0.066) and Asian patients (OR 3.61, p=0.07). In the time-dependent Cox modeling, ICI receipt was not independently associated with improved OS (HR 0.72, 95% CI 0.39–1.31, p=0.276), whereas advanced stage remained strongly associated with mortality (Stage III HR 2.19, p=0.042; Stage IV HR 3.05, p=0.003). Among ICI recipients, later initiation was associated with improved OS at the 120-day landmark (HR 0.07, p=0.024), with a similar but non-significant trend at 90 days (HR 0.27, p=0.058). Notably, patients initiating ICIs later consistently demonstrated lower baseline disease burden, including fewer Stage IV cases and greater representation of earlier-stage disease. Conclusions: In this real-world urban cohort, ICI adoption was driven primarily by disease stage rather than demographic characteristics. Apparent survival advantages associated with later ICI initiation likely reflect selection bias, as patients with lower disease burden and greater clinical stability survive long enough to initiate therapy later. These findings underscore the importance of accounting for disease severity and time-dependent biases when interpreting observation studies of immunotherapy outcomes. Overall (211) No ICI (139) ICI (72) p-value Age (median [IQR]) 63.00 [57.00, 68.00] 62.00 [57.00, 68.00] 63.50 [58.00, 70.00] 0.207 Male (%) 87 (41.2) 57 (41.0) 30 (41.7) 0.9 Race (%) 0.818 White 62 (29.4) 42 (30.2) 20 (27.8) Black 137 (64.9) 90 (64.7) 47 (65.3) Asian 12 (5.7) 7 (5.0) 5 (6.9) Smoking Status (%) 0.724 Never 29 (13.7) 21 (15.1) 8 (11.1) Former 113 (53.6) 73 (52.5) 40 (55.6) Current 69 (32.7) 45 (32.4) 24 (33.3) Stage (%) <0.001 I 45 (21.3) 38 (27.3) 7 (9.7) II 26 (12.3) 21 (15.1) 5 (6.9) III 56 (26.5) 39 (28.1) 17 (23.6) IV 84 (39.8) 41 (29.5) 43 (59.7) Histology (%) 0.321 NSCLC 179 (84.8) 121 (87.1) 58 (80.6) SCLC 28 (13.3) 15 (10.8) 13 (18.1) Other 4 (1.9) 3 (2.2) 1 (1.4)
Spatial Cascade Sites in Hierarchical COF‐Based Photocatalyst Enable C─C Coupling for Selective CO <sub>2</sub> Photoreduction to Ethylene
ABSTRACT The photoreduction of CO 2 into multi‐carbon (C 2+ ) products is a highly attractive route for CO 2 utilization; however, the yield and selectivity of C 2+ products are seriously limited by slow multi‐electron–proton transfer and sluggish C─C coupling kinetics. Herein, we construct a hierarchical tandem photocatalyst IS@COF‐Ni by growing imine‐pyridine covalent organic frameworks on non‐stoichiometric indium sulfide and introducing isolated Ni single‐atom sites at the interfacial edges. The synergistic effect between the spatially segregated sites promotes *CO dimerization, effectively lowering the kinetic barrier for high‐rate ethylene (C 2 H 4 ) generation. Thus, compared with its individual components, the IS@COF‐Ni heterojunction achieves exceptionally high C 2 H 4 productivity and selectivity in photocatalytic CO 2 reduction with water vapor in the absence of additives. In situ spectroscopic characterizations and theoretical calculations reveal that IS@COF‐Ni establishes a low‐energy pathway for electron and proton transfer, while the heterojunction interface effectively stabilizes the adsorbed CO (*CO) intermediate, facilitating C─C bond formation via coupling of adjacent *CO species to generate C 2 H 4 . This work provides a strategic approach for designing photocatalysts toward selective CO 2 ‐to‐C 2+ conversion.
Disproportionate reporting of severe gastrointestinal injury with ruxolitinib among JAK inhibitors in myeloproliferative neoplasms: A FAERS pharmacovigilance analysis.
e18600 Background: Janus kinase (JAK) inhibitors are cornerstone therapies for myeloproliferative neoplasms (MPNs). Although gastrointestinal (GI) adverse events have been reported with JAK inhibitors in other disease settings, comparative safety data evaluating severe GI injury among MPN-approved JAK inhibitors remain limited. We performed a pharmacovigilance analysis to assess whether reporting patterns for severe GI injury differ across individual JAK inhibitors used in MPNs. Methods: We conducted a retrospective disproportionality analysis using the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) from 2012 through 2025. Reports involving ruxolitinib, fedratinib, pacritinib, or momelotinib were identified using comprehensive generic and brand-name synonyms. To minimize duplicate reporting, analyses were restricted to the most recent report per case. The primary endpoint was a composite of severe GI injury defined by MedDRA preferred terms, including gastrointestinal hemorrhage, ischemic colitis or intestinal ischemia, peritonitis, bowel necrosis, and GI perforation. GI perforation was evaluated as a key secondary endpoint. Disproportionality was assessed using reporting odds ratios (RORs) with 95% confidence intervals (CIs), comparing ruxolitinib with all other MPN-approved JAK inhibitors. Sensitivity analyses were performed restricting to primary or secondary suspect (PS/SS) drug roles. Results: Among 16.7 million deduplicated FAERS reports, 67,318 involved MPN-approved JAK inhibitors. Ruxolitinib accounted for the majority of reported exposures (all-role: 62,233; PS/SS: 59,606). Across FAERS, 218,546 unique reports of severe GI injury were identified. In PS/SS analyses, severe GI injury was reported in 692 ruxolitinib-exposed cases compared with 47 cases among all other JAK inhibitors combined. Ruxolitinib was associated with a higher reporting frequency of severe GI injury relative to other agents (ROR >1.2, 95% CI excluding unity), with consistent findings across sensitivity analyses. GI perforation was rare but occurred almost exclusively with ruxolitinib (PS/SS: 41 cases), with no perforation signal observed for other agents. No temporal clustering was identified, and the signal persisted across multiple FAERS reporting periods. Conclusions: In this large FAERS pharmacovigilance analysis, ruxolitinib demonstrated a disproportionate reporting signal for severe GI injury, including rare but serious GI perforation, compared with other MPN-approved JAK inhibitors. These findings are hypothesis-generating and do not establish causality. Given the widespread use of ruxolitinib in MPNs, heightened clinical awareness and further prospective, patient-level studies are warranted to clarify underlying mechanisms and identify high-risk subgroups.
Burden and trends of cancer and attributable risk factors in women, 1990 to 2023, with projections of mortality to 2050.
e22624 Background: Cancer (CA) is a leading cause of premature mortality and disability in women worldwide. Women specific biological, hormonal, and reproductive factors intersect with rising exposure to metabolic and lifestyle risks—particularly in low- and middle-income countries (LMICs) to shape distinct CA patterns. However, global inequalities persist, and shifting risk profiles especially increasing metabolic exposures—underscore the need to track long-term trends in CA burden in women. Methods: This study applied the GBD 2023 analytical framework to quantify long-term trends in CA burden in women from 1990-2023 across 204 countries and territories. Using standardized GBD estimation pipelines, we evaluated deaths, disability-adjusted life years (DALYs), and incidence, stratified by CA type, Socio-demographic Index (SDI) region, attributable risk-factor, and age group. Temporal trends were summarized using estimated annual percent change (EAPC) derived from log-linear regression models. Future projections for cancer mortality were modeled through 2050. Results: From 1990-2023, global female CA deaths rose from 2.66 to 4.70 million (EAPC +1.55%); DALYs increased from 79.93 to 126.82 million (+1.19%); and incidence rose from 50.54 to 74.99 million (+1.25%). Prevalence counts increased across all SDI regions, with the steepest rise in Low SDI (+3.57%), followed by Low-middle SDI (+2.57%), Middle SDI (+2.19%), High-middle SDI (+1.37%), and High SDI (+1.33%). Age-standardized DALY rates declined across most cancers. However, other pharynx CA (+0.58%), lip and oral cavity CA (+0.26%), and non-melanoma skin CA (+0.07%) showed increasing trends. In contrast, the steepest declines were observed for stomach (−3.28%), esophageal (−2.36%), nasopharynx (−2.22%), Hodgkin lymphoma (−2.05%), and leukemia (−2.00%). Notably, breast (−0.31%), cervical (−0.63%), and liver CA (−0.91%) showed modest declines in age-standardized DALY rates. Across risk groups, age-standardized cancer DALY rates declined for all categories: Metabolic risks (EAPC −0.09%), Behavioral risks (−1.08%), Environmental/occupational risks (−1.32%). By age group, the largest DALY rate decline occurred in < 20 years (−1.71%), followed by ≥55 years (−1.04%), and 20–54 years (−0.39%). Incidence rates declined in < 20 years (−1.17%), 20–54 years (−0.51%), and were nearly stable in ≥55 years (−0.05%). Global CA deaths are projected to reach 6.62 million by 2050. Conclusions: Despite declining age-standardized DALY rates, the absolute burden of CA in women continues to rise, especially in low-SDI regions and younger age groups. While reductions in behavioral and environmental risk–related CA burden are encouraging, metabolic risks remain largely unchanged, calling for urgent, equity-driven investments in prevention, early detection, and risk-specific interventions.
Detection, quantification, and subtyping of adenoid cystic carcinoma (ACC) using methylation from liquid biopsies.
6118 Background: Despite the common indolent behavior of ACC, some patients will experience aggressive disease with short survival. These distinct clinical behaviors have been linked to two transcriptional profiles: ACC-I, defined by NOTCH1 activation, solid histology, and more aggressiveness, and ACC-II, characterized by predominant myoepithelial p63 expression and generally indolent behavior. DNA methylation has been widely used for tumor identification and subtyping; however, their relevance in distinguishing ACC-I and ACC-II has not yet been explored. Methods: Forty-one ACC tissue samples from the original subtype development cohort (41/54) and twenty-three cfDNA samples from a phase II trial of axitinib and avelumab in recurrent/metastatic ACC (NCT03990571) were profiled through target-enriched enzymatic methylation sequencing. ACC subtypes (ACC-I/II) in tissue samples were validated with unsupervised clustering of highly variable regions. To translate these signatures to liquid biopsy, we used the METER (METhylome AnalysER) pipeline to identify tumor-specific differentially methylated sites (DMS) and regions (DMRs) relative to 22 healthy donors, enabling ctDNA detection (METER-positive), tumor proportion score (TPS) quantification, and ACC subtype assignment in cfDNA. Methylation-derived TPS were validated with ichorCNA (copy number alterations) estimates, and cfDNA-assigned subtypes were benchmarked against matched tissue transcriptional signatures. Finally, we investigated if METER detection was associated with progression-free survival (PFS). Results: ACC tissue methylation identified two clusters that perfectly matched reference transcriptional subtypes (17/17 ACC-I and 24/24 ACC-II). ACC-I tumors were hypermethylated compared to ACC-II (11,205 hyper and 1,804 hypo DMR, p<0.05). Of 23 cfDNA samples, 15 (65.2%) were deemed as METER positive with a higher rate among ACC-I (8/9, 88.9%) compared to ACC-II (7/14, 50.0%), consistent with the more aggressive subtype of ACC-I. The median (range) estimated TPS among METER positive samples was (3.86%; range 0.7%-82%), with a numerical higher TPS among ACC-I (14.8% vs 2.0%, p=0.073). TPS had strong correlation with ctDNA estimation by ichorCNA (spearman 0.86, p<0.001). METER subtyping correctly identified ACC-subtypes in 15/15 (100%) METER positive and 4/8 (50%) METER negative samples. METER-positive patients had shorter PFS (HR = 4.32; 95%CI, 1.49–12.5, p=0.007), including among ACC-II patients (HR = 4.84; 95%CI, 1.20–19.5, p=0.026). Conclusions: Methylation-based subtyping of ACC is highly concordant with transcriptional profiles and enables detection and subtyping of ACC in liquid biopsies. Beyond subtyping, cfDNA detection may serve as an additional prognostic marker, identifying patients at high risk for progression even within traditionally indolent subgroups. Clinical trial information: NCT03990571 .
Implementation of a multi-cancer early detection (MCED) test in a private practice: Adoption, performance, and repeat-testing patterns.
10532 Background: Century City Medical Management (CCMM), a network of private primary care practices, implemented MCED testing as part of their cancer screening program. The MCED test detects a cancer signal in cell-free DNA isolated from peripheral blood and predicts the cancer signal origin (CSO). We describe CCMM’s MCED testing workflow and real-world performance. Methods: MCED testing is discussed with all CCMM patients as part of routine care, and shared decision-making guides test utilization. Pts with a positive MCED test result (cancer signal detected) are directed to pursue targeted diagnostic evaluation. This analysis reports on data from 1,949 pts who had 2,922 MCED tests, with results returned between 11/29/21 and 11/24/25. Diagnostic evaluations and clinical outcomes were collected through a Quality Assurance Program via follow-up with ordering providers. All data were deidentified in accordance with HIPAA using the Expert Determination method. The cancer signal detection rate (CSDR) was defined as the proportion of positive MCED tests among all tests administered during the first screening round. Among pts with diagnosed cancers and data available, CSO accuracy (top-two predictions) and median time to diagnosis were calculated. Results of MCED retesting after an inconclusive workup and the frequency of annual MCED screening were evaluated. Results: At the first MCED test (N = 1,949), median age was 63y (IQR: 55-71), 47.5% were female, and the CSDR was 1.03% (95% CI, 0.63–1.58%; 20/1,949). Across all screening rounds, 24 pts had a positive MCED result; 16/24 had ≥12mo follow-up at data cutoff. Among these 16, 3 had an active cancer at testing. Of the 13 pts with ≥12mo follow-up and no active cancer at testing, 7 were subsequently diagnosed with cancer. CSO accuracy was 100% (7/7). None of the diagnosed cancers had USPSTF A/B-recommended screening. 1 diagnosis (symptomatic pt) occurred before MCED results were returned; median time from test results to diagnosis of the other 6 was 62 days (IQR: 30.5-143). 5 pts with an initial positive test but negative CSO-directed workup had an MCED retest and ≥12-mo follow-up. Among these, 4 pts with a negative retest remained cancer-free through the end of follow-up, and 1 with a positive retest was diagnosed with a cancer matching the initial and retest CSO. Among 996 pts tested with >15mo of follow-up, 686 (69.6%) received ≥2 MCED screening tests. The most common interval between consecutive tests was annual (9-15mo), with <10% exceeding an 18-mo interval. Conclusions: MCED testing in a real-world population achieved test performance, CSO accuracy, and diagnostic resolution timelines consistent with prior studies, with high annual screening uptake and clinical utilization of repeat testing after inconclusive CSO-directed workups. The findings support feasible and scalable implementation of MCED testing in primary care.
Lung cancer navigation pilot: Preliminary results of nurse outreach support for lung cancer patients enrolled in the EOM.
e13580 Background: Oncology practices struggle with reducing unnecessary ED visits and hospitalizations. This is an important quality metric for EOM, MIPS, and many commercial plans. Practices can earn greater shared savings by reducing hospital and ED utilization. Objective: Determine whether nurse outreach can reduce admissions and emergency department (ED) visits among high-risk lung cancer patients enrolled in the EOM. About 2K lung cancer patients enrolled in EOM are attributed to TxO. This is the 2 nd highest group after breast cancer and represents 20% of our EOM population. Methods: Program began in September 2024. Patients were selected for the pilot using the list of EOM patients and the APEX report (an advanced analytic model which predicts 90-day mortality risk for patients with metastatic cancer). A mix of low to very-high patients were selected, enrolling moderate to very high-risk patients first until the case load of about 100 patients per navigator was reached. Calls ranged from weekly to every 4 weeks depending on patient need and treatment/OV schedule. Advance Care Planning introduction was promoted, with the goal to introduce at the initial call if appropriate or at least by the third call. Patients were removed (disenrolled) once they meet one of the following criteria: Death/Hospice; Transfers care outside TXO or decision to stop treatment and follow-up; EOM episode ends and no future treatment planned; Patient unresponsive to phone calls after 5 attempts. Results: Hospital and ED Utilization and re-visit rates were better for the Outreach group compared to the rest of the EOM Lung Cancer patient population. The expected cost for Admission and ED visits was calculated for the RN Outreach support patients. The expected cost = the quantity of patients x the utilization rate x the re-admit/visit rate x the average cost per event. Subtracting the actual cost for the Outreach group from the expected cost resulted in an estimated savings to total cost of care of $2.6M. 72% of patients received ACP support in the Outreach group. The Hospice Utilization rate for the Outreach group was 69% compared to 60% of the not followed group. Patient Experience was positive with patients grateful for the outreach. Conclusions: The EOM Lung Patient Pilot demonstrates that structured nurse outreach for high-risk metastatic lung cancer patients can meaningfully improve key quality and utilization outcomes. These findings suggest that targeted nurse navigation is an effective strategy to decrease unnecessary hospital admissions and ED visits. Improvements in ACP and hospice utilization were also seen. This may serve as a scalable approach to improving outcomes in other high-risk oncology populations. Outreach Not Followed Qty of Pts 199 2,008 % w/ Admits 31% 65% Avg LOS 5.92 6.08 ReAdmit Rate 1.63 2.12 % w/ ED Visits 27% 52% ReVisit Rate 1.74 2.06 ReAdmit & ReVisit Rate = Total Events / Qty of Pts with Events.
First-line lenvatinib versus dabrafenib plus trametinib (D+T) in <i>BRAF</i> -mutated differentiated thyroid cancer (DTC): Insights from real-world data.
6052 Background: BRAF mutations are present in approximately 40%–60% of patients with DTC. Lenvatinib is the NCCN Category 1 preferred first-line (1L) systemic therapy for progressive radioiodine-refractory (RAI-R) DTC—including BRAF -mutated disease. For patients with BRAF V600E–positive, RAI-R DTC, BRAF -targeted therapy is recommended for those unsuitable for lenvatinib or as a 2L agent for patients after disease progression or intolerance to one or more prior multikinase inhibitors. This analysis compares real-world effectiveness of 1L treatment with lenvatinib versus D+T in patients with BRAF -mutated tumors. Methods: Patients with DTC who initiated lenvatinib monotherapy (n=319; Dec 2014–Jul 2025) or D+T (n=131; June 2016–Sept 2025) were identified from the Tempus multimodal real-world database. Among them, 88 patients were treated with lenvatinib and 54 patients were treated with D+T as 1L therapy for DTC harboring BRAF V600E and/or K601E mutations. Retrospective analyses in this patient cohort (n=142) were performed to compare real-world progression-free survival (rwPFS) and overall survival (exploratory; rwOS) between patients treated with 1L lenvatinib and those treated with 1L D+T in BRAF -mutated DTC. Results: Of the 142 patients in the study cohort who were treated with lenvatinib or D+T, the median age was 66 years (interquartile range: 57–74); 45% of patients were female and 55% were male. The majority (75%) of patients were White, 5.7% were Asian, 3.4% were African American, and 16% had their race categorized as “Other.” Papillary, follicular, and Hürthle histology accounted for 92%, 0.7%, and 1.4% of patients, respectively, while the remaining 6% had uncommon or unspecified histology. Demographics and clinical characteristics were balanced between patients treated with 1L lenvatinib or D+T. At median follow-up of 33.6 months, median rwPFS indexed from treatment initiation dates was 17.0 months (95% confidence interval [CI], 12.2–27.8) in patients treated with lenvatinib and 6.2 months (95% CI, 4.9–8.7) in patients treated with D+T (hazard ratio [HR] for lenvatinib vs D+T, 0.44 [95% CI, 0.28–0.69]. Median rwOS was 70.7 months (95% CI, 51.8 – 90.8) in patients treated with lenvatinib, and 37.8 months (95% CI, 16.4–66.4) in patients treated with D+T (HR for lenvatinib vs D+T, 0.34 [95% CI, 0.18–0.62]). Sensitivity analyses showed that rwPFS and rwOS benefits associated with lenvatinib were maintained irrespective of age, sex, or race. Conclusions: In this retrospective study of real-world data, 1L therapy with lenvatinib demonstrated longer rwPFS and rwOS compared with 1L D+T in patients with DTC harboring BRAF mutations. While recognizing the inherent limitations of real-world data, this study has implications regarding the sequencing of agents used to treat patients with BRAF -mutated RAI-R DTC.
Pre-operative lymph node staging in localized esophageal cancer using a high-sensitivity ctDNA assay.
e16111 Background: Accurate lymph node (LN) staging remains essential in esophageal cancer (EC) treated with neoadjuvant chemoradiotherapy (CRT) followed by surgery. Circulating tumor DNA (ctDNA) offers a non-invasive measure of tumor burden and may complement imaging-based staging. We evaluated whether an ultrasensitive, tumor-informed ctDNA assay at diagnosis and its dynamics during CRT correlate with clinical and pathological nodal status, and whether ctDNA predicts recurrence-free survival (RFS). Methods: This retrospective analysis includes the first 32 patients from a multicenter prospective observational study (target n = 250). Patients had localized EC (ESCC n = 8; EAC n = 24), received CRT and were staged with PET-CT and, when indicated, laparoscopy. Plasma was collected at baseline, after CRT, and post-operatively. ctDNA was quantified using NeXT Personal, a whole-genome, tumor-informed assay leveraging up to ~1,800 patient-specific variants (limit of detection ~1 PPM). ctDNA dynamics were assessed using the post-CRT/baseline ratio. Associations with clinical staging (cT/cN), pathological outcomes (ypT/ypN), and RFS were evaluated using non-parametric tests and landmark Kaplan–Meier analyses. Results: ctDNA was detectable at diagnosis in all evaluable patients (100% pre-treatment sensitivity). Baseline ctDNA levels were significantly higher in cN-positive vs. cN-negative patients (median 1,580 vs. 58 ppM; p = 0.034). The ctDNA post-CRT/baseline ratio correlated with both pathological tumor stage (ypT; ρ = 0.38, p = 0.046) and pathological nodal stage (ypN; ρ = 0.45, p = 0.015), indicating that limited ctDNA decrease during CRT is associated with persistent nodal disease at surgery. Among ctDNA-negative patients after CRT (n = 18) who proceeded to surgery, only 1 patient harbored ypN+ disease (6%). 6/12 patients with persistently detectable ctDNA after CRT were ypN+ (50%). Notably, ctDNA negativity after CRT and before surgery was significantly prognostic of RFS (HR: 3.9, log-rank P = 0.04), and stratified patients by RFS more effectively compared with pathological complete response (pCR, HR: 2.5, log-rank P = 0.2). Furthermore, persistent ctDNA detection after surgery was also associated with inferior RFS (HR: 10.9, log-rank p≈2.2×10⁻⁶) with a median follow-up of 25.1 months. Conclusions: In this cohort of esophageal cancer patients treated with CRT and surgery, ctDNA PPM quantity at diagnosis correlated with clinical LN involvement and persistent ctDNA after CRT. CRT indicated a high likelihood of residual nodal involvement in the resection specimen. Conversly, ctDNA clearance after CRT was associated with a very low likelihood of ypN-positive disease and a limited residual primary tumor burden, which—if confirmed in larger cohorts—may open opportunities for organ-preserving strategies, addressing a current unmet need in esophageal cancer.
Neoadjuvant tislelizumab plus oxaliplatin and S-1 for locally advanced Siewert type II esophagogastric junction adenocarcinoma: A prospective multicenter phase II study.
4091 Background: Adenocarcinoma of the esophagogastric junction (AEG) is a malignant tumor with high morbidity and mortality globally. Among subtypes, Siewert type Ⅱ AEG remains a clinical challenge owing to its unique anatomical location and distinct clinicopathological features. This study aimed to evaluate the efficacy and safety of Tislelizumab in combination with oxaliplatin and S-1 as neoadjuvant therapy for patients with locally advanced Siewert type Ⅱ AEG. Methods: This is a prospective, multicenter clinical trial. Eligible patients were aged 19-75 years with pathologically confirmed Siewert type II adenocarcinoma of the esophagogastric junction, clinical stage T2-4N0-3M0, ECOG performance status 0-1, and adequate major organ function. All patients received 3 cycles of neoadjuvant therapy consisting of: tislelizumab (200 mg, D1, Q3W), oxaliplatin (130 mg/m², D1, Q3W), and S-1 (40-60mg based on BSA, po, bid, D1-14, Q3W). Following efficacy evaluation, patients underwent resection of the esophagogastric junction adenocarcinoma 4-8 weeks after completion of neoadjuvant therapy. The primary endpoint is pathologic complete response (pCR). Secondary endpoints include major pathologic response (MPR), objective response rate, R0 resection rate, overall survival (OS), disease-free survival (DFS), and safety. Results: As of June 2025, a total of 40 patients were enrolled in this study. The median age was 66 years (range, 37-76), 87.5% were male, and 92.5% had an ECOG performance status of 0. Among them, 87.5% (n = 35) of the patients completed 3 cycles of neoadjuvant therapy. The remaining 5 cases completed 1 (n = 1, due to grade≥3 TRAEs), 2 (n = 2, due to grade≥3 TRAEs and patient decision, respectively), and 4 (n = 2, patient decision) cycles of treatment. All patients underwent resection of the esophagogastric junction adenocarcinoma. The R0 resection rate was 97.5% (39/40), with no patients requiring re-operation or experiencing perioperative mortality. The pathologic complete response (pCR) rate was 25% (10/40). The primary tumor pCR rate reached 30% (12/40), and the major pathologic response (MPR) rate was 37.5% (15/40). Treatment-related adverse events (TRAEs) of any grade occurred in 95% (38/40) of patients, with grade ≥3 TRAEs observed in 15% (6/40). Conclusions: Neoadjuvant therapy with Tislelizumab in combination with oxaliplatin and S-1for locally advanced Siewert type Ⅱ esophagogastric junction adenocarcinoma yields a favorable pathological complete response (pCR) rate and shows a good safety profile. Clinical trial information: ChiCTR2300075638.
Trends in clinical characteristics, outcomes, and prognostic factors in melanoma: A retrospective real-world analysis.
e21586 Background: Melanoma remains a challenging heterogeneous malignancy with outcomes influenced by its stage at presentation and tumor biology. Understanding the institutional patterns at presentation, treatment utilization, and survival outcomes are essential, especially when real-world outcomes from our region remain underreported. This study aims to characterize the demographic features, explore the survival outcomes, and evaluate the factors associated with patient’s survival treated at a large academic medical center. Methods: We retrospectively analyzed adult (>18 years old) melanoma patients treated at AUBMC (1/2012–12/2024). Demographic data, tumor characteristics, treatment modalities, and clinical outcomes were extracted and analyzed from electronic health records. Overall survival (OS) and progression-free survival (PFS) were estimated using Kaplan-Meier methods, and associations between covariates and survival outcomes were evaluated via Cox proportional hazards regression. Multivariable Cox proportional hazards (CPH) models were used to adjust for confounding. Finally, OS was estimated using KM and compared across systemic treatment groups using the log-rank test. Multivariable Cox proportional hazards models were used to adjust for baseline prognostic factors. Results: Among 314 patients, median age at diagnosis was 56 years (IQR, 46–68), 51.6% were male, and median Charlson Comorbidity Index was 4. The most common histology was superficial-spreading melanoma (25.5%). At presentation, 31.1% had a Breslow thickness >2mm, and the median melanoma diameter was 15 mm. At diagnosis, 12.4% of patients had stage III disease, and 10.5% had stage IV disease. Overall, 80.3% were metastasis-free at diagnosis, while 2.9% had brain metastasis. At a median follow up of 27 months, 1-year and 2-year OS were 93% (95% CI 89%–96%) and 87% (95% CI, 81%–91%), respectively. In a multivariable CPH model, higher Breslow thickness was independently associated with worse OS (HR 3.83, 95% CI 1.06–13.87, p = 0.041). Kaplan–Meier analysis suggested improved OS with immunotherapy-containing regimens (p < 0.01); however, this association was not significant after multivariable adjustment (p = 0.2), suggesting that baseline disease characteristics likely explained the unadjusted differences. Conclusions: In this real world cohort, breslow thickness emerged as an independent predictor of worse overall survival. Survival outcomes were also favorable, with high 1- and 2-year OS rates. These findings support the value of real-world data in understanding treatment outcomes, and explore the need to develop risk-adapted personalized models to guide future therapeutic strategies.
Next-generation cancer organoids 2.0 in precision cancer medicine: Patient-derived pancreatic ductal adenocarcinoma organoids model to describe drug sensitivity and complexity of tumor biology.
4221 Background: Organotypic models of patient-specific tumors are revolutionizing our understanding of cancer heterogeneity and its implications for personalized medicine. The study's aims were as follows: 1) to establish a PDAC patient-derived organoid (PDO) model obtained from various PDAC specimens. 2) to find out clinicogenomic factors affecting patients’ outcomes. Methods: The SMC PDAC Cohort Patients were prospectively enrolled and underwent EUS-guided FNB, metastatic sites (such as liver, ascites, lung, and bone), and surgical resection. PDAC PDOs were comprehensively analyzed for histology, next generation sequencing (NGS), and high-throughput screening (HTS) drug sensitivity tests. Results: The 735 PDAC patients were prospectively enrolled in this study. PDAC PDO platform has been trying to establish from the following cancer specimens: ascites, biopsies from bone, liver, lung, and pancreas, or surgical resection. The success rate was as follows according to the source of obtaining site; ascites due to peritoneal seeding (3/3: 100%), bone mets (1/1: 100%), EUS-FNB (195/183: 87.8%), liver mets (7/8: 87.5%), lung mets (1/1: 100%), surgical specimens (394/500: 78.8%) and PDO was successfully established within 8.2±2.6 days. It took approximately 3 weeks to acquire each specimen and generate sufficient PDAC PDOs for the simultaneous HTS drug sensitivity test and NGS. Whole exome or genome sequencing (WES/WGS, n = 357) showed an almost identical concordance between original PDAC tissues and matched PDOs and the increased frequency of genetic alterations in PDOs. The HTS drug sensitivity test (n = 151) revealed the clinical correlation between the PDO response and the actual chemotherapeutic response of the study patients in both palliative and adjuvant in real-world settings (ranging from 84.0~ to 91.2%). In addition, whole transcriptome sequencing (n = 373) identified nab-paclitaxel resistance-associated genes such as ITGB7, ANPEP, and ST3GAL1, and also found early recurrence-related genes including COL2A1, CALB1, and CYP24A1 in the extracellular matrix, calcium signaling, and vitamin D metabolism. Conclusions: The PDAC PDO platform may become a valuable tool for personalized medicine and may give us insights into tumor biology.
Genomic signature–based risk score for PARP inhibitor response in <i>BRCA</i> -mutated ovarian cancer: A whole-exome sequencing analysis.
5547 Background: To identify whole-exome sequencing (WES)–based genomic signatures associated with response to poly(ADP-ribose) polymerase (PARP) inhibitors and develop a prediction model for PARP inhibitor effectiveness in BRCA -mutated advanced ovarian cancer, given the heterogeneous real-world treatment outcomes. Methods: This retrospective cohort study included patients with advanced ovarian cancer harboring BRCA 1/2 mutations who received olaparib or niraparib as first-line maintenance therapy between 2020 and 2022 at the National Cancer Center, South Korea. Among them, patients with at least one year of follow-up from the reference date (August 1, 2024) and sufficient additional tumor samples for WES were identified. Baseline characteristics were collected, and progression-free survival (PFS) and overall survival (OS) from the initiation date of PARP inhibitors were extracted from electronic medical records. Formalin-fixed, paraffin-embedded tumor samples of eligible patients underwent DNA extraction and WES. Prognostic modeling incorporated clinical and genetic variables using Cox proportional hazards regression with the least absolute shrinkage and selection operator (Lasso) method with leave-one-out cross-validation. Model performance was assessed using time-dependent area under the curve (AUC(t)). Results: Fifty-five patients were included. The median patient age was 57 years (range, 39–72). High-grade serous carcinoma comprised 96.4%. Neoadjuvant chemotherapy (NACT) was performed in 40.0%. Complete remission after adjuvant chemotherapy occurred in 87.3%. Olaparib was administered in 76.4% and niraparib in 23.6%. NACT-exposed samples had lower Mutant-Allele Tumor Heterogeneity (MATH) scores than non-NACT-exposed samples, indicating reduced intratumor heterogeneity (P < 0.001). Recurrence during PARPi maintenance occurred in 11 patients (20.0%) (olaparib: 5/42 [11.9%]; niraparib: 6/13 [46.2%]). No deaths were observed. The median PFS was 37.9 months (range, 4.2–54.5), and the median OS was 43.3 months (range, 20.2–60.1). Mutations significantly associated with progression-free survival included POLN (HR, 2.02; 95% CI, 1.12–3.67), ABCC10 (HR, 4.57; 95% CI, 1.13–18.5), GTPBP6 (HR, 2.07; 95% CI, 1.04–4.12), and TCTEX1D1 (HR, 0.19; 95% CI, 0.05–0.82). A genomic risk score model incorporating these mutations stratified patients ino low- and high-risk groups with strong prognostic discrimination for progression-free survival (HR, 4.44; 95% CI, 2.26–8.68; log-rank P < 0.001; AUC(t), 0.87). Conclusions: The genomic signature-based risk score model independently predicted PARP inhibitor effectiveness. External validation is warranted.
Precemtabart tocentecan (Precem-TcT, M9140), an anti-CEACAM5 antibody-drug conjugate (ADC), in patients with metastatic pancreatic ductal carcinoma (mPDAC): Results from the phase 1b/2 PROCEADE-PanTumor PDAC substudy.
4199 Background: CEACAM5 is overexpressed in many cancers, including PDAC, with limited expression in normal tissues. Precem-TcT, an anti-CEACAM5 ADC with an exatecan payload (topoisomerase 1 inhibitor [TOP1i]), showed a predictable, manageable safety profile and promising clinical activity (ORR: 31.0%; mPFS: 6.9 months) in the Phase 1 PROCEADE-CRC-01 study (NCT05464030) in patients with heavily pretreated metastatic colorectal cancer (mCRC). Methods: PROCEADE-PanTumor (NCT06710132) is an ongoing, Phase 1b/2, open-label study investigating the efficacy and safety of Precem-TcT monotherapy in advanced PDAC, gastric cancer, and non-small cell lung cancer. Here, we report interim results from the PDAC substudy, which enrolled patients with pretreated mPDAC (ECOG PS ≤1; and CEACAM5 high expression [≥50% tumor cells with immunohistochemistry ≥2+ staining]) in the US, Australia, France, Japan, and Korea, who were treated with Precem-TcT 2.8 mg/kg Q3W. Results: Between May and September 2025, 46 patients have been enrolled into the study (median age: 62.0 years; female: 65.2%; ECOG PS 0/1: 47.8%/52.2%). Most patients received one or two prior lines of therapy, and were irinotecan pretreated. As of November 2025, median duration of Precem-TcT treatment was 12.6 weeks. Treatment emergent adverse events (TEAEs) occurred in 95.7% of patients; grade ≥3: 78.3%. Most common grade ≥3 TEAEs were neutropenia (43.5%) and anemia (32.6%). Gastrointestinal TEAEs were mostly grade 1-2, and there were no cases of interstitial lung disease or ocular toxicities. TEAEs led to dose reductions in 15.2% and to permanent discontinuation in 4.3% of patients. There were no treatment-related deaths. As of January 2026, 10 patients had a PR (21.7%), 20 SD (43.5%), and 11 PD (23.9%); 14 patients remain on study. Updated data, including mPFS and CEACAM5 expression, will be presented at the congress. Preliminary pharmacokinetics of the conjugated antibody and unconjugated exatecan appeared generally consistent with previous data from patients with mCRC. Conclusions: Precem-TcT, the first anti-CEACAM5 ADC with a TOP1i-payload reporting clinical data in patients with mPDAC, showed a predictable safety profile in this population, consistent with data from patients with mCRC. Efficacy data in this highly refractory, mostly irinotecan-pretreated 2L/3L population benchmark favorably with the current SoC for 2L mPDAC. Clinical trial information: NCT06710132 .
Deep learning–derived MRI features for risk stratification in post-surgical <i>IDH</i> -mutant glioma.
e14006 Background: IDH-mutant gliomas undergo long periods of observation after radiochemotherapy before undergoing malignant transformation, leading to neurological morbidities and death. Unfortunately, the timing of disease progression is unpredictable. In this study, we developed a deep learning (DL)-based AI framework for extracting relevant sub-visual feature representation from routine structural post-surgical brain MRI scans for predicting progression-free survival (PFS) in IDH-mutant glioma patients. Our hypothesis is that AI-derived quantitative features from post-treatment structural MRI enable identification of high-risk patients who undergo earlier disease progression. Methods: A total of 188 IDH-mutant glioma patients from the University of Wisconsin Hospital were retrospectively analyzed, of whom 143 met the inclusion criteria of this study including (a) patients who underwent biopsy or surgical resection; availability of (b) chemo-radiation therapy information; and (c) post-operative MRI sequences (T1w, post-contrast T1w, T2w, FLAIR). Following co-registration, skull stripping and bias field correction, a Swin UNETR model for semantic segmentation was trained using the publicly available BraTS-Gli2024 challenge dataset, which includes ~1000 glioma post-treatment patients. The model provided tumor habitat segmentations, including enhancing and non-enhancing tumor, FLAIR hyperintensities, and tumor cavity. Using the Swin UNETR model, a total of 768 DL features per patient were extracted from the encoder as feature representations. For survival analysis, patients were randomly divided into training and validation sets in an 80:20 ratio. A LASSO-Cox regression model was then applied in a 10-fold cross validation scheme to identify independent features associated with PFS (defined as time to next intervention) on the training set. The discriminatory features identified on the training set by the Cox model were then applied on the test set to predict survival. Results: The Swin UNETR model achieved an overall dice score of 0.86 ± 0.13 for tumor habitat segmentation. Analysis of encoder-derived DL features using a Cox proportional hazards model yielded concordance indices of 0.76 and 0.78 for the training and testing cohorts, respectively, with statistically significant associations (p-value < 0.05) for both sets. Conclusions: Our analysis demonstrated that DL features extracted from routine post-surgical MRI may serve as non-invasive imaging biomarkers for the identification of IDH-mutant glioma patients at high risk of earlier disease progression.
Adenoid basal carcinoma: An NCDB study of demographic and socioeconomic factors.
e21550 Background: Adenoid basal carcinoma (ABC) is a rare, indolent variant of basal cell carcinoma that often presents as a pruritic, hemorrhagic lesion. Because the literature consists largely of isolated case reports and small series, its epidemiologic features remain poorly defined. Broader population-level data are needed to better understand its clinical context and diagnostic patterns. To address this gap, the National Cancer Database (NCDB) was analyzed to characterize the demographic and socioeconomic profile of patients diagnosed with ABC. Methods: A retrospective cohort analysis of the 2004–2020 NCDB identified patients with microscopically confirmed ABC (ICD-O-3 code 8098; N=237). Demographic variables, including sex, race, ethnicity, insurance status, facility type, and Charlson–Deyo score, were analyzed using descriptive statistics, and incidence trends were evaluated with regression analysis. Additional variables included treatment modalities, anatomical sites, and survival trends. Results: A total of 237 ABC cases were identified, with stable incidence from 2004–2020 (R²=0.0081). All patients were female, and a majority were White (65%), followed by Black (22.4%), and other races (12.7%). The mean age at diagnosis was 65.2 ± 13.8 years (range, 19–90), and most individuals were non-Hispanic (82%), lived in metropolitan counties with populations >1 million (65.4%), and had Charlson–Deyo scores of 0 (80.2%). Medicare (44.3%) and private insurance (36.3%) were the most common types of coverage. Treatment was concentrated in academic/research (42.6%) centers and comprehensive community cancer programs (28.3%). The most frequent primary site was the cervix uteri (177 cases, 74.7%), and the mean tumor size was 15.1 mm. Surgical management was performed in 92.0% of patients, with negative margins in 83.1%. Radiation (13.1%) and chemotherapy (6.3%) were used as secondary treatments in addition to surgery. Short-term mortality was low, with 99% 30-day survival and 97% 90-day survival. Mean survival was 160.3 months, with two-, five-, and ten-year survival rates of 96%, 90%, and 78%, respectively. Conclusions: This represents the first NCDB analysis of ABC, addressing a critical gap in its epidemiologic characterization. ABC predominantly affected older women and was commonly diagnosed among patients residing in metropolitan areas, with treatment frequently occurring at academic or research facilities. Most patients underwent surgery and exhibited favorable long- and short-term survival. Future studies should explore how demographic and socioeconomic factors influence diagnosis, treatment, and outcomes in this rare malignancy.