Browse Articles

Prognostic significance and healthcare resource utilization of emergency department visit etiology in patients with prostate cancer.

e23178 Background: Patients with prostate cancer (PC) frequently present to the emergency department (ED), yet the underlying reasons for these visits and their prognostic implications are not well characterized. We evaluated ED visit characteristics, outcomes, and survival according to visit etiology in patients with prostate cancer. Methods: Patients with PC presenting to the ED at our center were reviewed. ED visits were categorized as comorbidity-related (Co-R), disease-related (Ds-R), treatment-related (Tr-R), or other (Oth) based on ED clinical documentation and consultation notes. Classification was independently performed by an internist and a medical oncologist. Clinical characteristics, hospitalization rates, mortality, survival, and visit-related costs were analyzed. Results: Among 548 screened patients, 242 patients accounted for 510 ED visits. Visit etiologies were Co-R in 35.8%, Ds-R in 31.3%, Tr-R in 16.6%, and Oth in 18.3%. Metastatic disease at presentation was present in 48.9% of Co-R, 87.4% of Ds-R, 95.9% of Tr-R, and 53.8% of Oth visits. Common presenting symptoms included dyspnea (19.8%), chest pain (15.9%), and neurologic symptoms (15.4%) in Co-R; neurologic symptoms (15.7%), urinary retention (15.1%), and hematuria (15.1%) in Ds-R; nausea/vomiting (27.0%), fever (21.6%), and dyspnea (8.1%) in Tr-R; and falls (25.8%), abdominal pain (18.3%), and nausea/vomiting (10.8%) in Oth. Hospitalization rates were 39.0%, 37.7%, 36.5%, and 10.8%, while ED mortality rates were 3.3%, 1.9%, 4.1%, and 4.3% across Co-R, Ds-R, Tr-R, and Oth, respectively. Ninety-day mortality rates were 20.3%, 34.6%, 36.5%, and 17.2%. Compared with Co-R visits, Ds-R (HR:2.03, p < 0.001) and Tr-R (HR:1.93, p < 0.001) visits were independently associated with higher 90-day mortality, along with low hemoglobin (HR:0.85, p < 0.001), low albumin (HR:0.61, p < 0.001), and elevated liver function tests. Median survival after ED presentation was 28.5 months for Co-R, 7.2 months for Ds-R, 7.4 months for Tr-R, and 26.3 months for Oth (p < 0.001). Median ED visit costs were higher in Co-R and Ds-R visits than in Tr-R and Oth visits. Among hospitalized patients, total length of stay was longest in Ds-R (median 10 days, IQR 6–18; p = 0.03), whereas hospitalization costs were lowest in Tr-R (p = 0.022). Conclusions: ED visit etiology is a potent prognostic tool in PC. Ds-R and Tr-R visits represent clinical red flags for poor short-term survival, whereas comorbidity-related visits drive a substantial healthcare burden. Notably, one-quarter of non–disease-related visits were trauma-related, suggesting potentially preventable presentations. Proactive management of non-cancer comorbidities and fall-risk mitigation may reduce avoidable ED utilization, hospitalizations, and healthcare costs.

Baseline peripheral blood activated-to-total CD8⁺ central memory T cell ratio to identify melanoma patients at high risk for severe toxicity but limited clinical benefit.

2568 Background: Immune-related adverse events (irAEs) represent a major clinical challenge that limits the usage of immune checkpoint blockade (ICB) for cancer therapy. The occurrence of irAEs often correlates with improved therapeutic response to ICB therapy, creating a therapeutic dilemma. Many known irAEs predictors correlate with treatment efficacy, further complicating clinical decision-making. Biomarkers that decouple severe irAEs risk from therapeutic benefit are thus urgently needed to guide clinical decision-making. Methods: We performed comprehensive peripheral blood immune profiling in two HCI (Huntsman Cancer Institute) clinical cohorts with 201 melanoma patients treated with ICB, analyzing 488-626 immune cell subsets and 1,463-2,926 plasma proteins at two timepoints: pretreatment and 3-6 weeks on therapy. Activated CD8+ T central memory (CD8Tcm) cells were defined by the expression of CD38, HLA-DR, or Ki67. Associations with severe irAEs, survival and treatment outcomes were evaluated across the largest test cohort composed of nine independent, multi-institutional clinical cohorts, two HCI cohorts, five Swiss cohorts, and two publicly available datasets. Results: Elevated pretreatment activated-to-total CD8+Tcm ratio robustly predicts severe irAEs. Patients with high versus low activated-to-total CD8+Tcm ratio showed two-fold or higher odds of developing severe irAEs across nine independent melanoma cohorts. Remarkably, this immune signature is associated with worse patient survival following ICB. These results suggest that there is a dedicated toxicity predictor in the peripheral blood. Incorporation of pretreatment plasma CXCL9 levels into a machine-learning regression model further improved irAEs predictive performance (mean AUC=0.78 across four independent cohorts). These baseline biomarkers outperformed previously reported irAEs predictors. Conclusions: The pretreatment activated-to-total CD8+Tcm ratio, alone or in combination with plasma CXCL9, identifies melanoma patients at high risk for severe irAEs who are not likely to respond to ICB treatment. This first of blood-based biomarker was tested and validated on many independent patients’ cohorts and will allow to select patients based on the risk stratification prior to treatment initiation.

Opioid prescribing in cancer patients compared with non-cancer patients over two decades (2000-2021) among a national veterans population.

e23347 Background: Evaluating trends in opioid prescribing for veterans with cancer is essential for understanding the impact of national opioid stewardship policies designed to reduce opioid misuse. Methods: We used national Veterans Health Administration (VHA) electronic health record data from January 1, 2000, through December 31, 2021 for individuals with at least one year of continuous VHA care. Cancer patients were veterans who had a diagnosis of a primary malignancy at any point during the study period. Demographic characteristics including race and ethnicity categories were recorded at the time of first VHA encounter. We identified all outpatient pharmacy records for prescriptions of Schedule II opioid analgesics. Among patients diagnosed with cancer, we further determined the timing of the first opioid prescription relative to the cancer diagnosis date. For individuals who received at least one outpatient opioid prescription, we calculated the annual opioid exposure in morphine milligram equivalents (MMEs), as well as total days of opioid prescriptions and the number of prescriptions they received for each of the years they were on opioids. Results: Participants included 2,293,690 veterans diagnosed with cancer and 7,934,886 veterans without cancer. Veterans with cancer were more likely to be prescribed opioids compared to veterans without cancer (59.3% vs. 35.5%), received higher annual MMEs (1,842 vs. 785) and averaged 39 vs. 24 days of annual opioid prescription. Among cancer patients, there was a marked difference between those who started opioids before cancer diagnosis (3,652±8,151 MME) versus opioid-naïve cancer patients (915±3,732 MME; p<0.001). Opioid prescribing increased from 2000- 2012, when it started declining to 15.1% in cancer patients and 8.2% in non-cancer patients in 2021. The decline was particularly pronounced among cancer patients who had started opioids before cancer diagnosis. Similar patterns were observed for median MME, median days on opioids per year, and average number of opioid prescriptions per patient. Cancer patients who received opioids were significantly more likely to be diagnosed with substance use disorder or opioid use disorder compared with those who did not receive opioids. Veterans of Asian ancestry were least likely to receive opioids (40.4%) and had the lowest MME and number of days on opioids compared with all race and ethnicities. Conclusions: Opioid prescribing patterns changed substantially in the VHA during the era of national stewardship initiatives, especially among veterans with cancer. These changes highlight critical shifts in clinical practice but also emphasize the importance of evaluating how these changes have affected pain management and outcomes in veteran populations. Differences in opioid prescription across racial and ethnic groups demand attention to ensure equitable pain management.

Transforming evidence into answers: An agentic framework to support living clinical practice guidelines.

e17004 Background: Living clinical practice guidelines offer the opportunity to keep recommendations aligned with the latest evidence, but they also present a fundamental challenge: as evidence accumulates, panelists cannot sustainably query and synthesize a growing body of literature. In developing ASCO's living metastatic castration-resistant prostate cancer (mCRPC) guideline, we encountered this challenge firsthand. To address it, we developed and evaluated an interactive agentic framework that enables guideline panelists to query evidence in natural language and retrieve structured facts, tables, and figures—with the goal of extending this tool as a clinician-facing companion to enhance guideline accessibility. Methods: The mCRPC systematic review dataset supporting ASCO's clinical practice guideline was used, comprising 212 structured columns capturing trial characteristics, patient populations, and outcomes extracted from 188 references across 88 clinical trials. A large language model (LLM)-based agentic framework was developed employing a coding agent that translates natural language queries into executable scripts for structured evidence retrieval. Claude (Haiku 4.5, Sonnet 4.5, and Opus 4.5) served as the core LLM agents. We constructed a benchmark of 50 fact-based queries with gold-standard answers targeting trial identification, treatment modalities, and endpoint characteristics (e.g., "Which trials evaluated systemic monotherapy?"). Performance was evaluated using exact-match accuracy for final answers, and precision, recall, and F1 score for trial-level retrieval. Results: Across 50 queries, exact-match accuracy ranged from 82% to 86%. Haiku 4.5 and Sonnet 4.5 each achieved 82% (42/50), while Opus 4.5 achieved 86% (43/50). Precision ranged from 0.93–0.97, recall from 0.94–0.96, and F1 scores from 0.91–0.94. Conclusions: This agentic framework enables scalable, auditable, natural language interrogation of guideline evidence, addressing a critical bottleneck in living guideline maintenance. For panelists, it supports rapid, reproducible evidence surveillance during recommendation development. For clinicians, it offers a foundation for an interactive companion tool that transforms static guidelines into queryable resources—enabling personalized evidence retrieval at the point of care. LLM EM Precision Recall F1 claude-haiku-4-5 0.82(42/50) 0.95 0.94 0.91 claude-sonnet-4-5 0.82(42/50) 0.97 0.96 0.93 claude-opus-4-5 0.86(43/50) 0.93 0.96 0.94

Characterizing molecular and clinical features of systemic cancers that spread to the intradural spine.

e14042 Background: Metastatic cancers to the central nervous system (CNS) usually affect the brain. A small but clinically significant proportion also develop intradural spinal disease, though little is known about lesions to this compartment. Here, we characterize the molecular features of intradural spinal disease that spread from non-CNS cancers and describe the metastatic chronology such lesions go through to reach this space. Methods: All patients with MRI spine report findings of intramedullary or leptomeningeal lesions from 2020–2025 at a tertiary academic center were screened for intradural spinal disease. Patients with primary CNS cancer or non-cancer diagnoses were excluded. Demographic information, CSF cytopathology results, imaging characteristics of intradural spinal disease, dates of primary cancer and metastasis detection (CNS and non-CNS sites), and immunohistochemistry and next generation sequencing data of all disease sites were collected wherever available. The primary outcome of interest was overall survival (OS) from date of primary diagnosis. Results: 1401 patients were screened and 69 met inclusion criteria. The most common primary pathologies were lung (23/69, 33.3%) and breast (22/69, 31.9%). 30 (43.5%) patients had lumbar punctures for CSF cytopathology, of which 16 (23.2%) detected tumor cells. 8/69 (11.6%) patients underwent surgery for intradural disease. The distribution of intradural spine disease on initial detection was 38 (55.1%) cervical, 45 (65.2%) thoracic, and 53 (76.8%) lumbar. The mortality rate was 87.0% (60/69) within a median follow-up of 20.6 (IQR 42.0-72.9) months after primary cancer diagnosis. The median time from primary diagnosis to any metastasis was 0.5 (0-6.7), to the brain was 14.9 (0.6-42.7), to the osseous spine was 19.4 (0.7-38.9), and to the intradural spine was 29.5 (14.0-52.6) months. Mutations with common targetable therapeutics were noted in 32/69 (46.4%) primary lesions and patients with such actionable mutations trended towards improved OS (51.7 months, 31.2-105.2) compared to patients without (30.1 months, 12.6-58.9, p=0.072). The molecular profile of the intradural spine lesions were notably different from systemic markers for actionable mutations in 2/8 (25%) cases. Conclusions: Actionable mutations are present in nearly 50% of patients who develop spinal intradural disease, and so regular molecular characterization and sequencing of this space (e.g. CSF liquid biopsy) could have a clinically meaningful impact on these patients. Further, surveillance imaging of patients concerning for spinal intradural disease can miss >25% of intradural spinal disease if MRI of only one part of the spine is regularly imaged.

Rethinking MRI surveillance after breast cancer in the context of persistent racial disparities in mortality.

1675 Background: Detection of asymptomatic breast cancer recurrence by imaging is associated with improved survival compared to detection after symptoms develop. Currently, both the American Cancer Society and ASCO recommend against routine MRI for patients with a history of breast cancer and no clinical findings, unless they meet specific high-risk criteria. The American College of Radiology Appropriateness Criteria similarly characterizes surveillance MRI as conditionally appropriate, supporting its use only when additional risk factors are present. Since randomized trials to address surveillance strategies in diverse populations are unlikely to be feasible, we performed an observational study comparing MRI performance in intermediate and high-risk screening populations, and women with a history of breast cancer, stratified by race. Methods: We retrospectively analyzed 41,041 breast MRI exams performed in an urban health system from 1/2/2014 to 10/29/2025. In Pennsylvania, insurance covers supplemental MRI for women with dense breasts, and in our network, screening MRI is generally performed in asymptomatic patients current with mammography. Exams were categorized as high-risk screening (n=15,682, 38%), supplemental (n=10,905, 27%), or surveillance for a personal history of breast cancer (n=11,476, 28%). Logistic regression was used to estimate odds ratio for true and false positives by indication adjusted for age and race. Results: Cancer detection rates were higher in the surveillance group compared with high risk screening (true positive [TP] rate 21.0, n=241 vs 12.4, n=194), a 53% difference after adjusting for age and race (OR 1.53 [1.26–1.86], p<0.001), with similar false-positive rates (7.5%, n=857 vs 8.7%, n=1,359; OR 1.02 [0.93–1.11], p=0.7). MRI cancer detection rate was higher for supplemental than high-risk screening (OR 1.35 [1.24–1.46], p<0.001). Notably, in a subset comparison of screening MRI in Black vs. white women (n=37,383), TP rates were higher among Black women than white women across all MRI indications. The highest TP rate was observed in Black women undergoing surveillance MRI (31.9 vs. 20.5; OR 1.62 [1.11, 2.31], p=0.010), followed by supplemental (25.6 vs. 14,5; OR 1.84 [1.17, 2.77], p=0.005), and high-risk (21.6 vs. 11.4; OR 1.97 [1.22, 3.03], p=0.003). The TP rate remained higher for Black compared with white women for first MRI as well as subsequent MRI for surveillance, high-risk, and supplemental screening. Conclusions: Black women in the U.S. have 40% higher breast cancer mortality compared to non-Hispanic white women, and early detection offers the best chance for optimal survival. Our findings suggest surveillance MRI identifies a higher burden of clinically occult cancer than high-risk screening MRI, especially among Black women. Expanding access to surveillance MRI may help address persistent racial disparities in breast cancer outcomes.

BAYONET trial: A multicenter phase II trial of staged combination with encorafenib + binimetinib + cetuximab following encorafenib + cetuximab in patients with <i>BRAF</i> V600E-mutant metastatic colorectal cancer.

3559 Background: While the doublet combination of encorafenib (ENCO) + cetuximab (CET) is a standard treatment for patients with BRAF V600E-mutant metastatic colorectal cancer (mCRC), poor prognosis is expected after disease progression. As resistant mechanisms to BRAF + EGFR blockade, several MAPK pathway alterations, including RAS and RAF mutations, have been reported, suggesting that additional blockade of MAPK signaling may be an effective strategy. Furthermore, preclinical studies demonstrated that BRAF V600E-mutant colorectal cancer cell lines resistant to BRAF inhibitor + anti-EGFR antibody were sensitive to the triple combination of BRAF inhibitor + MEK inhibitor + anti-EGFR antibody. Methods: BAYONET is a single-arm, multicenter phase II trial designed to evaluate the efficacy and safety of the staged combination with ENCO, binimetinib (BINI), plus CET for patients with BRAF V600E-mutant mCRC who were refractory to ENCO plus CET. The main eligibility criteria are as follows: RAS wild-type/ BRAF V600E-mutant mCRC; age ≥ 20; ECOG PS 0 or 1; within 4 weeks from the last administration of previous ENCO or CET; no administration of other systemic therapy after refractoriness to ENCO + CET; complete response, partial response, or ≥4 months of stable disease was observed in the previous ENCO + CET. Included patients receive the combination treatment of ENCO (300 mg once a day) + BINI (45 mg twice a day) + CET (400 mg/m 2 initial dose and then 250 mg/m 2 once a week) in a 28-day cycle. The primary endpoint of this trial is 12-week progression-free survival (PFS) rate. The targeted sample size was calculated to be 30 on the basis of a power of 80%, a significance level of 10% (one-sided), a threshold of 20%, and an expected 12-week PFS rate of 40%. Pretreatment ctDNA analysis was performed using the Guardant360 assay. Results: Among 30 patients, all had microsatellite-stable or mismatch repair-proficient tumors, and 21 (70%) of patients received prior &lt;3 treatment lines. The locally assessed 12-week PFS rate was 33.3% (80% CI, 21.8–46.6). Since the lower limit of 80% CI exceeded 20%, the primary endpoint was met. With a median follow-up duration of 9.8 months (range, 2.1–35.8), the median PFS was 2.0 months (95% CI, 1.6–4.2), and the median overall survival was 14.7 months (95% CI, 7.1–18.7). Patients who achieved 12-week PFS had significantly lower ctDNA BRAF V600E VAF (median, 0.2%; range, 0-18%) compared to those who did not achieve it (median, 4.2%; range, 0-41%). Grade 3 or higher adverse events were observed in 50% of patients, most commonly anemia (20%), creatine kinase increase (13%), and diarrhea (10%). No new safety signal was identified. Conclusions: Staged combination with ENCO+BINI+CET showed potential to achieve disease control for patients with BRAF V600E-mutant mCRC who were refractory to ENCO+CET. Clinical trial information: jRCTs031210510.

Subsequent therapies and time to second progression-free survival events (PFS2) in treatment-naive chronic lymphocytic leukemia/small lymphocytic lymphoma (TN CLL/SLL) previously treated with zanubrutinib (zanu) or bendamustine-rituximab (BR) in SEQUOIA.

7046 Background: Cohort 1 of the phase 3 SEQUOIA trial assessed zanu, a next-generation BTKi, compared to BR in TN CLL/SLL without del(17p). Zanu previously demonstrated sustained superiority in progression free survival (PFS) and time to next treatment (TTNT) vs BR. However, evidence regarding outcomes on subsequent therapies (tx) after progressive disease (PD) on zanu remain limited. Here, we present results on follow-up (FU) tx after zanu, including PFS2. Methods: Pts without del(17p) were randomized to receive continuous zanu or six cycles of BR. Pts who received BR could crossover to receive zanu after PD. All pts were followed post-PD for details on any subsequent anti-CLL treatments. This analysis evaluated PFS, TTNT, TTNT or death (TTNT-D) and PFS2. P -values are descriptive. Results: In total 479 pts were randomized; 241 to zanu and 238 to BR. Baseline demographics and disease characteristics were well balanced between the arms. As of 31 October 2025, median FU was 78.7mo (range, 0.0-96.0). Sustained PFS superiority with zanu vs BR was observed (HR, 0.28; 95% CI, 0.21-0.38; P &lt;.0001) with 78mo PFS estimates of 70.9% (95% CI, 64.2-76.6) for zanu and 28.6% (21.7-35.8) for BR. Overall, 211/241 (87.6%) of pts treated with zanu had not received subsequent tx and 34 pts had died without subsequent tx. TTNT and TTNT-D favored zanu over BR (HR, 0.23 [95% CI, 0.15- 0.35; P &lt;.0001] and 0.37 [0.27-0.50; P &lt;.0001], respectively). Overall, 24/241 (10.0%) zanu pts and 82/238 (34.0%) BR pts had PD after study treatment and received subsequent tx. Of the 24 zanu pts receiving subsequent tx after PD, 13 (54.2%) received BCL2i-based regimens, 8 (33.3%) received chemoimmunotherapy (CIT) (incl. anti-CD20 antibody monotherapy), and 3 (12.5%) received BTKi. Of the 82 BR pts receiving subsequent tx after PD, 77 (93.9%) received BTKi (incl. 71 crossover pts), 3 (3.7%) received BCL2i-based regimens and 2 (2.4%) received CIT. Median time from PD following first tx to initiation of next tx was 2.1mo for zanu (range 0-25.7) and 7.4mo (0.3-48.0) for BR. Zanu pts who had PD and received subsequent BCL2i-based tx had a median FU of 33.5mo from initiation of next-line tx. Among these pts, 10 remain alive and without PD, two had died, and one had PD. Overall, PFS2 significantly favored zanu over BR (HR, 0.66; 95% CI, 0.45-0.98; P &lt;.05) with 78mo PFS2 estimates of 81.3% (95% CI, 75.6-85.8) vs 73.8% (67.2-79.3), respectively. Conclusions: PFS2 remained superior with zanu despite broad use of novel subsequent tx (including crossover) in BR pts. Importantly, BCL2i-based salvage after initial zanu achieved high PFS rates at almost 3 years of FU. Together, these findings support that initial zanu provides durable benefit while preserving sensitivity to effective subsequent tx. Clinical trial information: NCT03336333 .

Dynamic remodeling of the extracellular vesicle proteome during systemic therapy in advanced non–small cell lung cancer.

e15059 Background: Non–small cell lung cancer (NSCLC) remains associated with poor prognosis in advanced stages, and reliable biomarkers for monitoring treatment response are still limited. While liquid biopsy applications have largely focused on circulating nucleic acids, extracellular vesicles (EVs) represent a stable and biologically informative source of tumor- and microenvironment-derived proteins. EV proteomic profiling may provide dynamic insights into therapy-induced molecular changes and improve patient stratification. Methods: Plasma samples were collected from patients with advanced NSCLC at baseline (P0, pre-treatment) and at first radiological evaluation (P1), prior informed consent was obtained. Patients were stratified according to treatment regimen (chemotherapy, immunotherapy, or chemo-immunotherapy). EVs were isolated from plasma using an affinity-based enrichment method optimized for low sample volumes and characterized by nanoparticle tracking analysis and Western blotting. EV protein cargo was analyzed by data-independent acquisition (DIA) LC–MS/MS with label-free quantification. Multivariate analyses, functional enrichment (GO/KEGG), and protein–protein interaction network analyses were performed to identify treatment- and time-dependent proteomic signatures. Results: EV proteomic profiling identified 418 EV-associated proteins across all samples. Comparison between baseline (P0) and post-treatment (P1) revealed 108 differentially expressed proteins (43 enriched at P0, 65 at P1). In immunotherapy-treated patients, baseline EVs displayed heterogeneous profiles enriched in epithelial organization, lipid metabolism, complement activation, and coagulation pathways. Post-treatment EVs exhibited marked proteomic remodeling, with enrichment of immune modulation, oxidative stress response, metabolic adaptation, and cytoskeletal remodeling pathways. Functional analyses indicated increased complement regulation, protease inhibition, and immune–stromal interactions, consistent with therapy-induced tumor microenvironment reprogramming. Post-treatment EV profiles partially converged with other treatment groups while retaining immunotherapy-specific signatures. Conclusions: These findings support further study of plasma-derived EVs as they capture dynamic proteomic changes associated with immune checkpoint inhibition in NSCLC and represent a promising non-invasive tool for monitoring immunotherapy-induced molecular remodeling and for biomarker-driven patient stratification and follow-up.

Updated overall survival (OS) and long-term transfusion-independence (TI) from the phase 3 COMMANDS trial in erythropoiesis-stimulating agent (ESA)–naive patients (pts) with lower-risk myelodysplastic syndromes (LR-MDS).

6566 Background: In the phase 3 COMMANDS trial (NCT03682536), luspatercept (LUSPA) significantly improved RBC-TI ≥12 wks with concurrent mean hemoglobin (Hb) increase ≥1.5 g/dL during wks 1-24 (primary endpoint) vs epoetin alfa (EA) in pts with ESA-naive transfusion-dependent (TD) LR-MDS. Longer follow-up (&gt;2.5 yrs) showed a favorable OS trend and durable TI responses with LUSPA vs EA. We report updated results with 6 additional mos of follow-up (&gt;3 yrs). Methods: Eligible pts (≥18 yrs; ESA-naive; TD; LR-MDS) were stratified by baseline (BL) transfusion burden (TB), ring sideroblasts (RS), and serum erythropoietin (sEPO). Pts were randomized 1:1 to LUSPA (1.0-1.75 mg/kg; SC Q3W) or EA (450-1050 IU/kg; SC QW) for ≥24 wks. OS (from randomization), RBC-TI responses and Hb improvement (all from wk 1 to end of treatment), predictive BL biomarkers of OS, and safety were evaluated. Results: At cutoff (Oct 6, 2025), median (min-max) follow-up was 35.9 (1-73; LUSPA) and 30.8 (0-77; EA) mos. Median OS was not reached (NR) for LUSPA and 46.0 mos for EA (HR, 0.78; 95% CI, 0.56-1.09), with similar trends in subgroups (Table). RBC-TI ≥12 wks (LUSPA vs EA) occurred in 76.4% vs 55.8% of pts, and in 60.4% vs 39.2% with concurrent mean Hb ≥10 g/dL. Median cumulative duration (95% CI) of RBC-TI ≥12 wks was 184.4 (118.4-NE) wks for LUSPA and 95.1 (74.9-180.1) wks for EA (HR, 0.54; 95% CI, 0.36-0.80). RBC-TI ≥2.5 yrs (LUSPA vs EA) was achieved by 25.3% vs 10.5% of pts. Mean Hb increase ≥1.5 g/dL (LUSPA vs EA) was achieved by 80.2% vs 58.6% of pts; median duration (95% CI) of Hb increase ≥1.5 g/dL was 72.9 (53.9-89.9) and 47.9 (35.9-60.0) wks (HR, 0.61; 95% CI, 0.44-0.85). Improved OS was associated with BL biomarkers of favorable immune (LUSPA) and CV function (LUSPA and EA), less aggressive disease (EA), and preserved megakaryopoiesis (EA); these characteristics may enhance response to therapy. At cutoff, 17.6% (LUSPA) and 7.8% (EA) of pts remained on treatment; 84.6% and 83.2% had ≥1 dose escalation. No new safety concerns emerged. Deaths (any cause) occurred (LUSPA, 36.3%; EA, 43.0%), fewer LUSPA pts progressed to HR-MDS (3.3%; 7.3%), and AML progression was comparable (4.9%; 5.5%). Conclusions: Long-term follow-up demonstrated improved OS trends and sustained responses with LUSPA vs EA, overall and in subgroups. No new safety concerns emerged, reinforcing superior clinical benefit as first-line treatment in LR-MDS. Clinical trial information: NCT03682536 . Median OS, mos LUSPA EA HR (95% CI) Overall (n=182)NR (n=181)46.0 0.78(0.56-1.09) Stratification subgroup TB &lt;4 RBC U/8 wks (n=118)57.5 (n=111)47.2 0.87(0.57-1.33) TB ≥4 RBC U/8 wks (n=64)NR (n=70)39.9 0.62(0.37-1.05) RS+ (n=133)NR (n=130)47.2 0.70(0.47-1.04) RS− (n=49)54.8 (n=50)46.2 0.96(0.53-1.72) sEPO ≤200 U/L (n=145)NR (n=144)51.4 0.81(0.55-1.20) sEPO &gt;200 U/L (n=37)43.0 (n=37)35.4 0.67(0.35-1.28)

Updated overall survival, safety, and neurologic function outcomes from a phase 1 trial of bivalent chimeric antigen receptor (CAR) T cell therapy in recurrent glioblastoma (GBM).

2013 Background: We previously reported the safety, bioactivity, and preliminary efficacy of bivalent CAR T cells targeting EGFR and IL13Rα2 in 18 patients with recurrent GBM (Bagley, et. al, Nat Med 2025). Given that median follow-up time was only 8.1 months at the time of the data cut-off for that publication, we now report updated overall survival (OS) and safety data with longer follow-up time, as well as neurologic function outcomes. Methods: Eighteen patients with recurrent EGFR-amplified GBM were enrolled using a 3+3 design (dose levels: 5.0 x 10 6 , 1.0 x 10 7 , and 2.5 x 10 7 cells). Patients received a single intracerebroventricular (ICV) dose of CART-EGFR-IL13Rα2 cells. OS was defined as the time from the date of initial study treatment to the date of death from any cause, or censored at the last date of contact if the patient was not known to have died at the date of analysis cut-off (January 16, 2026). To capture changes in neurologic function over the course of the trial, patients were assessed prospectively using the Neurologic Assessment in Neuro-Oncology (NANO) scale. NANO is an objective clinician-reported outcome of neurologic function scored on nine domains, with a higher score indicating worse neurologic function. NANO scores were assessed at baseline prior to CART (day 0), on days 1, 4, 7, 10, 14, 21, and at the 1-month and 2-month visits post-CART. Results: With median follow-up time of 18.5 months, median OS was 12.0 months (95% confidence interval, 7.4 – 23.2 months) with 3 patients surviving greater than 18 months. Other than one patient with prolonged grade 1 neurotoxicity as previously described, no prolonged, late-onset, or unexpected toxicities were observed within this longer follow-up time, including no evidence of on-target off-tumor toxicity or secondary cancers. On day 1 post infusion, an increase in NANO scores was observed (mean increase in NANO score from baseline of 1.9; p=0.09), which resolved to baseline by day 4. Patient NANO scores were unchanged from baseline at one month (mean change 0.41; p=0.998) and two months (mean change 0.54; p=0.992) post infusion. Those who underwent retreatment with a second dose of CAR-T cells at the time of tumor progression (n=7) did not experience an increase in NANO score on day 1 (mean change 0.33, p=0.99), despite having a significant increase on day 1 following the first infusion (mean change 2.9, p=0.008). Conclusions: With over 18 months of median follow-up time, ICV delivery of CART-EGFR- IL13Ra2 cells in patients with recurrent GBM has not demonstrated long-term or delayed toxicities and is associated with promising OS outcomes that warrant continued clinical development. Neurologic function worsens immediately following infusion but recovers to pre-treatment baseline by one month following CAR T cell infusion. Clinical trial information: NCT05168423 .

Efficacy and safety results of tazemetostat in Japanese patients with advanced epithelioid sarcoma: The phase 2 TAZETTA trial (NCCH2107).

11560 Background: Epithelioid sarcoma is an ultra-rare soft tissue sarcoma with limited treatment options. Tazemetostat, a selective EZH2 inhibitor, demonstrated antitumor activity in a global phase 2 study (NCT02601950) and has been approved for the treatment of epithelioid sarcoma in the US. We conducted a Japanese phase 2 study (TAZETTA; NCCH2107) as a sub-study within the framework of the nationwide rare cancer registry project in Japan (MASTER KEY Project) to evaluate the efficacy and safety of tazemetostat in patients with advanced epithelioid sarcoma. Methods: The TAZETTA study is a non-randomized, single-arm, phase 2 trial. Eligible patients had locally advanced or metastatic epithelioid sarcoma with loss of INI1, had received at least one prior chemotherapy including doxorubicin, had an ECOG performance status of 0–2, and at least one measurable lesion according to RECIST v1.1. Tazemetostat was administered orally at 800 mg twice daily in 28-day cycles. The primary endpoint was the progression-free survival (PFS) rate at 18 weeks by blinded independent central review. The threshold 18-week PFS rate was set at 5% and the expected rate at 30%. With a one-sided alpha of 0.05 and 80% power, the required sample size was 12 patients. To allow for ineligibility or non-treatment patients, 15 patients were planned. Secondary endpoints included PFS, overall survival (OS), objective response rate (ORR), disease control rate, duration of response, and safety. In the ancillary study, biomarker samples were collected and subjected to exploratory analysis. Results: Given the smooth patient accrual, 17 patients were enrolled between July 2023 and July 2024. The median age was 43 years; 13 patients were men and 4 were women. Proximal-type disease was observed in 11 patients (64.7%). The centrally assessed 18-week PFS rate was 31.3% (90% CI, 14.0–50.2), and the lower limit of the 90% CI met the predefined efficacy threshold. The median PFS was 2.9 months (95% CI, 1.6–5.6), and the median OS was 12.3 months (95% CI, 4.9–not estimable). Treatment-related adverse event with Grade ≥3 included lymphopenia (1pt, 5.9%). No treatment-related deaths were observed. Conclusions: Tazemetostat demonstrated clinically meaningful activity and a manageable safety profile in Japanese patients with advanced epithelioid sarcoma, consistent with the previous reports, supporting its role as a therapeutic option in this population. Clinical trial information: jRCT2031220523.

Incidence, prevalence, and temporal trends of colorectal cancer across age, sex, race, and ethnicity: A large real-world analysis.

e15699 Background: Colorectal cancer (CRC) remains a leading cause of cancer-related morbidity and mortality in the United States. Recent epidemiologic data suggest rising CRC incidence among younger adults and persistent disparities across demographic groups. Using a large real-world database, we evaluated temporal trends in CRC incidence and prevalence from 2020 to 2024. Methods: We conducted a retrospective observational study using the TriNetX Global Collaborative Network, comprising data from 166 healthcare organizations and approximately 188 million patients. CRC cases were identified using ICD-10-CM codes C18–C20. Annual incidence proportion, prevalence, and incidence rates (cases per person-day) were calculated and stratified by age, sex, race, and ethnicity. Results: From 2020 to 2024, CRC incidence and prevalence increased steadily across all demographic groups. Overall incidence proportion rose from approximately 0.03% to 0.045–0.05%, while prevalence increased from roughly 0.10% to 0.14–0.16%. Incidence rates nearly doubled over the study period. CRC burden increased markedly with age, with the highest incidence rates observed among individuals aged ≥70 years. Although absolute rates remained low, adults aged 20–39 years demonstrated relative increases in incidence over time. Males consistently exhibited higher incidence and prevalence than females. Racial disparities persisted, with increasing incidence among Black, White, and Asian patients, and the highest prevalence observed in patients with unknown race. Conclusions: In this large real-world analysis, CRC incidence and prevalence increased steadily from 2020 to 2024 across all demographic groups. Rising incidence among younger adults, alongside persistent sex- and race-based disparities, underscores the need for equitable, risk-adapted screening strategies and improved access to preventive care.

Analytical characterization of the next-generation HER2 (CAL27) immunohistochemistry assay: A comparative multi-tumor study.

e15173 Background: HER2 protein expression has been observed in a variety of solid tumors. Advancements in antibody drug conjugates and other novel agents may benefit from a highly sensitive and specific assay to accurately detect a wide range of HER2 expression (1). A new VENTANA HER2 (CAL27) Assay was developed which shows high analytical specificity and sensitivity (2). To further characterize its analytical performance, this study compared the sensitivity and specificity of the VENTANA HER2 (CAL27) Assay with PATHWAY anti-HER-2/neu (4B5) Rabbit Monoclonal Primary Antibody and HercepTest mAb pharmDx (DG44) using a cohort of solid tumors. Methods: Formalin-fixed, paraffin-embedded (FFPE) tumor specimens (n=404) from biliary tract, bladder, cervical, colorectal, endometrial, gastric, non-small cell lung, ovarian, and pancreatic cancer were sourced from commercial vendors. Tissue sections were stained by immunohistochemistry (IHC) using three different HER2 assays (CAL27, 4B5, and DG44) per the manufacturer’s instructions. Stained sections were evaluated by a pathologist blinded to the assay for non-specific background, cellular and subcellular staining localization, and the percentage of tumor cells stained at each level. H-scores were calculated and paired t-tests were performed. Results: CAL27 demonstrated substantially increased analytical sensitivity compared to the other HER2 assays, as indicated by higher membrane H-score values at each quartile (Q1, median, and Q3). Differences in H-scores when comparing CAL27 vs. 4B5 or DG44 were statistically significant (p&lt;0.0001) based on paired t-tests. Furthermore, CAL27 exhibited greater specificity, demonstrating the lowest prevalence of cytoplasmic and off-target nuclear staining among the assays. In contrast with 4B5 and DG44, there were no instances where CAL27 cytoplasmic or nuclear staining interfered with interpretation of tumor cell membrane staining, the principal staining localization for HER2 IHC assays. Results are summarized in Table 1. Conclusions: The CAL27 assay demonstrated markedly higher analytical sensitivity compared to the 4B5 and DG44 assays across tumor types, with no interference from cytoplasmic or nuclear signals. These analytical data showing sensitive and specific tissue-based evaluation of the widest range of HER2 IHC expression with CAL27 may lead to improved HER2 IHC detection through future clinical investigations. Summary of staining results in tumor area by HER2 IHC assay. Staining Localization CAL27 4B5 DG44 Membrane H-score (quartile 1, median, quartile 3)* 35, 111, 190 3, 18, 71 4, 29, 110 Cytoplasmic 3.2% (13/404) 21.0% (85/404) 4.5% (18/404) Nuclear 0.2% (1/404) 20.0% (81/404) 0.5% (2/404) Cytoplasmic or nuclear interference with membrane staining interpretation 0.0% (0/404) 0.5% (2/404) 0.7% (3/404) *p&lt;0.0001 for CAL27 vs 4B5 or DG44 by paired t-tests.

Global trends in early-onset uterine cancer and the rising BMI-attributable burden: A population-based analysis, 1990–2023.

e17644 Background: Uterine cancer incidence has increased globally over recent decades, with growing concern regarding rising diagnoses among younger individuals. Obesity is a well-established risk factor for uterine cancer and may contribute to disease burden at earlier ages. However, global trends in early-onset uterine cancer incidence and the BMI-attributable burden have not been comprehensively characterized. Methods: We analyzed uterine cancer incidence and BMI-attributable disability-adjusted life year (DALY) rates using Global Burden of Disease 2023 estimates from 1990–2023 among individuals aged 15–49 years. Rates per 100,000 population were evaluated globally, by Socio-Demographic Index (SDI), and at the country level. Temporal trends were assessed using log-linear regression to estimate average annual percent change (APC) with 95% confidence intervals. Results: Between 1990 and 2023, an estimated 1,885,159 incident cases of uterine cancer occurred globally, with annual cases increasing from 35,102 in 1990 to 74,790 in 2023. Among individuals aged 15-49 years, incidence rates demonstrated a sustained upward trend over the study period. Globally, incidence rates increased from 3.56 in 1990 to 4.21 in 2023, corresponding to an APC of 0.15% (95% CI, 0.01%–0.28%). Substantial heterogeneity was observed across SDI regions, with High SDI regions consistently exhibiting the highest incidence rates, while low and middle SDI regions showed gradual increases over time. In 2023, the highest incidence rates were observed in Northern Mariana Islands (15.7), Barbados (14.6), and Bahamas (13.1). BMI-attributable DALY rates increased numerically from 8.05 to 10.0 between 1990 and 2023 (APC 0.14%; 95% CI, −0.07% to 0.34%). The highest BMI-attributable burdens in 2023 were observed in Nauru (98.7), Micronesia (68.7), and Guyana (65.0). Conclusions: Uterine cancer incidence among individuals aged 15–49 years has increased globally over the past three decades, alongside a numerically increasing BMI-attributable burden, suggesting a need for targeted interventions including obesity-focused strategies in younger populations.

<i>EGFR</i> -mutated lung cancer: A comparative analysis of common and uncommon variants.

8635 Background: Uncommon EGFR mutations represent a heterogeneous subset of EGFR-positive NSCLC with limited characterization. This study compares clinical, molecular, and survival outcomes between common and uncommon EGFR variants in a large real-world cohort. Methods: A retrospective analysis of 1,501 EGFR-positive NSCLC cases from MSK-CHORD was performed. Patients were classified as having common (Ex19del, L858R) or uncommon mutations (G719X, L861Q, S768I). Clinical and genomic variables were analysed using Wilcoxon and chi-square tests (significance p &lt; 0.05). Results: The study included 1,501 EGFR-mutated NSCLC patients, with 195 (13.0%) harboring uncommon and 1,306 (87.0%) harboring common EGFR mutations. Patients with uncommon mutations were older at diagnosis (median 73 vs 69 years; p = 0.003), while sex distribution remained similar (p = 0.868). Race distribution differed (p = 0.005), with more White patients in the uncommon EGFR group (80.3% vs. 69.5%) and fewer Asian patients (11.8% vs. 25.5%), while Black patient proportions were similar (7.9% vs. 5.0%). Smoking history also varied (p &lt; 0.005), with substantially fewer never-smokers among patients with uncommon mutations (25.6% vs 59.6%). Uncommon mutations were associated with higher genomic burden, including elevated TMB (median 4.92 vs 3.46 mut/Mb; p &lt; 0.001) and mutation count (median 6 vs 4; p &lt; 0.001). Stage at diagnosis did not differ (p = 0.3), nor did metastatic site distributions (p = 0.477). Median OS was significantly shorter for patients with uncommon EGFR mutations (36.69 vs 56.52 months), with HR 1.425 (p &lt; 0.005). In Stage IV patients receiving osimertinib, OS was numerically shorter for uncommon mutations but not statistically significant (29.98 vs 34.32 months; HR 1.351; p = 0.08). Among Stage IV patients with uncommon mutations, OS was comparable between afatinib and osimertinib (25.48 vs 26.70 months; p = 0.88). KRAS co-mutations were significantly more frequent in the uncommon group (4.62% vs 1.15%; p = 1.985×10⁻³). Conclusions: Uncommon EGFR mutations exhibit distinct clinical and genomic features and poorer survival compared with common variants, highlighting their heterogeneity and the need for tailored therapeutic strategies.

Domino Polymerization for the Synthesis of Reductively Degradable Poly(disulfide)s With Arbitrary Side‐Chain Structures

ABSTRACT The environmental persistence of conventional plastics has driven the development of degradable polymers with functional versatility. Poly(disulfide)s are particularly attractive due to their reductive degradability and environmentally triggered disassembly. In this study, we report a domino polymerization strategy that enables the rapid and modular synthesis of functional poly(disulfide)s with degradable main chains and diverse side‐chain structures. This method utilizes a single monomer, N ‐(2‐oxotetrahydrothiophen‐3‐yl)‐3‐(pyridin‐2‐yldisulfanyl)propanamide (PDTL), which contains both thiolactone (TL) and pyridyl disulfide (PDS) moieties. Polymerization is initiated by TL ring‐opening with commercially available amines, generating thiol groups that undergo disulfide exchange with PDS moieties of other monomers, resulting in chain propagation. The method accommodates primary, secondary amines, and ammonia, allowing incorporation of alkyl, allyl, propargyl, hydroxyl, carboxyl, amide, and tertiary amine functionalities. The reductive degradability of the poly(disulfide)s is confirmed using gel permeation chromatography and mass spectrometry. Furthermore, it is found that pH‐responsive properties appear in poly(disulfide)s possessing dimethylamino groups, where water solubility can be regulated in response to pH. Domino polymerization is further applied to synthesize reductively degradable poly(disulfide) gels using multi‐functional amines. This polymerization technique will lead to the development of reductively degradable polymers with various functions for use in various applications in the future.

Eco-friendly synthesis of zirconium dioxide nanoparticles from Peltophorum pterocarpum: Biological inertness assessment and molecular docking insights toward targeted drug delivery

Modulating Coplanarity of Polymer Acceptor Facilitates Hierarchical All‐Polymer Heterojunction for Efficient and Stable Semitransparent Solar Cells

ABSTRACT All‐polymer‐based semitransparent organic solar cells (ST‐OSCs) are better‐suited candidates than small‐molecule analogues for building‐integrated photovoltaics due to their superior operational stability. Their deployment is however hindered by low light utilization efficiency (LUE), mainly ascribed to intimating morphological control within bulk‐heterojunction architectures. To mitigate these constraints, here we synergize the polymer donor‐layer thinning strategy with a novel star‐shaped polymer acceptor in a quasi‐planar heterojunction (QPHJ) structure. Two such acceptors, PYBSe‐B and PYBSe‐L, are designed and synthesized with twisted π‐backbones based on a sterically hindered BTSe core flanked with tailorable side‐chains to regulate molecular planarity and self‐aggregation behavior. By inclusion into the bottom layer of PM6 donor in QPHJs, the relatively planar PYBSe‐L is aggregated between the PM6 domains whereas the amorphous PYBSe‐B is well‐mixed. PYBSe‐L is found to expand interdomain spacing and promote infiltration of the upper PYIT acceptor, strengthening optical transmittance and exciton dissociation. The champion devices achieved an efficiency of 11.16% and a record LUE of 4.80% for all‐polymer ST‐OSCs. Remarkably, the rigid hierarchical polymer–infiltration network formed by the QPHJ structure greatly reinforced storage stability of the unencapsulated device with an extrapolated T 80 lifetime (degradation to 80% of initial LUE values) of over 1310 h under ambient conditions.

The difference between subjective symptoms and objective symptom of edema during the menstrual cycle

Abstract This study aimed to elucidate the differences between subjective and objective symptoms of edema during the menstrual cycle in female university students with regular menstruation. The study included 29 female university students with regular menstrual cycles. The menstrual cycle was categorized into the early follicular, mid-luteal, and late-luteal phases, and measurements were taken during each phase. Subjective symptoms were evaluated using body charts and the Menstrual Distress Questionnaire (MDQ) to assess the presence of edema. Objective symptoms were measured using a body composition analyzer to assess body weight, body water content, and regional water content. In the evaluation of subjective symptoms of edema using the body chart, no significant differences were observed across cycles. However, when body regions were compared, there were significantly fewer subjective symptoms in the upper arm (Rt/Lt) than in the face and trunk across all phases, while the lower leg (Rt/Lt) had significantly more subjective symptoms. The MDQ (water retention item) showed no marked differences across cycles. For objective evaluations of edema, including body weight and total body water, no significant differences were observed across cycles; however, regional water content was significantly higher in the trunk and right arm during the mid-luteal and late-luteal phases than in the early follicular phase. Our findings suggest that the location and timing of subjective edema symptoms may differ from those of objective symptoms.