Browse Articles
Discover research articles across all indexed journals
A comparison of data-processing methods to support patient matching to oncology clinical trials.
e13651 Background: To pre-screen and match patients to trials at scale, technology tools are needed to support human workflows. These tools require numerous data inputs, including clinical concepts such as cancer diagnosis, stage, and histology. The more complete these data are, the better the accuracy of matches produced, and the more efficient human reviewers can be. While some clinical data are available from clinical system structured fields, most are within unstructured documentation. Methods including human abstraction, natural language processing (NLP) models, and large language model (LLM) agents can improve data completeness and accuracy beyond structured EHR and laboratory information management system (LIMS) data. We compared these methods to obtain diagnosis and stage data for Tempus Link, a tool that supports patient-trial matching at scale for a national oncology trials network. Methods: We randomly sampled patients with one of 4 previously abstracted cancer diagnoses: lung (LC), prostate (PC), colorectal (CRC), and breast cancer (BC). For LC we also examined histology (NSCLC vs SCLC). For each method (EHR, LIMS, NLP predicted, LLM agent), diagnosis, stage, and histology values were compared to the abstracted value, and accuracy was calculated as the percent of patients with a correct value among patients with an abstracted value available. Since these data are inputs into a tool used by RN screeners who confirm matches before notifying a site, diagnosis was focused on accuracy in broad cancer diagnostic categories (e.g. neoplasm of lung). In addition, we assessed completeness as presence of a usable value across all patients in a cohort. Results: Completeness for EHR, NLP, and LLM was higher for diagnosis compared to stage, ranging from 91.7 - 100% vs 20 - 95.8%. LIMS completeness was lower for all variables across cohorts, ranging from 0 - 87.8%. Conclusions: Variability exists across data sources and processing methods in generating data inputs for trial matching. All methods have high completeness and accuracy for diagnosis, while there are significant gains when applying NLP and LLMs for stage and histology. To balance accuracy, matching efficiency, and cost, the use of enhanced data processing methods are required for certain variables, but may not be needed for others. Accuracy by cohort, variable, and data processing method. Cohort EHR LIMS NLP LLM LCn=48 # diagnosis correct 48 (100%) 33 (68.8%) 48 (100%) 45 (93.8%) # NSCLC correctn=48 1 (2.1%) 27 (56.3%) 48 (100%) 39 (81.3%) # stage correctn=41 16 (39%) 0 (0%) 27 (65.9%) 30 (73.2%) PCn=48 # diagnosis correct 48 (100%) 43 (89.6%) 48 (100%) 44 (91.7%) # stage correctn=45 22 (48.9%) 0 (0%) 15 (33.3%) 38 (84.4%) CRCn=49 # diagnosis correct 47 (95.9%) 42 (85.7%) 47 (95.9%) 46 (93.9%) # stage correctn=46 21 (45.7%) 0 (0%) 36 (78.3%) 42 (91.3%) BCn=50 # diagnosis correct 50 (100%) 9 (18%) 47 (94%) 50 (100%) # stage correctn=36 5 (13.9%) 0 (0%) 23 (63.9%) 35 (97.2%)
Healthcare resource use (HCRU) intensity prior to cancer diagnosis and real-world outcomes in early-stage cancers.
1644 Background: Choosing Wisely initiatives aim to reduce futile care by emphasizing the role of clinicians in healthcare decision-making. However, HCRU is strongly shaped by patient preferences with some inherently favoring high-intensity (“maximalist”) care and others preferring low-intensity (“minimalist”) approaches. We hypothesized that HCRU in the years before an early-stage cancer diagnosis may, in part, represent the healthcare seeking preferences of patients rather than strictly reflecting their clinical needs. Methods: We conducted a population-based retrospective cohort study of adults diagnosed with stage I-II breast, lung or colon cancer from 2015 to 2020 in Alberta, Canada with follow-up to 2025. Pre-diagnosis HCRU was quantified using ambulatory, emergency, and inpatient encounters. Patients were categorized as high (≥75th percentile) or low (≤25th percentile) users; intermediate users were excluded. Cohorts of high vs. low users were matched based on age, sex, rurality, stage and Charlson Comorbidity Index (CCI) and stratified by tumor site. Outcomes included cancer-specific survival (CSS), overall survival (OS) and post-diagnosis HCRU. Results: The matched cohort included 7,384 patients (64% breast, 21% lung, 15% colon) with balance across all covariates. Median age was 67 years, 68% had stage I disease and 95% had CCI<2. For all cancer sites, CSS did not differ between high and low users (see Table). High pre-diagnosis HCRU was associated with worse OS in breast (HR 1.34, p<0.001) and colon cancer (HR 1.36, p=0.003), but not in lung cancer (HR 1.02, p=0.76). Importantly, HCRU patterns pre-diagnosis strongly persisted post-diagnosis, with high users continuing to consume significantly more healthcare resources than low users for 4 years after cancer diagnosis (p<0.001). Conclusions: In this relatively healthy cohort of early-stage, highly curable cancers, greater pre-diagnosis HCRU was associated with sustained post-diagnosis HCRU intensity without any significant differences in cancer-specific outcomes. Our findings suggest that utilization patterns may reflect patient-level behaviors as much as clinical necessity, thus contributing to volume-driven care that may have limited clinical value. Advancing future Choosing Wisely initiatives in oncology must extend beyond clinician stewardship to address patient expectations about their healthcare interactions. This strategy would better align with value-based care and promote health system sustainability. CSS based on pre-diagnosis HCRU. Cancer site Pre-diagnosis HCRU 10-year CSS (95% CI) HR for CSS (95% CI) P value Breast Low 89% (87%, 92%) Reference 0.55 Breast High 91% (90%, 93%) 0.93 (0.74,1.18) Lung Low 63% (58%, 69%) Reference 0.72 Lung High 68% (64%, 72%) 1.04 (0.85,1.26) Colon Low 90% (87%, 93%) Reference 0.39 Colon High 92% (89%, 95%) 0.82 (0.52,1.29)
Using stage-matched genomic analysis to support molecular similarity of unknown-primary Merkel cell carcinoma and identify ARID1B as a prognostic marker.
9584 Background: Merkel cell carcinoma (MCC) can present as nodal disease without an identifiable cutaneous primary, termed unknown-primary MCC (UP-MCC). UP-MCC has been reported to have distinct clinical outcomes compared with stage-matched known-primary MCC (KP-MCC), but its underlying biology is poorly understood. We compared the genomic landscapes of UP-MCC and stage IIIB KP-MCC and evaluated whether specific alterations are associated with survival in UP-MCC. Methods: A retrospective analysis was performed using the Dana-Farber Cancer Institute MCC repository (2012-2023). UP-MCC was defined as pathologically confirmed MCC without a detectable cutaneous primary; the comparator cohort comprised stage IIIB KP-MCC. Tumors underwent targeted next-generation sequencing (OncoPanel; 447 genes), and tumor mutational burden (TMB) and Merkel cell polyomavirus (MCPyV) status were recorded. Gene-level alteration frequencies were compared using Fisher’s exact test, and MCC-specific survival (MCCSS) and overall survival (OS) were assessed using Kaplan-Meier and Cox proportional hazards models. Results: Fifty-two patients were included (20 UP-MCC, 32 KP-MCC). UP-MCC and KP-MCC showed highly overlapping genomic landscapes, including similar distributions of MCPyV status, TMB, and recurrent driver alterations, with no gene-level differences detected. Within UP-MCC, SETBP1 alterations were associated with worse MCCSS (HR 6.74; 95% CI, 1.33–34.07; p=0.030), while ARID1B alterations identified a markedly high-risk UP-MCC subgroup with inferior MCCSS (HR 27.1; 95% CI, 3.02–243.11; p<0.001) and OS (HR 16.9; 95% CI, 1.9–152.9; p=0.004). All five ARID1B -altered UP-MCC patients died of MCC, compared with 4 of 15 ARID1B –wild-type; in multivariable analysis, ARID1B remained an independent predictor of worse MCCSS (p=0.003), whereas SETBP1 did not (p=0.313). In KP-MCC cohort, ARID1B was not significantly associated with MCCSS (p=0.1883). Conclusions: Stage-matched genomic profiling demonstrates molecular similarity between UP-MCC and KP-MCC, supporting the hypothesis that UP-MCC often represents regressed primary cutaneous tumors. Within UP-MCC, ARID1B alterations identify a high-risk subgroup with markedly inferior MCCSS, suggesting that targeted genomic profiling may refine risk stratification, surveillance strategies, and clinical trial design for patients with UP-MCC and these findings warrant validation in larger, multi-institutional cohorts. Baseline characteristics of UP-MCC and stage IIIB KP-MCC cohorts. Parameter UP-MCC (n=20) KP-MCC stage IIIB (n=32) p-value Age, mean (SD), years 68.2 (12.9) 74.2 (12.1) 0.204 Male sex, n (%) 13 (65.0) 26 (81.3) 0.323 MCPyV positive, n (%) 11 (55.0) 15 (46.9) 0.776 High TMB, n (%) 8 (40.0) 11 (34.4) 0.909 TMB, mean (SD), mut/Mb 14.1 (13.9) 18.7 (24.6) 0.572 SD= Standard Deviation.
Cardiovascular outcomes of durvalumab versus pembrolizumab in gastric cancer: A propensity-matched TriNetX analysis.
e24017 Background: Immune checkpoint inhibitors targeting the PD-1/PD-L1 axis are increasingly incorporated into gastric cancer treatment. Although rare, ICI-associated myocarditis is highly lethal and often presents early. PD-1 inhibition with pembrolizumab blocks both PD-L1 and PD-L2, resulting in broader immune activation, whereas PD-L1 inhibition with durvalumab preserves PD-1/PD-L2 signaling and may limit immune-mediated cardiotoxicity. In a population with substantial baseline cardiovascular risk, defining differential cardiovascular effects between these agents is clinically essential. Methods: A retrospective TriNetX U.S. Collaborative Network analysis compared cardiovascular outcomes among patients with gastric cancer treated with pembrolizumab versus durvalumab over follow-up periods of 1 year, 3 years, and lifetime. Propensity score matching was performed to balance demographics and baseline cardiovascular, secondary metastasis and metabolic comorbidities, including diabetes mellitus, hypertension, obesity, dyslipidemia, liver fibrosis, and cirrhosis. Outcomes assessed included all-cause mortality, stroke, major adverse cardiovascular events, heart failure, and arrhythmias, evaluated using time-to-event analyses. Results: The matched cohorts each comprised 327 individuals, with balanced baseline demographic and cardiovascular characteristics. Durvalumab was consistently associated with lower mortality and MACE risk across all time horizons (all p < 0.0001). (Table 1) Rates of stroke, heart failure, myocarditis, pericarditis, and arrhythmia were low across all time horizons and did not differ significantly between treatment groups. Conclusions: In a propensity-matched real-world cohort of patients with gastric cancer, durvalumab was associated with significantly lower all-cause mortality and MACE at 1-year, 3-year, and lifetime follow-up compared with pembrolizumab, Although limited by the observational design and potential residual confounding, these findings suggest a favorable survival and cardiovascular risk profile for durvalumab ( PD – L1) in this population and support the need for prospective and randomized studies to confirm these results. Cardiovascular outcomes in propensity-matched gastric cancer cohorts. Outcome Time Horizon Durvalumab Risk (%) Pembrolizumab Risk (%) Risk Ratio (95% CI) p-value Death 1 year 15.95 36.50 0.44 (0.33–0.58) <0.0001 3 years 18.10 49.08 0.37 (0.29–0.48) <0.0001 Lifetime 18.10 51.23 0.35 (0.27–0.46) <0.0001 MACE 1 year 17.91 38.64 0.46 (0.35–0.62) <0.0001 3 years 20.27 52.54 0.39 (0.30–0.50) <0.0001 Lifetime 20.27 54.58 0.37 (0.29–0.48) <0.0001
FLT3 inhibitors across the allogeneic transplant continuum in <i>FLT3</i> -mutated acute myeloid leukemia: A systematic review and meta-analysis.
e18523 Background: FLT3-mutated AML (30% of cases) carries a high relapse risk despite allogeneic hematopoietic stem cell transplantation (allo-HSCT). Currently, FLT3 inhibitor use is typically limited to the post-transplant maintenance phase. However, there may be substantial benefit to coordinated FLT3 inhibition throughout distinct treatment phases: from salvage therapy, which prepares patients for transplantation, to post-transplant consolidation after allo-HSCT. The integrated role in both the pre-transplant (bridge-to-transplant salvage therapy) and post-transplant (maintenance) phases has not been systematically evaluated. This study assesses the efficacy of FLT3 inhibitors in both phases. Methods: We conducted a systematic review of phase II–III trials and observational studies evaluating FLT3 inhibitors in adults with FLT3-mutated AML used as (1) salvage/bridge to allo-HSCT and/or (2) post-transplant maintenance. Primary outcomes were composite complete remission (CR/CRi) and proportion proceeding to allo-HSCT in the salvage setting, and relapse-free survival (RFS) in the maintenance setting. Random-effects meta-analysis was performed for randomized maintenance trials with comparator arms and extractable hazard ratios. Salvage and bridge outcomes were summarized descriptively due to heterogeneity. Results: Across included studies, FLT3 inhibitors used as salvage therapy in relapsed/refractory AML consistently achieved clinically meaningful remission and facilitated subsequent transplantation. In the phase III ADMIRAL trial, gilteritinib achieved a composite remission rate of 54% and enabled 26% of patients to proceed to allo-HSCT, compared with 22% remission and 15% transplant rates with salvage chemotherapy. For post-transplant maintenance, quantitative meta-analysis of randomized sorafenib-based trials demonstrated a significant reduction in relapse or death compared with control (pooled HR 0.45, 95% CI 0.31–0.66; I²=0%). In sensitivity analyses incorporating midostaurin maintenance, results were consistent, though underpowered. Exploratory pooled analysis of overall survival across maintenance studies showed improved survival (HR 0.55, 95% CI 0.36–0.86). Rates of acute and chronic graft-versus-host disease were generally comparable between maintenance and control arms, while treatment discontinuation varied by agent. Conclusions: FLT3 inhibitors establish standard-of-care benefit across the allo-HSCT continuum. Pre-transplant salvage FLT3 inhibition improves remission and enables transplantation. Post-transplant maintenance with sorafenib (all patients) and gilteritinib (MRD-positive) significantly reduces relapse and improves survival. These findings support integrated FLT3 inhibition across the transplant continuum as standard of care for FLT3-mutated AML.
Neoadjuvant PD-1 inhibitor plus paclitaxel, pingyangmycin, and platinum-based chemotherapy (4P therapy) as a bladder-sparing strategy for high-risk non–muscle-invasive/muscle-invasive bladder cancer: A retrospective cohort study.
e16606 Background: To evaluate the efficacy and safety of neoadjuvant PD-1 inhibitor plus paclitaxel, pingyangmycin, and platinum-based chemotherapy ("4P" therapy) in high-risk non-muscle-invasive/muscle-invasive bladder cancer (N/MIBC), with a focus on bladder preservation and neoadjuvant optimization. Methods: This single-center, observational, retrospective study enrolled patients with cT1-4N0M0 N/MIBC from July 2022 to June 2024. Thirty-three patients received "4P" therapy followed by Transurethral Precise Resection of Bladder Tumor (TUPRBT). Radiological and pathological assessments were conducted every two cycles. The primary endpoint was the objective response rate (ORR); secondary endpoints included the disease control rate (DCR), complete response (CR), partial response (PR), and bladder preservation rate. Safety was evaluated according to CTCAE version 5.0 criteria. Results: Among the 33 patients (median age: 65 years; 87.8% male), ECOG performance status scores were 0 (78.8%) and 1 (21.2%). Clinical stages were T1 in 5 patients (15.2%) and T2-4 in 28 patients (84.8%). Patients were stratified by treatment cycles: 1 cycle (6 patients, 3 CR, 3 PR), 2 cycles (24 patients, 10 CR, 10 PR, 4 SD), and 3 cycles (3 patients, 3 PR), and the average treatment cycles were 1.91. The ORR for all patients was 87.8% (29/33), with a CR rate of 39.3% (13/33). DCR was 100% (33/33), with no cases of progressive disease. Among the different PD-1 inhibitor subgroups, the toripalimab-treated patients achieved an ORR of 92.0% (23/25) and a CR rate of 44.0% (11/25). Among the 26 patients who underwent TUPRBT, pCR was achieved in 9 (34.6%). At a median follow-up of 29.3 months, the median EFS was 29.3 months (95% CI: 22.8 – NE), with 1- and 2-year EFS rates of 84.5% (95% CI: 72.0–97.0%) and 65.5% (95% CI: 44.8–86.2%). The median OS was not reached, with 1- and 2-year OS rates of 93.7% (95% CI: 85.3–100%) and 81.2% (95% CI: 66.1–96.3%). The overall bladder preservation rate was 93.9% (31/33), as 2 patients ultimately underwent radical cystectomy. Disease recurrence occurred in 11 patients. Four deaths occurred (1 case of brain metastasis, 1 case of severe pneumonia, 2 cases of unspecified causes), of which only 1 was a bladder cancer-specific death. Treatment-related adverse events were primarily grade 1–2 and manageable, supporting the tolerability of the “4P” therapy. Conclusions: The "4P" regimen, with high efficacy and manageable toxicity, achieved a 93.9% bladder preservation rate in patients undergoing TUPRBT. By retaining bladder function and reducing surgical morbidity, it is an attractive choice for high-risk non-muscle-invasive and muscle-invasive bladder cancer patients' bladder preservation and neoadjuvant optimization.
The stage-wise macromolecular assembly and structure evolution of silk along the silk gland
Quality improvement: Enhancing family history documentation for invasive cancer patients in a community oncology clinic.
e23284 Background: Accurate collection and documentation of family history in patients diagnosed with invasive malignancies is a vital metric for improving cancer patient outcomes. Unfortunately, comprehensive documentation in Electronic Health Records remains suboptimal. Methods: Our study included patients aged 18 and older with a diagnosis of invasive malignancy. Data was gathered by impartial abstractors and reported to the Michigan Oncology Quality Consortium (MOQC). The Abstractors extracted documentation information as yes or no from patient charts and submitted them to MOQC in a de-identified format. Patient data was collected on a rolling basis throughout the year and presented at semi-annual MOQC meetings. The Intervention: The QI initiative focused on provider education regarding accurate documentation of specific family history metrics in oncology clinic notes. These metrics were Invasive cancer diagnoses history in first degree relatives like biological parents, siblings and children and second-degree relatives like biological grandparents, aunts and uncles. Age at the time of diagnosis for affected relatives, or document age unknown A definitive closing statement: "No additional oncological history in first- or second-degree relatives." Intervention was initiated in early 2024 and education was completed via in-person lecture during quarterly staff meetings and supplemental informational pamphlets posted in provider workspaces. Data was collected from MOQC already de-identified over three years: 2023, 2024 and 2025. Results: Prior to initiation of the quality improvement, in the year 2023, 16% of the audited charts (99 patients) had complete documentation of Family History parametric in the notes. In year 1 of initiation of the quality improvement, in 2024, 22% of the audited charts (264 patients) had complete documentation of Family History parametric in the notes. In year 2 of initiation of the quality improvement, in 2025, 29% of the audited charts (235 patients) had complete documentation of Family History parametric in the notes. Following the initiation of the project, there was an absolute improvement of 13% in documentation compliance over two years. Conclusions: Suboptimal data collection can result in missed opportunities for hereditary genetic evaluations and cancer screenings. While our QI initiative demonstrated a steady upward trend in compliance, there remains significant room for growth. Barriers to sharper response include our large practice size and high provider turnover being a fellowship training site We have not yet reached the state target of 35%. To bridge this gap, future strategies may include: Increasing the frequency of provider education sessions. Implementing internal random chart audits. Targeted feedback to providers with consistently low documentation rates.
Clinical and genetic characteristics of MSI-high versus MSS gastric cancer patients: A single-institution experience.
e16091 Background: Microsatellite instability-high (MSI-H) gastric cancers (GC) are unique in genotype and have shown to achieve better clinical outcomes compared to microsatellite stable (MSS) GC. However, the distinct clinicopathologic features of MSI-H/MSS associated with genetic mutations remain poorly established outside of clinical trials. Our study aims to characterize the real-world clinicopathological and genetic characteristics of patients with MSI-H versus MSS GC. Methods: A single institution retrospective study was performed at City of Hope (COH) of 155 patients with reported MSI-H and MSS gastrointestinal cancers. Baseline characteristics were recorded including age, ethnicity, histology, stage, and overall survivals (OS). Next-generation sequencing (NGS) analysis was available for 35 samples with confirmed MSI-H/MSS GC. A two-sided Fisher’s exact test was performed to compare mutation frequencies between each group (MSI-H vs. MSS). Results: Clinically, patients with MSI-H GC were older with a median age of 67 years versus 64.5 years for MSS and more common for poorly differentiated adenocarcinoma. MSI-H tumors were more common for TMB-high (71%) compared to MSS (0%). MSI-H GC tumors showed greater PD-L1 positivity (100%) compared to MSS (0%), however, neither MSI-H nor MSS were positive for HER2 (0%). Among the top 36 mutated genes, patients with MSI-H showed a higher number of mutated genes per patient compared with patients with MSS (9.5 vs. 3.5).The most common mutated genes for patients with MSI-H were ARID1A (83%), RNF43 (50%), PIK3CA (42%), MSH3 (37%), TP53 (33%), MSH6 (33%), KMT2B (33%), and ACVR2A (33%). Meanwhile, most common mutated genes for patients with MSS were ARID1A (54%), TP53 (54%), and CDH1 (36%). Conclusions: Our real-world retrospective study identified significantly different clinicopathologic and genetic features of MSI-H versus MSS GC. Compared to patients with MSS GC, those with MSI-H GC were characterized by higher tumor mutational burden and PD-L1 positivity, both shown to predict response to immunotherapy and achieve favorable clinical outcome. Our findings provide valuable insight regarding real-world clinicopathological and genetic differences between MSI-H and MSS GC and support individualized biomarker-driven treatment strategies.
Neoadjuvant pyrotinib plus trastuzumab combined with chemotherapy for HER2+/HR+ early breast cancer: A multicenter phase 2 trail.
e12641 Background: HER2+/HR+ early breast cancer has a lower pathological complete response (pCR) rate after neoadjuvant therapy than HER2+/HR- subtype, underscoring an urgent need for new regimens. Pyrotinib is approved by National Medical Products Administration for neoadjuvant treatment of HER2+ breast cancer in China. Herein, we conducted this phase 2 multicenter trial to evaluate the efficacy and safety of pyrotinib plus trastuzumab combined with chemotherapy for this setting. Methods: This open-label multicenter phase 2 study enrolled untreated, histologically confirmed stage II-III HER2+/HR+ breast cancer patients. Patients received 6 cycles of neoadjuvant therapy: pyrotinib (initial dose: 240 mg qd) + trastuzumab + paclitaxel (nab-paclitaxel 260 mg/m² or docetaxel 75 mg/m², d1, q3w) + carboplatin (AUC = 5, d1, q3w). Diarrhea prophylaxis was permitted. Based on diarrhea severity, pyrotinib dose escalation was allowed, or the regimen could be switched to trastuzumab + pertuzumab (HP) + chemotherapy at physicians’ discretion if the patient cannot tolerate the diarrhea toxicity of pyrotinib. Patients completing or discontinuing neoadjuvant therapy underwent breast surgery. Adjuvant targeted/endocrine therapy was administered per physician judgment, and radiotherapy followed standard practice. Primary endpoint: pCR rate (ypT0/Tis ypN0). Secondary endpoints: objective response rate (ORR), breast pCR (bpCR), event-free survival (EFS), overall survival (OS), and safety. Results: As of December 2025, 16 patients completed neoadjuvant therapy and surgery. Median age was 52 years (range: 28-60 years); 10 (66.7%) had stage II and 6 (33.3%) stage III disease. 10 patients (66.7%) received pyrotinib + trastuzumab + chemotherapy (noted as Cohort 1), while 6 patients (33.3%) discontinued pyrotinib during treatment (after 1 to 4 cycles) and switched to HP + chemotherapy (noted as Cohort 2). Overall pCR rate was 50.0% (8/16): 60.0% (6/10) in Cohort 1 and 33.3% (2/6) in Cohort 2. With 1 additional patient achieving bpCR in Cohort 1, bpCR rate was 56.3% (9/16). ORR was 93.8% (15/16). Regarding treatment tolerance, 11 patients (68.8%) maintained 240 mg/day pyrotinib, 4 tolerated escalation to 320 mg/day, and 1 to 400 mg/day. Common treatment-related adverse events included diarrhea, nausea/vomiting, neutropenia, rash, and oral ulcers. Diarrhea occurred in 100% of patients for any grade and 43.8% (7/16) for grade ≥3. Only 1 patient required hospitalization for severe neutropenia with fever. No treatment-related deaths occurred. Conclusions: The neoadjuvant regimen of pyrotinib plus trastuzumab combined with chemotherapy showed promising efficacy and manageable safety in HER2+/HR+ early breast cancer, even with the low pyrotinib dose. This study is ongoing. Clinical trial information: ChiCTR2500102478.
Assessing the relationship between event-free and overall survival in the era of perioperative chemoimmunotherapy in gastroesophageal cancers.
e16135 Background: Perioperative chemoimmunotherapy is emerging as a global standard for non-metastatic gastroesophageal cancers, but data in phase III trials are mixed. Event-free survival (EFS) is a common endpoint in perioperative trials but whether EFS is a surrogate for overall survival in the era of chemoimmunotherapy is not well studied. Understanding EFS performance as a surrogate for overall survival is important to inform regulatory approval and early access to promising agents in this tumor type. Methods: We performed a systematic literature review to identify all randomized controlled trials (RCTs) evaluating neoadjuvant or perioperative systemic therapy for gastroesophageal adenocarcinomas. Using weighted linear regression of the hazard ratios extracted from the published literature, a meta-analysis was performed to evaluate the strength of correlation between EFS and overall survival (OS) across trials. Consistent with criteria from the Institute for Quality and Efficiency in Health Care, strong correlation was defined as a correlation coefficient (R) ≥0.85. Sensitivity analyses examining trials with heterogeneous tumor locations, perioperative systemic therapy only, and those with a clear definition of EFS, were performed to evaluate the consistency of the findings. Results: A total of 30 RCTs were included in the main analysis. Trial-level comparison identified a positive relationship between EFS and OS (estimated coefficient 0.68, 95% CI, 0.50–0.85; p < 0.001) corresponding to a correlation coefficient (R) of 0.83. Surrogate threshold effect (STE) analysis identified an HR for EFS of 1.18 or lower as a threshold at which the lower bound of the 95% confidence interval for HR_OS excluded 1.0.This correlation was preserved in sensitivity analyses, including 20 studies that defined EFS as beginning from cycle 1, day 1 of neoadjuvant therapy (slope relating log(HR_EFS) to log(HR_OS) was 0.6536 (95% CI, 0.4551–0.8521; p < 0.0001). Conclusions: EFS remains a good surrogate for OS in modern phase III perioperative chemoimmunotherapy trials including MATTERHORN, supporting the validity of intermediate endpoints for use in treatment regulatory decisions and drug development in operable gastroesophageal adenocarcinomas.
Enterprise-wide implementation of multicancer detection testing across a large academic health system.
10578 Background: Multicancer detection (MCD) testing using cell-free DNA methylation patterns is a novel blood-based approach to cancer screening. In 2022, Mayo Clinic implemented offering MCD testing into routine clinical practice. Prior real-world reports have been limited to single-site experiences. We evaluated enterprise-wide implementation of MCD testing across all three Mayo Clinic sites using the Mayo Clinic Platform. Methods: We performed a retrospective analysis of de-identified clinical data from Mayo Clinic Platform Discover, including GRAIL Galleri tests performed at Mayo Clinic sites in Rochester, MN; Scottsdale, AZ; and Jacksonville, FL. This activity did not require IRB review per 45 CFR 46.102. Demographics, test results, cancer diagnoses, and prior cancer history were extracted using a privacy-preserving protocol, with data de-identified by expert determination. ICD-10 diagnostic codes were used to identify subsequent cancer diagnoses. Results: A total of 8,201 Galleri tests were performed in 7,300 patients; some patients underwent repeat testing. Fifty-seven tests (0.7%) were positive. Among MCD-positive patients, cancer diagnostic codes were identified in 32 patients, yielding a positive predictive value (PPV) of 58.7%. The most common cancers identified were oropharyngeal (34.4%), lymphoid and hematopoietic (34.4%), and gastrointestinal malignancies (18.8%).Among 8,144 MCD-negative tests (99.3% of all testing performed), 423 patients (5.2%) subsequently received a cancer diagnostic code, yielding a negative predictive value of 92.5%. The most commonly identified cancers in this group were male genital organs (36.6%), urinary tract (10.6%), and thyroid or other endocrine glands (7.3%). Overall, 10.1% of patients had a prior cancer history, which was more common in MCD-positive than MCD-negative patients (22.6% vs 10.0%). Patients with positive tests were older (mean age 68.4 vs 61.5 years) and more often male (67% vs 62%). Race distribution was similar. Conclusions: Enterprise-wide implementation of MCD testing across diverse clinical settings is feasible within routine workflows. Over half of positive tests were associated with cancer diagnostic codes, supporting the clinical promise of MCD testing. Limitations include incomplete capture of externally diagnosed cancers and lack of cancer signal-of-origin data. The Mayo Clinic Platform enables scalable, secure analytics to support future cancer interception strategies.
Development and validation of a preoperative pathological risk prediction model for thymic tumors: A retrospective multicenter diagnostic study.
8122 Background: The preoperative diagnosis of thymic tumors currently relies on preoperative CT data and the experience of the surgeon, and there is still a lack of highly effective noninvasive radiomic prediction model to assist in clinical decision making. To develop and validate a preoperative CT-based radiomic model to accurately predict the pathological risk staging of thymic tumors. Methods: This is a retrospective diagnostic study that included patients from two independent centers Charité Universitätsmedizin Berlin (N = 74) as well as the First Hospital of Zhengzhou University (N = 143) between September 2003 and July 2024. WHO pathological type A, AB, and B1 of the postoperative thymic tumor were divided into low-risk groups (N = 110), and type B2, B3, and C (N = 107) were classified as high-risk groups. The patients were randomly divided into a training group (N = 130) and a validation group (N = 87) in a ratio of 6:4. Preoperative CT imaging of thymic tumor was used to define the tumor area and the peritumoral area (5 mm) and to extract the radiomic features using 3Dslicer software and the in-house pyradiomic software package. Feature selection and model development were conducted using the least absolute shrinkage and selection operator（LASSO）algorithm and logistic regression analysis. Results: A total of 217 patients undergoing thymectomy were included in this study. After univariate analysis, patients' age, smoking history, and the presence of combined myasthenia gravis (MG) were included in the clinical model, and the clinical model achieved AUC = 0.75 and 0.65 in the training and validation groups, respectively. The combined radiomic model has better performance than the two radiomic models alone, with AUC values of 0.86, 0.73, respectively. The nomogram model conducted by combined radiomic model with the clinical model achieved the best results in training (AUC = 0.89) and validation cohort (AUC = 0.84). Calibration curve and decision curve analysis (DCA) illustrated the clinical usability and reliability of the model. Conclusions: By combining the clinical model with a combined radiomic model, the nomogram model can effectively differentiate pathological risk staging of thymic tumors, providing surgeons with a potential preoperative decision support.
Fecal microbiota transplantation to reverse resistance to TKI plus PD-1 inhibitors in unresectable hepatocellular carcinoma: A phase 2 clinical trial.
e16205 Background: Tyrosine-kinase inhibitors (TKIs) and immune-checkpoint inhibitors (ICIs) improve survival in unresectable hepatocellular carcinoma (uHCC), but acquired resistance limits benefit. Fecal microbiota transplantation (FMT) has shown potential to reverse immunotherapy resistance. This study evaluated the efficacy of TKI + ICI in patients with uHCC and explored whether FMT post-progression overcome resistance. Methods: This prospective, single-arm phase 2 trial (ChiCTR2200058471) enrolled HCC patients with Child-Pugh A, BCLC-B/C, and ECOG PS 0–1. Treatment-naïve patients received first-line lenvatinib + PD-1 inhibitor, while previously treated patients received second-line regorafenib + tislelizumab. Patients who achieved CR, PR, or SD continued their respective regimens until radiologically confirmed disease progression. Upon progression, oral FMT capsules (30 g daily for 3 days, q3w×3 cycles) were added to the original treatment regimen. The primary endpoint was the DCR post-FMT. Secondary endpoints included ORR, PFS, OS, and safety post-FMT. Results: Between Aug 2022 and Mar 2025, 37 patients were enrolled. As of Jan 22, 2026, in the first-line cohort (n = 18), the ORR was 33.3%, DCR was 88.9%, median PFS was 7.9 months (95% CI: 4.5–16.0), and median OS not reached (95% CI: 16.7 In the second-line cohort (n = 19), the ORR was 0%, DCR was 31.6%, median PFS was 2.3 months (95% CI: 1.6–4.2), and median OS of 9.8 months (95% CI: 5.9–18.1). Across both cohorts, 13 patients received FMT post-progression. Among them, 8 achieved SD, resulting in a DCR of 61%. The median PFS was 5.1 months (95% CI: 2.3–8.3), and median OS remained unreached (95% CI: 8.5–NE). Treatment-related adverse events (AEs) occurred in 86.5 % of all patients, with grade ≥3 AEs observed in 11.8 %. FMT-related toxicity was mild, limited to one grade 1 fever and two grade 1 diarrhoea events. Longitudinal metagenomic analysis of HCC patients across different phases of targeted immunotherapy (with and without FMT) revealed that fluctuations in the abundance of Bacteroides caccae and Phocaeicola plebeius were closely synchronized with mirrored changes in therapeutic efficacy, showing a significant positive correlation. This dynamic association suggests that these microbial taxa contributes to sustaining treatment response and mediates the transient reversal of resistance following FMT. Conclusions: TKI combined with ICI demonstrated favorable efficacy and safety in advanced HCC. FMT post-progression achieved 61% DCR with tolerable toxicity, supporting its potential to overcome TKI-ICI resistance. Bacteroides caccae and Phocaeicola plebeius may be biomarkers of efficacy. Larger RCTs are needed to validate these findings and clarify microbiota-related mechanisms. Clinical trial information: ChiCTR2200058471.
From manual to AI-assisted: Improving infusion scheduling in a high-volume cancer center.
e23297 Background: Efficient use of infusion center resources is critical for high-volume cancer centers, yet infusion scheduling is often manual and labor-intensive. More effective scheduling has the potential to enhance patient-centered care through appointment consolidation, improve nursing continuity, and increase operational efficiency. We implemented Epic’s artificial intelligence (AI)-enabled, decision tree–based scheduling workflow as the standard process for infusion appointment booking at a large academic cancer center and evaluated its impact on operational outcomes. Methods: In September 2025, an automated infusion template generator that used historical operational data to optimize scheduling capacity was implemented, replacing manual electronic calendar-based scheduling. The workflow incorporated a decision tree–based tool guiding schedulers to align treatment requests with available resources and was informed by multidisciplinary stakeholder input. Statistical process control (SPC) charts evaluated changes in operational metrics during the three months before and after intervention. Primary outcomes included time spent scheduling appointments, nursing continuity defined by the proportion of appointments booked with a patient’s primary nurse, and chair utilization rate. A balancing measure was patient layover time between same-day appointments. Control limits were set at ±3 standard deviations (99.7%). Special cause variation was identified using Institute for Healthcare Improvement (IHI) criteria: points outside control limits, shifts of ≥7 consecutive points on one side of the center line, and trends of ≥6 consecutively increasing or decreasing points. Results: During the study period, 22,514 infusion appointments were scheduled. After implementation, SPC analysis demonstrated sustained shifts in scheduling time and patient layover time. Mean scheduler time per appointment decreased from 65 to 40 seconds, corresponding to an estimated 20 person-hours saved per month. Mean patient layover time between same-day appointments decreased from 53 minutes to 45 minutes. Following the process change, the proportion of chemotherapy appointments booked with a patient’s primary nurse rose above control limits. In the three months after implementation, the proportion of chemotherapy appointments booked with a patient’s primary nurse rose from 13 to 17%. No special cause variation was observed for chair utilization. Conclusions: AI-enabled decision tree scheduling was feasible in a high-volume oncology infusion center and was associated with improvements in select operational metrics including scheduling time, nursing continuity, and patient layover time, without adversely affecting chair utilization.
MUC5AC-driven transcriptional reprogramming for stratification of pancreatic ductal adenocarcinoma by immune suppression and stromal remodeling.
4194 Background: Mucin 5AC (MUC5AC) is aberrantly expressed in pancreatic ductal adenocarcinoma (PDAC) and associated with tumor progression; however, its role in altering the tumor immune microenvironment (TIME) and immunotherapy-relevant pathways is poorly understood. Thus, we evaluated whether MUC5AC expression delineates PDAC into distinct transcriptional and immunologic subtypes based on TIME. Methods: Bulk RNA-seq data from 730 PDAC tumors (Oncology Research Information Exchange Network, ORIEN) and 360 normal pancreatic tissues (GTEx) were analyzed. Differential gene expression and pathway enrichment were assessed using Gene Ontology and KEGG analyses. Immune infiltration and stromal states were inferred using multiple complementary deconvolution algorithms (CIBERSORT and xCell). Tumors in the lowest quartile of MUC5AC expression (≤25th percentile; MUC5AC-low, MUC5AC-L) were compared with the remaining samples (MUC5AC-high, MUC5AC-H). Associations between MUC5AC expression and immune cell populations and immunomodulatory genes were evaluated by using correlations and group-based analyses with false discovery rate (FDR) corrections. Results: Principal component analysis demonstrated clear separation between PDAC tumors and normal pancreas and revealed distinct transcriptional programs between MUC5AC-H (N = 544) and MUC5AC-L (N = 186) tumors. MUC5AC-H tumors exhibited a profoundly immunosuppressive and stromal-remodeled TIME, including enrichment of M2 macrophages, regulatory T cells, activated natural killer ells, fibroblast-associated signatures, and significantly higher immune and stromal scores (multiple FDR ≤10⁻⁶ to ≤10⁻²⁸). In contrast, MUC5AC-L tumors showed relatively enrichment of immune-activating populations, including naïve B cells, dendritic cell subsets, resting CD4⁺ memory T cells, and CD8⁺ T cells (multiple FDR < 10⁻¹⁵). MUC5AC expression strongly correlated with immune checkpoint genes CD274 (PD-L1) and TIGIT , and with multiple suppressive immune populations (p < 0.01 to < 10⁻¹⁰). Conclusions: MUC5AC defines a transcriptionally distinct, immunosuppressive PDAC subtype characterized by stromal activation, macrophage/Treg enrichment, and immune checkpoint engagement. MUC5AC may serve as a biomarker for immune stratification and supports rational combination strategies integrating stromal targeting with immune checkpoint modulation in PDAC.
Impact of SGLT2i and GLP1 agonists on prostate cancer incidence and outcomes: A real-world propensity score–matched cohort study.
e17148 Background: Prior studies have shown a decreased incidence of PCa in males with type two diabetes mellitus (TIIDM) receiving GLP1 agonists (GLP1-RA’s) as well as those receiving SGLT2 inhibitors (SGLT2i’s). To our knowledge, no studies have compared the incidence of PCa between these two classes of medications in males with TIIDM. Further, our study aims to compare their impact on PCa-specific outcomes. Methods: We utilized TriNetX’s US Collaborative Network. Males aged 18 to 90 with TIIDM were divided into three cohorts: those receiving SGLT2i’s, those receiving GLP1-RA’s, and those receiving neither. Patients with type one diabetes mellitus were excluded from all cohorts. Cohorts were propensity score-matched based on age, sex, race, social determinants of health, other hypoglycemic medications, endocrine disorders, genitourinary disorders, BMI, BPH, and BRCA1/2 mutation status. Using ICD-10 codes, we evaluated the following outcomes: mean PSA levels, PCa incidence, Gleason scores 7-10, and presence of osseous metastases. Generalized linear models were used to measure the association, and estimates were presented as mean values and odds ratios with 95% confidence intervals. Results: After matching, cohorts consisted of 309,556, 237,965, and 237,238 patients, respectively. Mean ages varied between 58.4 – 64.1 years of age. Caucasians accounted for 66.3-70.5% and black or African American for 13.9-15.1% of patients. Compared to controls, patients receiving SGLT2i had a statistically lower risk of developing PCa (OR 0.842, 95% CI 0.809 - 0.876, p < 0.0001). The SGLT2i. cohort had a significantly lower risk of developing PCa compared to the GLP1-RA cohort (OR 0.823, 95% CI 0.786-0.863, p < 0.0001). There was no difference in PCa incidence between the GLP1-RA cohort and control group (OR 1.024, 95% CI 0.978 - 1.072, p = 0.3071). Mean PSA scores were significantly lower in the SGLT2i group compared to control (4.10 vs 5.99, p = 0.0003) and the GLP1-RA group compared to control (3.203 vs. 4.206, p = 0.0066). There was a significantly lower risk of Gleason score 7 in the SGLT2i group vs. the GLP1-RA group (OR 0.618, 95% CI 0.43-0.89, p = 0.0089). The risk of osseous metastasis was also significantly lower in the SGLT2i group compared to control (OR 0.794, 95% CI 0.743-0.849, p < 0.0001), the GLP1-RA group compared to control (OR 0.763, 95% CI 0.703-0.828, p < 0.0001), and in the SGLT2i. group compared to the GLP1-RA group (OR 1.099, 95% CI 1.01-1.196, p = 0.0279). Conclusions: Overall, our data indicate that SGLT2i's may lower mean PSA levels, as well as the risk of PCa and osseous metastases in males with TIIDM compared to matched controls. Further, SGLT2i's appear to have superior effect on these outcomes compared to GLP1-RA's. GLP1-RA's may reduce mean PSA scores and osseous metastasis risk compared to matched-controls. However, more prospective studies are needed to confirm these findings.
Perceived bloggers’ competence as a moderator of the relationship between information quality, source credibility, and health information adoption in social media
Backgroud In the context of digital transformation, social media has become a key channel for the public to obtain health information, and bloggers play an important role in shaping public health behaviors. However, the internal mechanism of how PerceivedBloggers’ Competence (PBC) and information characteristics jointly affect Information Adoption (IA) has not been fully revealed. Aim This study integrates the Information Adoption Model (IAM) with Social Cognitive Theory (SCT) to examine the relationship between Perceived Bloggers’ Competence and Information Adoption, and to explore how Perceived Bloggers’ Competence moderates the impact of Information Quality (IQ) and Information Credibility (IC) on information adoption through Information Usefulness (IU). Method A total of 1219 participants (648 males and 571 females) were recruited using a cross-sectional online survey design. Data were collected through a structured questionnaire and analyzed using SPSS and AMOS. This study constructs a conceptual model with Information Usefulness as a mediator and perceived bloggers’ competence as a moderating variable and employs path analysis along with moderating effect tests to verify the hypotheses. Results In the study, Information Quality and Information Credibility had significantly positive effects on Information Usefulness, which in turn strongly predicted Information Adoption, also Perceived Bloggers’ Competence negatively moderates the relationship between information characteristics and adoption. Conclusion This study challenges the assumption that bloggers’ expertise necessarily promotes Information Adoption, also provides key insights for health communicators, showing that striking a balance between expertise and accessibility is essential for effective public health messaging on social media, exhibits an effect that can be called “competence discounting”.
Broadband terahertz conductivity and polarization dynamics of single-crystal diamond
We present broadband terahertz studies of high-purity single-crystal diamond grown by chemical vapor deposition over a frequency range of 0.2–4.5 THz. The experimentally characterized conductivity exhibits a pronounced non-monotonic frequency dependence that cannot be described by the conventional Drude model. To capture this behavior, we employ a constrained Drude–Lorentz (CDL) model, which incorporates free-carrier transport and an effective off-resonant polarization contribution beyond the measured spectral window. The real part of the dielectric function yields ε1 ≈ 5.64, consistent with the known static dielectric constant of diamond. The CDL model accurately reproduces the broadband response, from which a carrier mobility of ∼2963 cm2/V s and a carrier density of ∼1.05 × 1013 cm−3 are obtained. These results show that the broadband nature of terahertz time-domain spectroscopy is essential for resolving coupled carrier–polarization dynamics in high-purity diamond.