Citations

Background Predictive models employing machine learning algorithms are increasingly being used in clinical decision making, and improperly calibrated models can result in systematic harm. We sought to investigate the impact of class imbalance correction, a commonly applied preprocessing step in machine learning model development, on calibration and modelled clinical decision making in a large real-world context. Methods A histogram boosted gradient classifier was trained on a highly imbalanced national dataset of >1.8 million patients undergoing surgery, to predict the risk of 90-day mortality and complications after surgery. Class imbalance correction strategies including random oversampling, synthetic minority oversampling technique, random under-sampling, and cost-sensitive learning were compared to the natural distribution ('natural'). Models were tested and compared with classification metrics, calibration plots, decision curve analysis, and simulated clinical impact analysis. Results The natural model demonstrated high performance (AUROC 0.94, 95% CI 0.94--0.95 for mortality; 0.84, 95% CI 0.84--0.85 for complications) and calibration (log loss 0.05, 95% CI 0.04--0.05 for mortality; 0.23, 95% CI 0.23--0.24 for complications). Class imbalance mitigation (CSL, ROS, RUS, and SMOTE) did not improve AUROC or AUPRC but increased recall and F1 scores at the expense of precision and accuracy. However, these methods severely compromised model calibration, leading to significant over-prediction of risks (up to a 62.8 % increase) as further evidenced by increased log loss across all mitigation techniques. Decision curve analysis and clinical scenario testing confirmed that the natural model provided the highest net benefit. Conclusion Class imbalance correction methods result in significant miscalibration, leading to possible harm whenused for clinical decision making.

88 Background: The VISION (NCT03511664) and PSMAfore (NCT04689828) clinical trials have demonstrated the effectiveness of 177 Lu-PSMA-617 in patients with mCRPC who have previously received androgen receptor pathway inhibitors (ARPIs) with or without taxane-based chemotherapy. Sipuleucel-T is used for treatment of mCRPC in routine clinical practice, particularly among community urologists; however, to date, there are no data available to understand the effectiveness of 177 Lu-PSMA-617 in patients previously treated with sipuleucel-T. Methods: This was a retrospective, observational study of adult patients with a diagnosis of mCRPC who received 177 Lu-PSMA-617 between March 23, 2022, and June 27, 2025, and had a history of treatment with sipuleucel-T. Data were obtained from the PRECISION data platform, a harmonized dataset of patients with advanced prostate cancer in the US that integrates electronic health records and claims data from community, academic, urology, and medical oncology settings. The index date was the date of 177 Lu-PSMA-617 initiation. Patient characteristics and prostate-specific antigen (PSA) response rates were evaluated descriptively. Progression-free survival (PFS), defined as the time from 177 Lu-PSMA-617 initiation to disease progression or death, was analyzed using Kaplan–Meier curves. Results: A total of 290 patients met the inclusion criteria, of whom 72% were White and 12% were Black. At index, 32% of patients were treated in urology and 68% in oncology settings. The median age was 74 years, the median baseline PSA was 35.3 ng/mL (interquartile range [IQR] 10.0–99.2 ng/mL), and 57% of patients had a Gleason score of ≥8. In addition to sipuleucel-T, prior to 177 Lu-PSMA-617, 97% of patients had received ≥1 ARPI and 79% had received ≥1 taxane. The median follow-up for this cohort was 9.4 months (IQR 5.3–17.5 months). Among 144 patients with available PSA values both before and during 177 Lu-PSMA-617 treatment (representing 50% of the cohort), PSA response rates were as follows: a ≥50% reduction in PSA from baseline (PSA50) was observed in 60%, PSA80 in 42%, and PSA90 in 33% of patients. Overall, the median PFS was 15.2 months (95% confidence interval [CI] 11.6–19.4 months). Conclusions: In this real-world analysis of patients who received 177 Lu-PSMA-617 after sipuleucel-T treatment, the median PFS was similar to that observed in clinical trials, suggesting that 177 Lu-PSMA-617 can be sequenced after sipuleucel-T treatment in appropriate patients.

78 Background: 177 Lu-PSMA-617 is currently under investigation in combination with an ARPI in those with metastatic hormone-sensitive prostate cancer (PSMAddition trial: NCT04720157). This observational study was performed to assess the effectiveness of 177 Lu-PSMA-617 in combination with an ARPI in patients with mCRPC in the real-world setting. Methods: This was a retrospective cohort study of adults with mCRPC in the PRECISION data platform, a comprehensive database of prostate cancer patients compiled from diverse clinical settings in the US. Patients who received 177 Lu-PSMA-617 during March 23, 2022–June 27, 2025 were identified, among whom two cohorts were selected for inclusion: those who had evidence of ARPI use (apalutamide, darolutamide, abiraterone, or enzalutamide) at any time during 177 Lu-PSMA-617 treatment (concomitant ARPI cohort), and those who had no other treatment for mCRPC during 177 Lu-PSMA-617 treatment (no concomitant therapy cohort). The index date was initiation of 177 Lu-PSMA-617. Patient characteristics and prostate-specific antigen (PSA) responses were evaluated descriptively. Progression-free survival (PFS), defined as the time from 177 Lu-PSMA-617 initiation to disease progression or death, was estimated using Kaplan–Meier analysis. Results: A total of 1,880 patients treated with 177 Lu-PSMA-617 were identified, 140 in the concomitant ARPI cohort and 1,740 in the no concomitant therapy cohort. The median age was 74 years. Demographic characteristics were balanced between the groups, except that median baseline PSA was lower in the concomitant ARPI cohort when compared with patients in the no concomitant therapy cohort (18 ng/mL vs. 32 ng/mL). Before 177 Lu-PSMA-617 initiation, 94.3% and 73.6% of patients in the concomitant ARPI and 80.9% and 56.7% in the no concomitant therapy cohort had received ≥1 prior ARPI and ≥1 prior taxane, respectively. PSA response rates were broadly comparable between cohorts. The median (95% confidence interval) PFS was 14.7 (11.7–21.3) months among patients in the concomitant ARPI cohort vs 12.5 (11.5–13.6) months in the no concomitant therapy cohort. Ongoing follow-up aims to evaluate the impact of combining ARPI with 177 Lu-PSMA-617 on long-term clinical outcomes, including overall survival. Conclusions: The results of this study demonstrate that 177 Lu-PSMA-617 delivers clinical benefits inpatients with mCRPC with or without ARPI; however, addition of an ARPI may potentially improve these benefits. Prospective randomized controlled studies are needed to confirm this.

Cancer stem cells (CSCs) are increasingly recognized as central drivers of tumorigenesis, therapeutic resistance, and recurrence across diverse malignancies. This review synthesizes our current understanding of CSC biology across CNS tumors, with a focus on glioblastoma, where stem-like cells are sustained by specialized and overlapping tumor microenvironmental niches. Perivascular, hypoxic, invasive, immunosuppressive, and extracellular matrix-associated niches cooperatively enforce stemness, metabolic adaptability, immune evasion, and phenotypic plasticity, enabling CSC persistence despite maximal surgical resection and standard-of-care therapy. Notably, CSCs extend beyond radiographically defined tumor margins and populate peritumoral regions, providing a biological basis for near-universal recurrence. Advances in multiparametric imaging, stem cell-based ex vivo and in vivo models, and single-cell and spatial profiling have refined insight into CSC heterogeneity, niche dependence, and treatment resistance. Together, these findings reframe therapeutic strategies, highlighting the need for function-preserving maximal resection and multimodal therapies that target both CSC-intrinsic pathways and their supportive microenvironments.

CEUS is a safe and cost-effective means for characterization of indeterminate renal masses, with performance equal to or better than that of contrast-enhanced CT or MRI for stratification and surveillance of cystic renal masses and discrimination of benign avascular cysts from solid renal masses.

Abstract Background: Despite its widespread adoption, there is a lack of data as to whether breast MRI for local staging and surgical planning of newly diagnosed breast cancer patients improves oncologic outcomes. The Alliance A011104 phase III trial was designed to test the hypothesis that MRI detection of mammographically occult disease helps to tailor surgical intervention thereby reducing local recurrence and improving local regional control. Here we report on the primary outcome of 5-year local regional recurrence. Methods: Patients with newly diagnosed clinical stage I-II, ER/PR- and HER2- negative (triple- negative) or ER/PR-negative and HER2-positive breast cancer, deemed eligible for breast- conserving surgery (BCS) based on clinical exam, mammography +/- ultrasound were randomized to undergo staging breast MRI or no further local staging. Patients with multifocal tumors that could be encompassed in a single operative resection were eligible. Patients who were carriers of gBRCA 1/2, or who had bilateral breast cancer, or a prior history of breast cancer were excluded. Initial eligibility was limited to patients not scheduled to receive neoadjuvant chemotherapy; however, a subsequent amendment allowed for enrollment after completion of neoadjuvant chemotherapy. Categorical variables were compared between arms with a chi-square or Fisher’s exact test. LRR, distant recurrence and overall survival (OS) were estimated with Kaplan-Meier point estimates and 95% confidence intervals (CIs). Hazard ratios (HRs) and 95% CIs were obtained from Cox proportional hazards models. Results: Of 319 patients enrolled from 2/2014-12/2019,161 were randomized to undergo breast MRI and 158 to no breast MRI evaluation. The mean age at enrollment was 58.9 years (range 29-85). Most patients (72.1%) had T1 tumors, 93.4% were cN0 and 19.7% were HER2 positive. Systemic chemotherapy was utilized in 85% with 17.6% receiving treatment neoadjuvantly; receipt of systemic therapy was unknown for 15 patients (4.7%). Two hundred and ninety-eight patients (93.4%) underwent surgical intervention; 91.9% underwent BCS as initial surgical procedure with no significant differences noted between groups (92.7% in the no MRI group and 91.9% in the MRI group [p=0.62]). Of patients who underwent breast surgery, 293 underwent axillary evaluation with the majority (270 patients, 92%) undergoing sentinel lymph node biopsy as the only axillary intervention. Among patients who received neoadjuvant chemotherapy, the overall pathologic complete response rate was 39.3%, 29.0%in the MRI arm vs. 52.0% no MRI arm (p=0.10). Receipt of adjuvant radiation was comparable between arms, 85.4% and 85.0% receiving radiation in the no MRI and MRI arms, respectively. Among 298 patients evaluable for the primary endpoint, with median follow-up of 61.1 months (0.2, 68.1), there were no differences in rates of LRR; 5-year local regional control was 93.2% (89.0-97.6%) in the MRI group and 95.7% (92.3-99.1%) in the no MRI group (HR: 1.1, 95% CI: 0.3, 3.9). For the entire cohort, 5-year distant recurrence-free rate and OS were 94.3% (91.7, 97.1) and 92.2% (89.1, 95.4), respectively, with no differences between groups. Conclusions: With multimodality therapy, rates of 5-year LRR in patients with early stage triple- negative and HER2-positive breast cancer are very low. Breast MRI for local staging and surgical planning in this setting does not result in improved local regional control. Support: U10CA180821, U10CA180882; UG1CA189823; U10CA180868 (NRG Oncology); ClinicalTrials.gov Identifier: NCT01805076; https://acknowledgments.alliancefound.org Citation Format: I. Bedrosian, K. Ballman, L. M. McCall, C. E. Comstock, G. W. Unzeitig, T. Yen, A. Weiss, B. G. Haffty, J. M. Bensenhaver, S. M. DeSnyder, J. J. Liu, A. H. Partridge, L. A. Carey, K. K. Hunt.Effect of Preoperative Breast MRI Staging on Local Regional Recurrence (LRR) in Early Stage Breast cancer: Alliance A011104/ACRIN 6694 [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr GS2-07.

ABSTRACT Background Many older adults undergo gastrointestinal (GI) cancer surgery, yet the relative impacts of advanced age, comorbidity, and frailty on postoperative mortality and failure to rescue (FTR) remain unclear. We aimed to compare the impacts of these factors on postoperative outcomes in a large population‐based cohort. Method We conducted a national retrospective linked database study of patients aged ≥ 65 years who underwent resection for GI or hepatobiliary cancer in Aotearoa New Zealand between 2005 and 2020. Age, comorbidity (C3 Comorbidity Score), and frailty (Hospital Frailty Risk Score) were examined as independent predictors of 90‐day mortality, complications, and FTR (death following a complication). Logistic regression models were adjusted for demographic and clinical covariates. Results Among 21,729 patients (mean age 75.8 ± 6.8 years), 49.3% experienced one or more complications and 6.1% died within 90 days (FTR rate of 12.3%). In adjusted models, each additional 5 years of age increased 90‐day mortality odds by 40% (adjusted odds ratio [aOR] 1.40 and 95% CI 1.34–1.46). Although higher comorbidity (C3 > 3) and frailty (Hospital Frailty Risk Score > 15) also independently raised mortality risk (aOR 1.98 and 2.04, respectively), age exerted the largest effect on mortality and FTR. Conclusion Chronological age, comorbidity, and frailty each predict worse outcomes following GI cancer surgery, but advanced age remains the dominant driver of 90‐day mortality and FTR. These findings underscore the need for risk stratification, shared decision‐making, and preoperative optimization as wells as intense surveillance for complications and early definitive care of older adults undergoing major oncologic operations.

INTRODUCTION: Chronic gastroduodenal disorders are highly prevalent and encompass heterogeneous pathophysiology. Autonomic dysfunction may contribute, but its relative contribution to disease pathophysiology and symptoms is unclear. This study aimed to clarify the contributions of autonomic dysfunction in gastroduodenal disorders and symptom associations using multimodal testing, hypothesizing that distinct disorder subgroups would be revealed. METHODS: This prospective observational cohort study included 80 patients with chronic gastroduodenal symptoms. Participants underwent expert clinician-led autonomic evaluation, body surface gastric mapping (BSGM; Gastric Alimetry, New Zealand), and gastric emptying scintigraphy testing. BSGM phenotyping, gastric emptying cutoffs, and autonomic criteria followed standard criteria. RESULTS: Gastroparesis was diagnosed in 27 (33.8%) patients. Autonomic dysfunction was present in 21 (26.3%) patients, and abnormal BSGM in 20 (25%). Autonomic dysfunction and abnormal BSGM were both more prevalent in gastroparesis (51.9% vs 13.2%, P = 0.001; and 44.4% vs 15.1%, P = 0.005, respectively). However, these diagnoses tended not to overlap, with patients having autonomic dysfunction generally exhibiting normal BSGM tests. Autonomic dysfunction independently predicted delayed gastric emptying (OR 13.8, P < 0.001) and was associated with earlier meal responses (higher meal response ratio, P = 0.04) and worse postprandial distress symptoms ( P < 0.05). DISCUSSION: Autonomic dysfunction independently predicted delayed gastric emptying and revealed a gastroparesis mechanism distinct from gastric myoelectrical impairment. Autonomic dysfunction altered the postprandial gastric motor response and was associated with a greater symptom burden. Multimodal physiological testing including autonomic profiling reveals distinct mechanistic subgroups in patients with chronic gastroduodenal symptoms.

The NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Hodgkin lymphoma provide expert, multidisciplinary recommendations for the diagnosis and treatment of Hodgkin lymphoma. The panel convenes annually to update recommendations based on a review of recently published clinical trials. This selection from the NCCN Guidelines for Hodgkin Lymphoma focuses on the management of newly diagnosed and relapsed/refractory classic Hodgkin lymphoma in adults aged 18 to 60 years.

BACKGROUND: Bicuspid aortic valve (BAV) is a frequent congenital heart defect with a high heritability. Despite this, only a limited number of genes have been associated with the disease, and the molecular mechanisms remain unexplained in most cases. This study aimed to further understand the genetic architecture of BAV. METHODS: A genome-wide association study meta-analysis including 9631 cases among 65 677 participants was performed. Genes were prioritized using transcriptomic analyses based on RNA sequencing in relevant tissues, including human fetal and adult aortic valves. The impact of the knockdown or knockout of 4 candidate genes on cardiac development was verified in zebrafish. A polygenic risk score was developed, its association with BAV was evaluated in an independent cohort, and its association with a wide range of phenotypes (n=976) was evaluated in UK Biobank (n=355 618 individuals). RESULTS: Thirty-six genomic loci were identified, including 32 that were not described previously. Among the prioritized genes, KANK2 and ERBB4 were identified as potentially causal through transcriptomic analyses, colocalization, and Mendelian randomization based on gene expression in human aortic valves (n=484), whereas PRDM6 and STRN were prioritized using similar analyses from aortic (n=326) and left ventricular tissues (n=326), respectively. Targeting 4 candidate genes ( WNT4 , LEF1 , STRN , and KANK2 ) in zebrafish led to disruption in cardiac development. A polygenic risk score was associated with an odds ratio of 2.07 (95% CI, 1.90–2.25; P =5.43×10 -62 ) per SD for BAV and significantly associated with thoracic aortic aneurysm and atrial fibrillation in UK Biobank. CONCLUSIONS: This study supports a significant polygenic contribution to BAV, where the combination of multiple common variants in genes involved in heart morphogenesis disrupts aortic valve development.