Citations

Abstract Objective This study reports development and validation of a deep learning seizure detection algorithm for two‐channel subscalp electroencephalographic (EEG) recordings. Ultra‐long‐term monitoring of people with epilepsy may produce new insights into the timing and pattern of their seizures and may pave the way for novel therapeutic options. Although EEG is the accepted standard for epilepsy monitoring, ultra‐long‐term EEG recordings generate a massive quantity of data and are not reasonably reviewable in full by human readers. Methods The convolutional neural network‐bidirectional long short‐term memory (CNN‐BiLSTM) hybrid algorithm uses nine layers, operating on channel spectrograms 5 min in length with 50% overlap. Retrospective subscalp EEG data from 16 patients recorded at three centers were available for algorithm development and testing. EEG was recorded for a median of 63 days (range = 9–508), and a median of 17 seizures (range = 0–96) were recorded. Training data were augmented with scalp EEG seizures, and performance was benchmarked against a conventional spectral power classifier algorithm. We also evaluated an intrapatient training–testing approach where the algorithm was trained on the first 45% of data. Results The CNN‐BiLSTM trained on scalp EEG and subscalp EEG achieved an area under the receiver operating characteristic curve (AUROC) of .98 and an area under the precision–recall curve (AUPRC) of .50, which correspond to 94% sensitivity with 1.11 false detections per day. The same model trained on iEEG achieved only AUROC = .94 and AUPRC = .36. The conventional spectral band power detector achieved AUROC = .93 and AUPRC = .38. The CNN‐BiLSTM detector trained on the earliest half of the subscalp EEG data achieved AUROC = .93 and AUPRC = .37, corresponding to 87% sensitivity and 5.9 false detections per day. Significance High sensitivity and specificity are possible in automated seizure detection in two‐channel subscalp EEG data using a CNN‐BiLSTM framework. Performance of the detector is superior using subcutaneous EEG data for training rather than intracranial EEG, but addition of scalp EEG seizures for training was helpful.

Abstract Chimeric antigen receptor (CAR) T-cell (CAR-T) treatment for B-cell acute lymphoblastic leukemia (ALL) induces high initial response rates, but most patients relapse. Low disease burden (often defined as <5% blasts in the bone marrow) is associated with better outcomes. CAR HEMATOTOX (HT) is a score using prelymphodepletion hematologic and inflammatory parameters to predict outcomes in lymphoma. Here, we assess its prognostic utility in a large multicenter adult B-cell ALL cohort. Patients who received brexucabtagene autoleucel across 33 centers in North America were included as part of the Real-World Outcomes Collaborative of CAR-T in Adult ALL (ROCCA) consortium. An independent cohort of 61 patients with ALL treated with an investigational CD19 CAR-T therapy at 1 center was also described. Among 199 ROCCA consortium patients, 43 (22%) patients with HTlow scores had lower rates of delayed neutrophil recovery than those with HThigh scores (26% vs 52%, P = .002) and fewer severe infections (2.5% vs 18.8%, P = .011). They also had higher response rates, overall survival (OS), and event-free survival (EFS), as well as lower nonrelapse mortality and cumulative incidence of relapse. The survival differences remained significant after multivariable adjustment for disease burden and other covariates. In the investigational cohort of 61 patients, patients with HTlow scores had improved OS and EFS, as well as higher peak CAR-T expansion. In summary, CAR HT score is a prognostic factor independent of disease burden in adult ALL. HTlow score is associated with superior outcomes after CD19 CAR-Ts and higher CAR-T expansion in a single-center cohort. These trials were registered at www.clinicaltrials.gov as #NCT01044069 and #NCT01860937.

Abstract Background To ensure the Pediatric Infectious Diseases (ID) Certifying Examination remains aligned with current clinical practice, the American Board of Pediatrics (ABP) conducted a comprehensive practice analysis to update the exam content outline. Methods A panel of 9 pediatric ID physicians identified 34 core tasks across 5 performance domains to guide revisions to the existing outline. The panel identified relevant, testable knowledge areas aligned with these tasks and organized them into 22 domains and 186 initial subdomains. A nationwide validation survey of board-certified pediatric ID physicians collected relevance ratings and feedback on missing or outdated topics and suggested exam weights, which were used to guide further revisions. Subsequently, the pediatric ID subboard conducted an item reclassification process, which resulted in additional refinement. Results The final content outline comprises 22 domains and 196 subdomains, with updated domain weights to better reflect the clinical landscape. Notably, domains such as cardiovascular, bone and joint, and urologic infections increased in weight, while antimicrobial principles decreased. Conclusions This rigorous, inclusive process strengthens the exam’s content validity and supports the ABP’s goal of certifying physicians who possess the knowledge needed to deliver safe and effective care.

BACKGROUND AND OBJECTIVES: Brainstem cavernous malformations (BSCM) can result in spontaneous intracerebral hemorrhage (ICH), often resulting in significant morbidity. We aimed to assess the functional outcome and identify predictors of functional neurological outcome after single and multiple symptomatic hemorrhages. METHODS: As part of an international multicenter collaboration, institutional databases from 3 different tertiary referral centers included BSCM patients with complete baseline characteristics, MRI Data set, ≥1 ICH, and ≥1 follow-up examination followed at our institutions between 2003 and 2023. Functional neurological outcome was obtained using the modified Rankin Scale at diagnosis, before and after each ICH, and last follow-up. Patients were excluded after surgical removal of the lesion and/or loss of follow-up. RESULTS: A total of 383 patients (41.47 ± 15.51 age; 220 [57.4%] female) were included and followed for an average of 63.16 ± 85.75 months. Functional neurological outcome deteriorated in 47.2% ( P = .010) after the second ICH and in 46.5% ( P = .007) after the third ICH. Moreover, the functional neurological status was impaired in 22.3% ( P < .001) of patients at last available follow-up compared with the time of BSCM diagnosis. CONCLUSION: In our study, we observed that the chance of full recovery might decrease with each ICH. We observed a significantly associated neurological deterioration after each ICH compared with initial ICH.

Abstract Purpose: Isocitrate dehydrogenase (IDH) 1/2–isoform inhibitors have clinical efficacy in IDH1/IDH2-mutated neoplasms. However, primary resistance and secondary resistance limit their therapeutic potential. LY3410738, an oral, brain-penetrant, dual IDH1/IDH2-mutated isoform-selective inhibitor, was designed to overcome resistance. Patients and Methods: This global, multicenter, open-label, phase I study of patients with IDH-mutant solid tumors evaluated LY3410738 as monotherapy (dose escalation) for advanced solid tumors in combination with cisplatin–gemcitabine (CISGEM) for newly diagnosed cholangiocarcinoma or with durvalumab for relapsed/refractory cholangiocarcinoma (dose expansion; NCT04521686). Primary objectives were the maximum tolerated dose, recommended phase II dose, and preliminary antitumor activity. Safety, pharmacokinetics, inhibition of D-2-hydroxyglutarate, and ctDNA were assessed. Results: Overall, 119 patients received LY3410738 alone (N = 94) or in combination with CISGEM (N = 19) or durvalumab (N = 6). No dose-limiting toxicities were observed; the maximum tolerated dose was not determined. Common adverse events included nausea, vomiting, and decreased appetite. Overall response rates of 5.2% and 11.1% and disease control rates of 56.9% and 63.0% were observed for patients with relapsed/refractory IDH1- or IDH2-mutant cholangiocarcinoma or IDH1-mutant glioma, respectively. D-2-hydroxyglutarate normalization was rapid and durable. In dose-expansion cohorts, combination treatments were tolerable, with one dose-limiting toxicity in the durvalumab cohort. LY3410738 plus CISGEM demonstrated a response rate of 42.1%, a median duration of response of 8.1 months, and a median progression-free survival of 10.2 months for patients with newly diagnosed IDH-mutant cholangiocarcinoma. Conclusions: LY3410738 demonstrated largely cytostatic antitumor activity in IDH1- or IDH2-mutated cholangiocarcinoma and IDH1-mutated gliomas. LY3410738 plus CISGEM exhibited favorable antitumor activity in patients with treatment-naïve IDH-mutated cholangiocarcinoma, warranting further exploration as a treatment strategy.

Background: Cirrhosis and cirrhosis-related deaths have risen in the United States in recent years. Ascites is a common complication, often requiring large-volume paracentesis (LVP). The American Association for the Study of Liver Diseases (AASLD) recommends the administration of albumin in conjunction with LVP to prevent further complications of cirrhosis. Emerging research in cirrhosis care reveals significant variations in outcomes among different demographics. Therefore, we assessed the use of guideline-adherent albumin and outcomes in U.S. patients undergoing LVPs, particularly at the intersection of race, ethnicity, socioeconomic disparities, and cirrhosis. Methods: This retrospective study utilized Cerner Real World Data to identify adults with cirrhosis and ascites undergoing LVP between January 2016 and June 2022. We assessed albumin utilization patterns across racial and ethnic groups and payor types, and their overall impact on acute kidney injury (AKI)-related hospitalization using an adjusted generalized linear model (aGLM). Results: We identified 736 patients: 301 in the LVP + albumin group and 435 in the LVP-only group. Despite clinical recommendations, only 41% undergoing LVPs received albumin. White patients and commercially insured patients received albumin at higher rates (p=0.042 and p=0.009, respectively). The overall rate of AKI-related admissions within the 30-day post-procedure period was 26%. However, patients who received albumin during LVP had a 36% lower risk of short-term AKI-related hospitalization (OR: 0.64; p=0.03). Conclusions: These findings indicate a potential for broader albumin utilization in U.S. patients with refractory ascites undergoing repeated LVPs to reduce AKI-related admissions.

Abstract Background Spatially fractionated radiation therapy (SFRT) is a technique that delivers heterogenous dose distributions consisting of alternating regions of high dose “peaks” and low dose “valleys”. Current delivery methods for SFRT using megavoltage x‐rays usually treat large and bulky tumors with brass grid or volumetric modulated arc therapy (VMAT) lattice techniques. The size and spacing of high dose regions in these approaches are typically on the order of centimeters. However, multiple studies have suggested that decreasing these dimensions may improve the therapeutic ratio. Furthermore, a more compact approach to SFRT would allow for a greater number of high dose regions within the tumor, as well as application to smaller and more irregularly shaped targets thereby increasing the number of patients that could benefit from SFRT. Purpose This study describes the commissioning and first patient treatment using mini‐lattice radiation therapy (MLRT). MLRT uses a clinical linear accelerator and decreases the size and spacing of standard lattice SFRT by using individual multileaf collimator (MLC) leaves to deliver 5 mm wide high dose regions. Methods MLRT plans were created in the Varian Eclipse treatment planning system for a Varian Truebeam equipped with Millennium 120 MLCs. MLRT uses 6 MV Flattening Filter Free high dose rate and the width of individual MLCs to define 5 mm by 5 mm openings separated by closed MLCs to deliver alternating opened and blocked regions. Dynamic conformal arcs were used to conform MLCs to 4 mm spherical mini‐lattice structures in the gross tumor volume (GTV). A MLRT‐specific beam model was commissioned to accurately model the small MLRT fields. Film measurements were performed to assess the accuracy of MLRT plans calculations. Plans for seven treatment sites in different parts of the body for retrospective patient candidates were created with varying numbers of mini‐lattices and separation distances to assess the impact of varying these parameters on treatment dose metrics. MLRT was used for the first time to treat a patient with two fractions of MLRT. Results The AcurosXB calculation algorithm with modified x and y spot sizes, dosimetric leaf gap, transmission factor, and output factor table was used to generate a beam model for accurate MLRT calculations. The MLRT‐specific beam model resulted in gamma passing rates (1%/0.5 mm criteria) of 90%–99% for retrospective patient MLRT film measurements. Dose volume histogram statistics, equivalent uniform dose, and mean dose showed a higher number of mini‐lattices with smaller separation increased the dose to the GTV and surrounding tissue. Separation distances between mini‐lattices did not impact plan heterogeneity as measured by D10%/D90%. The first patient treated with the MLRT technique reported pain relief, had stable disease, and no acute toxicities following both fractions. Conclusions MLRT is feasible using clinical linear accelerators and existing radiation oncology infrastructure. It enables targeting smaller tumors with SFRT, allows for a greater number of high dose regions within the target compared to standard VMAT lattice techniques, and may be a useful technique for challenging treatment scenarios.

Abstract Background Delirium is often underdiagnosed in clinical practice and is not routinely coded for billing. While manual chart review can identify delirium, it is labor-intensive and impractical for large-scale studies. Natural language processing (NLP) can analyze unstructured text in electronic health records (EHRs) to extract meaningful clinical information. Methods To support national integration of NLP for EHR-based delirium identification across different institutions, we launched the Delirium Interest Group within the national Evolve to Next-Gen Accrual to Clinical Trials (ENACT) NLP Working Group. This paper outlines our initial efforts to standardize, evaluate, and translate an NLP-based delirium detection model into the i2b2/ENACT platform. Results Multisite contextual inquiry identified several key challenges, including variations in local screening practices (eg, tools used, documentation frequency, and quality control), the need for harmonized definitions in the context of EHRs, and the complexity of modeling temporal logic. Multisite NLP evaluation revealed variable performance degradation driven by differences in delirium screening practices, clinical documentation patterns and semantics, and note syntactic structures. Conclusion Our work represents an important first step toward enabling scalable and standardized NLP-based delirium detection across institutions. By engaging diverse institutions through the ENACT NLP Working Group, we identified shared challenges and site-specific variations that impact model implementation and performance. Our collaborative approach enabled the development of a more robust framework for delirium identification across heterogeneous EHR systems. Future efforts will build on this foundation to enhance the validity, usability, and translational impact of delirium detection.

Abstract Diffuse large B-cell lymphoma (DLBCL) poses a challenge in hematology given its varied symptoms, and the complex interplay between disease and treatment effects on health-related quality of life (HRQoL). The phase 3 POLARIX study demonstrated superior progression-free survival and a similar safety profile with polatuzumab vedotin plus rituximab, cyclophosphamide, doxorubicin, and prednisone (Pola-R-CHP) vs R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) in patients with previously untreated DLBCL. Here, we evaluate HRQoL through patient-reported outcome (PRO) instruments to fully characterize the patient experience in the POLARIX study. Changes from baseline in HRQoL, lymphoma symptoms, and gastrointestinal (GI) symptoms were assessed, as well as incidence and severity of common symptoms by PROs vs clinician-reported adverse events (AEs). Baseline characteristics of PRO-evaluable patients (N = 874) were consistent. Comparison between PROs and clinician-reported AEs revealed a notable discordance; patients generally reported a higher incidence of symptoms than clinicians, emphasizing the need for patient-centric tools to accurately capture the patient experience. Both treatments exhibited rapid and sustained improvements in HRQoL and lymphoma symptoms, with the most substantial improvements seen in global health status/QoL, lymphoma symptoms, fatigue, role, emotional, and social functioning. GI symptoms (diarrhea, constipation, nausea, and vomiting) were generally similar between treatment arms and returned to baseline levels after treatment completion. These HRQoL data underscore the complementarity of PROs, as an adjunct to clinician-reported AEs, in evaluating the efficacy and tolerability of new treatments, including Pola-R-CHP, which may represent a new benchmark for patient-reported HRQoL in previously untreated DLBCL. This trial was registered at www.clinicaltrials.gov as NCT03274492.

PURPOSE Ziftomenib—a potent, highly selective, oral menin inhibitor—was well tolerated and demonstrated encouraging clinical activity as monotherapy for relapsed/refractory NPM1 -mutated ( NPM1 -m) and KMT2A -rearranged AML in the KOMET-001 phase I trial. METHODS In the registration-enabling phase II part of KOMET-001, patients with relapsed/refractory NPM1 -m AML received ziftomenib 600 mg once daily. The primary end point was the rate of complete remission with full hematologic recovery (CR)/CR with partial hematologic recovery (CRh). RESULTS From January 26, 2023, to May 13, 2024, 92 patients (median age, 69 years [range, 33-84]) were treated. The primary end point was met, with a CR/CRh rate of 22% (95% CI, 14 to 32; P = .0058); 61% were negative for measurable residual disease. Overall response rate was 33% (95% CI, 23 to 43), with a median duration of 4.6 months (95% CI, 2.8 to 7.4). Prespecified subgroup analyses showed comparable CR/CRh regardless of previous therapy, including venetoclax, or type of comutations. Median overall survival was 6.6 months (95% CI, 3.6 to 8.6). Common grade ≥3 treatment-emergent adverse events were febrile neutropenia (26%), anemia (20%), and thrombocytopenia (20%). Differentiation syndrome occurred in 25% of patients (15% grade 3; no grade 4-5) and was manageable with protocol-defined mitigation. Three patients (3%) discontinued treatment because of ziftomenib-related adverse events. CONCLUSION Ziftomenib demonstrated significant clinical benefit and deep responses in patients with heavily pretreated, relapsed/refractory NPM1 -m AML. Ziftomenib was well tolerated with a safety profile consistent with previous studies, including manageable differentiation syndrome, lack of clinically significant QTc prolongation, and low rates of myelosuppression.