Citations

ABSTRACT Background Olfactory neuroblastoma (ONB) exhibits variable recurrence patterns, with significant late recurrences occurring years after treatment. We investigated cutoff periods to define late recurrence and identified predictors that distinguish patients at risk of late recurrence from those considered cured. Methods From a multi‐institutional cohort of 219 ONB patients treated at nine centers, we analyzed 77 patients with documented recurrence. Using a 48‐month cutoff to define early (< 48 months) versus late (≥ 48 months) recurrence, we evaluated 35 clinical, anatomic, and treatment variables using Firth penalized logistic regression across three comparisons: (1) late versus early recurrence timing, (2) late recurrence versus no recurrence, and (3) early recurrence versus no recurrence. Results Thirty‐four of the 77 (44.2%) recurrences were classified as late. Among the 77 patients who recurred, smaller tumor volume (OR 0.243, p = 0.035) and lower novel Dulguerov staging (OR 0.802, p = 0.038) were associated with late, rather than early recurrence. When patients with late recurrence were compared to those who remained disease‐free, radiographic skull base bone involvement was the strongest independent predictor (OR 2.90, 95% CI 1.14–8.64, p = 0.025), and this effect strengthened after adjusting for tumor volume (OR 3.80, p = 0.029). Pathologic dural invasion (OR 3.29, p = 0.013), Kadish stage (OR 2.02, p = 0.026), and T‐stage (OR 1.32, p = 0.040) were also significant. Conclusions A 48‐month cutoff effectively stratifies ONB recurrence risk. Skull base bone involvement independently predicts recurrence when compared to nonrecurrent cases and uniquely persists as a risk factor for late recurrence. These findings establish skull base bone involvement as a novel biomarker for risk‐stratified surveillance protocols.

Abstract Background The left anterior descending (LAD) coronary artery is a critical organ‐at‐risk in thoracic radiotherapy. In thoracic stereotactic body radiation therapy (SBRT), sharp dose gradients and poor LAD contrast on 4DCT scans introduce uncertainty in LAD dose assessment. Methods We evaluated the performance of an AI‐generated LAD region structure as a surrogate for manual LAD contours in early‐stage NSCLC. We used deep inspiration breath hold (DIBH) scans from 49 patients with manually delineated LADs; planning risk volume contours were expanded by 3 mm. Then, 96 lung SBRT dose maps were registered by matching the left anterior portion of the heart contours. Dice similarity coefficient (DSC), inclusion ratios (IR), dose–volume histogram (DVH) metrics, linear regression, and linear mixed models were used to compare the structures. The correlation of the AI LAD region and LAD dose to mean heart dose (MHD) was assessed. Results The AI LAD region showed strong correlation with manual LAD contours for maximum dose (D1%, R 2 = 0.93) and volume receiving ≥10 Gy (V10Gy, R 2 = 0.81). The planning risk volume also demonstrated high correlation to the AI LAD region. MHD was a poor predictor of LAD dose ( R 2 ≈0.4). Additional dose metric correlations were assessed. Discussion To investigate dose volume effects to the LAD during radiotherapy for stage I NSCLC, a robust segmentation paradigm must be established. It must be easily implementable on typical imaging acquired during treatment, with an emphasis on clinical translatability. Our model has shown that the AI LAD region has a strong correlation with LAD dose under typical lung SBRT doses and supports its use for further investigation into LAD dose volume constraints during lung SBRT. Conclusion AI LAD region provides a clinically feasible and scalable alternative to manual LAD contouring for dose estimation in lung SBRT. Its strong correlation with key dose metrics validates its use in large‐scale retrospective and prospective studies.

Introduction: This study aimed to develop and evaluate a deep learning (DL) deblurring algorithm for computed tomography (CT). Phantom testing assessed whether DL deblurring could improve CT image quality and coronary artery calcium (CAC) detectability. Materials and Methods: The algorithm consisted of 2 sequential convolutional neural networks (CNNs): one for denoising and one for deblurring. After deblurring, the estimated noise from the first model was reintroduced to preserve image texture. The CNN denoising model was trained to map high-noise images to routine low-noise ones. The CNN deblurring model was trained using natural images—blurred inputs paired with sharp targets. To assess performance, an anthropomorphic phantom with 100 small calcifications was scanned on a CT system using a CAC scoring protocol. Data were acquired at clinical and high doses, then reconstructed with and without DL deblurring, using 25% of the maximum denoising strength. Detectability was defined as the ability to calculate an Agatston score (≥3 adjacent voxels >130 HU). Results: At high dose, CAC detectability increased from 39% (standard reconstruction) to 48% with deblurring. Similar detectability (39%) was achieved at the routine dose with deblurring, representing a 6% improvement compared with the original routine dose data. Discussion: Quantitative results from phantom testing demonstrate that DL deblurring can improve image quality, based on a clinically relevant, task-based CAC detection metric at clinical dose levels.

Abstract Hemorrhagic shock is a common emergency that accounts for more than 10% of global mortality and up to 40% of trauma-related mortality. The Advanced Trauma Life Support (ATLS) guidelines outline resuscitation strategies in patients with massive hemorrhage based on clinical trial and observational data. Early intervention with fluids and/or blood products is recommended during resuscitation as key to maintaining vascular patency, stabilizing a patient’s blood pressure, maintaining tissue oxygenation, and limiting shock. From a transfusion perspective, there is wide heterogeneity in the quality of blood products due to donor and manufacturing variation. The optimal transfusion strategy for massive hemorrhage remains unclear. In silico models of resuscitation may offer a means to evaluate the effectiveness and safety of various resuscitation protocols. Here we describe a stochastic multicompartment model of fluid balance and resuscitation that includes (i) cardiovascular hemodynamics, (ii) body fluid compartments and capillary solute exchange, (iii) the ability to alter hemorrhage, resuscitation, and hemostasis parameters, and (iv) tissue oxygenation and capillary-alveolar gas exchange. Building upon deterministic frameworks, this stochastic model more faithfully reflects the clinical heterogeneity of bleeding patients at a level I trauma center. This allows for a proof of concept in silico trial comparing crystalloids, conventional component therapy (CCT), i.e. red cell, platelet and plasma components, to cold-stored low titer group O whole blood (LTOWB). In an in silico cohort of ATLS class III (>30% and ≤40% blood volume lost) hemorrhage, LTOWB resuscitation reduced the time spent in the critical hemostatic window (platelet count <50x109/L, INR≥2, hemoglobin (Hgb) <8 g/dL and fibrinogen <150 mg/dL) compared with CCT (90.59 vs. 147.62 minutes; p=0.04), with no difference in predicted event-free survival (t = 240 minutes; cardiac < 1.5, fluid overload > 10% or SBP > 150% of the starting SBP, or Hgb < 3.0 g/dL). In a separate cohort of ATLS class IV (>40% blood volume lost) hemorrhage, LTOWB yielded a higher predicted event-free survival (74%) versus the CCT arm (69%, p < 0.01). We demonstrate how this in silico platform can function as a digital twin for hemorrhagic trauma enabling precision transfusion strategies, and in parallel, an operational twin to support blood banks to forecast blood product demand and inform massive transfusion protocols and clinical trial design.

ABSTRACT Objectives Tongue base obstruction is a cause of persistent sleep apnea after adenotonsillectomy in pediatric patients. The internet is the most used tool by patients for gathering healthcare information. This study aims to assess the quality and readability of websites directed to patients regarding tongue base procedures for pediatric obstructive sleep apnea. Methods Queries for the terms “lingual tonsillectomy,” “posterior midline glossectomy,” “tongue reduction,” “tongue base suspension,” and “hypoglossal nerve stimulator” were entered into Google on September 11, 2023. Thirty sites were pulled for each, and up to 10 sites per query meeting inclusion criteria were included for analysis. The DISCERN tool was used to evaluate quality. Flesch Reading Ease Score (FRES) and Flesch–Kincaid Grade Level (FKGL) were used to measure readability. Sites were analyzed for mention of pediatrics. Results Thirty‐five unique sites were analyzed. The mean DISCERN score was 44.7 (range 28.5–64.0) or “fair.” The mean FRES was 46.3 (range 9.6–66.6) and FKGL 11.5 (range 7.4–19.2), indicating college and 12th‐grade reading levels. Only 8/35 sites mentioned pediatrics. The average number per query of relevant sites directed towards patients on the first search engine page was 3.6. Despite tongue base procedures being relevant to pediatrics, only one in five sites mentioned children. Conclusion Websites directed at patients regarding tongue base procedures are of “fair” quality and at readability grade levels far above what is recommended. The majority of sites available for tongue base procedures are geared to professionals rather than patients. This study demonstrates a need for high‐quality information for patients about these procedures.

Cercospora leaf spot (CLS) is a destructive disease of sugar beet (Beta vulgaris ssp. vulgaris L.) for which resistance conferred by the gene BvCR4 has been introgressed from sea beet (B. vulgaris ssp. maritima). BvCR4-mediated resistance has been broadly deployed to manage CLS globally in sugar beet; however, continuous implementation exerts strong selection pressure, increasing the risk of disease outbreaks due to resistance-breaking C. beticola populations. This study aimed to identify BvCR4 variants in public B. vulgaris spp. germplasm to assess their genetic and structural diversity. Polymerase chain reaction screening of 589 public B. vulgaris spp. germplasm detected BvCR4 homologs in 32 accessions with diverse responses to CLS including historically relevant sugar beet lines represented in breeding programs and Mediterranean sea beet populations of diverse genetic backgrounds. Analysis of the genomic sequences revealed higher diversity in sea beet compared to sugar beet. Additionally, population genetic analysis indicated that BvCR4 homologs in sugar beet likely originated from sea beet populations in Italy and Greece. We identified historical cultivated materials carrying full proteins encoded by BvCR4 alleles and delineated geographic regions where sea beet populations harbor BvCR4 homologs that may be useful in breeding programs. Structural variation was largely circumscribed within leucine-rich repeat domains, although CLS resistance phenotyping was not associated with specific polymorphisms, suggesting BvCR4-mediated resistance may rely on interactions with other co-receptors. This study provides resources for sugar beet improvement by enabling the identification of BvCR4 homologs in diverse germplasm and disclosing materials with potentially novel resistance.

Abstract Improved understanding of how chimeric antigen receptor (CAR) and endogenous T‐cell phenotypes associate with progression‐free survival (PFS) is needed to individualize and improve immunotherapy outcomes in multiple myeloma (MM). We characterized and investigated the prognostic value of immune dynamics defined with multidimensional and computational flow cytometry in 258 bone marrow aspirates from 107 MM patients treated with idecabtagene vicleucel in the KarMMa trial. At screening, there was an association between longer PFS and increased percentages of PD1 + cytotoxic T cells and ICOS − TIGIT + Tregs. At Month 1 after infusion, the presence of >1% CAR‐T cells, a CD4/CD8 CAR‐T‐cell ratio >0.09, and greater abundance of non‐activated non‐cytolytic CAR‐T cells identified patients with longer PFS. Notably, 47 of the 115 endogenous T‐cell clusters were prognostic at the latest assessment performed after CAR‐T infusion. Lower percentages of ICOS + TIGIT − Tregs showed the strongest association with longer PFS. Altogether, this large dataset of dynamic immune profiling throughout treatment with anti‐B‐cell maturation antigen (anti‐BCMA) CAR‐T cells uncovered that the T‐cell composition prior to apheresis as well as CAR‐T and endogenous T‐cell phenotypes after infusion associates with PFS. These findings have potential for translation into routine laboratory practice for improved understanding of response and resistance to CAR‐T cells in relapsed/refractory MM.

Abstract Background In children with non-metastatic atypical teratoid/rhabdoid tumor (ATRT), proton therapy is increasingly utilized worldwide due to the radiosensitivity of afflicted young patients. General questions persist, however, surrounding dose and volume of radiotherapy. Methods Between 2009-2024, 39 children from a single institution with non-metastatic intracranial ATRT received focal proton therapy at a median age of 2.4 years (range 0.7-19). Twenty-two patients had supratentorial tumors and 27 had gross/near total resection prior to radiation. Most common chemotherapy regimens were ACNS0333 and SJYC07 (n = 10 each) and Dana-Farber (n = 6). Four patients received consolidative high-dose chemotherapy after radiation. The median radiation dose was 50.4 Gy (range, 50.4-54). The cumulative incidence method assessed freedom from selected endpoints.The Fine-Gray method assessed the level of statistical significance between levels of selected prognostic factors. Results With a median followup of 5.2 years (range 0.5-16.2), the 5-yr overall survival, progression-free survival, and local control was 63%, 64%, and 97%, respectively. Thirteen patients developed isolated leptomeningeal failure as first site of recurrence. Only subtotal resection was significantly associated with local recurrence (100% vs 92%, p &lt; 0.05). Patient age, tumor site, extent of resection, radiation dose, and chemotherapy regimen were otherwise not significantly associated with disease control or survival. Four patients experienced symptomatic radiation necrosis, including 1 death: Three of 4 cases received 54 Gy and consolidative intrathecal methotrexate (n = 2) or high dose ACNS0333 intravenous chemotherapy (n = 1, fatal) after radiation. The other patient received 50.4 Gy and necrosis was immediately preceded by ventriculoperitoneal shunt failure. Six patients developed hearing loss requiring aids. One patient developed radiation-induced meningioma a decade following treatment. Conclusion This represents the largest experience reporting 5-year outcomes following focal radiation in children with non-metastatic intracranial ATRT. When applied with a variety of modern chemotherapy regimens in real-world settings, focal proton therapy at a dose of 50.4 Gy may offer the best therapeutic ratio but does not eliminate serious toxicity. If avoidance of craniospinal radiation remains an ongoing priority, further improvement in survival will depend upon systemic therapy.

Abstract Background Cholangiocarcinoma (CCA) is an aggressive hepatobiliary malignancy with marked molecular heterogeneity, limited systemic treatment options, and poor prognosis. Chemoresistance and KRAS mutations remain major therapeutic challenges. Dysregulation of the Hippo pathway, particularly YAP/TAZ hyperactivation, drives CCA progression and resistance. TEAD, the essential transcriptional partner of YAP/TAZ, represents a promising target. IAG933 is a novel oral pan-TEAD inhibitor designed to disrupt YAP/TAZ–TEAD interactions. RMC-6236, a pan-KRAS inhibitor, has also gained attention. Combining TEAD and KRAS inhibition may overcome resistance and improve efficacy in CCA. Methods We evaluated IAG933 in human (HuCCT1, RBE) and murine (SB1, FAC) CCA cell lines. Cells were treated with IAG933 alone or in combination with GemCis or RMC-6236. Cell viability assays were used to determine IC50 values and calculate combination index (CI). TEAD activity was assessed by luciferase reporter and qRT-PCR of YAP/TAZ target genes (CTGF, CYR61, NUAK2). In vivo efficacy was evaluated in syngeneic orthotopic and patient-derived xenograft (PDX) models. Results IAG933 suppressed TEAD-dependent transcription (IC50≈500 nM) and downregulated YAP/TAZ target genes. Combination with GemCis or RMC-6236 produced synergistic effects (CI &lt; 1) in all CCA lines. In vivo, IAG933 monotherapy suppressed tumor growth similarly to GemCis. Dual TEAD and KRAS inhibition achieved marked tumor regression and significantly delayed tumor regrowth after treatment cessation in KRAS-mutant PDX models. Conclusion IAG933 effectively suppresses YAP/TAZ–TEAD signaling and inhibits CCA growth. While the combination with GemCis showed limited additional benefit due to strong single-agent efficacy, synergy with RMC-6236 produced durable antitumor effects, particularly in run-out studies, delaying regrowth in KRAS-mutant models. These findings highlight IAG933 as a promising therapeutic candidate targeting the Hippo–TEAD pathway, with potential benefit in KRAS-mutant CCA.

Abstract Large Language Models (LLMs) demonstrate strong capabilities in healthcare applications. However, recruiting human evaluators to assess usability and effectiveness becomes challenging in the large-scale LLM-generated outcomes. Existing metrics are often outdated for evaluating LLM-based healthcare technologies. This paper proposes a Research through Evaluation (RtE) approach that can refine metrics and improve assessments of LLM-generated outputs. We deploy RtE in a retrospective comparative study of prostate cancer patient inquiries, comparing responses from human clinical care teams with those generated by a GPT-4-based in-basket bot. Through three rounds of co-evaluations with clinical professionals and an LLM-as-graders study with four LLMs, the RtE method shows that starting with small, iterative evaluations helps human graders adjust metrics to suit their daily practices better. Additionally, in the LLM-as-graders study, LLMs exhibit similar behavior, preferring the in-basket bot's responses over the clinical care team. Our methodology and findings demonstrate that the RtE method can enhance the transparency and robustness of evaluation metrics for LLM-generated content, with the potential to generalize to evaluations in other clinical practices.