Clinical and Translational Radiation Oncology

Purpose: Glioblastoma (GBM) remains associated with poor outcomes, with most recurrences occurring within the high-dose radiation field, suggesting persistent radioresistance. Exportin 1 (XPO1) inhibition with Selinexor has demonstrated radiosensitizing effects in preclinical models. We conducted a phase I trial to evaluate the safety, tolerability, and preliminary efficacy of Selinexor in combination with standard chemoradiation for newly diagnosed GBM. Methods: This investigator-initiated phase I dose-escalation trial (3+3 design) enrolled adults with newly diagnosed GBM or gliosarcoma. Patients received standard radiotherapy (60 Gy in 30 fractions) with concurrent temozolomide and escalating doses of Selinexor. Three dose levels were evaluated: 80 mg weekly (weeks 1, 2, 4, 5); 60 mg twice weekly (weeks 1, 2, 4, 5); and 60 mg twice weekly (weeks 1-6) throughout radiotherapy. The primary endpoint was determination of the maximum tolerated dose (MTD) based on dose-limiting toxicities (DLTs). Secondary endpoints included progression-free survival (PFS), overall survival (OS), patterns of failure, and patient-reported outcomes (MDASI-BT). Results: Eleven patients were enrolled. Median age was 58 years, and median KPS was 90. The MTD was established at Selinexor 60 mg twice weekly during weeks 1, 2, 4, and 5 of chemoradiation. Dose level 3 exceeded the MTD with two DLTs. Treatment compliance was high, with minimal missed radiotherapy fractions. Median PFS was 15.9 months (95% CI, 6.2 28.5), and median OS was 17.4 months (95% CI, 14.1 not reached). Most recurrences were central (5/6 evaluable patients). Notably, multiple cases of delayed pseudoprogression were observed at 5, 9, 10, and 23 months post-radiotherapy. Patient-reported symptom burden remained stable over time. Conclusions: Selinexor can be safely combined with standard chemoradiation in patients with newly diagnosed GBM, with an MTD of 60 mg twice weekly during select treatment weeks. Preliminary efficacy signals and an increased incidence of delayed pseudoprogression suggest a potential radiosensitizing effect. These findings support further investigation of Selinexor in larger, prospective studies.

Background and purpose. Patients with oropharyngeal cancer (OPC) treated with radiotherapy (RT) exhibit heterogeneous courses of radiation-associated dysphagia (RAD) during recovery, yet most survivorship models typically treat RAD uniformly. This study aimed to identify distinct, data-driven RAD longitudinal Phenotypes based on imaging-graded swallow function from pre-treatment to 30 months post-RT and to characterize their baseline predictors. Materials and methods. Heterogeneous linear mixed-effects latent class trajectory modeling was applied to longitudinal DIGEST scores from the Stiefel MDA-OPC prospective registry. Eligible patients had [≥]3 Modified Barium Swallow (MBS) assessments between baseline and 30 months post-RT. Models were evaluated across functional forms and 1-5 latent classes; final selection used the Bayesian Information Criterion. Baseline predictors of class membership were identified via binary logistic regression. Results. The cohort comprised 650 OPC patients (2,116 MBS assessments; mean age 61 years, 89% male, 93% HPV-positive). Four RAD Phenotypes were identified: No/Minimal RAD (n=385/650, 59%), Mild/Moderate RAD (n=104/650, 16%), Moderate/Severe Transient RAD (n=94/650, 15%), and Moderate/Severe Progressing RAD (n=67/650, 10%). Classification quality was acceptable (mean posterior probabilities 0.78-0.89; entropy 0.69). Baseline DIGEST impairment, base-of-tongue primary, advanced T stage, and age [≥]60 independently predicted membership in higher-burden Phenotypes (AUC=0.845; 10-fold CV-AUC=0.835). Conclusion. RAD following RT for OPC comprises four biologically and clinically distinct longitudinal Phenotypes, predictable from pre-treatment characteristics. These findings support trajectory phenotyping as outcome framework for RAD research and risk-adaptive survivorship care.

Accurate segmentation of thin-wall organs-at-risk (OARs)-the cochlea, vestibular semicircular canals, internal auditory canal, tympanic cavity, and middle ear-is clinically relevant for head-and-neck radiotherapy planning, yet these small, thin-wall structures remain among the most challenging targets for automated delineation. Dual-frequency feature fusion is a promising direction for boundary-sensitive representation, but under the investigated FP16 FFT-FcaNet setting, we observe an approximately 863-fold activation-scale mismatch between the FFT and FcaNet branches, causing a nominal 5 percent residual coefficient to behave as an approximately 43-fold dominant term. We propose FreqFuseNet, which resolves this mismatch by normalizing the FcaNet branch to the FFT activation scale before residual injection with a fixed low-amplitude coefficient (beta = 0.05), restoring beta as an interpretable 5 percent residual-amplitude coefficient relative to the FFT feature scale. Under a controlled binary per-OAR ROI protocol on the SegRap2023 head-and-neck CT benchmark across 10 clinically prioritized thin-wall OARs, FreqFuseNet achieves Dice of 0.849, HD95 of 0.824 mm, and SDice@1mm of 0.959 in the primary seed, with comparable performance in an independent second seed (Dice 0.843, HD95 0.823 mm). FreqFuseNet yields statistically significant case-level aggregate improvements over 3D U-Net and MedNeXt-S (Wilcoxon p < 0.01 and p < 0.05, respectively), using only 29.7 million parameters versus 414.6 million for the full wavelet baseline.

Background: Independent end-to-end verification of adaptive radiotherapy on MR-Linac systems is limited by the lack of patient-specific phantoms able to reproduce imaging and dosimetric properties from CT and MRI scanners. We present a method for automated generation of 4D, patient-specific, multi-material 3D-printable phantoms for quality assurance of adaptive radiotherapy on a 1.5T MR-Linac. Methods: Patient images were automatically segmented using a pretrained deep learning model. The segmented structures were converted into high-resolution 3D meshes and assembled into printable phantoms. A dosimeter holder was inserted at user-defined anatomical locations, with orientation optimized to avoid traversal across heterogeneous tissue interfaces. Physiological motion was incorporated by generating phantoms from images at different timepoints and interpolating deformation fields to create continuous 4D models. Multi-material organs designed by mixing a set of six polymers at various proportions were used to reproduce tissue-specific imaging properties. The properties of material mixtures were evaluated in a clinical CT simulator and a 1.5T MR-Linac. Results: The proposed workflow enables automated generation of anatomically realistic phantoms with several types of embedded dosimeters. A discrete search method was designed for placement and immobilization of OSLD, film, and ion chamber dosimeters. Calibration curves for Hounsfield units were derived through variations in radiopaque material content, while MR signal intensity was modulated by gel and tissue matrix mixtures. Patient-derived abdominal phantoms were fabricated at multiple scales while replicating internal anatomical detail. Multi-dimensional phantom generation enabled continuous representation of motion states with consistent mesh topology across phases. Conclusions: We demonstrate an end-to-end workflow for automated generation of 4D patient-specific phantoms for MR-Linac quality assurance. The method combines realistic anatomy, embedded dosimetry, multimodal imaging properties, and physiological motion within a single fabrication framework. This approachmay enable an improved validation of adaptive radiotherapy workflows in MR-guided treatment devices.

Abstract Background: The lack of effective biomarkers and therapeutic targets to overcome radioresistance in cervical cancer remains a major clinical challenge. Tumor necrosis factor receptor-associated factor 6 (TRAF6), an E3 ubiquitin ligase pivotal in immune and inflammatory signaling, has been implicated in various malignancies. However, its role in radioresistance in cervical cancer remains unclear. Methods: TRAF6 expression was evaluated in cervical cancer tissues from 162 patients who underwent postoperative radiotherapy at our institution and in 304 cases from the TCGA-CESC cohort. The prognostic significance of TRAF6 was assessed using Kaplan-Meier and Cox regression analyses. A nomogram integrating TRAF6 expression with clinicopathological factors was constructed to predict overall survival (OS) and progression-free survival (PFS). The functional role of TRAF6 in malignant phenotypes and radiosensitivity was investigated using shRNA-mediated knockdown in HeLa and C33A cervical cancer cells. Immune cell infiltration patterns associated with TRAF6 expression were analyzed using ssGSEA and xCELL algorithms based on TCGA data. Results: TRAF6 expression was significantly elevated in cervical cancer tissues compared with adjacent normal tissues (70.99% vs. control, P < 0.001) and was higher in radioresistant than in radiosensitive patients (P < 0.001). High TRAF6 expression was associated with shorter OS (HR = 18.73, P = 0.004) and PFS (HR = 8.44, P < 0.001) and was identified as an independent risk factor for radiotherapy resistance (OR = 8.44, P < 0.001). The TRAF6-integrated nomogram demonstrated good predictive accuracy for OS (C-index = 0.7351) and PFS (C-index = 0.7444). TRAF6 knockdown in cervical cancer cells significantly suppressed proliferation, migration, and invasion, while substantially enhancing radiosensitivity of tumor cells. Functional enrichment analysis revealed that TRAF6-related genes were enriched in autophagy, mitophagy, and HPV infection pathways. Immune cell infiltration analysis showed that TRAF6 expression correlated with distinct immune cell profiles, characterized by enrichment of activated dendritic cells, M1 macrophages, and regulatory T cells, alongside depletion of cytotoxic effectors such as CD8+ T cells and {gamma}{delta} T cells. Conclusions: TRAF6 could be a prognostic biomarker associated with poor outcomes and indicator of radiotherapy resistance in cervical cancer, TRAF6 represents a potential therapeutic target for overcoming radioresistance in cervical cancer.

Background: Nuclear medicine and radiopharmaceutical development require coordinated radiochemistry, dosimetry, molecular imaging, radiation-safety and clinical decision processes. Current workflows remain fragmented, difficult to audit and poorly standardised for evaluating domain-specific AI support. Methods: We developed RadGuide AI, a nuclear medicine agent built around a traceable data-model-tool loop. Patent, literature and clinical-trial records were converted into 15,596 initial QA items; relevance screening, completeness checks, semantic deduplication and cross-validation retained 5,474 core QA items. MedGemma-27B-Instruct served as the foundation model and was adapted with LoRA. The system incorporated 55 MCP-wrapped tools covering radiopharmaceutical R&D, clinical decision support, imaging analysis and radiation-safety/dosimetry. Evaluation used a locked N=200 benchmark with predefined denominators, leakage control, expert scoring, statistical procedures, factuality audits and tool-execution metrics. Results: RadGuide-LLM achieved 88.5% answer accuracy (177/200; 95% CI, 83.3-92.2%) and a Macro-Average score of 21.5/25 (bootstrap 95% CI, 20.9-22.0), exceeding GPT-4o, DeepSeek-V3.2 and the base MedGemma model in this technical evaluation. Supplementary audits reported guideline compliance, terminology recall, knowledge coverage, tool-routing success and preclinical/phantom dosimetry agreement with explicit denominators and confidence intervals. Interpretation: RadGuide AI converts nuclear medicine queries into auditable retrieval, tool selection, calculation, verification and reporting workflows. The findings support technical feasibility, not definitive patient-level clinical validation; prospective multicentre studies and external benchmark release remain required before clinical deployment.

Background: Three-dimensional tumor spheroids are an established radiobiology model, but scalable, reproducible readouts of dose-dependent radiation response are lacking. We evaluated whether optical coherence tomography (OCT) radiomics can quantify dose-associated response in spheroids, and how it compares with conventional brightfield morphology. Methods: This in vitro, cross-sectional study used SAS oral squamous cell carcinoma spheroids seeded at two densities (5000 and 10000 cells), irradiated at 0 to 12 Gy, and imaged on days 1 to 11 post-irradiation. Each OCT acquisition yielded co-registered structural-intensity and speckle-variance volumes. Radiomic features (shape, first-order, texture) were extracted with Radiomics.jl, filtered for repeatability, correlation-pruned, and ensemble-ranked. Dose correlation was assessed by repeated 5-fold cross-validation across five regressors, comparing brightfield-only (BF), OCT-only, and combined OCT+BF feature sets with paired Wilcoxon tests. Results: OCT-only models consistently outperformed the BF baseline (median R2 0.77 to 0.85 versus 0.61 to 0.69; p<0.001 for all regressors). Adding brightfield to OCT gave no consistent benefit, reaching significance only for Random Forest (p=0.026, power 0.62). A compact shared feature subset combined brightfield area dynamics with OCT texture, shape, and speckle-variance descriptors, all showing low repeat-scan variability relative to cohort variability. Conclusions: OCT radiomics provides a sensitive, reproducible, label-free high-throughput readout of spheroid radiation dose response that outperforms the current brightfield-based approach, without requiring concurrent brightfield acquisition.

Background: Cutaneous squamous cell carcinoma (cSCC) accounts for a significant proportion of skin malignancies in India, yet data on patterns of failure, particularly for extremity and truncal primaries remain scarce. We audited a decade of surgically treated cSCC at a tertiary cancer center to characterize failure patterns and associated risk factors. Methods: This retrospective study included 161 patients with histopathologically confirmed cSCC treated surgically between January 2013 and December 2023, comprising 127 upfront/residual and 34 recurrent presentations. Primary sites were extremities (64%), head and neck (26%) and torso (10%). 21 patients had Marjolin's ulcer. Outcomes included local, regional and distant failure, recurrence-free survival and overall survival. Brigham and Women's Hospital (BWH) staging was applied to assess prognostic utility. Statistical analysis was done using Kaplan-Meier and competing-risk methods. Results: Median follow-up was 2.4 years. Regional recurrence was the predominant failure pattern seen in 26 patients, local recurrence was seen in 14 patients and distant metastasis in 13. The 3-year cumulative incidences of local, regional and distant failure were 11%, 19% and 8.4% respectively. Rates of regional recurrence were substantially higher than Western series. Extremity primaries accounted for 19/26 regional recurrences. BWH T2b disease showed the highest regional failure rate (27.6%), exceeding T3 (17.8%) and T2a (6%) with perineural invasion significantly associated with regional failure in T2b/T3 tumors (p<0.001). Median time to regional metastasis was 8.4 months. At 3 years, overall survival was 77% and progression-free survival was 64%. Conclusion: Regional recurrence is the dominant mode of failure in this cohort, at rates higher than most published series, with extremity primaries and BWH T2b staging identifying particularly high-risk subgroups. These findings highlight the need for a comprehensive staging system encompassing non head and neck cSCC and support prospective evaluation of elective nodal staging and adjuvant radiotherapy in high-risk patients, alongside intensified surveillance.

Radiation exposure induces systemic and cellular damage, contributing to acute radiation syndrome and long-term effects such as premature aging and carcinogenesis. At the cellular level, radiation triggers apoptosis, mutation, and transformation through oxidative damage and activation of pathways including ER stress-mediated autophagy. Autophagy plays a context-dependent dual role in stressed cells, but its contribution to intestinal recovery after acute radiation remains unclear. Here, we evaluated combinatorial radiomodification using gamma radiation (8 Gy) and autophagy modulators in whole-body irradiated C57BL/6 mice (8-10 weeks old, n = 10). Mice were treated with autophagy inducers or inhibitors and euthanized at 3-, 8-, and 30-day post-irradiation. The jejunal-ileal region was analyzed via antioxidant assays, immunoblotting, H&E staining, and immunohistochemistry. Radiation significantly altered oxidative stress and autophagy markers, including increased LC3-II and decreased SQSTM1/p62. Autophagy induction enhanced intestinal proliferation (as measured by Ki-67), whereas inhibition impaired regeneration. Rapamycin pretreatment improved survival and reduced markers of intestinal injury following 8 Gy total body irradiation (TBI), whereas chloroquine exacerbated several injury-associated parameters. Overall, our findings suggest that targeted modulation of autophagy is a promising strategy for alleviating radiation-induced gastrointestinal injury and provide mechanistic insights relevant to therapeutic development.

Objective: Preclinical small-animal irradiators such as the FLASH-SARRP can support the advancement of photon-FLASH toward the clinic. This study aimed at characterizing the FLASH-SARRP and established a robust quality assurance (QA) workflow to enable accurate and reproducible preclinical experiments. Approach: Custom 3D-printed spacers were designed to ensure reproducible X-ray tube alignment, sample positioning and mounting of the dosimetric tools. Beam characteristics were evaluated using a combined dosimetric approach. High spatially resolved dose distributions were obtained from Gafchromic films, whereas a plastic scintillating fiber was employed to monitor in real-time the temporal pulse structure and synchronization between the two X-ray tubes. Day-to-day variability of the delivery was evaluated over several sessions. Main results: The FLASH-SARRP achieved dose-rates of around 80 Gy/s when both tubes were used simultaneously and provided a homogeneous irradiation field suitable for small-animal studies. A desynchronization between the two tubes was observed with an average delay of 10 ms, resulting in temporal dose-rate heterogeneity. Additionally, a substantial inter-session variability (~11%) was found, whereas the intra-session variability was relatively low (~4%). Inter-session variability was reduced to 5%, approaching the intra-session variability, by adding Gafchromic films/scintillator-based quality assurance (QA) workflow into the irradiation routine. Significance: This work highlights the importance of temporal dosimetry for preclinical FLASH studies. Additionally, a practical QA framework is proposed integrating real-time monitoring with reference dosimetry. The proposed work enables adaptive dose delivery, thereby enhancing the reproducibility of the irradiations, which is crucial for reliable preclinical studies on the FLASH effect.

Three audiences -- the family of a newly diagnosed Ewing sarcoma patient, the long-term survivor, and the cooperative-group trial statistician -- receive cohort-mean answers to patient-level questions because the patient-level data machine learning requires do not exist for rare cancers. We present a framework producing patient-level predictions from published aggregate trial data. A six-stage discrete-event Monte Carlo simulation integrates genetic risk factors, serial biomarker dynamics with genotype-conditional weighting, post-surgical ctDNA-based minimal residual disease (ctDNA-MRD) assessment, and treatment-related mortality as a separable competing risk. Adverse-effects modules project 30-year incidence across five organ systems from chemotherapy and radiation exposures. Its four structural ingredients are instantiated in Ewing sarcoma and validated against trial data from more than 3,400 patients. The framework achieves 3.2% mean absolute error across 23 efficacy endpoints (none exceeding 6%) and falls within published confidence intervals for all 20 toxicity endpoints. ctDNA-MRD stratification separates candidate populations -- 5.5% recurrence (de-escalation) versus 87.8% (intensification) -- and multi-timepoint integration produces 16-fold five-year EFS resolution spanning 5-96%, exceeding the 3- to 5-fold ranges of single-timepoint approaches. The 16.1-fold recurrence risk ratio emerges from simulation, not as a supplied parameter. Genotype-conditional weighting improves discrimination over equal-weight scoring in every subgroup (Pearson r +0.060 to +0.129), with largest gains where biological rationale is strongest. A Monte Carlo framework calibrated to published aggregate data turns cohort-mean answers into patient-level predictions as exemplified in the rare cancer Ewing sarcoma, where the conventional patient-level machine-learning pathway is structurally unavailable; transfer to other rare cancers remains a hypothesis for future validation. Survivorship-surveillance refinement is the most concrete current use; trial-design and prognostic counseling are next-decade pathways.

Abstract Introduction: The optimal examined lymph node (ELN) count after resection for pancreatic body/tail ductal adenocarcinoma (PDAC) remains uncertain. Guidelines recommend 12-15 nodes, but the value of higher thresholds is unclear. Method: SEER patients with pancreatic body/tail PDAC undergoing resection from 2000 to 2020 were analysed. Survival-anchored ELN thresholds were assessed using log-rank cut-point search, segmented Cox analysis, adjusted restricted cubic splines, and overlap-weighted restricted mean survival time (OW-RMST). A structured synthesis of 17 studies compared threshold attainment after conventional distal pancreatectomy (DP), radical antegrade modular pancreatosplenectomy (RAMPS), and posterior/artery-first approaches. Results: Among 5107 patients, 3630 deaths occurred (71.1%). Log-rank analysis identified ELN = 12 as the optimal binary cut-point; segmented Cox analysis identified ELN = 21 as a change point (bootstrap 95% CI 6.0-35.0). Adjusted splines showed a nonlinear inverse association between ELN and mortality, with attenuation beyond approximately 21 nodes. Each 5-node increase in ELN was associated with lower mortality (HR 0.964, 95% CI 0.949-0.980; P < 0.001). At 60 months, OW-RMST gains for ELN >= 12, >= 14, and >= 21 were 2.59, 2.31, and 2.60 months. Estimated probabilities of achieving ELN >= 21 were 16.5% after conventional DP, 40.0% after RAMPS, and 82.7% after posterior/artery-first approaches, with lowest certainty for the latter. Conclusion: ELN >= 12 is a minimum quality benchmark after resection for pancreatic body/tail PDAC, whereas approximately 21 nodes may be a higher-yield target. RAMPS may improve target attainment, but survival superiority remains unproven.

Abstract Objective To evaluate the adherence to guidelines and readability of large language model-generated sexual health information related to female sexual dysfunction following cystectomy, and to determine whether adherence differs across models and prompt formats. A secondary objective was to introduce an analytic strategy using principal component analysis to examine the dimensions of readability metrics. Methods Three large language models (LLMs), ChatGPT, Gemini, and Perplexity were prompted with six clinical questions related to sexual function after cystectomy. Questions were phrased in long-form and short-form language. Responses were independently graded by two reviewers, derived from guideline recommendations. Linear mixed-effects models predicted adherence as functions of LLM, prompt, and reviewer, with clinical questions as a random intercept. Readability was assessed using five metrics, and principal component analysis (PCA) was used to determine latent structure. Results ChatGPT demonstrated the highest (estimated marginal mean [emm] = 0.769), outperforming Gemini (0.499) and Perplexity (0.457). Shorter, less complex prompts elicited higher adherence than more complex, clinical prompts. All models produced content that exceeded recommended reading levels. PCA demonstrated that a single dominant component accounted for 76.7% of variance across readability indices, indicating a shared underlying construct. Conclusion ChatGPT produced the most guideline-concordant information overall. High linguistic complexity was seen across models, highlighting a barrier to patient comprehension. These findings characterize large language models as variable medical information systems whose outputs rely heavily on prompt structure and model type.

The histone variant H2A.J is implicated in radiation-induced senescence by promoting the transcription of inflammatory genes. While H2A.J expression increases in renal tubular epithelial cells (TECs) following ionizing radiation (IR), its functional role remains poorly understood. To investigate this, constitutive H2A.J knock-out (KO) mice and wild-type (WT) controls were subjected to CT-guided IR (fractionated whole-body or localized kidney IR). Kidneys were analyzed at acute, intermediate, and chronic stages using immunofluorescence, histochemistry, automated image analysis, and electron microscopy. In WT TECs, IR induced rapid chromatin incorporation and C-terminal serine phosphorylation of H2A.J. Conversely, KO kidneys exhibited significantly more severe histopathological damage, including tubular dilation, flattened epithelium, associated with increased apoptosis, and premature senescence, characterized by persistent DNA damage with lamin B1 loss. Notably, KO TECs displayed disrupted mitochondrial networks and reduced brush borders even at baseline, which were further exacerbated by IR. Unlike WT controls, KO kidneys developed progressive tubular atrophy and incipient fibrosis, indicating a failure in regenerative capacity. Our findings suggest that H2A.J loss impairs tubular regeneration due to defective mitochondrial activation, resulting in insufficient energy supply for coordinated repair. Collectively, these results identify H2A.J as a critical stress-adaptive histone variant essential for the epigenetic regulation of tissue repair following radiation-induced damage. One Sentence SummaryIn irradiated kidney, the loss of histone variant H2A.J impairs the chromatin-mediated adaptation of mitochondrial function in tubular epithelial cells, thereby exacerbating cellular stress - characterized by increased induction of apoptosis and senescence - and ultimately leading to tubular atrophy. Translational RelevanceAcute and chronic kidney injury are frequent complications of genotoxic cancer therapies. Chemo- and radiotherapy induce DNA lesions that trigger cell death and senescence, often leading to irreversible renal damage. However, renal regeneration can occur through the dedifferentiation, proliferation, and redifferentiation of surviving tubular epithelial cells (TECs). This repair process is governed by epigenetic mechanisms that regulate the DNA damage response (DDR) and adapt gene expression programs. Following ionizing radiation (IR), epigenetic remodeling involves the incorporation of histone variants that modulate chromatin accessibility for stress-responsive transcription factors. We identify the histone variant H2A.J as a constitutive component of renal TECs, significantly upregulated after exposure to ionizing radiation (IR). Using H2A.J knock-out (KO) mice, we demonstrate that its absence disrupts acute damage responses and prevents coordinated repair, severely impairing regeneration. Mechanistically, H2A.J deficiency compromises mitochondrial function under postirradiation metabolic stress, driving the transition from acute injury to chronic kidney disease via persistent inflammation and maladaptive tubulointerstitial repair. Targeting these epigenetic drivers offers a promising strategy for regenerating damaged kidney tissue in oncology.

Background: High ambient temperature in early pregnancy has been linked to an increased risk of childhood acute lymphoblastic leukemia (ALL). To better understand biological mechanisms, the current study evaluated potential interaction between temperature and genetic characteristics. Methods: We used data from California birth records (1982-2008) and California Cancer Registry (1988-2011) to identify ALL cases (n=3,353) diagnosed <=14 years of age and non-cancer controls (n=3,530) matched 1:1 on sex, race, ethnicity, and birth year and month. Weekly ambient temperatures throughout pregnancy were assessed on a 1-km grid around the birth address, while genetic data were available from a genome-wide association study using neonatal blood spots. We evaluated the association between ambient temperature and ALL risk by quartiles of established genetic risk score for ALL. Next, we formally tested gene-temperature interactions in the association with ALL, correcting for multiple testing, for genes previously identified with epigenetic changes due to both temperature and ALL. All analyses were adjusted for potential confounders. Results: The elevated risk of ALL per 5 degrees C increase of weekly mean ambient temperature, confined to early pregnancy, was more pronounced among children with the lowest genetic susceptibility to ALL, especially among Latino children (first quartile: odds ratio [OR] = 1.50, 95% confidence interval [CI]: 1.14-1.97); fourth quartile: OR=1.03, 95% CI: 0.83-1.28). There were significant interactions (p<0.002) between ambient temperature and polymorphisms in BNC1 among non-Latino White children, and suggestive interactions (p<0.05) with TBPL2 and NRXN1 in the full population. Conclusions: Our findings suggest that there may be interactions between ambient temperature in early pregnancy and offspring genotype in the risk of childhood ALL. Impact: If replicated, these findings could help elucidate the biological mechanisms linking high ambient temperature in early pregnancy and the risk of childhood ALL.

Introduction: For oncology patients with limited treatment options, clinical trials may be a critical lifesaving pathway. Identifying relevant trials, however, is a time-consuming and difficult task. Several patient-trial matching processes incorporating large language models (LLMs) have been proposed to alleviate the burden on patients and oncologists. We aim to explore the benefits and practical challenges of zero-shot LLM-assisted trial matching processes by analyzing the results for a single pancreatic cancer patient. Materials and Methods: The results of a simple zero-shot LLM-assisted clinical trial matching process for our patient were compared to those of a "human benchmark," which was developed manually by two of the authors interfacing directly with ClinicalTrials.gov. Performance metrics -- sensitivity, specificity, precision, and accuracy -- were calculated. In addition, a qualitative content analysis (QCA) of LLM reasoning text was done to identify patterns in "errors," which we define as a human-LLM discrepancy in final patient eligibility. Implications and severity of errors are discussed. Results: The zero-shot LLM-assisted process returned potential trials with a sensitivity, specificity, and precision of 81.1%, 89.3%, and 86.5% respectively compared to the human benchmark. Qualitative error analyses revealed that about 73% of errors could potentially be alleviated with improved prompting and information access. Overall performance seemed comparable to that of human reviewers. Conclusion: The results from this preliminary real-world case study provide additional evidence to the literature in support of the integration of LLMs in clinical trial matching to provide benefit to patients with metastatic cancer with limited options.

Background. Duodenal adenocarcinoma (DA) is a rare gastrointestinal malignancy with limited DA-specific evidence on the prognostic role of perineural invasion (PNI) and the performance of AJCC 8th-edition staging. We described PNI, AJCC 8 discrimination, and adjuvant-chemotherapy subgroup associations in a single-center cohort. Methods. We retrospectively analyzed 51 patients with curatively resected, histologically invasive DA (2013-2025). Overall survival (OS) and disease-free survival (DFS) were estimated by Kaplan-Meier analysis; prognostic associations by Cox regression. Model discrimination was quantified by the C-index; adjuvant-chemotherapy associations were explored across pre-specified subgroups. Analyses were exploratory and hypothesis-generating. Results. Median follow-up was 34.9 months; 26 patients (51%) died and 32 (63%) recurred, with the highest recurrence frequency 6-12 months after surgery. AJCC 8 staging discriminated modestly (C-index 0.595); no adjacent stage pair differed significantly after Bonferroni correction (minimum raw pairwise P = 0.051). Among candidate factors, PNI had the largest effect estimate (univariable HR 2.08, 95% CI 0.88-4.90, P = 0.095) and the largest incremental discrimination when added to a stage + T + N base model (delta C-index +0.062; likelihood-ratio P = 0.145), exceeding the increments from tumor size (+0.046), differentiation (+0.034), and sex (+0.016). Adjuvant chemotherapy was associated with hazard-ratio reductions in PNI-positive (HR 0.19, 95% CI 0.02-1.59), stage III (HR 0.36, 95% CI 0.12-1.15), and node-positive (HR 0.36, 95% CI 0.12-1.15) subgroups; none reached statistical significance, consistent with limited power in subgroups of 8-22 patients. Conclusion. In this small single-center cohort, PNI showed the largest prognostic effect estimate among candidate variables and the largest incremental discrimination beyond AJCC 8 stage, although neither reached statistical significance and AJCC 8 discriminated only modestly. These hypothesis-generating findings - directionally concordant with larger published DA series - support formal evaluation of PNI as a stratification variable in multicenter cohorts and in individual-patient-data meta-analyses of duodenal adenocarcinoma, rather than a change in current practice.

Purpose: To evaluate the latest open-weight vision-language models (VLMs) on the Japanese Diagnostic Radiology Board Examination (JDRBE), assessing overall accuracy and the effects of image input, reasoning, and language. Materials and Methods: In this retrospective study, 29 open-weight VLMs from 13 developers, released in or after January 2025, were evaluated on 327 image-bearing questions from four years of the JDRBE, a non-public benchmark with low risk of data leakage. Each question was answered by each model with and without the image(s), under three language conditions and with reasoning enabled and disabled. Accuracy was the primary outcome, and within-model differences were tested with paired bootstrap confidence intervals and sign-flip permutation tests with Benjamini-Hochberg correction. Results: In the Japanese condition with image input and reasoning, the leading models reached 73.7% (gemma-4-31B-it), 73.1% (Qwen3.5-397B-A17B), and 72.1% (Kimi-K2.6). On the 2025 subset, these three models (74.1%-75.5%) scored above the mean accuracy of five newly board-certified radiologists who passed the 2025 examination (72%; range, 65%-83%). Accuracy broadly scaled with model size, although compact gemma-4-31B-it matched larger models. Enabling reasoning improved accuracy in nearly all models and the contribution of image input was larger when reasoning was enabled, particularly in higher-performing models. English prompts generally outperformed Japanese prompts. Conclusion: Several open-weight VLMs, without medical adaptation, performed at or above the mean of newly board-certified radiologists on the JDRBE, with both model size and reasoning contributing. The highest Japanese-language accuracy came from a compact model suitable for parameter-efficient fine-tuning and serving on a single graphics processing unit.

Objectives. To determine whether medical machine olfaction via tracking the activation of mammalian G-Protein Coupled odorant Receptors (GPCR) stabilized by proprietary co-polymers on a photonic MZI chip can be used to diagnose prostate cancer (PCa) via urine scent. Specifically, scent character is compared against the current diagnostic PCa screening gold-standard in the US: the serum level of prostate specific antigen (PSA). The device is an artificial nose sensor built on a commercial photonic platform that reads interchangeable Mach-Zehnder interferometer (MZI) chips. These chips were functionalized with a stabilised panel of mammalian olfactory G-protein-coupled receptors (GPCRs). These samples had been characterized into POSITIVE or CONTROL for PCa six to eight years prior by standard hospital diagnostic procedures and by trained medical detection dogs, then stored at -80 Celcius. A subset of 80 patients urine samples was subsequently thawed and used for training and testing the medical machine olfaction system of RealNose as an initial validation of the novel technology and methodological approach. We posed two primary research questions: (a) whether the cancer-associated odor profile would remain detectable by machine-based systems following long-term storage and with what accuracy could it be used to cluster (YES and AUC 0.79 from scent character alone), and (b) what technical and procedural requirements would be necessary to translate such a signal into a clinically useful diagnostic assay (more training samples (500 predicted to yield 0.93) and increased breadth of receptors per chip and/or more chips per device in next iteration seen as helpful). Design, setting, participants. Retrospective diagnostic-accuracy feasibility study on 80 biobanked urine samples (40 PCa, 40 non-cancer; 368 sensor runs; a subset of unknown Gleason grade) from a single UK NHS urology service, the same collection used to train canine detectors. Main outcome measures: Patient-level Receiver Operating Characteristic (ROC) area under the curve (AUC) under patient-grouped cross-validation with a fold-honest pooled-control reference (reconstructed from training-partition controls only); sensitivity, specificity and predictive values at pre-specified operating points; 1000-fold whole-procedure label-permutation significance; patient bootstrap 95% CIs; and leave-one-day-out / leave-one-chip-out generalisation. Results. An L2-regularised linear classifier when allowed to see between three and six chips outcome on a patient sample extracted within-instrument AUC 0.79 (95% CI 0.69 to 0.88; 1000-permutation p = 0.001) from urine scent alone, exceeding this cohort own serum prostate-specific antigen (PSA) discrimination (AUC 0.645; itself within the population range for PSA 0.67) and obtained without a blood draw (at the Youden point, sensitivity 0.75, specificity 0.78, PPV 0.77, NPV 0.76). Upon allowing PSA the total AUC rose to 0.82. This was not a plateau: AUC rose from chance at 30 training samples, passed the serum-PSA range at 40, and reached 0.79 at 80 patients (0.82 if PSA was included), with an inverse-power fit projecting 0.93 by n = 500 and 0.96 by n = 1000. The discriminant was a genuine multivariate receptor pattern, independent of patient age (Spearman 0.09; the cohort is not age-matched). So at least for these data, neither age, nor collection day, ambient humidity/temperature, or overall signal amplitude (sometimes thought of as intensity of smell) were predictive of prostate cancer status, yet the scent character was. Transfer to a new sensor chip fell to AUC 0.57 without calibration, meaning the remaining obstacles are hardware portability rather than signal existence: much as a detection dog acclimatizes to a new setting, the system improves with on-site calibration prior to use. Conclusions: A genuine, confound-controlled olfactory PCa signature is recoverable from 80 samples, surpasses this cohort serum PSA (0.645) and exceeds the population PSA range, and improves monotonically with training-set size. We present this as a small-sample feasibility benchmark, not yet a validated diagnostic; the dominant remaining factor is training-set size, and the path to clinical-utility and improved AUC is clearly found to be a larger, multi-site, age-matched, and ideally prospective training cohort. A transferable small-sample lesson is also reported: adaptive feature searches (evolutionary and self-calibrating-protocol handle search) artificially inflate cross-validation and collapse under whole-procedure permutation, whereas non-adaptive averaging survives, giving a robust scent signal obtainable from the headspace of urine samples and recordable by the RealNose device that keeps improving with expanding sample training set.

Abstract Aim: Low chemotherapy response is a major risk factor for early mortality in cancer patients; it is one of the biggest challenges in cancer treatment. Main aim of this study is to identify chemotherapy non-responder, prognostic significance of pre-chemotherapy baseline variables in survival, distinguish most effective anti-cancer drug classes and formulation. Methods: In this multi-center retrospective cohort (n=2459) patients deceased with NSCLC and received anti-cancer drugs were included for analyses. To identify chemotherapy non-responder, patient population was divided into three sub-groups based on chemotherapy prescription frequency [1-15] as group-A, [16-30] as group-B, and [[&ge;]31] as group-C. Multivariate analysis was performed to identify risk of 1-year mortality in these groups. To prognose chemotherapy response in resected and unresected NSCLC patients, 0-7 days pre-chemotherapy white blood cell (WBC) count total five-ranges were compared as per overall survival in abnormal Vs normal WBC counts. Results: Post-stratification in group-A there were (n=1289) patients, in group-B (n=648) patients, and in group-C (n=522) patients. In group-A (n=301) patients 23% were found to have no new metastasis post-diagnosis significantly less p-value (0.004) compared to Group-B (n=125) 19.3%, and group-C (n=110) 19.2% patients p-value (0.008). Metastasis during chemotherapy was found significantly less in 20% patients of group-A, compared to (33%) in group-B, and (43%) in group-C p-value (<0.001). Post-chemotherapy initiation OS in group-A patients were significantly less 9 months (95% CI 9.3 - 9.6) compared to group-B 19 months (95% CI 17.7 - 20.2) and group-C 36.6 months (95% CI 34.6 - 38.5) patients p-value (<0.0001). Despite of low new metastasis and post chemo metastasis, group-A patients survived significantly less based on these outcomes group-A patients were considered as chemotherapy non-responder. Males and NSCLC stage III/IV patients were at higher risk; clinical benefits are corelated to surgery and radiotherapy for chemotherapy non-responder. Leukocytosis in both resected/unresected NSCLC group-A (13%) patients were found to be bad prognostic factor of survival in unresected group-B (5%) patients. Oral formulation of receptor tyrosine kinase inhibitors (RTKI) was effective in non-responders. Conclusion: Stratification of patient population based on chemotherapy prescriptions could be a useful method to find chemotherapy response in retrospective analysis. Patients with pre-chemotherapy leukocytosis should be closely monitored prior to selection of chemotherapy dose and formulation.