IntroductionAccurate prediction of neoadjuvant chemotherapy (NAC) response in breast cancer is important for treatment planning and individualized therapy selection. Breast MRI contains subtle patterns related to treatment response. These patterns vary across image regions and may be difficult for standard deep learning models to capture. In this retrospective clinical prediction model development and validation study, we propose GRACE-ViT, a Vision Transformer-based framework for three-class NAC response prediction from pre-treatment breast MRI. The three classes are partial response, complete response, and stable disease.MethodsGRACE-ViT uses a pretrained ViT backbone with a lightweight token recalibration module called Grouped Recalibration with Adaptive Contextual Emphasis (GRACE). The GRACE module refines patch-token features using dual-mode token scoring, a spatial coherence prior, and cross-group feature integration. This helps the model focus on important breast regions while keeping global image context. The model was evaluated on 736 axial T1-weighted breast MRI images. It was compared with convolutional, transformer-based, and medical-imaging models under the same preprocessing, training, validation, and testing protocol. Performance was measured using accuracy, mean F1-score, precision, and recall where bootstrap resampling was used to estimate 95% confidence intervals.ResultsGRACE-ViT reached a mean accuracy of 95.50% (95% CI: 90.99-99.10%) and a mean F1-score of 95.47% (95% CI: 91.10-99.06%). It outperformed the strongest baseline with only a small increase in computational cost. The per-class results were balanced across the three response categories, suggesting that the model did not favor one clinical outcome over another. Visual results showed that the model mainly focused on clinically relevant breast regions rather than background areas.ConclusionsThese results show that targeted token recalibration can improve ViT-based breast MRI analysis without greatly increasing model complexity, leading GRACE-ViT to provide an efficient and interpretable imaging-based approach for predicting NAC response and supporting individualized treatment decisions in breast cancer care.

Manual chart review (MR) of electronic health records (EHRs) is time-consuming, error-prone, and limits the reproducibility and scalability of real-world data (RWD) research. Automation and standardization using natural language processing (NLP) could improve efficiency and scalability. CTcue is an NLP-based software platform designed to extract structured and unstructured data from EHRs. This study evaluated the accuracy and efficiency of CTcue versus MR in patients with early-stage resectable non-small cell lung cancer (NSCLC).

Included were all patients with stage I to III NSCLC who underwent lung resections between January 2018 and December 2021 at the Leiden University Medical Center, the Netherlands. Demographics, tumor characteristics, treatment, and outcomes were collected. CTcue performance was compared with MR using weighted F1-scores, accuracy, precision, and recall for categorical variables and Bland-Altman analysis for continuous variables.

Eighty-five patients (70.2% of patients from the manual cohort) were identified by both methods and included in the comparison. CTcue achieved weighted F1-scores >0.85 for seven of 15 categorical variables, including sex, tumor location, and deceased status, although some scores were based on low number of observations in both cohorts. Lower performance was observed for variables with varying terminology in documentation, such as Eastern Cooperative Oncology Group status and pathological N-stage. Continuous variables showed negligible mean differences, indicating good agreement. Survival outcomes were identical in both data sets.

CTcue performance for patient selection was lower than anticipated. However, it enables accurate, efficient extraction of structured and unstructured EHR data in early-stage NSCLC. Manual validation remains necessary for variables with varying terminology. Further development of artificial intelligence-based tools-particularly for free-text data extraction-will be crucial to enhance the accuracy and scalability of future RWD research.

The incidence of cancer is rising in India, and radiotherapy plays a vital role in the management of most cancers. More radiotherapy centers are being set up across the country to cater to this need. Quality assurance (QA) of radiotherapy is crucial to minimize risks and ensure best outcomes for all patients receiving radiotherapy. We outline an effort to develop a standardized radiotherapy QA program implementable across India through a consensus process among stakeholders.

The team structured a Standardised Unified Radiotherapy QA model (SURAQSA), a program of pan-India radiotherapy QA incorporating elements of radiotherapy trials QA, adapted for the Indian health care system. Engagement of a cross-section of radiation oncologists working across India to achieve consensus on the issue was essential. This paper describes the consensus development process among the diverse Indian centers/professionals in support of the SURAQSA program. The statement was introduced to clinicians through the vehicle of an abbreviated Delphi consensus process, with guidance from the Royal College of Radiologists (United Kingdom). We outline the elements of the SURAQSA program and describe the consensus adopted.

Thirteen centers from across India were represented in the workshop. Three of the consensus statements on external QA of each center, prospective peer review of contours, plan quality measures, and protocol-driven management, as the basic requirements received unanimous support. Consensus was also achieved on the pilot subsite for this exercise with minor disagreement.

This report showed feasibility of a wide stakeholder engagement to deliver focused outputs and is a huge step toward formulating site-specific radiotherapy QA programs. The process allowed oncologists to highlight the need to take essential steps for radiotherapy safety across centers with varying availability of trained manpower and infrastructure.

Head and neck cancers (HNCs) remain a major contemporary public health challenge, yet few population-based studies have examined the long-term incidence trends of HNCs in low- and middle-income countries.

Using International Classification of Diseases, 10th Revision code data from the Hanoi Cancer Registry (1991-2020), we examined 30-year trends in HNC incidence rates using age-standardized incidence rates (ASIRs). Incidence rates were stratified by sex, primary site, 5-year age groups, and within six 5-year diagnostic periods spanning 1991-1995 through 2016-2020. Ratio-based imputation was applied to account for expansions in Hanoi's incorporated administrative areas. Temporal trends were assessed using joinpoint (JP) regression to estimate average annual percent changes and age-period-cohort models were used to evaluate independent age, period, and birth cohort effects.

A total of 19,589 HNC cases were identified (74% men). ASIRs increased from 10.9 per 100,000 in 1991-1995 to 18.8 in 2006-2010, and then declined to 11.8 in 2016-2020. The proportion of male cases of HNCs increased from 62.2% in 1991-1995 to 79% in 2016-2020. Nasopharyngeal and sinonasal cancers decreased from 68% to 21% of male cases and from 59% to 36% of female cases. Oral cavity cancers increased proportionally among men. JP and age-period-cohort analyses indicated a strong period effect peaking in 2006-2010 and declining risk among post-1950 birth cohorts.

Over the 30-year study period, we observed a growing male predominance of HNCs, a shift in burden toward oral cavity tumors over time, and a widening male-to-female disparity in several HNC primary site cancers. Our findings highlight an evolving epidemiologic transition in Northern Vietnam, underscoring the need for strengthened surveillance, screening, and targeted prevention and control strategies.

The purpose of this study was to identify participation determinants among high-risk individuals in risk-stratified upper GI (UGI) cancer screening.

This sequential mixed-methods study was conducted in a population-based UGI cancer screening program in Ningxia Province, China. It integrates quantitative and qualitative methods to comprehensively examine participation determinants. Between April and December 2023, quantitative data were collected from 2,571 high-risk individuals identified by validated diagnostic models. Subsequently, 42 high-risk individuals (median age: 63.0 years; 81.0% male) were interviewed by stratified purposive sampling from March to April 2024. Logistic regression explored factors associated with screening adherence. Thematic analysis of qualitative interviews integrated quantitative findings using the Capability, Opportunity, and Motivation-Behavior framework.

Of the high-risk individuals, 52.8% declined endoscopy. Nonparticipation was significantly associated with older age (adjusted odds ratio [aOR], 1.28 [1.06-1.54]), male sex (1.31 [1.01-1.70]), Hui ethnicity (1.35 [1.12-1.62]), nonsmoking (1.23 [1.01-1.50]), nondrinking (1.36 [1.09-1.71]), no family history of UGI cancer (1.48 [1.20-1.84]), fewer UGI symptoms (2.64 [2.19-3.18]), and no prior UGI examination (1.67 [1.37-2.05]). In the qualitative phase, seven key barriers were identified, including insufficient knowledge about UGI cancer and risk assessment, physical limitations, low perceived risk, gastroscopy-related fears, cost concerns, time constraints, and inadequate information.

Addressing limited knowledge and program accessibility may improve participation and real-world implementation of risk-stratified UGI cancer screening.

Patients with lymphoma with bone marrow infiltration have a poor prognosis, and reliable bone marrow infiltration confirmation methods are critical for clinical management. This study aimed to construct machine learning models to identify bone marrow infiltration-influencing factors and predict lymphoma bone marrow infiltration using multimodal clinical data.

A retrospective study included 243 bone marrow infiltration and 541 non-bone marrow infiltration patients with lymphoma. Eight machine learning algorithms were used to build models based on clinical characteristics, laboratory tests, and fluorine-18fluorodeoxyglucose Positron emission computed tomography (PET/CT) data. Model performance was evaluated by AUC and accuracy. A sensitivity analysis was performed in the pathology-confirmed bone marrow infiltration cohort to assess potential incorporation bias.

Platelet count, hemoglobin, serum lactate dehydrogenase, lymph node involvement distribution, bone infiltration characteristics, and maximum standardized uptake value (SUV_max) of sternum, SUV_max of thoracic vertebra, SUV_max of lumbar vertebra, SUV_max of sacrum, SUV_max of femur correlated with lymphoma bone marrow infiltration. The random forest (RF) model showed the best performance (AUC = 0.842, accuracy = 80.9%) in the testing set. In pathology-confirmed subgroup analysis, the RF model maintained stable performance (AUC = 0.846, accuracy = 82.5%).

The RF model is a powerful tool for lymphoma bone marrow infiltration prediction (AUC = 0.774, accuracy = 80.3%) using routine clinical, laboratory and PET/CT data, providing a promising approach for bone marrow infiltration assessment, PET/CT interpretation and quantitative risk stratification.

Lebanon's ongoing economic collapse has severely disrupted access to cancer care and increased out-of-pocket financial burden for patients. Financial toxicity and psychological distress are recognized barriers to treatment experience and adherence, yet little is known about how these factors interact within collapsing health systems. This study assessed whether psychological distress is indirectly associated between financial toxicity and treatment satisfaction among Lebanese adults with cancer.

A cross-sectional survey was conducted among 308 adult patients with cancer receiving or recently completing treatment across Lebanon. Participants completed validated Arabic versions of the Comprehensive Score of Financial Toxicity Functional Assessment of Chronic Illness Therapy (COST-FACIT; financial toxicity; lower scores indicate greater toxicity), PHQ-4 (anxiety/depression), FACIT-TS-G (treatment satisfaction), and Multidimensional Scale of Perceived Social Support (social support). Descriptive statistics, bivariate analyses, and multivariable linear regression were performed. Indirect associations were tested using Hayes' PROCESS Model 4 with 5,000 bias-corrected bootstrap samples, adjusting for relevant sociodemographic and clinical covariates including age, insurance status, and treatment variables.

Participants had a mean age of 47.9 years; 58.4% was female. Lower COST scores (greater financial toxicity) were significantly associated with higher psychological distress (P < .001). Higher distress was significantly associated with lower treatment satisfaction (P < .001). In multivariable models, financial toxicity, psychological distress, social support, and insurance coverage were independent predictors of treatment satisfaction. Indirect analysis indicated a significant indirect association between financial toxicity and treatment satisfaction through psychological distress. The direct association remained significant, indicating mediation.

Among patients with cancer living through Lebanon's economic crisis, greater financial toxicity was associated with higher psychological distress and lower treatment satisfaction. Psychological distress accounted for this relationship. These findings underscore the need for integrated psychosocial support and financial navigation within resource-constrained oncology settings.

Outcomes in adults age 40 years and older with ALL are inferior to younger patients, with limited real-world data.

This is a retrospective analysis of adults age 40 years and older with newly diagnosed ALL enrolled in the Hematology Cancer Consortium registry from January 2019 to January 2021. Baseline variables, treatment patterns, response rates, toxicities, and survival were assessed. Overall survival (OS) and event-free survival (EFS) were estimated using Kaplan-Meier methodology.

Of 349 registered patients, 61.3% (n = 214/349) initiated therapy, while 38.6% (n = 135/349) refused therapy. B-ALL represented 81% of cases (n = 174/214). Berlin-Frankfurt-Münster was used in 61.6% (n = 132/214) and German ALL in 30.3% (n = 65/214). Induction mortality was 18.6% (n = 40/214), mainly infection-related. Complete remission rates were 59.8% among assessed patients (n = 128/214), with minimal residual disease negativity in 45.8% (n = 61/133). Treatment de-escalation due to intolerance occurred in 30.8% (n = 66/214), and 35% (n = 75/214) were lost to follow-up. 3.7% (n = 8) patients underwent allotransplantation. At a median follow-up of 34 months, 3-year OS and EFS rates for the entire cohort were 31% and 29.6%, respectively (n = 214). Median OS and EFS were 13 and 11 months, respectively.

Adults age 40 years and older with ALL in India have poor outcomes, driven by early mortality, treatment intolerance, and limited access to advanced therapies.

Lymphoma affecting the central and peripheral nervous systems in equids is rarely reported.

To describe the clinical presentation, neuroanatomical localization, laboratory and pathologic features, type of lymphoma, and outcome to further current knowledge.

Fourteen equids (13 horses, 1 mule).

Retrospective study. Medical records from a veterinary teaching hospital were reviewed (1993-2025). The inclusion criteria consisted of a definitive diagnosis of lymphoma of the nervous system. Data collected included signalment, neurologic findings, laboratory results, imaging, cerebrospinal fluid (CSF) analysis, immunophenotyping, and necropsy examination. Descriptive statistics were used.

Nervous system lymphoma, either primary or concurrent with multicentric lymphoma, accounted for 11.6% of lymphoma cases. Multifocal neurologic disease was the most common presentation. Common hematologic findings included leukocytosis with neutrophilia, normal plasma fibrinogen concentration, high neutrophil-to-lymphocyte and red cell distribution width-to-platelet ratios, and hypocalcemia. Cytology of the CSF provided a definitive diagnosis in 56% of sampled cases. Prognosis was poor with a median survival time of 4.5 days from diagnosis. All lymphomas were of B-cell origin, including diffuse large B-cell lymphoma, small lymphocytic lymphoma, and T-cell-rich large B-cell lymphoma.

Nervous system lymphoma in equids might be more common than previously reported. Primary nervous system small lymphocytic B-cell lymphoma in a horse and multifocal nervous system lymphoma in a mule have not been reported previously. Analysis of CSF is useful in the antemortem diagnosis of central nervous system lymphoma. Neurologic examination, CSF analysis, and necropsy should be routinely performed in cases of multicentric lymphoma to rule out neurologic involvement.

L-CHOP (L-asparaginase, cyclophosphamide, doxorubicin, vincristine, prednisone) is the standard therapy for lymphoma in dogs, but relapse and limited long-term survival are common. Given the radiosensitivity of lymphoma, sequential half-body radiotherapy (HBRT) has been investigated to improve outcomes.

Evaluate progression-free survival (PFS), overall survival (OS), and adverse events in dogs treated with a no-maintenance L-CHOP protocol combined with low-dose-rate (LDR, 6 Gy; 8-14 cGy/min) HBRT compared to L-CHOP alone.

Client-owned dogs with intermediate- to high-grade multicentric lymphoma treated with L-CHOP-LDR-HBRT or a 6-month L-CHOP protocol (controls).

Medical records were retrospectively reviewed in this case-control study. Adverse events were graded using VCOG-CTCAE v2 (Veterinary Cooperative Oncology Group-Common Terminology Criteria for Adverse Events guidelines) criteria.

Seventy-five dogs received L-CHOP-LDR-HBRT and were compared with 115 controls. Median PFS and OS for the L-CHOP-LDR-HBRT group were 532 days (95% confidence interval [CI], 389-782) and 752 days (95% CI, 526-963), significantly longer than in controls (296 days [95% CI, 260-331] and 457 days [95% CI, 382-553], respectively, P < .0001). A delay in irradiating the caudal half negatively affected PFS and OS. Adverse event data were available for 30 dogs treated with L-CHOP-LDR-HBRT; most were low-grade, with grade 3 and 4 toxicities occurring in 4.5% of all administered treatments. One dog required hospitalization after HBRT. Dose reductions and treatment delays were observed in 39 and 74 instances, respectively, over the 28-week protocol in this cohort.

Combining L-CHOP with LDR-HBRT, without maintenance chemotherapy, might substantially extend PFS and OS over L-CHOP alone in dogs with multicentric lymphoma, with predominantly mild, manageable toxicities.