Conventional volumetric modulated arc therapy (VMAT) optimization relies on gradient-based methods in which analytical gradients must be explicitly derived for each objective function and machine constraint. This dedicated mathematical derivation imposes substantial barriers to extending the optimization framework with new clinical objectives or delivery constraints, and the resulting CPU-based implementations do not natively exploit modern GPU hardware.

To develop an end-to-end differentiable VMAT optimization framework using computation graphs that enable automatic differentiation-based gradient computation for both fluence map optimization (FMO) and direct aperture optimization (DAO), while incorporating VMAT machine delivery constraints within a unified, GPU-accelerated pipeline.

A two-stage pipeline-fluence map optimization (FMO) followed by DAO-was developed. FMO reformulates dose calculation as a differentiable computation graph, enabling gradient computation for arbitrary objective functions including non-smooth Heaviside step terms. For DAO, a novel analytically derived computation graph maps MLC leaf positions to delivered fluence through closed-form time-averaged bixel exposure integration with finite-width boundary corrections, ensuring exact differentiability. Machine delivery constraints-including leaf collision, speed limits, and dose rate bounds-are incorporated as differentiable penalty terms for joint end-to-end optimization. Twenty lung cancer cases from the GDP-HMM AAPM Challenge dataset (60 Gy/30 fractions) were optimized using objectives derived from clinical Eclipse reference plans and compared against MatRad and Eclipse.

The proposed method achieved PTV D98 of 58.88 ± 0.33 Gy, comparable to Eclipse (59.07 ± 0.55 Gy, p = 0.26) and significantly higher than MatRad (57.27 ± 0.48 Gy, p < 0.001). The Paddick conformity index was 0.81 ± 0.06, comparable to Eclipse (0.84 ± 0.22, p = 0.48) and superior to MatRad (0.64 ± 0.11, p < 0.001). OAR sparing closely matched Eclipse: heart mean dose 10.74 ± 4.89 Gy vs. 11.55 ± 4.76 Gy; lung mean dose 15.33 ± 3.50 Gy vs. 16.70 ± 3.29 Gy; LAD mean dose 9.55 ± 4.97 Gy vs. 9.05 ± 3.94 Gy (p = 0.18). MatRad achieved lower heart doses at the cost of degraded PTV coverage and conformity. Optimization time was significantly shorter (59.28 ± 43.95 s vs. 390.30 ± 184.90 s, p < 0.001), while delivery times were equivalent (p = 0.92).

The proposed differentiable VMAT framework demonstrated superior goal-directed optimization fidelity compared with MatRad-more reliably translating objectives into intended dosimetric outcomes-achieving plan quality approaching clinical Eclipse plans with faster optimization.

Palliative care (PC) is guideline recommended to improve symptoms and quality of life concurrent with cancer-directed therapy for patients with advanced cancer, yet utilization remains low among Medicare beneficiaries. While Medicare Advantage (MA) enrollment has surpassed traditional Medicare (TM), how PC and other health care utilization differs between MA and TM for beneficiaries with advanced cancer is unknown.

To compare PC, systemic therapy, and hospice utilization among patients with advanced cancer covered by MA vs TM and evaluate whether differences are associated with physician network composition.

This retrospective cohort study used linked Surveillance, Epidemiology, and End Results (SEER) and Medicare data for beneficiaries aged 66 years or older diagnosed with distant-stage breast, colorectal, lung, pancreatic, or prostate cancer from January 1, 2016, to December 31, 2021, who had continuous MA or TM enrollment and at least 2 months' survival. Beneficiaries were followed up from diagnosis until the earlier of death or December 31, 2021. Each beneficiary was assigned a treating oncologist; MA beneficiaries within each plan were matched 1:1 to TM beneficiaries based on treating oncologists. Data were analyzed from November 18, 2025, to July 26, 2026.

Medicare plan type.

Time to first PC, systemic therapy, and hospice utilization. Multivariable models estimated the associations between MA enrollment and outcomes before and after matching on treating oncologist, adjusting for sociodemographic and clinical characteristics.

Among 135 402 eligible beneficiaries, 32.3% were enrolled in MA and 67.7% in TM; mean (SD) age was 76.2 (6.7) years, and 53.3% were male. Medicare Advantage beneficiaries had higher 6-month cumulative incidence of PC compared with TM beneficiaries (13.3% vs 9.3%; adjusted hazard ratio [AHR], 1.39 [95% CI, 1.35-1.44]; P < .001), particularly those in a health maintenance organization (HMO) plan (AHR, 1.66 [95% CI, 1.60-1.73]; P < .001). After matching, there were no overall differences between MA and TM in PC receipt (AHR, 1.05 [95% CI, 0.98-1.12]; P = .19), and the AHR was attenuated for HMO plans vs TM (AHR, 1.16 [95% CI, 1.08-1.24]; P < .001). Medicare Advantage beneficiaries had lower systemic therapy receipt (adjusted probability difference, -5.38 percentage points [pp]; 95% CI, -6.02 to -4.74 pp; P < .001) and higher hospice enrollment (3.42 pp; 95% CI, 2.82-4.03 pp; P < .001) than TM beneficiaries; differences persisted after matching (systemic therapy: -5.70 pp [95% CI, -6.89 to -4.52 pp]; P < .001; hospice enrollment: 1.68 pp [95% CI, 0.56-2.80 pp]; P = .003).

In this retrospective cohort study of Medicare beneficiaries with advanced cancer, MA enrollment was associated with increased PC utilization, primarily attributable to differential oncologist network composition. Persistent differences in systemic therapy use and hospice enrollment after matching by treating oncologists suggest that MA plan-level features beyond physician networks and patient characteristics may be associated with these outcomes.

Panethnic categorization of Asian American individuals masks sociodemographic diversity. Whether this aggregation obscures heterogeneity in cancer-related financial toxicity remains unknown due to a lack of disaggregated data.

To evaluate associations between disaggregated Asian American identity and financial toxicity among cancer survivors compared with non-Hispanic White survivors.

This cross-sectional study analyzed National Institutes of Health All of Us Research Program data. Participants were Asian American and non-Hispanic White adults with an electronic health record-confirmed malignant neoplasm diagnosed before completing financial toxicity surveys. Surveys were completed between June 2018 and December 2024; analyses concluded July 2026.

Self-reported Asian and non-Hispanic White race; self-reported Chinese, Filipino, Indian or Pakistani, Japanese, Korean, and Vietnamese ethnicity.

Three financial toxicity domains were assessed: cost-related nonadherence, foregone medical care, and financial worry. Adjusted odds ratios (AORs) were estimated using multivariable logistic regression adjusting for clinical covariates, with sequential adjustment for socioeconomic factors and nativity.

The cohort included 55 511 survivors (median [IQR] age, 66.0 [56.0-73.0] years): 54 230 (97.7%) non-Hispanic White and 1281 (2.3%) Asian American individuals. Overall, 27 316 (49.2%) experienced at least 1 financial toxicity domain. In the clinical baseline model, Chinese survivors had lower odds of cost-related nonadherence than non-Hispanic White survivors (AOR, 0.68; 95% CI, 0.53-0.88). Vietnamese (AOR, 2.15; 95% CI, 1.30-3.56), Filipino (AOR, 1.58; 95% CI, 1.22-2.07), and Japanese (AOR, 1.44; 95% CI, 1.10-1.89) survivors had higher odds of financial worry than non-Hispanic White survivors. In aggregate, Asian American identity was associated with lower cost-related nonadherence odds (AOR, 0.79; 95% CI, 0.69-0.92) and higher financial worry odds (AOR, 1.31; 95% CI, 1.16-1.47). Worry associations attenuated after nativity adjustment.

In this cross-sectional study, disaggregation revealed heterogeneity in financial toxicity across Asian American subgroups, which was obscured by panethnic classification. Elevated financial worry may reflect immigration factors and US wealth accumulation rather than income alone. Disaggregated data are needed to effectively target financial navigation.

Chimeric antigen receptor (CAR) T-cell therapy is one of the most promising types of immunotherapies for the treatment of cancer. Unlike conventional T cells, which recognize antigens only when presented by the major histocompatibility complex (MHC), CARs use an antibody-derived binding domain to recognize native tumor-associated surface antigens directly, enabling MHC-independent target recognition. The two main cytotoxic mechanisms used by CAR T cells are the quick release of lytic granules containing granzymes and perforin to cause target cell apoptosis and the activation of death-receptor pathways (FasL/Fas or TRAIL) to cause caspase-dependent cell death. Four generations of CAR designs have been created to improve therapeutic efficacy and an advanced fifth generation is under development. Although CAR-T therapies have been very successful in treating hematological malignancies, their efficacy in solid tumors is limited because of immunosuppressive tumor microenvironments, tissue architecture, and antigen expression patterns that prevent CAR-T cell infiltration, activation, and persistence. With encouraging but early results, CAR T-cell therapy is moving forward into many clinical studies aimed at different solid tumors. Antigens including epidermal growth factor receptor (EGFR), mesothelin (MSLN), Disialoganglioside 2 (GD2), and B7-H3 were the subject of important clinical research conducted between 2021 and 2025. Emerging cellular treatments such as Chimeric Antigen Receptor Natural Killer cells (CAR-NK), CAR-macrophages and T-cell receptor (TCR)-engineered T cells offer alternate options to circumvent the limitations of CAR T-cells in solid tumors. This review aims to examine the evolution of next-generation CAR-T cell therapy for solid tumors by highlighting recent innovations, key clinical advancements, and persisting therapeutic challenges. It further seeks to analyze the biological and translational barriers limiting efficacy in solid malignancies and to explore emerging strategies and future directions that may enhance clinical outcomes.

To determine the best combination of reconstruction parameters on a spectral photon-counting CT (SPCCT) system to improve detection of peritoneal metastases (PMs) on virtual monoenergetic images (VMIs) at 40 keV.

A spectral phantom was scanned using SPCCT and dual-energy CT (EID-DECT) at 8.5 mGy (CTDI_vol). 40 keV VMIs were reconstructed on EID-DECT using a level 4 Spectral algorithm and kernel B. For SPCCT, two reconstruction kernels (STD-B and HRB) and four iDose⁴ levels (i5/i7/i9/i11) were used. Noise power spectrum (NPS) and task-based transfer function (TTF) were analyzed. Detectability indices (d') for non-mucinous and mucinous peritoneal metastases (nmPMs and mPMs) were calculated.

Noise magnitude (-22.4 ± 0.8%) and noise texture (f_av) values were lower (-21.3 ± 4.4%), but spatial resolution (f₅₀) was higher (5.8 ± 3.4%) with STD-B than with HRB at all iDose⁴ levels. For both simulated PMs, d' values were higher with STD-B than with HRB (22.3 ± 0.3%). Compared to EID-DECT, noise magnitude and d' values were similar or better with STD-B at i7/i9/i11 and i9/i11 for HRB. At all iDose⁴ levels and with both kernels, f_av or f₅₀ values were higher with SPCCT, except for f₅₀ at i11 with HRB.

Optimized reconstruction settings on SPCCT offer better image quality and lesion detectability on 40 keV VMIs than EID-DECT. Combining the STD-B kernel with level i9 achieved the best compromise between noise texture, spatial resolution, and detectability for simulated PMs.

Question Can the performance of a spectral photon-counting CT (SPCCT) improve detection of peritoneal metastases (PM on 40 keV images compared with a conventional dual-energy CT (EID-DECT) system? Findings SPCCT with optimized reconstruction settings improved image noise, spatial resolution, and lesion detectability compared with EID-DECT at 40 keV. Relevance Statement SPCCT may enhance the preoperative assessment of colorectal cancer patients by better visualizing small or low-contrast peritoneal lesions. Validation in larger clinical cohorts is required to confirm these findings and define standardized reconstruction protocols.

Well-differentiated liposarcoma (WDLPS) is a low-grade malignancy characterized by an indolent clinical course and a strong tendency for local recurrence. Unlike other sarcoma subtypes, pure WDLPS has a very low metastatic potential. However, in the course of the disease, dedifferentiation may occur and is associated with a substantially higher risk of metastases and mortality. Surgery remains the cornerstone of treatment for both primary and recurrent abdominal WDLPS, when a complete resection is possible. The role of systemic therapy in WDLPS is less clearly defined. Conventional cytotoxic chemotherapy generally has limited activity in pure WDLPS, and most clinical trials have included both WDLPS with dedifferentiated liposarcoma (DDLPS) or other liposarcoma subtypes, making the data for pure WDLPS difficult to interpret. Therefore, in clinical practice, we consider systemic treatment for patients with unresectable, symptomatic disease and in cases of repeated recurrences in which further local treatment is not possible. For selected patients with indolent and asymptomatic disease, active surveillance remains an appropriate strategy. The recent development of treatments directed against molecular alterations in WDLPS, namely CDK4 and MDM2 amplifications, have opened up a new era in the systemic treatment of this disease. Among these, anti-CDK4 agent palbociclib has shown the most mature clinical evidence to date, including recent results of overall survival data. MDM2 inhibition and combinations targeting both CDK4 and MDM2 represent promising areas which are currently under investigation. Due to the rarity of WDLPS and the lack of evidence from prospective studies, enrollment in clinical trials should be prioritized whenever possible.

The low 5-year survival rate for Oral Squamous Cell Carcinoma (OSCC) is largely attributed to the limitations of current invasive diagnostic procedures and diagnosis at a late stage. Biomarkers from saliva and exfoliated oral cells could be used to diagnose oral cancer as the malignancy begins in the epithelium and would be particularly useful for screening patients in high-risk categories like tobacco users. However, single biomarker use can be deceptive. This study investigates the candidature of onco-miRNAs (miR-21-5p, miR-31-5p, and miR-155-5p) and tumor suppressor miRNAs (miR-125-3p and miR-133a), along with their targets p53 and EGFR, in oral exfoliated cells, as noninvasive biomarkers for timely detection of oral cancer in high-risk tobacco users.

We analyzed salivary miRNAs and mRNAs in 40 individuals comprising oral cancer patients, tobacco consumers, and healthy controls using quantitative real-time PCR (qRT-PCR). This study focused on selected oncogenic and tumor-suppressor miRNAs. The miRNA target mRNAs, p53 and EGFR, were analyzed by qRT-PCR and immunocytochemistry. Diagnostic potential was assessed using receiver operating characteristic (ROC) curve analysis.

The relative gene expression levels of all studied miRNAs and their target mRNA were found deregulated among the study participants. miR-31-5p increased 23-fold in tobacco consumers (p = 0.002) and 38-fold in oral cancer patients (p = 0.002). EGFR expressions increased 12-fold in tobacco consumers (p = 0.001) and 20-fold in oral cancer patients (p = 0.0003).

The results of this pilot study suggest that combined salivary miR-31-5p and oral exfoliated-cell EGFR expression is a candidate biomarker panel for noninvasive screening in high-risk tobacco users. Given the small, single-cohort design, these preliminary, hypothesis-generating findings require validation in larger, independent cohorts before clinical application.

The present study aimed to evaluate changes in QoL and psychological symptoms (anxiety and depression) in women with BC between the first days (T0) and 1 month after surgery (T1). In addition, the study examined longitudinal associations of perceived pain, upper limb disability, and disease acceptance with QoL and psychological symptoms across T0 and T1.

A prospective, observational cohort study was conducted on 101 women with BC. At T0 and T1, the EuroQol Five-Dimension Five-Level, the Hospital Anxiety and Depression Scale, the Numeric Rating Scale, the Quick Disability of the Arm, Shoulder and Hand, and an adapted version of the Chronic Pain Acceptance Questionnaire were administered. Data were analyzed using paired-samples t-tests and linear mixed models (LMMs).

The results showed increased QoL and decreased psychological symptoms between T0 and T1. The LMMs revealed that assessment time was not significantly associated with QoL and psychological symptoms once perceived pain, upper limb disability, and disease acceptance were considered. Notably, perceived pain and upper limb disability were negatively associated with QoL. In addition, upper limb disability was positively associated with depressive symptoms. Lastly, disease acceptance was negatively associated with both anxiety and depressive symptoms.

Improvements in QoL and psychological symptoms may occur within the first month after BC surgery. At equivalent levels of perceived pain, upper limb disability, and disease acceptance, the timing of assessment seems to provide no additional information about QoL and psychological symptoms. The findings suggest the importance of timely, multidisciplinary rehabilitation care for women with BC during the early post-operative period.

To critically review the contemporary evidence on single-port robot-assisted partial nephrectomy (SP-RAPN), focusing on comparative outcomes, patient selection, anatomy-based surgical approaches, technical implementation, and its current role relative to multi-port RAPN (MP-RAPN).

Available comparative studies, multicenter registries, and recent meta-analyses indicate that SP-RAPN achieves perioperative, functional, and early oncological outcomes comparable to MP-RAPN in appropriately selected patients. Its principal added value appears to lie not in universal platform superiority, but in facilitating anatomy-adapted strategies, particularly retroperitoneal and low anterior access approaches in posterior tumors, confined operative spaces, and selected patients with hostile abdomen. However, current evidence remains predominantly retrospective and derives largely from high-volume expert centers. SP-RAPN should be considered a complementary robotic platform. Its greatest clinical value emerges when platform-specific engineering characteristics are matched to patient anatomy, surgical access, institutional experience, and surgeon expertise. Future studies should validate objective selection criteria and assess long-term, patient-reported, and efficiency outcomes.

To develop and internally evaluate an interpretable machine learning model for distinguishing current gastric cancer (GC) from intestinal metaplasia (IM) among patients with psychological symptoms.

This retrospective, single-center, cross-sectional study included 302 patients with psychological symptoms, comprising 185 with IM and 117 with GC. Patients were randomly divided into a training cohort (n = 212) and an independently retained validation cohort (n = 90). Candidate features were evaluated using LASSO, Boruta, and recursive feature elimination. Seven machine learning algorithms were compared using nested cross-validation in the training cohort. The final model was assessed in the validation cohort using discrimination, calibration, decision curve analysis, and Shapley Additive Explanations (SHAP).

Seven features were retained: age, albumin, sex, psychological symptom type, total bilirubin, smoking status, and monocyte count. XGBoost achieved the highest mean outer-fold AUC (0.839 ± 0.029) and was selected as the final model. In the validation cohort, XGBoost achieved an AUC of 0.793 (95% CI 0.670-0.895), accuracy of 0.800, sensitivity of 0.657, specificity of 0.891, and F1 score of 0.719. The Brier score was 0.167. Decision curve analysis suggested potential clinical net benefit. SHAP analysis identified age and psychological symptom type as the leading contributors to classification.

The interpretable XGBoost model showed moderate discrimination for distinguishing current GC from IM and may provide supplementary information for clinical assessment. It should not replace endoscopic or histopathological diagnosis, and external multicenter validation is required before broader clinical implementation.