Radiation-induced inflammatory responses are known to contribute to lung injury, but how these responses occur under ultra-high dose rate FLASH irradiation (FLASH) remains poorly characterized. The present study aimed to compare inflammation-mediated cancer progression in FLASH versus conventional (CONV) irradiation-induced lung injury in mice. Using a modified Varian 23CX clinical accelerator to deliver electron beam FLASH, we performed whole thoracic irradiation in healthy C57BL/6J mice with both FLASH and CONV modalities. While both triggered similar acute inflammatory responses, FLASH resulted in significantly less pulmonary fibrosis at 12 weeks post-irradiation. Acute radiation-induced lung inflammation promoted cancer progression in both groups, with neutrophil recruitment contributing to tumor cell metastasis. During chronic inflammation, however, FLASH led to fewer metastatic colonies than CONV. Further analysis revealed that FLASH accelerated macrophage polarization toward the M2 phenotype during chronic inflammation, whereas CONV promoted M1 polarization. Importantly, FLASH maintained tumor control comparable to CONV while markedly reducing normal tissue toxicity. These results demonstrate that radiation-induced inflammation facilitates cancer progression and suggest that differential macrophage polarization may help explain the distinct lung responses to CONV versus FLASH irradiation. Further studies are needed to elucidate the tissue changes that occur prior to tumor cell seeding in irradiated lungs.

PurposeThis study aimed to quantify the statistical cure in patients with unresectable locally advanced esophageal squamous cell carcinoma (LA-ESCC) treated with definitive radiotherapy (RT)-based strategies and to explore the independent prognostic factors of cure.Methods and MaterialsThis retrospective study analyzed 801 patients with unresectable LA-ESCC at Affiliated Cancer Hospital of Zhengzhou University from 2014 to 2023. All patients received definitive RT and were stratified into chemoradiotherapy (CRT, n=689) or CRT combined with immunity checkpoint inhibitors (CRT+ICIs, n=112) based on treatment received. A relative survival (RS)-based mixture cure modeling methodology was used to estimate the cure fraction and cure point. The model was externally validated using a dataset of 5,000 matched patients from the SEER database.ResultsOur model estimated a cure fraction of 10.4% and a cure point of 6.7 years. Subset analysis indicated that patients treated with CRT+ICIs had a higher cure fraction than those treated with CRT (30.6% vs 10.9%), with a shorter time to cure (3.9 vs 7.1 years). Validation with the SEER dataset showed a comparable cure fraction (10.8%) but a longer cure point (9.3 years). Multivariable analysis suggested that the favorable independent prognostic factors of statistical cure included a higher BMI (β, 0.10; 95% CI, 0.01 to 0.19; p=0.039) and the CRT+ICIs regimen (β, 2.18; 95% CI, 0.91 to 3.44; p<0.001). Factors associated with a lower probability of cure were age ≥65 years (β -0.04; 95% CI, -0.09 to 0.00; p=0.032) and chronic comorbidities (β, -1.10; 95% CI, -2.14 to -0.05; p=0.040).ConclusionsStatistical cure is achievable in unresectable LA-ESCC patients receiving RT-based regimens. However, the traditional 5-year overall survival (OS) insufficiently reflects long-term survival, indicating that follow-up in the CRT group should be at least 7 years. Incorporation of ICIs facilitated curative potential and shortened cure point, supporting a 4-year OS surrogate endpoint for CRT+ICIs. These findings emphasize the value of integrating cure models into clinical practice to optimize individualized treatment and surveillance strategies.

PURPOSE:Anatomy image-driven lung function imaging methods have been explored for thoracic radiotherapy, but most contrast-free approaches rely on unimodal surrogates. This study aimed to develop a multimodal contrast-free pulmonary perfusion reconstruction framework (MCF-Q) that integrates computed tomography (CT) and magnetic resonance imaging (MRI) to leverage complementary anatomical and functional information from routinely acquired CT and non-contrast MRI, improve agreement with single-photon emission computed tomography perfusion (SPECT-Q), and explore its potential to support functional lung avoidance radiotherapy (FLART). METHODS AND MATERIALS:This prospective analysis included 21 patients with lung cancer who underwent pulmonary SPECT-Q, CT, and 1H MRI. MCF-Q adopted a dual-branch deep learning architecture to extract complementary features from CT and MRI and fuse them into pulmonary perfusion maps. Seven-fold cross-validation was performed to evaluate voxel-wise and function-wise agreement between MCF-Q and SPECT-Q, including Spearman's correlation coefficient (R) and the Dice similarity coefficient (DSC). The dosimetric analysis was also conducted by comparing a conventional radiotherapy (ConvRT) plan with FLART plans guided by different perfusion maps. RESULTS:For voxel-wise assessment, the MCF-Q achieved an R value of 0.7831 ± 0.0821. For function-wise similarity, the MCF-Q gained the DSC value of 0.8396 ± 0.0379 in high-functional regions, and 0.7680 ± 0.0555 in low-functional regions. All metrics calculated from MCF-Q showed significant improvement over single-modality-based lung function imaging methods. In dosimetric performance, the MCF-Q-guided FLART achieved better dose sparing in high-functional regions, while maintaining comparable whole-lung and organ-at-risk dose metrics. CONCLUSIONS:In this study, the proposed MCF-Q demonstrated the feasibility of multimodal perfusion reconstruction from CT and MRI, with improved agreement with SPECT-Q, and provided radiotherapy-planning-relevant functional information that may facilitate functional lung avoidance strategies. These findings support the value of integrating routinely acquired CT and MRI for contrast-free, planning-relevant perfusion estimation, warranting validation in larger cohorts.

Purpose Recent studies indicate that lung function can change significantly during radiation therapy (RT) course. However, additional function imaging scans are not part of routine RT workflow, limiting the use of functional avoidance for adaptive therapy. To bridge this gap, we aimed to develop a deep learning model to synthesize functional maps directly from fractional cone-beam computed tomography (CBCT) images, enabling potential functional image-guided adaptive radiotherapy (FIGART). Methods Data were prospectively collected from 60 lung cancer patients who underwent intensity-modulated radiation therapy. In addition to standard planning CT and fractional cone-beam computed tomography (CBCT) scans, all patients received a baseline single-photon emission computed tomography (SPECT) perfusion scan before radiotherapy. A subset of 16 patients also underwent a follow-up SPECT scan after completing RT. A three-dimensional Gated-Attention U-Net (GAU-Net) was developed to synthesize perfusion maps directly from CBCT images. To address the challenges of inherent CBCT noise and artifacts, the network architecture was augmented by integrating gated attention modules and 3D deformable convolutions within the skip-connection pathways. This design enhances multi-scale feature fusion for more robust image synthesis. The synthesized perfusion images were quantitatively compared to the reference SPECT scans using cross validation. Voxel-wise agreement was assessed using the Spearman correlation coefficient (R), structural similarity index (SSIM) and mean squared error (MSE), while functional region agreement was evaluated using the Dice similarity coefficient (DSC). The potential clinical benefit was assessed through dosimetric evaluation. Results Quantitative analysis demonstrated CBCT-based functional images strong agreement with ground-truth SPECT, with a R of 0.68±0.09, SSIM of 0.75±0.09, MAE of 0.15±0.03, MSE of 0.04±0.02, and DSC values of 0.76±0.07 and 0.85±0.05 for high- and low-functional regions, respectively. Regarding dose sparing of the lung high-function region, CBCT-based functional imaged guided plan significantly reduced the mean dose by 6.97±4.22 Gy and V20 by 13.54±8.66% compared to anatomical plan, while achieving a better target dose homogeneity index of 5.26±0.65. Conclusion We developed a CBCT-based lung function imaging method using a deep learning model. The evaluation demonstrated its feasibility for functional guidance planning in FIGART. A larger cohort study is warranted in the future.

PURPOSE:Contrast-enhanced magnetic resonance image (MRI) imaging via administration of contrast agents is critical for diagnosis, staging, and treatment of nasopharyngeal carcinoma (NPC). However, gadolinium-based contrast agents can lead to severe adverse effects, especially in patients with compromised kidney function, necessitating a safer alternative for contrast enhancement. In this study, we aim to develop and assess the clinical feasibility of a federated learning model for synthesizing virtual contrast-enhanced MRI (VCE-MRI) images from contrast-free scans for patients with NPC. METHODS AND MATERIALS:In this multicenter, retrospective study, we developed and clinically evaluated a federated learning-based VCE-MRI synthesis model (FL-VCE-MRI) using pretreatment contrast-free T1-weighted and T2-weighted MRI scans. The model was trained using data from 14 centers involving 2061 patients. External validation was performed with an independent dataset from 25 centers comprising 126 patients. Additionally, 10 clinicians from 8 centers assessed the quality of the synthetic images. RESULTS:The FL-VCE-MRI model demonstrated high generalizability in the external validation dataset, achieving a mean absolute error of 45.85 (95% CI, 43.73-47.96). For centers facing challenges in developing well-performing single-center models, the FL-VCE-MRI improved the average mean absolute error from 53.93 (95% CI, 49.21-58.65) to 45.93 (95% CI, 41.64-50.22). Clinical evaluations indicated that the synthesized VCE-MRI images are both reliable and clinically valuable, with no significant differences compared with gadolinium-enhanced MRI in disease diagnosis, tumor staging, and delineation. CONCLUSIONS:The FL-VCE-MRI model shows potential as a noninvasive alternative to gadolinium-enhanced MRI for patients with NPC.

This study aimed to assess the dosimetric benefits of functional avoidance proton therapy planning using anatomy-wise computed tomography (CT)-derived lung ventilation imaging (VIaw), which helps spare both high-functioning lung volume (HFV) and recoverable low-functioning lung volume (rLFV) for the first time in proton therapy. In a cohort of 18 lung cancer patients, VIaw was generated from planning CT scans. For each patient, we created anatomical and functional-guided intensity-modulated radiation therapy (aIMRT, fIMRT) and proton therapy (aIMPT, fIMPT) plans. For a subset of eight patients with tracheal obstruction by the tumor, we generated two extra rLFV-sparing plans (rfIMRT, rfIMPT) that specifically incorporated constraints to spare both HFV and rLFV. For the 18 patients, functional IMPT (fIMPT) demonstrated superior HFV sparing, achieving statistically significant reductions in all HFV dose parameters compared to fIMRT (V5: 34.9

Background:Brain metastases (BM) significantly affect both prognosis and quality of life in patients with metastatic non-small cell lung cancer (NSCLC), highlighting the need for innovative therapeutic strategies. In this systematic review, we evaluated the prevalence and clinical significance of key genomic alterations in BM from NSCLC. Methods:Comprehensive searches of PubMed, Web of Science, Embase, and the Cochrane Library identified 19 studies encompassing 3028 patients, 930 of whom had BM. Results:Among BM patients, 50% harbored TP53 mutations, while 49% exhibited PD-L1 overexpression, and 30% demonstrated a tumor mutational burden (TMB) ≥10 mut/Mb. These alterations were associated with tumor progression, immune evasion, and increased mutational load. Additional actionable biomarkers, though less frequent, included ERBB2, BRAF, and NTRK alterations. Importantly, our findings revealed substantial intertumor heterogeneity between primary tumors and BM, underscoring the need of site-specific genomic profiling to inform treatment decisions. Conclusions:Overall, this study underscores the critical role of genomic profiling in advancing precision medicine approaches and supports the integration of novel targeted and immunotherapeutic strategies to improve outcomes in NSCLC patients with brain metastases. The dynamic evolution of biomarker expression during disease progression and tumor heterogeneity remains major challenges to the development of durable treatment strategies.

Emerging evidence highlights the critical role of metabolic pathways in breast cancer (BC) progression. Here, we developed a butyrate metabolism-specific gene (BMRG) signature to predict clinical outcomes and immunotherapy responses in BC, providing a novel pathway-focused prognostic tool. Using data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO), we identified 102 butyrate metabolism-related differentially expressed genes (DEGs) through the intersection of DEGs, WGCNA-derived key module genes, and BMRGs. Univariate Cox followed by least absolute shrinkage and selection operator (LASSO) analysis identified nine genes to construct a prognostic signature, which served as an independent prognostic factor. Risk stratification revealed distinct immune microenvironment and mutation landscapes between subgroups, with risk scores strongly correlating with immune checkpoint expression. The signature exhibited robust prognostic performance, with AUC values for 3-, 5-, and 7-year overall survival ranging from 0.65-0.69 in TCGA and 0.57-0.77 in independent GEO cohorts. Protein-protein interaction analysis identified ACSL1 as a key hub gene, and functional validation confirmed that ACSL1 knockdown suppressed BC cell proliferation and migration. Our findings establish this novel nine-gene butyrate metabolism-specific signature as a promising prognostic biomarker and potential therapeutic target for BC, providing a metabolism-focused perspective for personalized BC management.

Images often suffer from intensity inhomogeneity and noise during the imaging process, which poses considerable challenges for image segmentation. This paper proposes a novel adaptive structure variational model integrating Retinex theory for image segmentation, which effectively addresses intensity inhomogeneity and noise in the segmentation process. Grounded in Retinex theory, we decompose the image and impose two constraints: a piecewise-constant constraint on the reflectance to delineate homogeneous regions amid inhomogeneity, and a spatial smoothness constraint on the illumination to model the bias field. The primary novelty of the work lies in the introduction of an adaptive weighted matrix, comprising a rotation matrix and a scaling matrix, coupled with the gradient operator. This design enables the proposed model to perform anisotropic regularization, allowing it to adaptively capture directional structural features, preserve complex boundaries and textures, and simultaneously suppress noise effectively. We establish the existence of a solution for the proposed model and employ an efficient alternating minimization algorithm for numerical solution. Numerical experiments on synthetic, natural, and medical images demonstrate the desirable performance of the proposed model.

Neddylation contributes to cancer progression, but its relationship with proteasome-related genes and the immune microenvironment in laryngeal squamous cell carcinoma (LSCC) remains unclear. We aimed to identify neddylation-associated genes in LSCC and characterize their molecular, metabolic, and immune features. Transcriptomic data from The Cancer Genome Atlas and Gene Expression Omnibus were analyzed using differential expression, functional enrichment, and machine-learning approaches. Key genes were further evaluated by survival, pathway, immune infiltration, and somatic mutation analyses and validated by quantitative real-time PCR (qRT-PCR) in five paired LSCC and adjacent tissues. Seventy-four differentially expressed neddylation-related genes were identified. Four genes, KCTD6, PSMB2, PSMB4, and PSMD2, were selected by both LASSO and SVM-RFE. KCTD6-associated genes were enriched in bile acid metabolism, while macrophage M0 and resting CD4 + memory T cells were associated with prognosis. Lower PSMB4 and PSMD2 expression was associated with immune escape-related features. qRT-PCR showed concordant expression patterns, with KCTD6 downregulated and PSMB2, PSMB4, and PSMD2 upregulated in tumors; only PSMB2 reached statistical significance. We identified a distinct set of proteasome-related neddylation-associated genes associated with metabolic and immune characteristics in LSCC. These findings suggest potential links between neddylation-associated proteasome genes and the immune microenvironment in LSCC.

e16139 Background: The prognosis for patients with locally advanced esophageal squamous cell carcinoma (LA-ESCC) remains poor. While immune checkpoint inhibitors have shown efficacy in advanced disease, their synergistic potential with concurrent chemoradiotherapy (CRT) in the neoadjuvant setting for LA-ESCC requires prospective validation. This study aimed to prospectively evaluate the efficacy and safety of the PD-L1 inhibitor adebrelimab combined with concurrent CRT as a novel neoadjuvant strategy for resectable LA-ESCC. Methods: This was a prospective, single-arm trial. Eligible patients had resectable thoracic ESCC, clinically staged as cT2N+ or cT3-4a any N M0. The regimen comprised adebrelimab (1200 mg intravenously every 3 weeks for 2 cycles) plus concurrent CRT (weekly paclitaxel/carboplatin and radiotherapy totaling 41.4 Gy in 23 fractions). Radical esophagectomy was scheduled 4-8 weeks after neoadjuvant therapy completion. The primary endpoint was the pathological complete response (pCR; ypT0N0) rate. Key secondary endpoints included the major pathological response (MPR; ≤10% residual viable tumor) rate, R0 resection rate, and safety. Results: Between June 2024 and October 2025, 32 patients were enrolled (median age 64.6 years; 75.0% male; 75.0% stage III). Of these, 24 patients proceeded to surgery, achieving a 100% (24/24) R0 resection rate. The primary endpoint was met, with a pCR rate of 58.3% (14/24). The MPR rate was 91.7% (22/24). Pathological downstaging was profound: 87.5% (21/24) of patients achieved primary tumor (T-stage) downstaging (ypT < cT), and 66.7% (16/24) achieved nodal clearance (ypN0). Regarding safety, grade ≥3 leukopenia occurred in 43.8% (14/32) of patients, all of whom had concomitant grade ≥3 lymphopenia. Common non-hematological toxicities included nausea (58.1%), constipation (54.8%), decreased appetite (48.4%), and vomiting (41.9%), which were all grade 1-2 and manageable with supportive care. Conclusions: Neoadjuvant adebrelimab plus concurrent CRT demonstrated remarkable efficacy in LA-ESCC, yielding high rates of pCR (58.3%) and MPR (91.7%), alongside significant pathological downstaging. While hematologic toxicity requires vigilant management, the profound tumor regression achieved offers a promising new strategy for improving outcomes in this population. These compelling results warrant further validation in a phase III randomized controlled trial. Future studies should explore optimization of the chemotherapy component to potentially enhance tolerability. Clinical trial information: ChiCTR2400084445.

Objectives To evaluate the impact of tumor volume segmentation variability on the repeatability of radiomic features (RFs) and to determine how RF repeatability influences the generalizability of radiomic models for predicting overall survival (OS) in patients with oropharyngeal carcinoma (OPC).Methods We retrospectively analyzed CT images from 1017 patients with oropharyngeal carcinoma across three institutions. Perturbation methods were applied to simulate variations in gross tumor volume segmentation. RFs were extracted from both the original images and Laplacian of Gaussian-filtered images using different perturbation masks. RF repeatability was quantified using intra-class correlation coefficients (ICC). Repeatable RFs were progressively incorporated into the modeling process according to different ICC thresholds to assess the influence of feature repeatability on model generalizability.Results Incorporation of RFs with ICC values between 0.7 and 0.8 improved the AUC index of the two-year and three-year OS models in external validation cohorts. Using an ICC threshold of 0.7, RFs were classified into high- and low-repeatability groups, and OS models were trained and validated using the training, internal testing, and external validation cohorts. Across all cohorts, the OS model trained with high-repeatability RFs demonstrated significantly superior performance compared to the model trained with low-repeatability RFs.Conclusion The findings demonstrate that selecting RFs with ICC values greater than 0.7 substantially enhances both the generalizability and predictive performance of CT-based radiomic models for patients with OPC. This study further underscores the importance of considering RF repeatability, particularly in the presence of tumor volume segmentation variability, to improve the robustness and clinical reliability of radiomic models.

Lung cancer remains the leading cause of cancer-related mortality worldwide, with poor prognosis in advanced stage diseases. Although early diagnosis has the potential to improve patient outcomes, current diagnostic methods remain suboptimal, highlighting the need for accurate molecular biomarkers. We systematically reviewed 49 studies to evaluate the diagnostic performance of SHOX2 methylation, RASSF1A methylation, and their combined panel for lung cancer detection. The combined SHOX2/RASSF1A methylation panel demonstrated a pooled sensitivity of 77.8% (95% CI: 72.3%–82.5%) and specificity of 89.0% (95% CI: 86.6%–91.1%), with an HSROC area under the curve (AUC) of 0.916. SHOX2 methylation alone yielded a sensitivity of 69.4% and specificity of 91.7%, whereas RASSF1A methylation showed lower sensitivity (45.7%) but the highest specificity (93.8%). Pairwise comparisons demonstrated that the combined panel significantly improved sensitivity compared with either SHOX2 or RASSF1A alone while maintaining specificity comparable to SHOX2, although lower than that of RASSF1A. Subgroup analyses showed that assay method and pathological subtype contributed to differences in pooled sensitivity, whereas leave-one-out sensitivity analyses confirmed the robustness of the pooled estimates. In conclusion, the combined SHOX2/RASSF1A methylation panel provides a more balanced diagnostic performance than either biomarker alone and represents a promising adjunctive approach for lung cancer detection. Future studies should focus on standardizing detection methods, integrating these biomarkers with other diagnostic modalities, and evaluating their diagnostic performance in early-stage lung cancer to further enhance their clinical utility.

Background:The efficacy of immunochemotherapy (ICT) remains poor in patients with polymetastases from esophageal squamous cell carcinoma (ESCC). While radiation therapy (RT) has shown promise in oligometastatic settings, its role when combined with ICT for polymetastatic ESCC (with >5 metastatic lesions) remains unclear. Objective:This study evaluated the efficacy and value of RT in patients with polymetastatic ESCC who received ICT as first-line treatment. Design:This multi-center cohort study was conducted at 20 hospitals in China. Methods:In total, 331 patients who received at least one cycle of first-line ICT between January 2019 and December 2021 were enrolled. Among them, 88 received ICT plus RT (RT group), and 243 received ICT alone (non-RT group). Propensity score matching (PSM) was performed to control for potential confounders (75 patients/group). Outcomes included overall survival (OS), progression-free survival, objective response rate (ORR), symptom control, and safety. This study was registered at the Clinicaltrials.gov registry (identification number NCT05142709). Results:Both before and after PSM, no significant OS benefit was observed with RT group (median OS: 15.2 vs 12.2 months, hazard ratio (HR) 0.80 (0.60-1.07), p = 0.14; 15.0 vs 11.0 months, HR 0.80 (0.55-1.15), p = 0.23, respectively), though pre-PSM ORR favored RT (59.1% vs 40.3%, p = 0.003). RT demonstrated superior symptom control, with significantly higher rates of dysphagia improvement (63.3% vs 36.4%, p = 0.0006) and meaningful pain reduction (59.4% vs 40.0%, p = 0.007). Grade ⩾3 treatment-related adverse events were comparable between groups (38 vs 40 cases post-PSM, p = 0.74), with equivalent grade 5 toxicities (1.3% each). Conclusion:For polymetastatic ESCC, RT combined with ICT enhanced symptom control without severe toxicity, though it did not improve survival. This supports its personalized use for quality of life in symptomatic patients.

Objective.This study aims to develop a multi-modality-guided dose prediction (MMDP)-based auto-planning algorithm for functional lung avoidance radiotherapy (FLART) guided by voxel-wise lung function images.Approach.The proposed auto-planning algorithm consists of a novel MMDP model and a function-guided dose mimicking algorithm. The MMDP model features extracting complementary features from multi-modality images for predicting dose distributions close to FLART plans. An instance-weighting anatomy-to-function training strategy is tailored to enhance prediction accuracy. A function-guided voxel-wise dose mimicking algorithm is developed to convert predicted dose into FLART (MMDP-FLART) plans. We retrospectively collected data from 163 lung cancer patients across three institutions, comprising 114/28 cases for training/validation and 21 cases with SPECT ventilation (V) images for testing. Furthermore, we prospectively collected 33 cases with SPECT perfusion (Q) images for evaluation. MMDP-FLART plans were compared against conventional radiotherapy (ConvRT) and FLART plans manually created by senior clinicians.Main results.MMDP achieved accurate dose predictions, with median prediction errors for all assessed dose-volume histogram (DVH) metrics within ±1 Gy/±1%. The MMDP model reduced prediction absolute errors for functionally weighted mean lung dose (fMLD) by 12.77% compared to an anatomy-guided dose prediction model and the instance-weighting anatomy-to-function training strategy reduced prediction absolute errors for fMLD by 22.64%. Compared to manual ConvRT plans, MMDP-FLART plans effectively reduced fMLD by 0.80 Gy (11.9%,p< 0.01) and 0.46 Gy (6.0%,p< 0.01) on SPECT V and Q datasets respectively. Compared to manual FLART plans, MMDP-FLART plans exhibited lower and comparable fMLD on SPECT V and Q datasets respectively with lower dose to heart and esophagus.Significance. The MMDP model with instance-weighting anatomy-to-function training can achieve accurate dose prediction for FLART. The MMDP-based auto-planning algorithm can produce FLART plans leveraging voxel-wise lung function information from V/Q images. It shows promise in promoting FLART planning efficiency, consistency, and quality.