Technology in Cancer Research & Treatment最新文献

BackgroundDeep generative models can improve the generalization of deep learning in medical imaging by enriching limited training data with diverse, realistic synthetic images.PurposeTo assess whether Denoising Diffusion Probabilistic Models (DDPM) generated synthetic MRI, with and without mutual information (MI) regularization, enhances brain tumor classification across heterogeneous datasets.Study TypeRetrospective.PopulationA total of 559 patients with low and high grade brain tumors (LGG, HGG) were included from two datasets: public dataset (BraTS, n = 335) and clinical dataset (TASMC, n = 224), used exclusively to evaluate model generalization.Field Strength/Sequence1.5 T/3.0T-MR / T1WI, T1WI + C, T2WI, and FLAIR images.AssessmentDDPM models were trained to generate synthetic MR images of low grade glioma (LGG) and high grade glioma (HGG), with a variant incorporating MI. Image quality was assessed using Pearson-correlation, Frechet-Inception-Distance (FID) and Inception-Score (IS). For classification purposes. For classification, a 2D ResNet-152 was trained under four setups: (1) real images (baseline), (2) +augmentation, (3) +DDPM, and (4) +DDPM + MI. Performance was assessed by accuracy and F1-score. Robustness was tested through cross-dataset evaluation using a 5-fold ensemble.ResultsThe DDPM models, with and without MI, generated high-quality synthetic images, achieving FID = 31.47, 45.00, and IS = 1.50, 1.25, respectively. Lower FID and higher IS indicate enhanced realism and diversity, suggesting that MI improved both the quality and variability of the generated images. Cross-dataset evaluation demonstrated that DDPMs with MI achieved superior generalization performance in brain tumor classification task, with accuracies of 0.89 and 0.85 for BraTS-to-TAMSC and TAMSC-to-BraTS evaluations, respectively. These results outperform the baseline model (0.87, 0.80), traditional data augmentation (0.85, 0.78), and the standard DDPM without MI (0.82, 0.83).Data ConclusionDDPM + MI with ensemble learning significantly improves brain tumor generalization across diverse datasets, consistently outperforming baseline, traditional augmentation, and standard DDPM. This combination offers a robust solution for cross-institutional clinical applications.

IntroductionOwing to the limitation in the field size of the magnetic resonance (MR)-Linac, currently, tumors with a length of >20 cm cannot be treated. Thus, the present study aimed to develop an expanded magnetic resonance imaging-guided adaptive radiotherapy (MRIgART) workflow for long, continuous planning target volumes (PTVs).MethodsThe PTVs were divided into two sub_target volumes (PTV_sub1 and PTV_sub2). We established two isocenters and defined a field overlap region. By adjusting the MR scan range, devising the online and offline adaptive procedures, synchronizing the online adaptive processes, and constructing a pretreatment dose evaluation, a new MRIgART workflow for long PTVs was established. The new workflow was validated using an in-house-made MR phantom. Additionally, the ArcherQA Monte Carlo-based method, ArcCHECK phantom, and ionization chamber measurement method were used for dose verification.ResultsTwo clinical scenarios were established: (1) both PTV_sub1 and PTV_sub2 followed the adapt-to-position (ATP) workflow, and (2) PTV_sub1 followed the adapt-to-shape (ATS) workflow, whereas PTV_sub2 followed the ATP workflow. The feasibility of the proposed MRIgART workflow for long, continuous PTVs was demonstrated through three independent rounds of testing and validation for each scenario. When field overlaps were utilized, the PTV length that can be treated is 40 cm minus the length of field overlap region. The average gamma pass rates for the PTV_sub1 and PTV_sub2 adaptive plans were 95.74% and 98.63%, respectively (ArcherQA vs TPS). For the field overlap region, the average gamma pass rate was 95.50% (ArcCHECK vs TPS). The difference between the ionization chamber measurements and calculated results was smaller than 2%.ConclusionThis study demonstrated the feasibility, safety, and accuracy of the MRIgART workflow for long PTVs. This workflow provides an effective solution for expanding the application of MRIgART to patients with long, continuous PTVs.

Salvage treatment for locally recurrent prostate cancer after primary radiotherapy remains a clinical challenge, with multiple modalities- including stereotactic body radiotherapy (SBRT), high-dose-rate (HDR) brachytherapy, and low-dose-rate (LDR) brachytherapy-competing for optimal use. The recent UroGEC expert review in Radiotherapy & Oncology provides a timely synthesis of available evidence and underscores the potential role of brachytherapy in this setting. Here, we contextualize these findings with recently published meta-analyses that expand the evidence base and refine our understanding of salvage outcomes. Updated analyses highlight significant differences across modalities: HDR brachytherapy achieves favorable disease control with low gastrointestinal toxicity, whereas LDR appears to offer superior relapse- free survival in selected subgroups at the cost of higher late genitourinary morbidity. By contrast, SBRT, although attractive for its non-invasiveness, demonstrates lower long-term relapse-free survival when scrutinized in broader pooled cohorts, despite acceptable toxicity. Collectively, these findings emphasize that the "one-size-fits-all" paradigm is inadequate. Clinical decision-making must instead be individualized, integrating oncologic efficacy, toxicity risks, patient comorbidities, and personal preferences. Looking forward, prospective trials and harmonized outcome reporting will be essential to strengthen the comparative evidence. Until then, a nuanced, patient-centered approach-anchored in multidisciplinary discussion-remains the cornerstone of salvage treatment planning. This perspective complements and extends the UroGEC review, underscoring the need to balance efficacy with quality of life in managing radio- recurrent prostate cancer.

BackgroundDiffuse large B-cell lymphoma (DLBCL) is the most common form of non-Hodgkin-lymphoma. Although it can be cured in many patients, a significant proportion of patients fail the primary treatment and require second-line treatment. Currently, only limited data on real-world outcomes with standard therapies in Austrian patients with DLBCL are available, and while novel therapies are emerging, no historical benchmarks have been established to serve as a reference for these novel treatments.MethodsWe performed a retrospective, single-center analysis of patients with DLBCL diagnosed between 2010 and 2018 who had been treated with standard therapies. To establish efficacy benchmarks for novel therapies, we applied both clinical-trial and real-world-derived criteria to analyze the outcomes of patients potentially eligible for novel or future treatments.ResultsAlthough many patients can be cured with frontline therapy, outcomes are poor, especially in high-risk patients. Patients failing frontline therapy, especially those fulfilling the chimeric antigen-receptor (CAR) T-cell eligibility criteria, had dismal outcomes, and very few patients achieved long-term remission. Our data provide benchmark outcomes for patients eligible for novel treatments such as antibody-drug-conjugate (ADC) or CAR T-cell therapy-based treatments for potential future comparative analyses.ConclusionsPatients with DLBCL treated in Austria showed comparable outcomes to those reported in other real-world studies. Overall, standard chemotherapy-based approaches provide unsatisfactory outcomes in high-risk patients and patients in whom frontline therapy fails. Because many patients are now eligible for alternative first- and second-line treatments, such as ADC-based or CAR T-cell therapy, our efficacy benchmarks can serve for the future evaluation of these therapies in the Austrian healthcare environment.

IntroductionDaily anatomical variations in prostate cancer radiotherapy, particularly due to pelvic organ motion and filling, can compromise target coverage and increase exposure to organs at risk (OARs). Conventional image-guided radiotherapy (IGRT) uses fixed safety margins and daily couch corrections to account for these variations, potentially leading to overtreatment of healthy tissue or insufficient tumor coverage. Online adaptive radiotherapy (oART), based on cone-beam computed tomography (CBCT), enables daily plan adaptation to the patient's anatomy, offering improved precision, enhanced target coverage, and better OAR sparing. This retrospective study compares oART to conventional IGRT in prostate cancer treatment.MethodsA total of 153 treatment fractions from six consecutive prostate cancer patients treated with oART on a Varian Ethos system were analyzed. For each fraction, three plans were evaluated: the scheduled plan (initial plan recalculated on daily CBCT), the adapted plan (reoptimized based on daily anatomy), and the verification plan (applied dose recalculated on a post-adaptation CBCT). Dose-volume metrics for target volumes and OARs were assessed, and clinical acceptability was evaluated. Interfractional prostate volume changes and treatment times were examined.ResultsCTV D_98% improved significantly with adaptation (median 97.85% to 98.55%; p < 0.01) and further increased in the verification plan (98.8%; p < 0.01), alongside reduced interquartile ranges. PTV D_98% rose from 90.1% to 97.1% with adaptation and to 96.9% after verification (p < 0.01). Bowel and bladder doses showed dosimetrical advantage. Clinically acceptable plans increased from 24.8% (scheduled) to 98% (adapted) and 85.6% (verification). Scheduled plans were not used clinically. Median prostate volume remained stable despite inter-individual variation. oART required about twice the treatment time of IGRT.ConclusionAlthough more time-consuming, oART improved target dose coverage and optimized OAR sparing, while simultaneously reducing dose variability for both the target and some OARs compared to IGRT. The plan acceptability improved significantly.

IntroductionAxillary nodal burden reflects the biological aggressiveness and prognostic behavior of breast cancer. This study aimed to develop a subregional habitat radiomics model based on multiparametric magnetic resonance imaging (MRI) and to evaluate its performance in predicting high axillary nodal burden in patients with breast cancer.MethodsIn this retrospective study, a total of 221 patients who underwent axillary lymph node dissection were categorized as having limited (0-2 metastatic nodes) or high (≥3 metastatic nodes) nodal burden based on pathological findings. Morphological MRI features were visually evaluated by experienced radiologists. A clinical model was established using univariate and multivariate logistic regression analyses. Conventional radiomics (C-radiomics) and habitat radiomics features were extracted from the whole tumor and its subregions, respectively, based on multiparametric MRI. The clinical, C-radiomics, and habitat radiomics models were then integrated into a comprehensive nomogram for quantitative prediction of axillary nodal burden.ResultsIn predicting axillary nodal burden, the habitat radiomics model outperformed both the C-radiomics and clinical models, achieving areas under the curve (AUCs) of 0.791 (0.712-0.870) and 0.798 (0.686-0.911) in the training and validation cohorts, respectively. The C-radiomics model achieved AUCs of 0.733 (0.631-0.836) and 0.738 (0.612-0.865), while the clinical model achieved AUCs of 0.753 (0.663-0.843) and 0.733 (0.596-0.870). The combined nomogram demonstrated the highest diagnostic performance, with AUCs of 0.895 (0.839-0.951) and 0.885 (0.802-0.969) in the training and validation cohorts, respectively.ConclusionsThe integrated nomogram combining clinical, C-radiomics, and habitat radiomics models demonstrated strong predictive efficacy for preoperative assessment of axillary nodal burden in breast cancer. Future multicenter prospective studies are warranted to validate these results and refine the model's clinical applicability.

IntroductionNiraparib and bevacizumab are two principal maintenance therapies for newly diagnosed advanced ovarian cancer (AOC) patients with BRCA wild-type (BRCAwt) status, regardless of homologous recombination deficiency (HRD). In China, however, a considerable proportion of BRCAwt patients have unknown or untested HRD status, complicating treatment selection.MethodsTo evaluate and compare the efficacy of niraparib and bevacizumab as maintenance therapy for BRCAwt AOC, we conducted a retrospective cohort study using real-world clinical data. Descriptive statistics were used to summarize clinical and demographic characteristics. Progression-free survival (PFS) was estimated using Kaplan-Meier analysis and compared using a stratified Cox proportional hazards model. A multivariable Cox regression was performed to adjust for potential confounding variables. Exploratory subgroup analyses were conducted, and propensity score matching (PSM) was applied as a sensitivity analysis.ResultsA total of 94 patients were included, with 51 receiving niraparib and 43 receiving bevacizumab. The median PFS was not reached in the niraparib group versus 13.77 months (95% CI, 4.12-23.41) in the bevacizumab group (HR = 0.240, 95% CI, 0.128-0.451; P < .001). After covariate adjustment, the median PFS was 19.55 months (95% CI, 9.40-NA) with niraparib and 8.64 months (95% CI, 4.53-NA) with bevacizumab, with an adjusted HR of 0.282 (95% CI, 0.136-0.587; P = .001). In the PSM sensitivity analysis, the median PFS was not reached (95% CI, 19.55-NR) in the niraparib group and was 18.33 months (95% CI, 8.90-25.26) in the bevacizumab group (HR = 0.360, 95% CI, 0.176-0.736; P = .005).ConclusionThis analysis suggests that niraparib may provide a progression-free survival advantage compared with bevacizumab in BRCAwt AOC patients, with both regimens appearing to be generally well tolerated in the real-world setting. These findings offer preliminary reference value for maintenance treatment selection in patients with newly diagnosed BRCAwt AOC.

BackgroundBy integrating Digitally Reconstructed Radiograph (DRR) images of pulmonary tumors with Electronic Portal Imaging Device (EPID) images to assist in target segmentation, and subsequently comparing morphological changes in segmented targets across different radiotherapy stages, this approach enables precise quantification of dynamic variations in target volume and shape. This methodological integration provides objective evidence for treatment response evaluation and dynamic optimization of treatment plans, thereby significantly enhancing the precision of radiotherapy delivery.MethodsThe proposed multimodal segmentation framework, named EPIDSeg-Net, comprises an encoder, a multi-scale feature layer, and a decoder. The encoder utilizes a dual-branch architecture: a CNN branch for extracting local texture features and a Swin-Transformer branch for capturing global semantic features. The model first calibrates multimodal input features through a Dual Attention Mechanism (DAM) to adaptively adjust modality-specific weights, thereby enhancing tolerance to missing image information in multi-sequence segmentation. Subsequently, two key modules are implemented within the multi-scale feature layer: a Large-Kernel Grouped Attention Gating (LKG-Gate) module to strengthen local contextual awareness, and a Multi-Path Feature Extraction (MPFE) module to improve feature robustness via a parallel structure. These designs enable the model to effectively focus on lung tumor target regions, optimize segmentation accuracy, and achieve high-performance reconstruction.ResultsThe framework effectively integrates multimodal features, enabling high-precision localization and sharp boundary delineation while preserving anatomical details. Quantitative evaluations demonstrate superior performance: DICE = 93.2 (92.4∼93.9), CE = 0.352, HD95 = 9.42 (6.03∼12.8), IOU = 86.0 (84.1∼87.9), and SENCE = 0.828. Overall, the model excels at preserving gradient information, regional integrity, and fine details; effectively suppresses feature loss; and reduces missed segmentation rates, leading to improvements in both subjective and objective performance metrics.ConclusionThe proposed segmentation method effectively integrates information from EPID and DRR images, enabling more precise localization and segmentation of lesion regions within EPID images while enhancing segmentation accuracy.