Technology in Cancer Research & Treatment最新文献

IntroductionTumor-infiltrating lymphocytes (TILs) are key indicators of immune response and prognosis in breast cancer (BC). Accurate prediction of TIL levels is essential for guiding personalized treatment strategies. This study aimed to develop and evaluate machine learning models using ultrasound-derived radiomics and clinical features to predict TIL levels in BC.MethodsThis retrospective study included 256 BC patients between January 2019 and August 2023, who were randomly divided into training (n = 179) and test (n = 77) cohorts. Radiomics features were extracted from the intratumor and peritumor regions in ultrasound images. Feature selection was performed using the "Boruta" package in R to iteratively remove non-significant features. Extra Trees Classifier was used to construct radiomics and clinical models. A combined radiomics-clinical (R-C) model was also developed. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, and decision curve analysis (DCA) to assess clinical utility. A nomogram was created based on the best-performing model.ResultsA total of 1712 radiomics features were extracted from the intratumor and peritumor regions. The Boruta method selected five key features (four from the peritumor and one from the intratumor) for model construction. Clinical features, including immunohistochemistry, tumor size, shape, and echo characteristics, showed significant differences between high (≥10%) and low (<10%) TIL groups. Both the R-C and radiomics models outperformed the clinical model in the test cohort (area under the curve values of 0.869/0.838 vs 0.627, P < .05). Calibration curves and Brier scores demonstrated superior accuracy and calibration for the R-C and radiomics models. DCA revealed the highest net benefit of the R-C model at intermediate threshold probabilities.ConclusionUltrasound-derived radiomics effectively predicts TIL levels in BC, providing valuable insights for personalized treatment and surveillance strategies.

IntroductionThoracic SMARCA4-deficient tumors, which are rare and aggressive malignancies found in the lung or thoracic cavity, present a challenge in treatment standardization. This challenge arises from their resistance to chemotherapy and the absence of targeted therapy options.MethodsThoracic SMARCA4-deficient tumors were identified retrospectively using pathology databases. The clinicopathological characteristics of these tumors are outlined, and the clinical outcomes of advanced patients treated with immune checkpoint inhibitors (ICIs) in combination with chemotherapy and chemotherapy alone are reviewed.ResultsThirty-nine patients had thoracic SMARCA4-deficient tumors, with a median age of 62 years. The cohort consisted of 92.3% males, and 89.7% had a history of smoking. Within this group, 94.9% had stage III/IV disease at diagnosis. SMARCA4-deficient non-small cell lung cancer (SMARCA4-DNSCLC) and SMARCA4-deficient undifferentiated tumors (SMARCA4-DUT) display distinct histological and immunohistochemical features. Thirty-five patients underwent systemic therapy, achieving an ORR of 51.4%, a DCR of 82.9%, and a median OS of 20.9 months. Patients were categorized into chemotherapy (28.6%) and ICIs plus chemotherapy (71.4%) groups. The ICIs plus chemotherapy group exhibited an ORR of 64.0% and a DCR of 96.0%, while the chemotherapy group had an ORR of 20.0% and 50.0%, respectively (P < .0001 for ORR and DCR). The median OS for ICIs plus chemotherapy and chemotherapy groups were 20.9 months and 6.5 months, and median PFS were 9.6 months and 3.5 months, respectively, all statistically significant (P < .05). Multivariate COX regression analysis indicated that treatment was an independent prognostic factor for OS.ConclusionThoracic SMARCA4-deficient tumors exhibit a lack of SMARCA4 expression, displaying high malignancy and aggressiveness while exhibiting poor response to standard chemotherapy. The combination of ICIs with chemotherapy could potentially serve as an effective treatment approach for thoracic SMARCA4-deficient tumors.

PurposeAutomating quality assurance (QA) for contours generated by automatic algorithms is critical in radiotherapy treatment planning. Manual QA is tedious, time-consuming, and prone to subjective experiences. Automatic segmentation reduces physician workload and improves consistency. However, an effective QA process for these automatic contours remains an unmet need in clinical practice.Materials and MethodsThe patient data used in this study was derived from the AAPM Thoracic Auto-Segmentation Challenge dataset, including left and right lungs, heart, esophagus, and spinal cord. Two groups of organ-at-risk (OAR) were generated. A ResNet-152 network was used as a feature extractor, and a one-class support vector machine (OC-SVM) was employed to classify contours as 'high' or 'low' quality. To evaluate the generalizability, we generated low-quality contours using translation and resizing techniques and assessed correlations between detection limits and metrics such as volume, Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and mean surface distance (MSD).ResultsThe proposed OC-SVM model outperformed binary classifiers n metrics such as balanced accuracy and area under the receiver operating characteristic curve (AUC) . It demonstrated superior performance in detecting various types of contour errors while maintaining high interpretability. Strong correlations were observed between detection limits and contour metrics.ConclusionOur proposed model integrates an attention mechanism with a one-class classification framework to automate QA for OAR delineations. This approach effectively detects diverse types of contour errors with high accuracy, significantly reducing the burden on physicians during radiotherapy planning.

Introductionin radiotherapy, fiducial markers improve the accuracy of radiation delivery, and their use has become increasingly important in the treatment of various cancers, particularly those in the prostate and lung. This work aims to determine, via Monte Carlo simulations and numerical ultrasound transport, the feasibility of fiducial marker localization via single-shot x-ray acoustic computed tomography.Methodspatient data from CT scans for two treatment sites, prostate and lung, were used to model the fiducial marker localization process. Monte Carlo simulation was used to calculate the absorbed dose distribution in each patient resulting from the irradiation with a 120 kVp x-ray imaging source, assuming that the dose is imparted in a short pulse. Ultrasound transport through each patient was modeled with the numerical ultrasound transport package k-Wave. For the image reconstruction process, as the exact internal patient structure will not be known at the time of treatment, a homogenous medium with the patient external contour and dimensions was used.ResultsIt is shown that the use of a homogeneous model to approximate the actual patient material composition during the reconstruction process, necessary as the geometry of the internal structures is not known at the time of the treatment, severely degrades the quality of the x-ray acoustic tomography images, but that it is still possible to determine the fiducial marker position with an accuracy of or better than 1 mm. The largest errors are observed for the lung patient when the lung is in an inflated state.Conclusionsit has been shown that single-shot x-ray acoustic tomography can be an effective tool for the tracking and localization of radiotherapy fiducial markers, exhibiting an accuracy of better than 1 mm, despite the poor visual quality of the resultant images.

Papillary thyroid cancer (PTC), the most prevalent form of thyroid malignancy, is generally indolent but poses a recurrence risk of 10%-15%, leading to a clinical paradox: the need to mitigate recurrence while avoiding overtreatment. Current prognostic frameworks, reliant on anatomical and histopathological factors, often result in inefficient treatment pathways, unnecessary surgical interventions, and increased patient burden. The advent of molecular diagnostics presents a paradigm shift in risk stratification. Implementing preoperative molecular tests could transform PTC management by enabling tailored therapeutic strategies, reducing the need for completion thyroidectomies, optimizing the selection of patients for active surveillance, and refining the use of adjuvant therapies such as radioactive iodine. While genomic alterations such as BRAF and TERT mutations have been explored as prognostic markers, their predictive utility remains limited. In contrast, transcriptomic profiling has emerged as a powerful tool for identifying aggressive PTC subtypes with greater precision. Transcriptomic-based prognostic tests, like the novel Thyroid GuidePx^® classifier, effectively stratify PTCs into distinct molecular subgroups with differing recurrence risks, surpassing traditional clinicopathological models in predictive accuracy. By shifting toward biologically informed decision-making, we can enhance clinical efficiency, minimize patient morbidity, and improve overall healthcare resource utilization.

ObjectiveThis study analyzed the dosimetric impact of residual errors (rotational and deformation errors) in left-sided breast cancer radiotherapy after cone-beam CT (CBCT)-based translational errors correction.MethodsTwenty patients treated with intensity-modulated radiotherapy (IMRT) were retrospectively analyzed. Virtual CT images were generated by deforming and registering CBCT images with planning CT images. The accumulated dose was calculated to assess residual errors effects on target and organs at risk (OARs). A phantom test was conducted to evaluate rotational errors impacts.ResultsResults showed significant dose differences: for 4005 cGy, D98 and D95 of the breast (PTV_b) decreased, and mean dose, V30, and V20 of the left lung reduced; for 5000 cGy, D98 of the supraclavicular lymph nodes (PTV_sc) and PTVb, D95 of PTV_b, and mean dose and V20 of the heart differed significantly. Phantom simulations revealed that pitch angles ≤-1.8° and roll/yaw angles >2° caused overdosing in the left lung and heart, with maximum dose differences of 31.89% (heart) and 19.19% (lung) for 4005 cGy, and 26.32% (heart) and 20.92% (PTV_sc) for 5000 cGy.ConclusionResidual errors significantly affect dose distribution despite CBCT-based translational correction. Improved immobilization techniques or 6DOF couch correction are recommended to mitigate rotational errors.

IntroductionThis study sought to develop a predictive model using CT-based habitat radiomics to forecast pathological complete response (pCR) and progression-free survival (PFS) in esophageal squamous cell carcinoma (ESCC) patients receiving standardized neoadjuvant chemoradiotherapy (nCRT) followed by curative surgery.MethodsWe retrospectively analyzed baseline CT imaging data from 228 ESCC patients in a prospective cohort database. Patients were randomly divided into training and validation sets (7:3 ratio). Whole-tumor and habitat-derived radiomic features were extracted from pretreatment CT scans. For pCR prediction, habitat signatures were developed using Logistic Regression (LR), RandomForest (RF), and XGBoost models, optimized via grid search. PFS prediction employed Cox proportional hazards modeling with selected features. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), Hosmer-Lemeshow calibration curves, and decision curve analysis.ResultsThe habitat models retained 10 features for pCR prediction and 12 for PFS prediction. For pCR, the habitat-derived RF model demonstrated superior performance (training AUC: 0.821; validation AUC: 0.826), outperforming both other habitat models and the whole-tumor radiomics model (training AUC: 0.645). Similarly, the habitat-based RF model for PFS achieved higher AUCs (training: 0.759, 95% CI: 0.627-0.889; validation: 0.810, 95% CI: 0.653-0.966) compared to whole-tumor radiomics (training: 0.623; validation: 0.519).ConclusionOur analyses indicated a trend where habitat radiomics might outperform whole-tumor radiomics in predicting pCR and PFS for resectable ESCC after nCRT. While this merits further investigation, current evidence is insufficient to confirm its clinical utility for personalized treatment guidance.

ObjectivesIrreversible Electroporation (IRE) is both open surgery and minimally invasive cancer therapy used in the treatment of liver tumors. The therapy demands precision and accuracy to ensure complete tumor ablation. Reliable simulation tools can help achieve this goal by predicting the tissue regions that will reach the required electric field threshold and by suggesting correcting actions when the predicted outcome is inadequate. This article retrospectively compares segmented ablations from intra-procedural computed tomography (CT) scans with computer simulations to check their validity in predicting the operation outcome and the required electric field threshold.Methods10 patient ablation procedures were retrospectively analyzed using a detailed computational model of electroporation, informed by the patient-specific geometry of each case. CT scans were analyzed by three physicians over two sessions to assess intra- and inter-observer variability. Same day postoperative images were used for accuracy. The resulting measured ablations from the patient's data were compared to simulation predictions, both in terms of ablated volumes and 3D similarity scores (Dice coefficient).ResultsSimulated ablation volumes were computed across electric field thresholds (465-750 V/cm), showing highest volumes at 465 V/cm and lowest at 750 V/cm. Comparison with physician segmented volumes showed best match for 500-600 V/cm ablation threshold: this result was consistent across different patients despite differences among patient's conditions and characteristics. 3D analysis revealed Dice scores between 0.63 and 0.77 (mean: 0.71), indicating moderate to good agreement. Visual and statistical comparisons further validated the reliability of the simulation model within this threshold range.ConclusionThis study highlighted the accuracy of IRE ablation volume predictions by comparing retrospective CT based ablation volume segmentations with electric field simulations. The best match occurred at 500 to 600 V/cm thresholds, with post-procedure measurements. Despite observer variability and modeling limitations, Dice scores showed moderate to good agreement, validating the simulation model and emphasizing timely imaging for accuracy.

IntroductionSarcomatoid renal cell carcinoma (sRCC) is rare but highly aggressive and is associated with poor prognosis and limited treatment responsiveness. Despite several studies investigating its clinicopathological features, existing research is often limited by small sample sizes and short follow-up periods, and currently, no prognostic risk model is specific to patients with non-metastatic sRCC. This study aimed to investigate the clinicopathological characteristics of patients with non-metastatic sRCC and develop a predictive model for postoperative mortality risk.MethodsIn this retrospective study, we analyzed the clinical data of 45 patients diagnosed with non-metastatic sRCC who underwent surgical treatment at our institution's Department of Urology, between January 2008 and June 2024. These patients were compared with 527 patients with non-sarcomatoid renal cell carcinoma (non-sRCC). The primary endpoint was death, and the exact cause of death was recorded. Routine postoperative examinations and treatment details were documented through outpatient and inpatient electronic medical record systems.ResultsThe results indicated significant differences in body mass index, hypertension, surgical approach, nephrectomy type, surgical duration, maximum tumor diameter, tumor necrosis, T stage, and Ki-67 expression between patients with sRCC and those with non-sRCC (P < 0.05). Survival analysis revealed that the cancer-specific survival (CSS) for patients with sRCC was significantly lower than that for patients with non-sRCC (P < 0.001). Cox univariate and multivariate analyses identified maximum pathological tumor diameter, T stage, and high Ki-67 expression as independent risk factors. Based on these factors, we developed a postoperative mortality risk prediction model for patients with sRCC, with the calibration curves demonstrating a good fit of the model.ConclusionsThe proposed model is designed for patients with non-metastatic sRCC. It has potential clinical application value, aiding in the identification of high-risk patients and providing guidance for individualized treatment and close follow-up.

Lymphoma is a highly heterogeneous malignancy, demanding accurate and precise diagnosis to guide the selection of the appropriate treatment for optimal outcome. Copy number aberration (CNA) has been suggested to play an important role in the occurrence and development of lymphoma and thus can be explored as biomarker to improve disease management. It is believed that CNAs in variable forms and complexities can be triggered by both exogenous (eg viral infection and ionizing radiation) and endogenous factors (eg genetic predisposition and evolutionary forces). However, conventional cytogenetic methods have limitations to detect all types of CNAs with accuracy and adequate details. The emergence of new technologies, including fluorescence in situ hybridization (FISH), chromosome microarray analysis (CMA), and especially next-generation sequencing (NGS) has made significant progress in the identification and characterization of CNAs or CNA-related genomic aberrations. Accumulating data addressing molecular insights and clinical implications have provided us more theoretical and experimental support for its clinical translation. Currently, while only limited number of CNAs or CNA-related genomic variation, such as deletion/amplification of DNA segments, have been documented in major guidelines or consensus for their clinical potential in lymphoma, more CNAs remain to be further characterized and/or discovered for their clinical relevance. Taking together, with available and upcoming evidence, CNA should play an important role as a diagnostic and prognostic biomarker while integrated with the current settings in lymphoma.