The study included 100 patients with LACC who underwent definitive CCRT with IMRT/VMAT technique followed by 3D-IGABT. MRI-based contouring included T2WI and DWI images for primary tumor (GTVp) and lymph nodes (GTVn). The contours were imported to MIM software to extract first-order radiomic features. Radiomic values from pre-treatment (PreRx), pre-brachytherapy (PreBT), differences between PreRx and PreBT (Diff) radiomic and clinical factors were analyzed using univariate and multivariate Cox regression analysis. Predictive models of PFS, LRFS, DMFS, and OS were created along with the optimism index and calibration plot.

The median follow-up time was 24.5 months. The 2-year of PFS, LRFS, DMFS, and OS rates were 71, 88.6, 83.1, and 83.5 %, respectively. For all clinical outcomes, CF + RF combined from PreRx and PreBT resulted in the highest Harrell's C-index compared with the CF or RF alone. Compare with Diff models, models from PreRx and PreBT resulted in higher Harrell's C-index. The C-indexes from the CF + RF model from PreRx and PreBT for PFS, LRFS, DMFS, and OS were 0.739, 0.873, 0.830 and 0.967 with the optimism indexes of 0.312, 0.381, 0.316, and 0.242, respectively.

Radiomic features from the first-order statistics added values to clinical factors to predict the outcomes after CCRT. The highest prediction model performance was for the combined clinical and radiomics from PreRx and PreBT.

The assessment of neoadjuvant treatment's response is critical for selecting the most suitable therapeutic options for patients with breast cancer to reduce the need for invasive local therapies. Breast magnetic resonance imaging (MRI) is so far one of the most accurate approaches for assessing pathological complete response, although this is limited by the qualitative and subjective nature of radiologists' assessment, often making it insufficient for deciding whether to forgo additional locoregional therapy measures. To increase the accuracy and prediction of radiomic MRI with the aid of machine learning models and deep learning methods, as part of artificial intelligence, have been used to analyse the different subtypes of breast cancer and the specific changes observed before and after therapy. This review discusses recent advancements in radiomic MRI models for presurgical response assessment for patients with early breast cancer receiving preoperative treatments, with a focus on their implications for clinical practice.

To establish and validate deep learning (DL) models based on pre-treatment multiparametric magnetic resonance imaging (MRI) images of primary rectal cancer and basic clinical data for the prediction of synchronous liver metastases (SLM) in patients with Rectal cancer (RC).

In this retrospective study, 176 and 31 patients with RC who underwent multiparametric MRI from two centers were enrolled in the primary and external validation cohorts, respectively. Clinical factors, including sex, primary tumor site, CEA level, and CA199 level were assessed. A clinical feature (CF) model was first developed by multivariate logistic regression, then two residual network DL models were constructed based on multiparametric MRI of primary cancer with or without CF incorporation. Finally, the SLM prediction models were validated by 5-fold cross-validation and external validation. The performance of the models was evaluated by decision curve analysis (DCA) and receiver operating characteristic (ROC) analysis.

Among three SLM prediction models, the Combined DL model integrating primary tumor MRI and basic clinical data achieved the best performance (AUC = 0.887 in primary study cohort; AUC = 0.876 in the external validation cohort). In the primary study cohort, the CF model, MRI DL model, and Combined DL model achieved AUCs of 0.816 (95% CI: 0.750, 0.881), 0.788 (95% CI: 0.720, 0.857), and 0.887 (95% CI: 0.834, 0.940) respectively. In the external validation cohort, the CF model, DL model without CF, and DL model with CF achieved AUCs of 0.824 (95% CI: 0.664, 0.984), 0.662 (95% CI: 0.461, 0.863), and 0.876 (95% CI: 0.728, 1.000), respectively.

The combined DL model demonstrates promising potential to predict SLM in patients with RC, thereby making individualized imaging test strategies.

Accurate synchronous liver metastasis (SLM) risk stratification is important for treatment planning and prognosis improvement. The proposed DL signature may be employed to better understand an individual patient's SLM risk, aiding in treatment planning and selection of further imaging examinations to personalize clinical decisions.

Not applicable.

The aim of this study was to thoroughly analyze the reproducibility of radiomics feature extraction across three Image Biomarker Standardization Initiative (IBSI)-compliant platforms using a digital phantom for benchmarking. It uncovers high consistency among common features while also pointing out the necessity for standardization in computational algorithms and mathematical definitions due to unique platform-specific features.

We selected three widely used radiomics platforms: LIFEx, Computational Environment for Radiological Research (CERR), and PyRadiomics. Using the IBSI digital phantom, we performed a comparative analysis to extract and benchmark radiomics features. The study design included testing each platform's ability to consistently reproduce radiomics features, with statistical analyses to assess the variability and agreement among the platforms.

The results indicated varying levels of feature reproducibility across the platforms. Although some features showed high consistency, others varied significantly, highlighting the need for standardized computational algorithms. Specifically, LIFEx and PyRadiomics performed consistently well across many features, whereas CERR showed greater variability in certain feature categories than LIFEx and PyRadiomics.

The study findings highlight the need for harmonized feature calculation methods to enhance the reliability and clinical usefulness of radiomics. Additionally, this study recommends incorporating clinical data and establishing benchmarking procedures in future studies to enhance the role of radiomics in personalized medicine.

Osteosarcoma (OS) is the most common primary bone malignancy, and neoadjuvant chemotherapy (NAC) improves survival rates. However, OS heterogeneity results in variable treatment responses, highlighting the need for reliable, non-invasive tools to predict NAC response. Radiomics-based machine learning (ML) offers potential for identifying imaging biomarkers to predict treatment outcomes. This systematic review and meta-analysis evaluated the accuracy and reliability of radiomics models for predicting NAC response in OS.

A systematic search was conducted in PubMed, Embase, Scopus, and Web of Science up to November 2024. Studies using radiomics-based ML for NAC response prediction in OS were included. Pooled sensitivity, specificity, and AUC for training and validation cohorts were calculated using bivariate random-effects modeling, with clinical-combined models analyzed separately. Quality assessment was performed using the QUADAS-2 tool, radiomics quality score (RQS), and METRICS scores.

Sixteen studies were included, with 63 % using MRI and 37 % using CT. Twelve studies, comprising 1639 participants, were included in the meta-analysis. Pooled metrics for training cohorts showed an AUC of 0.93, sensitivity of 0.89, and specificity of 0.85. Validation cohorts achieved an AUC of 0.87, sensitivity of 0.81, and specificity of 0.82. Clinical-combined models outperformed radiomics-only models. The mean RQS score was 9.44 ± 3.41, and the mean METRICS score was 60.8 % ± 17.4 %.

Radiomics-based ML shows promise for predicting NAC response in OS, especially when combined with clinical indicators. However, limitations in external validation and methodological consistency must be addressed.

The aim of this study was to map multicollinearity of the radiomic feature set in five independent [18F]FDG-PET cohorts with different tumour types and identify generalizable non-redundant features.

Five [18F]FDG-PET radiomic cohorts were analysed: non-small cell lung carcinomas (N = 35), pheochromocytomas and paragangliomas (N = 40), head and neck squamous cell carcinomas (N = 54), [18F]FDG-positive thyroid nodules with indeterminate cytology (N = 84), and gastric carcinomas (N = 206). Lesions were delineated, and 105 radiomic features were extracted using PyRradiomics. In every cohort, Spearman's rank correlation coefficient (ρ) matrices of features were calculated to determine which features showed (very) strong (ρ > 0.7 and ρ > 0.9) correlations with any other feature in all five cohorts. Cluster analysis of an averaged correlation matrix for all cohorts was performed at a threshold of ρ = 0.7 and ρ = 0.9. For each cluster, a representative, non-redundant feature was selected.

Seventy-two and 90 out of 105 features showed a (very) strong correlation with another feature in the correlation matrix in all five cohorts. Cluster analysis resulted in 35 and 15 non-redundant features at thresholds of ρ = 0.9 and ρ = 0.7, including 6 and 3 shape features, 4 and 2 intensity features, and 25 and 10 texture features, respectively. Seventy or 90 redundant features could be omitted at these thresholds, respectively.

At least two-thirds of the radiomic feature set could be omitted because of strong multicollinearity in multiple independent cohorts. More redundant features could be identified using a less conservative threshold. Future research should indicate whether multicollinearity of the radiomic feature set is similar for other radiopharmaceuticals and imaging modalities.

Question Radiomic feature sets contain many strongly correlating features, which results in statistical challenges. Findings Analysis of the correlation matrices showed that the same radiomic features were strongly correlated in five independent [18F]FDG-PET cohorts with different tumour types. Clinical relevance At least two-thirds of the radiomic feature set could be omitted, because of strong multicollinearity. More redundant features could be identified using a less conservative threshold.

Background/Objectives: Small bowel tumors (SBTs) encompass a diverse range of tumor types, with benign tumors being the most prevalent. However, the incidence of malignant SBTs is increasing, particularly small bowel adenocarcinoma; this poses a diagnostic challenge for clinicians and radiologists due to the varied and nonspecific clinical and radiological presentations associated with SBTs. In fact, SBTs can present differently in emergencies, often mimicking inflammatory diseases or manifesting as complications such as intussusception, small bowel obstruction (SBO), intestinal ischemia, perforation, gastrointestinal bleeding, or metastatic disease. These tumors can remain asymptomatic for extended periods. Methods: We present a pictorial review on the role of imaging in evaluating SBTs, focusing on the emergency setting where diagnosis can be incidental. We also include some representative cases that may be useful for radiologists and residents in clinical practice. Results: Despite these challenges, contrast-enhanced computed tomography (CECT) is usually the best modality to use in emergencies for evaluating SBTs, and in some cases, a diagnosis can be made incidentally. However, when possible, multimodal imaging through cross-sectional imaging remains crucial for the non-invasive diagnosis of SBTs in stable patients, as endoscopic procedures may also be impractical. A complementary CT study with distension using negative oral contrast media, such as water, polyethylene glycol, or mannitol solutions, can improve the characterization of SBTs and rule out multiple SBT locations, particularly in small bowel neuroendocrine tumor (NET) and gastrointestinal tumor (GIST) localization. Positive water-soluble iodine-based oral contrast, such as Gastrografin (GGF), can be used to evaluate and monitor the intestinal lumen during the nonsurgical management of small bowel obstruction (SBO) or in suspected cases of small bowel perforations or the presence of fistulas. Magnetic resonance enterography (MRE) can aid in improving the characterization of SBTs through a multiplanar and multisequence study. Positron emission tomography combined with CT is generally an essential modality in evaluating metastatic disease and staging and assessing tumor prognosis, but it has limitations for indolent lymphoma and small NETs. Conclusions: Therefore, the integration of multiple imaging modalities can improve patient management and provide a preoperative risk assessment with prognostic and predictive indicators. In the future, radiomics could potentially serve as a "virtual biopsy" for SBTs, allowing for better diagnosis and more personalized management in precision medicine.

Gliomas are the most prevalent and aggressive form of primary brain tumors. The clinical challenge in managing patients with this disease revolves around the difficulty of diagnosis, both at onset and during treatment, and the scarcity of prognostic outcome indicators. Radiomics involves the extraction of quantitative features from medical images with the help of artificial intelligence, positioning it as a promising tool to be integrated into the care of glioma patients. Using data from 52 studies and 12,482 patients over two years, this review explores how radiomics can enhance the initial diagnosis of gliomas, especially helping to differentiate treatment stages that may be difficult for the human eye to do otherwise. Radiomics has also been able to identify patient-specific tumor molecular signatures for targeted treatments without the need for invasive surgical biopsy. Such an approach could lead to earlier interventions and more precise individualized therapies that are tailored to each patient. Additionally, it could be integrated into clinical practice to improve longitudinal diagnosis during treatment and predict tumor recurrence. Finally, radiomics has the potential to predict clinical outcomes, helping both patients and providers set realistic expectations. While this field is continuously evolving, future research should conduct such studies in larger, multi-institutional cohorts to enhance generalizability and applicability in clinical practice and focus on combining radiomics with other modalities to improve its predictive accuracy and clinical utility.

Purpose To build a deep learning framework using contrast-enhanced MRI for lesion segmentation and automatic molecular subtype classification in breast cancer. Materials and Methods This retrospective multicenter study included patients with biopsy-proven invasive breast cancer between January 2015 and January 2021. An automatic breast lesion segmentation model was developed using three-dimensional (3D) ResU-Net as the backbone, and its accuracy was evaluated in an internal and two external testing datasets using the Dice score. An ensemble model for classification of breast cancer into four molecular subtypes (Ensemble ResNet) was then developed by combining both two-dimensional and 3D lesion features. The performance of Ensemble ResNet was evaluated in the three testing datasets using the area under the receiver operating characteristic curve (AUC). Results A total of 687 female patients (mean age ± SD, 48.70 years ± 8.97) were included, with 289, 61, 73, and 264 patients included in the training, internal testing, and two external testing datasets, respectively. The proposed segmentation model achieved high accuracy in internal testing dataset 1, external testing dataset 2, and external testing dataset 3 (Dice scores: 0.86, 0.82, 0.85) and luminal A, luminal B, human epidermal growth factor receptor 2 (HER2)-enriched, and triple-negative breast cancer (TNBC) subtypes (Dice scores: 0.8571, 0.8323, 0.8199, 0.8481). Ensemble ResNet demonstrated high performance for the prediction of luminal A subtypes (AUC range, 0.74-0.84), luminal B subtypes (AUC range, 0.68-0.72), HER2-enriched subtypes (AUC range, 0.73-0.82), and TNBC (AUC range, 0.80-0.81) in the three testing datasets. Conclusion The proposed novel deep learning framework based on MRI achieved high, robust performance in fully automatic classification of breast cancer molecular subtypes. Keywords: MR-Imaging, Breast, Oncology, Breast Cancer, Molecular Subtype, Deep Learning Framework Supplemental material is available for this article. © RSNA, 2025.

BackgroundNon-invasive methods with universal prognostic guidance for detecting breast cancer (BC) survival biomarkers need to be further explored.ObjectiveThis study aimed to investigate C-C motif chemokine receptor type 5 (CCR5) prognosis value in BC and develop a radiomics model for noninvasive prediction of CCR5 expression in BC.MethodsA total of 840 cases with genomic information were included and divided into CCR5 high- and low-expression groups for clinical characteristic differences exploration. Bioinformatics and survival analysis including Kaplan-Meier (KM) survival analysis, Cox regression, immunoinfiltration analysis, and tumor mutation load (TMB) were performed. For radiomics model development, 98 cases with dynamic contrast-enhancement magnetic resonance imaging (DCE-MRI) scans were used. Radiomics features extracted were using Pyradiomics and filtered by maximum-relevance minimum-redundancy (mRMR) and recursive feature elimination (REF) algorithms. Support vector machine (SVM) and logistic regression (LR) models were developed to predict CCR5 expression, with the radiomics score (Rad_score) representing the predicted probability of CCR5 expression. The models' performance was compared using the Delong test, and the model with the superior area under the curve (AUC) values was selected to analyze the correlation between CCR5 expression, Rad_score, and immune genes.ResultsThe CCR5 high-expression group exhibited better overall survival (OS) (p < 0.01). Six radiomics features were selected for model development. The AUCs of the SVM model predicting CCR5 were 0.753 and 0.748 in the training and validation sets, respectively, while the AUCs of the LR model were 0.763 and 0.762. Calibration curves and decision curve analysis (DCA) validated the models' calibration and clinical utility. The SVM_Rad_score showed a strong association with immune-related genes.ConclusionsThe DCE-MRI radiomics model presents a novel, non-invasive tool for predicting CCR5 expression in BC and provides valuable insights to inform clinical decision-making.

This study aimed to investigate the value of radiomics analysis in the precise diagnosis of triple-negative breast cancer (TNBC) based on breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and apparent diffusion coefficient (ADC) maps.

This retrospective study included 326 patients with pathologically proven breast cancer (TNBC: 129, non-TNBC: 197). The lesions were segmented using the ITK-SNAP software, and whole-volume radiomics features were extracted using a radiomics platform. Radiomics features were obtained from DCE-MRI and ADC maps. The least absolute shrinkage and selection operator regression method was employed for feature selection. Three prediction models were constructed using a support vector machine classifier: Model A (based on the selected features of the ADC maps), Model B (based on the selected features of DCE-MRI), and Model C (based on the selected features of both combined). Receiver operating characteristic curves were used to evaluate the diagnostic performance of the conventional MR image model and the 3 radiomics models in predicting TNBC.

In the training dataset, the AUCs for the conventional MR image model and the 3 radiomics models were 0.749, 0.801, 0.847, and 0.896. The AUCs for the conventional MR image model and 3 radiomics models in the validation dataset were 0.693, 0.742, 0.793, and 0.876, respectively.

Radiomics based on the combination of whole volume DCE-MRI and ADC maps is a promising tool for distinguishing between TNBC and non-TNBC.

Glioblastoma (GBM) and solitary brain metastasis (SBM) exhibit similar radiomics features on magnetic resonance imaging (MRI), yet their treatment strategies and prognoses significantly differ. Therefore, accurate differentiation between these two types of tumors is crucial for clinical decision-making. This study aims to establish and validate an efficient diagnostic model based on the radiomic features of the T1-weighted contrast-enhanced (T1CE) sequence in the 10 mm brain-tumor interface region to achieve precise differentiation between GBM and SBM.

This study retrospectively collected contrast-enhanced T1-weighted imaging data from 226 GBM patients and 206 SBM patients at three centers between January 2010 and October 2024. Samples from centers 1 and 2 were used as the training set, while samples from center 3 were used as the test set. Two observers manually delineated the tumor edges on the T1CE images layer by layer to obtain the Region of Interest (ROI) covering the entire tumor volume. A 10 mm brain-to-tumor interface (BTI) was extracted using Python code. Radiomic features were extracted from the 10 mm BTI region, followed by feature selection and model construction. Finally, SHAP (SHapley Additive exPlanations) was used to visualize the model. Three radiologists with 2, 6, and 18 years of diagnostic experience independently evaluated the test set samples without knowing the patient information or pathology results, establishing three diagnostic models. The DeLong test was used to compare these models with the radiomic model.

Ultimately, ten radiomic features were used for modeling. The model established using the logistic regression (LR) algorithm had an AUC of 0.893 on the training set and 0.808 on the test set. The AUCs of the three radiologists with different diagnostic experiences on the test set were 0.699, 0.740, and 0.789, respectively, all lower than that of the radiomic model. The DeLong test showed that ModelBTI performed significantly better than Doctor 1 (p<0.05) in the test set, but there was no statistically significant difference in performance between ModelBTI and Doctors 2 and 3.

The radiomic model constructed based on the 10 mm brain-tumor interface can effectively differentiate between GBM and SBM, capturing tumor heterogeneity from a new perspective, thereby significantly improving diagnostic performance and providing assistance for clinical diagnosis.

The original contributions presented in the study are included in the article/Supplemental material, further inquiries can be directed to the corresponding authors.

To evaluate the feasibility of radiomics analysis using dual-layer detector spectral CT (DLCT)-derived iodine maps for the preoperative prediction of the Ki-67 proliferation index (PI) in pancreatic ductal adenocarcinoma (PDAC).

A total of 168 PDAC patients who underwent DLCT examination were included and randomly allocated to the training (n = 118) and validation (n = 50) sets. A clinical model was constructed using independent clinicoradiological features identified through multivariate logistic regression analysis in the training set. The radiomics signature was generated based on the coefficients of selected features from iodine maps in the arterial and portal venous phases of DLCT. Finally, a radiomics-clinical model was developed by integrating the radiomics signature and significant clinicoradiological features. The predictive performance of three models was evaluated using the Receiver Operating Characteristic (ROC) curve and Decision Curve Analysis. The optimal model was then used to develop a nomogram, with goodness-of-fit evaluated through the calibration curve.

The radiomics-clinical model integrating radiomics signature, CA19-9, and CT-reported regional lymph node status demonstrated excellent performance in predicting Ki-67 PI in PDAC, which showed an area under the ROC curve of 0.979 and 0.956 in the training and validation sets, respectively. The radiomics-clinical nomogram demonstrated the improved net benefit and exhibited satisfactory consistency.

This exploratory study demonstrated the feasibility of using DLCT-derived iodine map-based radiomics to predict Ki-67 PI preoperatively in PDAC patients. While preliminary, our findings highlight the potential of functional imaging combined with radiomics for personalized treatment planning.

Texture features can capture microstructural patterns and tissue heterogeneity, playing a pivotal role in medical image analysis. Compared to deep learning-based features, texture features offer superior interpretability in clinical applications. However, as conventional texture features focus strictly on voxel-level statistical information, they fail to account for critical spatial heterogeneity between small tissue volumes, which may hold significant importance. To overcome this limitation, we propose novel 3D patch-based texture features and develop a radiomics analysis framework to validate the efficacy of our proposed features. Specifically, multi-scale 3D patches were created to construct patch patterns via k-means clustering. The multi-resolution images were discretized based on labels of the patterns, and then texture features were extracted to quantify the spatial heterogeneity between patches. Twenty-five cross-combination models of five feature selection methods and five classifiers were constructed. Our methodology was evaluated using two independent MRI datasets. Specifically, 145 breast cancer patients were included for axillary lymph node metastasis prediction, and 63 cervical cancer patients were enrolled for histological subtype prediction. Experimental results demonstrated that the proposed 3D patch-based texture features achieved an AUC of 0.76 in the breast cancer lymph node metastasis prediction task and an AUC of 0.94 in cervical cancer histological subtype prediction, outperforming conventional texture features (0.74 and 0.83, respectively). Our proposed features have successfully captured multi-scale patch-level texture representations, which could enhance the application of imaging biomarkers in the precise prediction of cancers and personalized therapeutic interventions.

To enhance the prognostic assessment and management of lung cancer with cystic airspaces (LCCA) by integrating temporal clinical and phenotypic dimensions of tumor growth.

A retrospective analysis was conducted on LCCA patients treated at two hospitals. Clinical and imaging characteristics were analyzed using the independent samples t-test, Mann-Whitney U test, and χ2 test. Features with significant differences were further analyzed using multivariate Cox regression to identify independent prognostic factors. Radiomic features were extracted from CT images, and volume doubling time (VDT) was calculated from two follow-up scans. Separate predictive models were constructed based on radiomic features and VDT. A fusion model integrating radiomic features, VDT, and independent clinical prognostic factors was developed. Model performance was evaluated using receiver operating characteristic curve and the area under the curve, with DeLong's test used for comparison.

A total of 193 patients were included, with an average survival time of 48.5 months. Significant differences were found between survivors and non-survivors in age, smoking status, chronic obstructive pulmonary disease, and tumor volume (P < 0.05). Multivariate Cox analysis identified smoking and chronic obstructive pulmonary disease as independent risk factors (P = 0.028 and P = 0.013). The VDT for survivors was 421 (298 582.5) days compared to 334.5 ± 106.1 days for non-survivors (Z = -3.330, P = 0.001). In the validation set, the area under the curve for the VDT model was 0.805, for the radiomic model 0.717, and for the fusion model 0.895, demonstrating the highest predictive performance (P < 0.05).

Integrating VDT, radiomics, and clinical imaging features into a fusion model improves the accuracy of predicting the five-year survival rate for LCCA patients, enhancing personalized and precise cancer treatment.

Microvascular invasion (MVI) affects the prognosis and treatment of hepatocellular carcinoma (HCC); however, its preoperative diagnosis is challenging. Analysis of computed tomography (CT) images using radiomics can detect MVI, but its effectiveness depends on the imaging conditions. We compared the efficacies of radiomics, clinical, and combined models for predicting MVI in HCC using nonstandardized scanning protocols.

This multicenter study included 533 patients who underwent hepatic resection for HCC. Patients were divided randomly into training (n = 426) and test groups (n = 107). We manually extracted 3D CT features in hepatic arterial, portal venous, and venous phases. The radiomics model was trained by machine learning. A logistic regression model was developed based on clinical information, and a fused model was created integrating clinical information and radiomics prediction score (Rad_Score). We calculated areas under the receiver operating characteristic curves (AUCs) for the radiomics, clinical, and mixed models in the test groups.

The clinical model incorporated hepatitis B virus surface antigen, tumor diameter, and log-transformed α-fetoprotein and des-gamma-carboxyprothrombin. The AUCs of the radiomics and clinical models were comparable (p = 0.76). Rad_Score was not an independent significant factor in the fused model (p = 0.40) and its addition did not improve the accuracy of the clinical model alone (p = 0.51).

A clinical model is as effective as a CT radiomics model for predicting MVI status in patients with HCC based on real-world scanning data, and integration of both models does not improve the predictive performance compared with a clinical model alone.

Invasive lung adenocarcinoma (LUAD) with the high-grade patterns (HGPs) has the potential for rapid metastasis and frequent recurrence. Therefore, accurately predicting the presence of high-grade components is crucial for doctors to develop personalized treatment plans and improve patient prognosis.

To develop a CNN-transformer fusion network based on radiomics and clinical information for predicting the HGPs of LUAD.

A total of 288 lesions in 288 patients with pathologically confirmed invasive LUAD were enrolled. Firstly, radiomics features were extracted from the entire tumor region on lung computed tomography (CT) images and then fused with clinical patient characteristics. Secondly, a structure was proposed that concatenated a convolutional neural network (CNN) and Transformer encoding blocks to mine and extract more comprehensive information. Finally, a classification prediction was performed through fully connected layers.

Accuracy, sensitivity, specificity, precision, and area under the receiver operating characteristic (ROC) curve (AUC) were utilized for evaluation of the model's classification prediction performance. Delong's test was used to compare the AUCs of different models for significance. The proposed model was effective with an accuracy of 0.86, sensitivity of 0.67, specificity of 0.94, precision of 0.74, and AUC of 0.91.

The CNN-transformer fusion network, based on radiomics and clinical information, demonstrates good performance in predicting the presence of HGPs and can be employed to assist in the development of personalized treatment plans for patients with invasive LUAD.

Aims/Background Colorectal cancer (CRC) is one of the major global health problems, with high morbidity and mortality, underscoring the need for new diagnostic and prognostic tools. Therefore, this study aims to evaluate the significance of integrating radiomics with genetic data in CRC prognostic assessment and improve the accuracy of prognosis prediction. Methods This study included computed tomography (CT) images from 225 CRC patients and RNA-seq information from 654 patients, obtained from the TICA database. Key radiomics features and genes were identified through radiomics feature extraction, least absolute shrinkage and selection operator (LASSO) regression analysis, and Kaplan-Meier survival analysis. Furthermore, a CRC prognostic model was constructed using these key genes and radiomics features. Results This study identified 170 key radiomics features. Out of them, five were significantly associated with CRC prognosis. Transcriptome data analysis identified 8 key genes, among which the high expressions of Inhibin Subunit Beta B (INHBB), Potassium Voltage-Gated Channel Subfamily Q Member 2 (KCNQ2), and Ubiquilin Like (UBQLNL) were significantly correlated with poor prognosis. Age, tumor stage, pathological T stage, and pathological N stage were determined as independent prognostic factors. Moreover, immune infiltration analysis demonstrated that the immune score of the low-risk group was higher than that of the high-risk group, with significant differences in some immune cells, and key genes were correlated with immune cells. Additionally, the constructed CRC prognostic model incorporating three genes, INHBB, KCNQ2, and UBQLNL, exhibited high prediction accuracy in the validation set, with area under the curve (AUC) values of 0.80, 0.87, and 0.84 at 1-year, 3-year, and 5-year, respectively, indicating good prediction performance and reliability of the model. Conclusion The multimodal data combining radiomics features and gene expression data can improve the accuracy of CRC prognostic assessment, providing a valuable prognostic prediction tool for clinical practice and helping to guide the selection and optimization of treatment regimens.

Adult patients with diffuse lower-grade gliomas (dLGG) show heterogeneous survival outcomes, complicating postoperative treatment planning. Treating all patients early increases the risk of long-term side effects, while delayed treatment may lead to impaired survival. Refinement of prognostic models could optimize timing of treatment. Conventional radiological features are prognostic in dLGG, but MRI could carry more prognostic information. This study aimed to investigate MRI-based radiomics survival models and compare them with clinical models.

Two clinical survival models were created: a preoperative model (tumor volume) and a full clinical model (tumor volume, extent of resection, tumor subtype). Radiomics features were extracted from preoperative MRI. The dataset was divided into training set and unseen test set (70:30). Model performance was evaluated on test set with Uno's concordance index (c-index). Risk groups were created by the best performing model's predictions.

207 patients with mutated IDH (mIDH) dLGG were included. The preoperative clinical, full clinical and radiomics models showed c-indexes of 0.70, 0.71 and 0.75 respectively on test set for overall survival. The radiomics model included four features of tumor diameter and tumor heterogeneity. The combined full clinical and radiomics model showed best performance with c-index = 0.79. The survival difference between high- and low-risk patients according to the combined model was both statistically significant and clinically relevant.

Radiomics can capture quantitative prognostic information in patients with dLGG. Combined models show promise of synergetic effects and should be studied further in astrocytoma and oligodendroglioma patients separately for optimal modelling of individual risks.

The purpose of this study was to develop three predictive models utilising clinical factors, radiomics features, and habitat features, to distinguish between nonclinically significant prostate cancer (csPCa) and clinically significant PCa (non-csPCa) on the basis of biparametric MRI (bp-MRI).

A total of 175 patients were enrolled, including 134 individuals with csPCa and 41 with non-csPCa. The clinical model was developed using optimal predictive factors obtained from univariable logistic regression and modelled through a random forest approach. Image acquisition and segmentation were performed first in the creation of both the radiomics model and the habitat model. The K-means clustering algorithm was then used exclusively for habitat generation in the development of the habitat model. Finally, feature selection and model construction were performed for both models. Model comparison and diagnostic efficacy assessment were conducted through receiver operating characteristic curve analysis, decision curve analysis (DCA), and calibration curve analysis.

The habitat model outperformed both the radiomics model and the clinical model in distinguishing csPCa from non-csPCa patients. The AUC values of the habitat model in the training and test sets were 0.99 and 0.93, respectively. Furthermore, DCA and the calibration curves highlighted the superior clinical utility and greater predictive accuracy of the habitat model in comparison with the other two models.

We developed a habitat-based radiomics model with a greater ability to distinguish between csPCa and non-csPCa on the basis of bp-MRI than a traditional radiomics model and clinical model. This introduces a novel approach for assessing the heterogeneity of PCa and offers urologists a quantitative, noninvasive method for preoperatively evaluating the aggressiveness of PCa.

To investigate the predictive value of serum alpha - fetoprotein (AFP), lectin-reactive alpha-fetoprotein (AFP-L3), and multimodal magnetic resonance imaging (MRI) radiomics in forecasting therapeutic efficacy and prognosis following radiofrequency ablation (RFA) in patients with hepatocellular carcinoma (HCC).

A retrospective analysis was conducted on HCC patients who underwent RFA between January 2019 and December 2023. Clinical and radiologic features of HCC were analyzed. A predictive model was developed using clinical data and radiomic features collected before surgery, with the goal of predicting prognosis after RFA. The predictive performance of the model was evaluated using AUC values in both training and validation cohorts.

A total of 298 HCC patients were included in the study, divided into a good prognosis group (n=145) and a poor prognosis group (n=153). Serum AFP and AFP-L3 levels were significantly higher in the poor prognosis group (P=0.007 and P=0.02, respectively). Independent predictive factors included: AFP-L3 (95% CI -1.228, -1.1.61; P<0.001), AFP (95% CI 0.017, 0.036; P<0.001), intratumoral hemorrhage (95% CI 0.380, 0.581; P<0.001), peritumoral arterial tumor enhancement (95% CI 0.193, 0.534; P<0.001) and low signal intensity around liver and gallbladder tumors (95% CI 0.267, 0.489; P<0.001). The combined clinical-radiological-radiomics model demonstrated superior predictive performance, with AUC value of 0.897 in the training set and 0.841 in the validation set, outperforming individual models and sequences.

The integrated clinical-radiological-radiomics model showed excellent predictive performance for the prognosis of HCC patients undergoing RFA, surpassing individual models. Key predictors included serum AFP, AFP-L3 levels, intratumoral hemorrhage, and peritumoral low signal intensity. This multimodal approach offers a promising tool for individualized prognostic assessment and improved clinical decision-making.

Cervical cancer remains a major cause of mortality among women globally, with lymph node metastasis (LNM) being a critical determinant of patient prognosis.

In this study, MRI scans from 153 cervical cancer patients between January 2018 and January 2024 were analyzed. The patients were assigned to two groups: 103 in the training cohort; 49 in the validation cohort. Radiomic features were extracted from T2-weighted imaging (T2WI) and apparent diffusion coefficient (ADC) maps. The ITK-SNAP software enabled three-dimensional manual segmentation of the tumor regions in cervical cancer to identify regions of interest (ROIs). The collected data was divided for the training and validation of the Support Vector Machine (SVM) model.

The combined T2WI and ADC-based radiomics model exhibited robust diagnostic capabilities, achieving an area under the curve (AUC) of 0.804 (95% CI [0.712-0.890]) in the training cohort and an AUC of 0.811 (95% CI [0.721-0.902]) in the validation cohort. The nomogram that includes radiomic features, International Federation of Gynecology and Obstetrics (FIGO) stage, and LNM has a C-index of 0.895 (95% CI [0.821-0.962]) in the training cohort and a C-index of 0.916 (95% CI [0.825-0.987]) in the validation cohort. The C-statistics are all above 0.80, and the predicted variables are nearly aligned with the 45-degree line, consistent with the results shown in the calibration plot. This indicates that our model demonstrates good discrimination ability and satisfactory calibration.

The MRI radiomics model, leveraging T2WI combined with ADC maps, offers an effective method for predicting LNM in cervical cancer patients.

Gastric cancer (GC) is a leading cause of cancer-related deaths worldwide, with delayed diagnosis often limiting effective treatment options. This study introduces a novel, noninvasive radiomics-based approach using [18F] FDG PET/CT (fluorodeoxyglucose positron emission tomography/computed tomography) to predict vascular endothelial growth factor (VEGF) status and survival in patients with GC. The ability to noninvasively assess these parameters can significantly influence therapeutic decisions and outcomes.

We conducted a retrospective study involving patients diagnosed with GC, stratified into training, validation, and test groups. Each patient underwent a [18F] FDG PET/CT scan, and radiomic features were extracted using dedicated software. A Radiomics Score (RS) was calculated, serving as a predictor for VEGF status. Statistical analyses included logistic regression and Cox proportional hazards models to evaluate the predictive power of RS on survival outcomes.

The developed radiomics model demonstrated high predictive accuracy, with the RS formula achieving an area under the receiver operating characteristic curve of 0.861 in the training cohort and 0.857 in the validation cohort for predicting VEGF status. The model also identified RS as an independent prognostic factor for survival, where higher RS values correlated with poorer survival rates.

The findings underscore the potential of [18F] FDG PET/CT radiomics in transforming the management of GC by providing a noninvasive means to assess tumor aggressiveness and prognosis through VEGF status. This model could facilitate earlier and more tailored therapeutic interventions, potentially improving survival outcomes in a disease marked by typically late diagnosis and limited treatment success.

This study was designed to develop and validate a radiomic nomogram for the differential diagnosis of myelin oligodendrocyte glycoprotein antibody-related disease (MOGAD) and aquaporin-4 immunoglobulin G-positive neuromyelitis optica spectrum disorder (AQP4+NMOSD).

We retrospectively analysed data from a primary cohort consisting of 21 MOGAD and 63 AQP4+NMOSD patients and an external validation cohort comprising 10 MOGAD and 34 AQP4+NMOSD patients. Radiomic features were extracted from lesions of the cervical spinal cord and brainstem from sagittal T2-weighted MR images. We constructed a prediction model by integrating radiomic features with clinical data and evaluated its performance using calibration curves and decision curve analysis (DCA).

We developed a comprehensive nomogram that combines clinical and radiomic features to distinguish MOGAD from AQP4+NMOSD. The discriminative ability of the nomogram was quantified by the area under the receiver operating characteristic (ROC) curve (AUC), achieving values of 0.915 (95 % CI, 0.859-0.970) in the primary cohort and 0.837 (95 % CI, 0.715-0.959) in the validation cohort, indicating high diagnostic accuracy. The calibration analyses showed good concordance between the model predicted and actual outcomes.

This study successfully validated the radiomic feature model, demonstrating its superior performance in differentiating MOGAD from AQP4+NMOSD. The nomogram, integrating radiomic features with conventional imaging characteristics of brainstem and cervical cord lesions, significantly enhanced differentiation capability. Both models proved valuable in improving diagnostic accuracy, with radiomic features contributing most significantly.

This scoping review systematically maps externally validated machine learning (ML)-based models in cancer patient care, quantifying their performance, and clinical utility, and examining relationships between models, cancer types, and clinical decisions. By synthesizing evidence, this study identifies, strengths, limitations, and areas requiring further research.

The review followed the Joanna Briggs Institute's methodology, Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines, and the Population, Concept, and Context mnemonic. Searches were conducted across Embase, IEEE Xplore, PubMed, Scopus, and Web of Science (January 2014-September 2022), targeting English-language quantitative studies in Q1 journals (SciMago Journal and Country Ranking > 1) that used ML to evaluate clinical outcomes for human cancer patients with commonly available data. Eligible models required external validation, clinical utility assessment, and performance metric reporting. Studies involving genetics, synthetic patients, plants, or animals were excluded. Results were presented in tabular, graphical, and descriptive form.

From 4023 deduplicated abstracts and 636 full-text reviews, 56 studies (2018-2022) met the inclusion criteria, covering diverse cancer types and applications. Convolutional neural networks were most prevalent, demonstrating high performance, followed by gradient- and decision tree-based algorithms. Other algorithms, though underrepresented, showed promise. Lung and digestive system cancers were most frequently studied, focusing on diagnosis and outcome predictions. Most studies were retrospective and multi-institutional, primarily using image-based data, followed by text-based and hybrid approaches. Clinical utility assessments involved 499 clinicians and 12 tools, indicating improved clinician performance with AI assistance and superior performance to standard clinical systems.

Interest in ML-based clinical decision-making has grown in recent years alongside increased multi-institutional collaboration. However, small sample sizes likely impacted data quality and generalizability. Persistent challenges include limited international validation across ethnicities, inconsistent data sharing, disparities in validation metrics, and insufficient calibration reporting, hindering model comparison reliability.

Successful integration of ML in oncology decision-making requires standardized data and methodologies, larger sample sizes, greater transparency, and robust validation and clinical utility assessments.

Financed by FCT-Fundação para a Ciência e a Tecnologia (Portugal, project LA/P/0063/2020, grant 2021.09040.BD) as part of CSS's Ph.D. This work was not registered.

The amygdala is a subcortical region in the mesiotemporal lobe that plays a key role in emotional and sensory functions. Conventional neuroimaging experiments treat this structure as a single, uniform entity, but there is ample histological evidence for subregional heterogeneity in microstructure and function. The current study characterized subregional structure-function coupling in the human amygdala, integrating post-mortem histology and in vivo MRI at ultra-high fields. Core to our work was a novel neuroinformatics approach that leveraged multiscale texture analysis as well as non-linear dimensionality reduction techniques to identify salient dimensions of microstructural variation in a 3D post-mortem histological reconstruction of the human amygdala. We observed two axes of subregional variation in this region, describing inferior-superior as well as mediolateral trends in microstructural differentiation that in part recapitulated established atlases of amygdala subnuclei. Translating our approach to in vivo MRI data acquired at 7 Tesla, we could demonstrate the generalizability of these spatial trends across 10 healthy adults. We then cross-referenced microstructural axes with functional blood-oxygen-level dependent (BOLD) signal analysis obtained during task-free conditions, and revealed a close association of structural axes with macroscale functional network embedding, notably the temporo-limbic, default mode, and sensory-motor networks. Our novel multiscale approach consolidates descriptions of amygdala anatomy and function obtained from histological and in vivo imaging techniques.

Variability in treatment response may be attributable to organ-level heterogeneity in tumor lesions. Radiomic analysis of medical images can elucidate non-invasive biomarkers of clinical outcome. Organ-specific radiomic comparison across immunotherapies and targeted therapies has not been previously reported. We queried the UPMC Hillman Cancer Center registry for patients with metastatic melanoma (MEL) treated with immune checkpoint inhibitors (ICI) (anti-programmed cell death protein-1 (PD-1)/cytotoxic T-lymphocyte associated protein 4 (CTLA-4) (ipilimumab+nivolumab; I+N) or anti-PD-1 monotherapy) or BRAF-targeted therapy. The best overall response was measured using Response Evaluation Criteria in Solid Tumors V.1.1. Lesions were segmented into discrete volume-of-interest with 400 radiomics features extracted. Overall and organ-specific machine-learning models were constructed to predict disease control (DC) versus progressive disease (PD) using XGBoost. 291 patients with MEL were identified, including 242 ICI (91 I+N, 151 PD-1) and 49 BRAF. 667 metastases were analyzed, including 541 ICI (236 I+N, 305 PD-1) and 126 BRAF. Across cohorts, baseline demographics included 39-47% women, 24%-29% M1C, 24-46% M1D, and 61-80% with elevated lactate dehydrogenase. Among ICI patients experiencing DC, the organs with the greatest reduction were liver (-66%±8%; mean±SEM) and lung (-63%±5%). For patients with multiple same-organ target lesions, the highest interlesion heterogeneity was observed in brain among patients who received ICI while no intraorgan heterogeneity was observed in BRAF. 221 ICI patients were included for radiomic modeling, consisting of 86 I+N and 135 PD-1. Models consisting of optimized radiomic signatures classified DC/PD across I+N (area under curve (AUC)=0.85) and PD-1 (0.71) and within individual organ sites (AUC=0.72~0.94). Integration of clinical variables improved the models' performance. Comparison of models between treatments and across organ sites suggested mostly non-overlapping DC or PD features. Skewness, kurtosis, and informational measure of correlation (IMC) were among the radiomic features shared between overall response models. Kurtosis and IMC were also used by multiple organ-site models. In conclusion, differential organ-specific response was observed across BRAF and ICI with within organ heterogeneity observed for ICI but not for BRAF. Radiomic features of organ-specific response demonstrated little overlap. Integrating clinical factors with radiomics improves the prediction of disease course outcome and prediction of tumor heterogeneity.

The objective of this study is to develop an automated method for segmenting spleen computed tomography (CT) images using a deep learning model. This approach is intended to address the limitations of manual segmentation, which is known to be susceptible to inter-observer variability. Subsequently, a prediction model of gastric cancer (GC) serosal invasion was constructed in conjunction with radiomics and deep learning features, and a nomogram was generated to explore the clinical guiding significance.

This study enrolled 311 patients from two centers with pathologically confirmed of GC. we employed a deep learning model, U-Mamba, to obtain fully automatic segmentation of the spleen CT images. Subsequently, radiomics features and deep learning features were extracted from the entire spleen CT images, and significant features were identified through dimensionality reduction. The clinical features, radiomic features, and deep learning features were organized and integrated, and five machine learning methods were employed to develop 15 predictive models. Ultimately, the model exhibiting superior performance was presented in the form of a nomogram.

A total of 18 radiomics features, 30 deep learning features, and 1 clinical features were deemed valuable. The DLRA model demonstrated superior discriminative capacity relative to other models. A nomogram was constructed based on the logistic clinical model to facilitate the usage and verification of the clinical model.

Radiomics and deep learning features derived from automated spleen segmentation to construct a nomogram demonstrate efficacy in predicting serosal invasion in GC. Concurrently, fully automated segmentation provides a novel and reproducible approach for radiomics research.

To study the application of radiomics in cancer imaging with a focus on lung cancer, renal cell carcinoma, gastrointestinal cancer, and head and neck cancer.

Different electronic databases were considered. Articles published in the last five years were analyzed (January 2019 and December 2023). Papers were selected by two investigators with over 15 years of experience in Radiomics analysis in cancer imaging. The methodological quality of each radiomics study was performed using the Radiomic Quality Score (RQS) by two different readers in consensus and then by a third operator to solve disagreements between the two readers.

19 articles are included in the review. Among the analyzed studies, only one study achieved an RQS of 18 reporting multivariable analyzes with also non-radiomics features and using the validation phase considering two datasets from two distinct institutes and open science and data domain.

This informative review has brought attention to the increasingly consolidated potential of Radiomics, although there are still several aspects to be evaluated before the transition to routine clinical practice. There are several challenges to address, including the need for standardization at all stages of the workflow and the potential for cross-site validation using heterogeneous real-world datasets. It will be necessary to establish and promote an imaging data acquisition protocol, conduct multicenter prospective quality control studies, add scanner differences and vendor-dependent characteristics; to collect images of individuals at additional time points, to report calibration statistics.

COVID-19 remains widespread and poses a threat to people's physical and mental health, especially middle-aged and elderly individuals. Early identification of COVID-19 patients at high risk of progressing to critical disease helps improve overall patient outcomes and healthcare efficiency.

To develop a radiomics nomogram to predict the risk of newly admitted middle-aged and elderly COVID-19 patients progressing to critical disease.

A total of 794 patients (aged 40 years or above) were retrospectively included in the study from two institutions, all of them were with non-critical COVID-19 on admission. At follow-up, patients were divided into non-critical group and critical group. About 443 patients (384 non-critical and 59 critical) from the first hospital were randomly assigned to the training (n = 311) and internal validation (n = 132) set in a 7:3 ratio. Additionally, an independent external cohort of 351 patients (292 non-critical and 59 critical) from another hospital was evaluated. Radiomics signatures and clinical indicators were used to build a radiomics model and a clinical model after computed tomography (CT) image processing, CT whole-lung segmentation, feature extraction, and feature selection. The radiomics nomogram model integrated radiomics model and clinical model. The receiver operating characteristic curve (AUC) was used to assess the performance of the proposed models. Calibration curves and decision curve analysis were used to assess the performance of the radiomics nomogram.

For the training, internal validation, and external validation sets, the AUC values of the radiomic nomogram for the prediction of COVID-19 progression were 0.916, 0.917, and 0.890, respectively. Calibration curves indicated that there was no significant departure between prediction and observation in three sets. The decision curve image demonstrated the clinical utility of the nomogram model.

Our nomogram model incorporates radiomics features and clinical indicators, it provides a new pathway to increase predictive accuracy or clinical utility, further helping to provide personalized management for middle-aged and elderly patients with COVID-19.

Our study aims to assess the robustness of myocardial radiomic texture features (RTF) to segmentation variability and variations across scanners with different field strengths, addressing concerns about reliability in clinical practices. We conducted a retrospective analysis on 45 pairs of CMR T1 maps from 15 healthy volunteers using 1.5 T and 3 T Siemens scanners. Manual left ventricular myocardium segmentation and a deep learning-based model with Monte Carlo Dropout generated masks with different levels of variability and 1023 RTFs extracted from each region of interest (ROI). Reproducibility: the extent to which RTFs extracted from 1.5 T and 3 T images are consistent, and repeatability: the extent to which RTFs extracted from multiple segmentation runs at the same field strength agree with each other, were measured by the intraclass correlation coefficient (ICC). We categorized ICC values as excellent, good, moderate, and poor. We reported the proportion of RTFs that fell in each category. The proportion of RTFs with excellent repeatability decreased as the proportion of ROI pixels in congruence across segmentation runs decreased. Up to 31% of RTFs showed excellent repeatability, while 35% showed good repeatability across segmentation runs from the manually generated masks. Across scanners (i.e., 1.5 T vs 3 T), only 7% exhibited good reproducibility. While our results demonstrate RTF sensitivity to differences in field strength and segmentation variability, we identified certain preprocessing filters and feature classes that are less sensitive to these variations and, as such, may be good candidates for imaging biomarkers or building machine-learning models.

Glioblastoma (GBM) is a highly aggressive brain tumor associated with a poor patient prognosis. The survival rate remains low despite standard therapies, highlighting the urgent need for novel treatment strategies. Advanced imaging techniques, particularly magnetic resonance imaging (MRI), are crucial in assessing GBM. Disruptions in various oncogenic signaling pathways, such as Receptor Tyrosine Kinase (RTK)-Ras-Extracellular signal-regulated kinase (ERK) signaling, Phosphoinositide 3- Kinases (PI3Ks), tumor protein p53 (TP53), and Neurogenic locus notch homolog protein (NOTCH), contribute to the development of different tumor types, each exhibiting distinct morphological and phenotypic features that can be observed at a microscopic level. However, identifying genetic abnormalities for targeted therapy often requires invasive procedures, prompting exploration into non-invasive approaches like radiogenomics. This study explores the utility of radiogenomics and machine learning (ML) in predicting these oncogenic signaling pathways in GBM patients.

We collected post-operative MRI scans (T1w, T1c, FLAIR, T2w) from the BRATS-19 dataset, including scans from patients with both GBM and LGG, linked to genetic and clinical data via TCGA and CPTAC. Signaling pathway data was manually extracted from cBioPortal. Radiomic features were extracted from four MRI modalities using PyRadiomics. Dimensionality reduction and feature selection were applied and Data imbalance was addressed with SMOTE. Five ML models were trained to predict signaling pathways, with Grid Search optimizing hyperparameters and 5-fold cross-validation ensuring unbiased performance. Each model's performance was evaluated using various metrics on test data.

Our results showed a positive association between most signaling pathways and the radiomic features derived from MRI scans. The best models achieved high AUC scores, namely 0.7 for RTK-RAS, 0.8 for PI3K, 0.75 for TP53, and 0.4 for NOTCH, and therefore, demonstrated the potential of ML models in accurately predicting oncogenic signaling pathways from radiomic features, thereby informing personalized therapeutic approaches and improving patient outcomes.

We present a novel approach for the non-invasive prediction of deregulation in oncogenic signaling pathways in glioblastoma (GBM) by integrating radiogenomic data with machine learning models. This research contributes to advancing precision medicine in GBM management, highlighting the importance of integrating radiomics with genomic data to understand tumor behavior and treatment response better.

Tumour-infiltrating lymphocytes (TILs) are strongly correlated with the prognosis and immunotherapy response in bladder cancer. The TIL status is typically assessed through microscopy as part of tissue pathology. Here, the authors developed Rad-TIL model, a novel radiomics model, to predict TIL status in patients with bladder cancer.

The authors enrolled 1089 patients with bladder cancer and developed the Rad-TIL model by using a machine-learning method based on computed tomography (CT) images. The authors applied a radiogenomics cohort to reveal the key pathways underlying the Rad-TIL model. Finally, the authors used an independent treatment cohort to evaluate the predictive efficacy of the Rad-TIL model for Bacillus Calmette-Guérin (BCG) immunotherapy.

The authors developed the Rad-TIL model by integrating tumoral and peritumoral features on CT images and obtained areas under the receiver operating characteristic curves of 0.844 and 0.816 in the internal and external validation cohorts, respectively. Patients were stratified into two groups based on the predicted radiomics score of TILs (RSTIL). RSTIL exhibited prognostic significance for both overall and cancer-specific survival in each cohort (hazard ratios: 2.27-3.15, all P<0.05). Radiogenomics analysis revealed a significant association of RSTIL with immunoregulatory pathways and immune checkpoint molecules (all P<0.05). Notably, BCG immunotherapy response rates were significantly higher in high-RSTIL patients than in low-RSTIL patients (P=0.007).

The Rad-TIL model, a noninvasive method for assessing TIL status, can predict clinical outcomes and BCG immunotherapy response in patients with bladder cancer.