Radiomics Prediction Model Research Articles

MRI based radiomics nomogram for predict recurrence of non functioning pituitary macroadenomas post surgery

Objective: To establish and validate a comprehensive predictive model combining clinical data and radiomics features to improve the accuracy of predicting recurrence within five years after surgery in patients with non-functioning pituitary macroadenomas (NFMA). Methods: This retrospective study included 292 NFMA patients who underwent surgery between January 2012 and January 2018, with an additional 123 patients as an external test set. Clinical, pathological, and conventional imaging features were collected and analyzed using univariate and multivariate logistic regression to identify independent risk factors for postoperative recurrence. Radiomic features were extracted from preoperative T1-weighted (T1WI), T2-weighted (T2WI), and T1-enhanced images using 3D Slicer software. A radiomics prediction model was developed, and a combined model integrating clinical and radiomics features was established. The predictive performance of the models was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Results: The clinical model (Cli-score), radiomics model (Rad-score) and combined model were developed. The diagnostic performance of the clinical model in the external test set, showed an AUC of 0.757 (95%CI: 0.671–0.830), with SEN, SPE, and ACC of 82.5%, 59.04%, and 71.54%, respectively. The diagnostic performance of the radiomics model in the external test set showed an AUC of 0.835 (95% CI: 0.757–0.896), with 80%, 79.52% and 63.41% for SEN, SPE and ACC%, respectively. The diagnostic performance of the combined model in the external test set showed an AUC of 0.863 (95% CI: 0.790–0.919), with SEN, SPE, and ACC of 80%, 81.93%, and 68.30%, respectively. The calibration curve indicated good predictive performance, and DCA confirmed the high clinical utility of the combined model. Conclusion: The combined model provides a more accurate prediction of NFMA recurrence. This model can guide postoperative follow-up strategies and aid in early initiation of adjuvant therapy for high-risk patients.

Radiomics in the differential diagnosis of glioblastoma under the primary neurooncoimaging conditions

Introduction. According to the 2021 WHO Classification of Tumors of the Central Nervous System (CNS) and the 2023 Clinical Practice Guidelines on the Drug Management of Primary CNS Cancers, the first step of molecular genetic testing to identify the morphological type and malignancy of adult-type diffuse gliomas is the detection of isocitrate dehydrogenase (IDH) mutation status. However, tumor tissue biopsy as the conventional diagnostic standard has a number of limitations that can potentially be mitigated by applying the principles of radiomics to the interpretation of magnetic resonance (MR) images. The aim of our study is to develop a radiomics model for IDH mutation status prediction, which can be applied to primary diagnostic imaging in patients with suspected adult-type diffuse gliomas. Materials and methods. We conducted a retrospective comparative statistical analysis of radiomic features extracted from 46 conventional brain MR images of the patients with adult-type diffuse gliomas and identified IDH mutation status using the Random Forest algorithm of machine learning in combination with various preprocessing methods of the source imaging data and a semi-automated LevelTracing tool used for segmentation of the regions of interest (ROI). Results. The most effective combination of tools for preprocessing, segmentation, and classification was found to be ScaleIntensity, LevelTracing, and Random Forest, respectively. Using this combination, we verified the reliability of six radiomic predictors identified at the previous study stage. These features were all associated with IDH mutation status, and most of them capture texture heterogeneity in the ROIs at the voxel level. We were also able to improve the prognostic performance of our classification model up to AUC = 0.845 ± 0.089 (p 0.05). Conclusion. Based on a small, technically heterogeneous sample of routine MR imaging data, we developed a multiparametric model of IDH mutation status prediction in the patients with adult-type diffuse gliomas. Our conclusion is that relatively uniform preprocessing techniques based on uniform voxel intensity changes, which allow to preserve the structural detail, are feasible in clinical practice. The identified radiomic, likely voxel-based, features reflect the severity of perifocal vasogenic edema and the measure of intratumor morphological heterogeneity. We plan to assess the reproducibility of the study results using similar medical imaging data from open sources and to develop a color mapping technique for the ROIs to facilitate visual interpretation of quantitative radiomic data.

Contrast-Enhanced Computed Tomography Radiomics Predicts Colony-Stimulating Factor 3 Expression and Clinical Prognosis in Ovarian Cancer

Contrast-Enhanced Computed Tomography Radiomics Predicts Colony-Stimulating Factor 3 Expression and Clinical Prognosis in Ovarian Cancer

CBCT radiomics features combine machine learning to diagnose cystic lesions in the jaw.

The aim of this study was to develop a radiomics model based on cone beam computed tomography (CBCT) to differentiate odontogenic cysts (OC), odontogenic keratocysts (OKC) and ameloblastomas (AB). In this retrospective study, CBCT images were collected from 300 patients diagnosed with OC, OKC and AB who underwent histopathological diagnosis. These patients were randomly divided into training (70%) and test (30%) cohorts. Radiomics features were extracted from the images, and the optimal features were incorporated into Random Forest model, Support Vector Classifier (SVC) model, Logistic Regression model and a soft VotingClassifier based on the above three algorithms. The performance of the models was evaluated using a receiver operating characteristic (ROC) curve and the area under the curve (AUC). The optimal model among these was then used to establish the final radiomics prediction model, whose performance was evaluated using the sensitivity, accuracy, precision, specificity and F1 score in both the training cohort and the test cohort. The six optimal radiomics features were incorporated into a soft VotingClassifier. Its performance was the best overall. The AUC values of the One-vs-Rest (OvR) multiclassification strategy were AB-vs-Rest 0.963; OKC-vs-Rest 0.928; OC-vs-Rest 0.919 in the training cohort and AB-vs-Rest 0.814; OKC-vs-Rest 0.781; OC-vs-Rest 0.849 in the test cohort. The overall accuracy of the model in the training cohort was 0.757, and in the test cohort was 0.711. The VotingClassifier model demonstrated the ability of the CBCT radiomics to distinguish the multiple types of diseases (OC, OKC and AB) in the jaw and may have the potential to diagnose accurately under non-invasive conditions.

Breast density prediction model in digital versus synthetic mammograms from a radiomic point of view: A retrospective study.

Breast density prediction model in digital versus synthetic mammograms from a radiomic point of view: A retrospective study.

Development of a clinical prediction model for acute myocarditis using coronary computed tomography angiography-based radiomics.

Both acute myocarditis patients and normal cohort usually present with normal coronary computed tomography angiography (CCTA) performance, and the performance of CCTA radiomics on the prediction for myocarditis is still unclear. This study aims to build a clinical prediction model for acute myocarditis using CCTA-based radiomics. A total of 215 consecutive patients from the Affiliated Jinhua Hospital, Zhejiang University School of Medicine (Center 1) and Sir Run Run Shaw Hospital, Zhejiang University School of Medicine (Center 2) who underwent CCTA and were diagnosed as normal or acute myocarditis were enrolled. All CCTA images of myocardium were automatically segmented to extract radiomics features. Pearson correlation analysis was used to identify features that were highly correlated with others. The application of the 5-fold cross-validation test reduced reliance on a single training set and provided more robust performance estimation. The best radiomics prediction model was chosen and combined with the clinical labels to construct a clinical-radiomics model for the classification of patients as with or without myocarditis. Pearson's correlation and least absolute shrinkage and selection operator (LASSO) regression analyses identified 10 radiomics features and 7 clinical features which demonstrated the best correlation. The receiver operating characteristic curves of the three models that used the support vector machine (SVM) demonstrated the best performance. The area under the curves (AUCs) of Model 1 (Rad-score model) using training and test datasets were 0.970 (0.949-0.991) and 0.912 (0.832-0.992), respectively. The AUCs of Model 2 (clinical factors model) for the training and test datasets were 0.992 (0.983-1.000) and 0.943 (0.875-1.000), respectively. Model 3 (clinical factors and Rad-score model) demonstrated the best results, with AUCs of 1.000 (0.999-1.000) and 0.951 (0.880-1.000) in the training and test datasets, respectively. The CCTA-based radiomics model constructed using machine learning demonstrated good performance for predicting myocarditis.

Predictive classification of lung cancer pathological based on PET/CT radiomics.

To develop and validate a combined clinical and radiomics model for non-invasive prediction of lung cancer (LC) pathological types (lung adenocarcinoma, lung squamous cell carcinoma, and small cell lung cancer) based on patients' pre-treatment FDG PET/CT images and clinical data, as a complementary tool to aid in the diagnosis of LC pathological histological classification. In total, 896 patients with pathological confirmation of lung cancer were part of this retrospective study. The training and test groups included 819 patients who underwent scanning using scanner 1. The independent validation group included 77 patients who using scanner 2. The optimal features were retained by least absolute shrinkage and selection operator algorithm dimensionality reduction screening of the collected radiomics features, clinical parameters, and PET metabolic parameters. Five models were established to predict the lung cancer pathological types by the k-nearest neighbor classification (KNN) algorithm. The performance of the prediction model was assessed by calculating the area under the curve (AUC) from the receiver operator characteristic curve (ROC). Of all five predictive models (the PET-only radiomics model, the CT-only radiomics model, the PET/CT radiomics model, the clinical-only model and the combined clinical and PET/CT radiomics model), the clinical combined PET/CT radiomics model exhibited best performance. The macro-AUC for the training, test and independent validation groups were 0.974, 0.931, 0.960, the micro-AUC were 0.976, 0.940, 0.970, and the accuracy were 0.963, 0.914, and 0.961, respectively. Our model combined radiomics and clinical data and showed higher performance in non-invasively predicting the LC pathological types, which suggesting that PET/CT radiomics may be a promising technique for predicting LC histopathology.

Clinical value of the nomogram model based on endoscopic ultrasonography radiomics and clinical indicators in identifying benign and malignant lesions of the pancreas.

Based on endoscopic ultrasonography (EUS) radiomics and clinical data, we constructed a radiomics model and a nomogram model for identifying benign and malignant pancreatic lesions, and explored the diagnostic performance of these two prediction models. Images and clinical data of 151 patients with pancreatic lesions detected by EUS from January 2018 to September 2023 were retrospectively collected. The patients were randomly divided into a training set and a validation set at a ratio of 7:3. Through feature extraction and feature screening of EUS images, we calculated the radiomics score (rad-score) to realize the construction of the radiomics model. Collecting the clinical data, laboratory test results, and rad-scores from patients, univariate and multivariate logistic regression analyses were used to screen statistically significant influencing factors that could help identify benign and malignant lesions of the pancreas, and a nomogram model was constructed. The diagnostic performance and clinical utility of the two prediction models were evaluated using the receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Through feature extraction and screening, eight non-zero coefficient features were finally selected to calculate the rad-score. Multivariate logistic regression analysis showed that rad-score, age, and CA199 were the influencing factors in predicting benign and malignant pancreatic lesions. A nomogram model was constructed based on the three factors. In the validation set, the nomogram model exhibited superior performance with an AUC = 0.865 (95% CI 0.761-0.968) compared to the radiomics prediction model. The calibration curve and DCA depicted that the nomogram model demonstrated superior accuracy and yielded a higher net benefit for clinical decision-making compared to the radiomics prediction model. Based on EUS radiomics and clinical indicators, we constructed a promising nomogram model to accurately identify benign and malignant pancreatic lesions.

CT-based radiomics predictive model for spread through air space of IA stage lung adenocarcinoma.

BackgroundSpread through air spaces (STAS) in lung adenocarcinoma means different treatment and worse prognosis.PurposeTo construct a radiomics model based on CT scans to predict the presence of STAS in stage IA lung adenocarcinoma, compared with the traditional clinical model.Material and MethodsThe study included 317 patients (median age = 57.21 years; age range = 45.84-68.61 years) with pathologically confirmed stage IA lung adenocarcinoma. In total, 122 (38.5%) patients were diagnosed with STAS by pathology after the operation. Two experienced radiologists independently segmented the lesions using MITK software and extracted 1791 radiomics features using Python. Single-factor t-test or Mann-Whitney U-test and LASSO were used to screen for radiomics signatures related to STAS. This study constructed a radiomics model, a clinical model, and a combined model, combining radiomics and clinical features. Model performance was evaluated using the area under the curve (AUC).ResultsBy single-factor analysis, four clinical features and 13 radiomics features were significantly associated with STAS. The three models (the clinical, radiomics, and combine models) achieved predictive efficacy, with an AUC of 0.849, 0.867, and 0.939, respectively, in the training set and 0.808, 0.848, and 0.876, respectively, in the testing set.ConclusionThe combined model based on the radiomics and clinical features of preoperative chest CT could be used to preoperatively diagnose the presence of STAS in stage IA lung adenocarcinoma and has an excellent diagnostic performance.

The Value of MRI and Radiomics for the Diagnostic Evaluation of Thyroid-Associated Ophthalmopathy.

Thyroid-associated ophthalmopathy (TAO) is a vision-threatening autoimmune disease that involves the extraocular muscles (EOMs) and periorbital fat. Typical signs of TAO include eyelid recession, proptosis, diplopia, and decreased visual acuity. As a self-limited disease, there is major bipolarity in clinical outcomes in TAO population. The early diagnosis and prediction of these refractory and relapsed patients is essential. Unfortunately, commonly used tools such as CAS/NOSPECTS, are based on clinical symptoms and signs alone, have significant limitations. Some imaging techniques or examinations, such as magnetic resonance imaging (MRI), can be very effective in assisting TAO assessment, from exhaustive whiteboard notes to optimized patient outcomes. Being one of the most commonly used and accurate objective examinations for TAO assessment, MRI boosts no ionizing radiation, high soft tissue contrast, better reflection of tissue water content, and the ability to quantify multiple parameters. In addition, novel MR sequences are becoming increasingly more familiar in TAO and other areas of clinical and scientific research. Moreover, radiomics, a method involving the extraction of a large number of features from medical images through algorithms, is a more recent approach used in the analysis and characterization of TAO data. Thus, this review aims to summarize and compare the value of routine and novel functional MRI sequences and radiomics prediction models in the diagnosis and evaluation of TAO.

Development of a clinical prediction model for benign and malignant pulmonary nodules with a CTR ≥ 50% utilizing artificial intelligence-driven radiomics analysis

ObjectiveIn clinical practice, diagnosing the benignity and malignancy of solid-component-predominant pulmonary nodules is challenging, especially when 3D consolidation-to-tumor ratio (CTR) ≥ 50%, as malignant ones are more invasive. This study aims to develop and validate an AI-driven radiomics prediction model for such nodules to enhance diagnostic accuracy.MethodsData of 2,591 pulmonary nodules from five medical centers (Zhengzhou People’s Hospital, etc.) were collected. Applying exclusion criteria, 370 nodules (78 benign, 292 malignant) with 3D CTR ≥ 50% were selected and randomly split 7:3 into training and validation cohorts. Using R programming, Lasso regression with 10-fold cross-validation filtered features, followed by univariate and multivariate logistic regression to construct the model. Its efficacy was evaluated by ROC, DCA curves and calibration plots.ResultsLasso regression picked 18 non-zero coefficients from 108 features. Three significant factors—patient age, solid component volume and mean CT value—were identified. The logistic regression equation was formulated. In the training set, the ROC AUC was 0.721 (95%CI: 0.642–0.801); in the validation set, AUC was 0.757 (95%CI: 0.632–0.881), showing the model’s stability and predictive ability.ConclusionThe model has moderate accuracy in differentiating benign from malignant 3D CTR ≥ 50% nodules, holding clinical potential. Future efforts could explore more to improve its precision and value.Clinical trial numberNot applicable.

Utility of Multiparametric Breast MRI Radiomics to Predict Cyclin D1 and TGF-β1 Expression.

To develop a machine learning model that integrates clinical features and multisequence MRI radiomics for noninvasively predicting the expression status of prognostic-related factors cyclin D1 and TGF-β1 in breast cancer, providing additional information for the clinical development of personalized treatment plans. A total of 123 breast cancer patients confirmed by surgical pathology were retrospectively enrolled in our Hospital from January 2016 to July 2022. The patients were randomly divided into a training group (87 cases) and a validation group (36 cases). Preoperative routine and dynamic contrast-enhanced magnetic resonance imaging scans of the breast were performed for treatment subjects. The region of interest was manually outlined, and texture features were extracted using AK software. Subsequently, the LASSO algorithm was employed for dimensionality reduction and feature selection to establish the MRI radiomics labels. The diagnostic efficacy and clinical value were assessed through receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA). In the cyclin D1 cohort, the area under the receiver operating characteristic (ROC) curve in the clinical prediction model training and validation groups was 0.738 and 0.656, respectively. The multisequence MRI radiomics prediction model achieved an AUC of 0.874 and 0.753 in these respective groups, while the combined prediction model yielded an AUC of 0.892 and 0.785. In the TGF-β1 cohort, the ROC AUC for the clinical prediction model was found to be 0.693 and 0.645 in the training and validation groups, respectively. For the multiseries MRI radiomics prediction model, it achieved an AUC of 0.875 and 0.760 in these respective groups; whereas for the combined prediction model, it reached an AUC of 0.904 and 0.833. Decision curve analysis (DCA) demonstrated that both cohorts indicated a higher clinical application value for the combined prediction model compared with both individual models-clinical prediction model alone or radiomics model. The integration of clinical features and multisequence MRI radiomics in a combined modeling approach holds significant predictive value for the expression status of cyclin D1 and TGF-β1. The model provides a noninvasive, dynamic evaluation method that provides effective guidance for clinical treatment.

Overlooked and underpowered: a meta-research addressing sample size in radiomics prediction models for binary outcomes

ObjectivesTo investigate how studies determine the sample size when developing radiomics prediction models for binary outcomes, and whether the sample size meets the estimates obtained by using established criteria.MethodsWe identified radiomics studies that were published from 01 January 2023 to 31 December 2023 in seven leading peer-reviewed radiological journals. We reviewed the sample size justification methods, and actual sample size used. We calculated and compared the actual sample size used to the estimates obtained by using three established criteria proposed by Riley et al. We investigated which characteristics factors were associated with the sufficient sample size that meets the estimates obtained by using established criteria proposed by Riley et al.ResultsWe included 116 studies. Eleven out of one hundred sixteen studies justified the sample size, in which 6/11 performed a priori sample size calculation. The median (first and third quartile, Q1, Q3) of the total sample size is 223 (130, 463), and those of sample size for training are 150 (90, 288). The median (Q1, Q3) difference between total sample size and minimum sample size according to established criteria are −100 (−216, 183), and those differences between total sample size and a more restrictive approach based on established criteria are −268 (−427, −157). The presence of external testing and the specialty of the topic were associated with sufficient sample size.ConclusionRadiomics studies are often designed without sample size justification, whose sample size may be too small to avoid overfitting. Sample size justification is encouraged when developing a radiomics model.Key PointsQuestionSample size justification is critical to help minimize overfitting in developing a radiomics model, but is overlooked and underpowered in radiomics research.FindingsFew of the radiomics models justified, calculated, or reported their sample size, and most of them did not meet the recent formal sample size criteria.Clinical relevanceRadiomics models are often designed without sample size justification. Consequently, many models are too small to avoid overfitting. It should be encouraged to justify, perform, and report the considerations on sample size when developing radiomics models.

A nomogram predicts the new high-grade patterns of pulmonary invasive non-mucinous adenocarcinoma based on the radiomics and clinical features.

To develop a nomogram for the prediction of the new high-grade patterns of invasive non-mucinous adenocarcinoma (INMA) based on the radiomics and clinical features to provide accurate individualized treatment for patients. We collected patients pathologically diagnosed with INMA at our hospital. The study's endpoint, defined as 'new high-grade', was characterized by the presence of micropapillary patterns at ≥5% or high-grade patterns (including solid, micropapillary, and complex glandular) at ≥20%. Patients were randomly divided into training and validation cohorts in a ratio of 8:2. The region of interest (ROI) of chest plain scan images was sketched using 3D slicer software. The image and clinical features were analyzed by Least Absolute Shrinkage and Selection Operator (LASSO), univariate, and multivariate regression to construct the radiomics signature and nomogram model. The nomogram model was validated using the validation cohort. A total of 226 patients were divided into training (n = 180) and validation (n = 46) cohorts. From the ROI of these patients, 107 image features were extracted. LASSO regression analysis identified 16 image features that were used to construct the radiomics signature. The area under the curve values for the radiomics signature in the training and validation cohorts were 0.803 and 0.772, respectively. The Harrell's concordance index for the model, with 95% confidence intervals (CI), was 0.815 (CI: 0.806-0.824) for the training cohort and 0.802 (CI: 0.761-0.843) for the validation cohort. The radiomics prediction model demonstrates strong predictive capabilities and could serve as a valuable tool for guiding personalized surgical treatment strategies for patients with INMA.

Development and validation of a radiomic prediction model for TACC3 expression and prognosis in non-small cell lung cancer using contrast-enhanced CT imaging

Development and validation of a radiomic prediction model for TACC3 expression and prognosis in non-small cell lung cancer using contrast-enhanced CT imaging

Deep Learning-Based Automatic Segmentation Combined with Radiomics to Predict Post-TACE Liver Failure in HCC Patients.

To develop and validate a deep learning-based automatic segmentation model and combine with radiomics to predict post-TACE liver failure (PTLF) in hepatocellular carcinoma (HCC) patients. This was a retrospective study enrolled 210 TACE-trated HCC patients. Automatic segmentation model based on nnU-Net neural network was developed to segment medical images and assessed by the Dice similarity coefficient (DSC). The screened clinical and radiomics variables were separately used to developed clinical and radiomics predictive model, and were combined through multivariate logistic regression analysis to develop a combined predictive model. The area under the curve (AUC), calibration curve, and decision curve analysis (DCA) were applied to compare the performance of the three predictive models. The automatic segmentation model showed satisfactory segmentation performance with an average DSC of 83.05% for tumor segmentation and 92.72% for non-tumoral liver parenchyma segmentation. The international normalized ratio (INR) and albumin (ALB) was identified as clinically independent predictors for PTLF and used to develop clinical predictive model. Ten most valuable radiomics features, including 8 from non-tumoral liver parenchyma and 2 from tumor, were selected to develop radiomics predictive model and to calculate Radscore. The combined predictive model achieved the best and significantly improved predictive performance (AUC: 0.878) compared to the clinical predictive model (AUC: 0.785) and the radiomics predictive model (AUC: 0.815). This reliable combined predictive model can accurately predict PTLF in HCC patients, which can be a valuable reference for doctors in making suitable treatment plan.

Machine learning-based radiomics for guiding lymph node dissection in clinical stage I lung adenocarcinoma: a multicenter retrospective study.

Preoperative assessment of lymph node status is critical in managing lung cancer, as it directly impacts the surgical approach and treatment planning. However, in clinical stage I lung adenocarcinoma (LUAD), determining lymph node metastasis (LNM) is often challenging due to the limited sensitivity of conventional imaging modalities, such as computed tomography (CT) and positron emission tomography/CT (PET/CT). This study aimed to establish an effective radiomics prediction model using multicenter data for early assessment of LNM risk in patients with clinical stage I LUAD. The goal is to provide a basis for formulating lymph node dissection strategies before surgery in early-stage lung cancer patients. A total of 578 patients with LUAD from three medical centers [Cancer Hospital, Chinese Academy of Medical Sciences (CCAM), the First Affiliated Hospital of Chongqing Medical University (1CMU), and Beijing Chao-Yang Hospital (BCYH)] who underwent preoperative chest CT were divided into three groups, the training group (n=336), the testing group (n=167), and the independent validation group (n=75). The records of 1,316 radiomics features of each primary tumor were extracted. The least absolute shrinkage and selection operator (LASSO) analysis and multivariable logistic regression were used to reduce the data dimensionality, select features, and construct the prediction models. In the training group, the area under the curve (AUC) for the clinical model, radiomics model, and composite model were 0.820, 0.871, and 0.883, respectively. In the testing group, the AUC for the clinical model, radiomics model, and composite model were 0.897, 0.915, and 0.934, respectively. In the validation set, the AUC of the radiomics model was the highest at 0.870, while the composite model and clinical model had AUCs of 0.841 and 0.710, respectively. The results of the Delong test showed that the AUCs of the radiomics model and composite model were significantly higher than those of the clinical model in both the training and validation groups. The decision curve analysis showed that the radiomics nomogram was clinically useful. This study developed and validated a radiomics prediction model, which enables easy LNM prediction in stage I LUAD patients. This model provides a basis for formulating lymph node dissection strategies before surgery and helps to better determine the tumor node metastasis stage of the early-stage LUAD.

Intranodular and perinodular ultrasound radiomics distinguishes benign and malignant thyroid nodules: a multicenter study.

Ultrasound based radiomics prediction model can improve the differentiation ability of benign and malignant thyroid nodules to avoid overtreatment. This study evaluates the role of predictive models based on intranodular and perinodular ultrasound radiomics in distinguishing between benign and malignant thyroid nodules. A total of 1,076 thyroid nodules were enrolled from three hospitals between 2016 and 2022, forming the training, validation and test cohorts. The clinical signature (Clinic_Sig) was developed based on clinical information and conventional morphological features of ultrasound. Expanding 1 pixel, 3 pixels, 5 pixels, 7 pixels, and 9 pixels outward from the thyroid nodule, six radiomics models were constructed using intranodular (intra) and combined radiomics (intranodular and perinodular: +p1,+p3,+p5,+p7,+p9) features. The model with the best area under the curve (AUC) was defined as radiomics signature (Rad_Sig). The combined model was constructed from Clinic_Sig and Rad_Sig. AUC and calibration curves were used to evaluate the predictive performance of the model. Decision curve analysis (DCA) was used to evaluate the clinical net benefit of the model. The intra+p1 radiomics model exhibited the highest efficacy (AUC =0.863) in the test cohort, which was combined with Clinic_Sig to construct the combined model. Compared with Clinic_Sig and Rad_Sig, the combined model showed the higher predictive performance, with AUCs of 0.942 (training), 0.894 (validation), and 0.933 (test). The calibration curve showed that the predicted probabilities of the combined model were in good agreement with the actual probabilities, and DCA indicated that it provided more net benefit than the treat-none or treat-all scheme. The combined model based on clinical signatures, intranodular and perinodular ultrasound radiomics has the potential to effectively predict benign or malignant thyroid nodules.

An Application of Machine-Learning-Oriented Radiomics Model in Clear Cell Renal Cell Carcinoma (ccRCC) Early Diagnosis.

Aims/Background Clear cell renal cell carcinoma (ccRCC) is a common and aggressive form of kidney cancer, where early diagnosis is crucial for improving prognosis and treatment outcomes. Radiomics, which utilizes machine learning techniques, presents a promising approach in medical imaging for the early detection and characterization of such conditions. This study aims to explore the clinical utility of a machine-learning-based radiomics model in the early diagnosis of ccRCC. Methods Case data and abdominal computed tomography (CT) tumour images of patients with ccRCC were obtained from The Cancer Imaging Archive (TCIA) database. The dataset included 31 cases in the training set (19 males and 12 females, with an average age of 58.1 years) and 13 cases in the validation set (8 males and 5 females, with an average age of 69.6 years). The volume of interest (VOI) was manually delineated, slice by slice, along the tumour's edge in cross-sectional images of ccRCC. Radiomics features were extracted from each region of interest (ROI) using the "PyRadiomics" plug-in in 3D Slicer software (version 5.1.0, Massachusetts Institute of Technology and Brigham and Women's Hospital, Boston, MA, USA). Feature selection was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis, followed by 10-fold cross-validation. The selected radiomics features were then used to construct prediction models based on two different supervised machine learning algorithms: logistic regression and random forest. The diagnostic performance of these models was evaluated using receiver operating characteristic (ROC) curves and calibration curves. Finally, clinical data were integrated with the radiomics features to enhance the prediction model. Results A total of 44 radiomics features were ultimately selected to establish the prediction model based on the training set results. Among the two machine learning models, the logistic regression model demonstrated superior diagnostic performance. An evaluation of model establishment, considering both individual radiomics features (DifferenceVariance, JointEnergy.1, JointEntropy.2, MeanAbsoluteDeviation.7, SmallAreaHighGrayLevelEmphasis.7) and clinical data, indicated that the logistic regression model was stable and exhibited strong diagnostic performance, good calibration, and clinical applicability in patients with ccRCC. When clinical data were combined with radiomics features in the model, the area under the curve (AUC) reached 0.969, with an optimal threshold of -2.290, and sensitivity and specificity values of 89.3% and 95.2%, respectively. The calibration curve also confirmed that the logistic regression model had high calibration accuracy and greater clinical application value. Conclusion This machine-learning-based radiomics prediction model demonstrated significant value in the early diagnosis of clear cell renal cell carcinoma (ccRCC).

Histopathological correlations of CT-based radiomics imaging biomarkers in native kidney biopsy

BackgroundKidney biopsy is the standard of care for the diagnosis of various kidney diseases. In particular, chronic histopathologic lesions, such as interstitial fibrosis and tubular atrophy, can provide prognostic information regarding chronic kidney disease progression. In this study, we aimed to evaluate historadiological correlations between CT-based radiomic features and chronic histologic changes in native kidney biopsies and to construct and validate a radiomics-based prediction model for chronicity grade.MethodsWe included patients aged ≥ 18 years who underwent kidney biopsy and abdominal CT scan within a week before kidney biopsy. Left kidneys were three-dimensionally segmented using a deep learning model based on the 3D Swin UNEt Transformers architecture. We additionally defined isovolumic cortical regions of interest near the lower pole of the left kidneys. Shape, first-order, and high-order texture features were extracted after resampling and kernel normalization. Correlations and diagnostic metrics between extracted features and chronic histologic lesions were examined. A machine learning-based radiomic prediction model for moderate chronicity was developed and compared according to the segmented regions of interest (ROI).ResultsOverall, moderate correlations with statistical significance (P < 0.05) were found between chronic histopathologic grade and top-ranked radiomic features. Total parenchymal features were more strongly correlated than cortical ROI features, and texture features were more highly ranked. However, conventional imaging markers, including kidney length, were poorly correlated. Top-ranked individual radiomic features had areas under receiver operating characteristic curves (AUCs) of 0.65 to 0.74. Developed radiomics models for moderate-to-severe chronicity achieved AUCs of 0.89 (95% confidence interval [CI] 0.75–0.99) and 0.74 (95% CI 0.52–0.93) for total parenchymal and cortical ROI features, respectively.ConclusionSignificant historadiological correlations were identified between CT-based radiomic features and chronic histologic changes in native kidney biopsies. Our findings underscore the potential of CT-based radiomic features and their prediction model for the non-invasive assessment of kidney fibrosis.