Computed Tomography Radiomic Features Research Articles

Learning and depicting lobe-based radiomics feature for COPD Severity staging in low-dose CT images

BackgroundChronic obstructive pulmonary disease (COPD) is a prevalent and debilitating respiratory condition that imposes a significant healthcare burden worldwide. Accurate staging of COPD severity is crucial for patient management and treatment planning.MethodsThe retrospective study included 530 hospital patients. A lobe-based radiomics method was proposed to classify COPD severity using computed tomography (CT) images. First, we segmented the lung lobes with a convolutional neural network model. Secondly, the radiomic features of each lung lobe are extracted from CT images, the features of the five lung lobes are merged, and the selection of features is accomplished through the utilization of a variance threshold, t-Test, least absolute shrinkage and selection operator (LASSO). Finally, the COPD severity was classified by a support vector machine (SVM) classifier.Results104 features were selected for staging COPD according to the Global initiative for chronic Obstructive Lung Disease (GOLD). The SVM classifier showed remarkable performance with an accuracy of 0.63. Moreover, an additional set of 132 features were selected to distinguish between milder (GOLD I + GOLD II) and more severe instances (GOLD III + GOLD IV) of COPD. The accuracy for SVM stood at 0.87.ConclusionsThe proposed method proved that the novel lobe-based radiomics method can significantly contribute to the refinement of COPD severity staging. By combining radiomic features from each lung lobe, it can obtain a more comprehensive and rich set of features and better capture the CT radiomic features of the lung than simply observing the lung as a whole.

An unsupervised learning model based on CT radiomics features accurately predicts axillary lymph node metastasis in breast cancer patients: diagnostic study.

The accuracy of traditional clinical methods for assessing the metastatic status of axillary lymph nodes (ALNs) is unsatisfactory. In this study, the authors propose the use of radiomic technology and three-dimensional (3D) visualization technology to develop an unsupervised learning model for predicting axillary lymph node metastasis in patients with breast cancer (BC), aiming to provide a new method for clinical axillary lymph node assessment in patients with this disease. In this study, we retrospectively analyzed the data of 350 patients with invasive BC who underwent lung-enhanced computed tomography (CT) and axillary lymph node dissection surgery at the Department of Breast Surgery of the Second Xiangya Hospital of Central South University. The authors used 3D visualization technology to create a 3D atlas of ALNs and identified the region of interest for the lymph nodes. Radiomic features were subsequently extracted and selected, and a prediction model for ALNs was constructed using the K-means unsupervised algorithm. To validate the model, the authors prospectively collected data from 128 BC patients who were clinically evaluated as negative at our center. Using 3D visualization technology, we extracted and selected a total of 36 CT radiomics features. The unsupervised learning model categorized 1737 unlabeled lymph nodes into two groups, and the analysis of the radiomic features between these groups indicated potential differences in lymph node status. Further validation with 1397 labeled lymph nodes demonstrated that the model had good predictive ability for axillary lymph node status, with an area under the curve of 0.847 (0.825-0.869). Additionally, the model's excellent predictive performance was confirmed in the 128 axillary clinical assessment negative cohort (cN0) and the 350 clinical assessment positive (cN+) cohort, for which the correct classification rates were 86.72 and 87.43%, respectively, which were significantly greater than those of clinical assessment methods. The authors created an unsupervised learning model that accurately predicts the status of ALNs. This approach offers a novel solution for the precise assessment of ALNs in patients with BC.

The relationship between contrast-enhanced computed tomography radiomics features and mitosis karyorrhexis index in neuroblastoma

ObjectiveMitosis karyorrhexis index (MKI) can reflect the proliferation status of neuroblastoma cells. This study aimed to investigate the contrast-enhanced computed tomography (CECT) radiomics features associated with the MKI status in neuroblastoma.Materials and methods246 neuroblastoma patients were retrospectively included and divided into three groups: low-MKI, intermediate-MKI, and high-MKI. They were randomly stratified into a training set and a testing set at a ratio of 8:2. Tumor regions of interest were delineated on arterial-phase CECT images, and radiomics features were extracted. After reducing the dimensionality of the radiomics features, a random forest algorithm was employed to establish a three-class classification model to predict MKI status.ResultsThe classification model consisted of 5 radiomics features. The mean area under the curve (AUC) of the classification model was 0.916 (95% confidence interval (CI) 0.913–0.921) in the training set and 0.858 (95% CI 0.841–0.864) in the testing set. Specifically, the classification model achieved AUCs of 0.928 (95% CI 0.927–0.934), 0.915 (95% CI 0.912–0.919), and 0.901 (95% CI 0.900–0.909) for predicting low-MKI, intermediate-MKI, and high-MKI, respectively, in the training set. In the testing set, the classification model achieved AUCs of 0.873 (95% CI 0.859–0.882), 0.860 (95% CI 0.852–0.872), and 0.820 (95% CI 0.813–0.839) for predicting low-MKI, intermediate-MKI, and high-MKI, respectively.ConclusionsCECT radiomics features were found to be correlated with MKI status and are helpful for reflecting the proliferation status of neuroblastoma cells.

Development and validation of a 18F-FDG PET/CT radiomics nomogram for predicting progression free survival in locally advanced cervical cancer: a retrospective multicenter study

BackgroundThe existing staging system cannot meet the needs of accurate survival prediction. Accurate survival prediction for locally advanced cervical cancer (LACC) patients who have undergone concurrent radiochemotherapy (CCRT) can improve their treatment management. Thus, this present study aimed to develop and validate radiomics models based on pretreatment 18Fluorine-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)-computed tomography (CT) images to accurately predict the prognosis in patients.MethodsThe data from 190 consecutive patients with LACC who underwent pretreatment 18F-FDG PET-CT and CCRT at two cancer hospitals were retrospectively analyzed; 176 patients from the same hospital were randomly divided into training (n = 117) and internal validation (n = 50) cohorts. Clinical features were selected from the training cohort using univariate and multivariate Cox proportional hazards models; radiomic features were extracted from PET and CT images and filtered using least absolute shrinkage and selection operator and Cox proportional hazard regression. Three prediction models and a nomogram were then constructed using the previously selected clinical, CT and PET radiomics features. The external validation cohort that was used to validate the models included 23 patients with LACC from another cancer hospital. The predictive performance of the constructed models was evaluated using receiver operator characteristic curves, Kaplan Meier curves, and a nomogram.ResultsIn total, one clinical, one PET radiomics, and three CT radiomics features were significantly associated with progression-free survival in the training cohort. Across all three cohorts, the combined model displayed better efficacy and clinical utility than any of these parameters alone in predicting 3-year progression-free survival (area under curve: 0.661, 0.718, and 0.775; C-index: 0.698, 0.724, and 0.705, respectively) and 5-year progression-free survival (area under curve: 0.661, 0.711, and 0.767; C-index, 0.698, 0.722, and 0.676, respectively). On subsequent construction of a nomogram, the calibration curve demonstrated good agreement between actually observed and nomogram-predicted values.ConclusionsIn this study, a clinico-radiomics prediction model was developed and successfully validated using an independent external validation cohort. The nomogram incorporating radiomics and clinical features could be a useful clinical tool for the early and accurate assessment of long-term prognosis in patients with LACC patients who undergo concurrent chemoradiotherapy.

CT radiomics based on different machine learning models for classifying gross tumor volume and normal liver tissue in hepatocellular carcinoma

Background & aimsThe present study utilized extracted computed tomography radiomics features to classify the gross tumor volume and normal liver tissue in hepatocellular carcinoma by mainstream machine learning methods, aiming to establish an automatic classification model.MethodsWe recruited 104 pathologically confirmed hepatocellular carcinoma patients for this study. GTV and normal liver tissue samples were manually segmented into regions of interest and randomly divided into five-fold cross-validation groups. Dimensionality reduction using LASSO regression. Radiomics models were constructed via logistic regression, support vector machine (SVM), random forest, Xgboost, and Adaboost algorithms. The diagnostic efficacy, discrimination, and calibration of algorithms were verified using area under the receiver operating characteristic curve (AUC) analyses and calibration plot comparison.ResultsSeven screened radiomics features excelled at distinguishing the gross tumor area. The Xgboost machine learning algorithm had the best discrimination and comprehensive diagnostic performance with an AUC of 0.9975 [95% confidence interval (CI): 0.9973–0.9978] and mean MCC of 0.9369. SVM had the second best discrimination and diagnostic performance with an AUC of 0.9846 (95% CI: 0.9835– 0.9857), mean Matthews correlation coefficient (MCC)of 0.9105, and a better calibration. All other algorithms showed an excellent ability to distinguish between gross tumor area and normal liver tissue (mean AUC 0.9825, 0.9861,0.9727,0.9644 for Adaboost, random forest, logistic regression, naivem Bayes algorithm respectively).ConclusionCT radiomics based on machine learning algorithms can accurately classify GTV and normal liver tissue, while the Xgboost and SVM algorithms served as the best complementary algorithms.

Improving the Reproducibility of Computed Tomography Radiomic Features Using an Enhanced Hierarchical Feature Synthesis Network

Improving the Reproducibility of Computed Tomography Radiomic Features Using an Enhanced Hierarchical Feature Synthesis Network

Impact of slice thickness on reproducibility of CT radiomic features of lung tumors

Background Radiomics posits that quantified characteristics from radiographic images reflect underlying pathophysiology. Lung cancer (LC) is one of the prevalent forms of cancer, causing mortality. Slice thickness (ST) of computed tomography (CT) images is a crucial factor influencing the generalizability of radiomic features (RF) in oncology. There is scarcity of research that how ST affects variability of RF in LC. The present study helps in identifying the specific RF categories affected by variations in ST and provides valuable insights for researchers and clinicians working with RF in the field of LC.Hence, aim of the study is to evaluate influence of ST on reproducibility of CT-RF for lung tumors. Methods This is a prospective study, 32 patients with confirmed histopathological diagnosis of lung tumors were included. Contrast Enhanced CT (CECT) thorax was performed using a 128- Incisive CT (Philips Health Care). The image acquisition was performed with 5-mm and 2 mm STwas reconstructed retrospectively. RF were extracted from the CECT thorax images of both ST. We conducted a paired t-test to evaluate the disparity in RF between the two thicknesses. Lin’s Concordance Correlation Coefficient (CCC) was performed to identify the reproducibility of RF between the two thicknesses. Results Out of 107 RF, 66 (61.6%) exhibited a statistically significant distinction (p<0.05) when comparing two ST and while 41 (38.3%) RF did not show significant distinction (p>0.05) between the two ST measurements. 29 features (CCC ≥ 0.90) showed excellent to moderate reproducibility, and 78 features (CCC ≤ 0.90) showed poor reproducibility. Among the 7 RF categories, the shape-based features (57.1%) showed the maximum reproducibility whereas NGTDM-based features showed negligible reproducibility. Conclusions The ST had a notable impact on the majority of CT-RF of lung tumors. Shape based features (57.1%). First order (44.4%) features showed highest reproducibility compared to other RF categories.

CT radiomics in the identification of preoperative understaging in patients with clinical stage T1-2N0 esophageal squamous cell carcinoma.

Predicting preoperative understaging in patients with clinical stage T1-2N0 (cT1-2N0) esophageal squamous cell carcinoma (ESCC) is critical to customizing patient treatment. Radiomics analysis can provide additional information that reflects potential biological heterogeneity based on computed tomography (CT) images. However, to the best of our knowledge, no studies have focused on identifying CT radiomics features to predict preoperative understaging in patients with cT1-2N0 ESCC. Thus, we sought to develop a CT-based radiomics model to predict preoperative understaging in patients with cT1-2N0 esophageal cancer, and to explore the value of the model in disease-free survival (DFS) prediction. A total of 196 patients who underwent radical surgery for cT1-2N0 ESCC were retrospectively recruited from two hospitals. Among the 196 patients, 134 from Peking University Cancer Hospital were included in the training cohort, and 62 from Henan Cancer Hospital were included in the external validation cohort. Radiomics features were extracted from patients' CT images. Least absolute shrinkage and selection operator (LASSO) regression was used for feature selection and model construction. A clinical model was also built based on clinical characteristics, and the tumor size [the length, thickness and the thickness-to-length ratio (TLR)] was evaluated on the CT images. A radiomics nomogram was established based on multivariate logistic regression. The diagnostic performance of the models in predicting preoperative understaging was assessed by the area under the receiver operating characteristic curve (AUC). Kaplan-Meier curves with the log-rank test were employed to analyze the correlation between the nomogram and DFS. Of the patients, 50.0% (67/134) and 51.6% (32/62) were understaged in the training and validation groups, respectively. The radiomics scores and the TLRs of the tumors were included in the nomogram. The AUCs of the nomogram for predicting preoperative understaging were 0.874 [95% confidence interval (CI): 0.815-0.933] in the training cohort and 0.812 (95% CI: 0.703-0.912) in the external validation cohort. The diagnostic performance of the nomogram was superior to that of the clinical model (P<0.05). The nomogram was an independent predictor of DFS in patients with cT1-2N0 ESCC. The proposed CT-based radiomics model could be used to predict preoperative understaging in patients with cT1-2N0 ESCC who have undergone radical surgery.

Development and validation of a radiomics nomogram for diagnosis of malignant pleural effusion

ObjectiveWe aimed to develop a radiomics nomogram based on computed tomography (CT) scan features and high-throughput radiomics features for diagnosis of malignant pleural effusion (MPE).MethodsIn this study, 507 eligible patients with PE (207 malignant and 300 benign) were collected retrospectively. Patients were divided into training (n = 355) and validation cohorts (n = 152). Radiomics features were extracted from initial unenhanced CT images. CT scan features of PE were also collected. We used the variance threshold algorithm and least absolute shrinkage and selection operator (LASSO) to select optimal features to build a radiomics model for predicting the nature of PE. Univariate and multivariable logistic regression analyzes were used to identify significant independent factors associated with MPE, which were then included in the radiomics nomogram.ResultsA total of four CT features were retained as significant independent factors, including massive PE, obstructive atelectasis or pneumonia, pleural thickening > 10 mm, and pulmonary nodules and/or masses. The radiomics nomogram constructed from 13 radiomics parameters and four CT features showed good predictive efficacy in training cohort [area under the curve (AUC) = 0.926, 95% CI 0.894, 0.951] and validation cohort (AUC = 0.916, 95% CI 0.860, 0.955). The calibration curve and decision curve analysis showed that the nomogram helped differentiate MPE from benign pleural effusion (BPE) in clinical practice.ConclusionThis study presents a nomogram model incorporating CT scan features and radiomics features to help physicians differentiate MPE from BPE.

The value of radiomics-based CT combined with machine learning in the diagnosis of occult vertebral fractures

PurposeTo develop and evaluate the performance of radiomics-based computed tomography (CT) combined with machine learning algorithms in detecting occult vertebral fractures (OVFs).Materials and methods128 vertebrae including 64 with OVF confirmed by magnetic resonance imaging and 64 corresponding control vertebrae from 57 patients who underwent chest/abdominal CT scans, were included. The CT radiomics features on mid-axial and mid-sagittal plane of each vertebra were extracted. The fractured and normal vertebrae were randomly divided into training set and validation set at a ratio of 8:2. Pearson correlation analyses and least absolute shrinkage and selection operator were used for selecting sagittal and axial features, respectively. Three machine-learning algorithms were used to construct the radiomics models based on the residual features. Receiver operating characteristic (ROC) analysis was used to verify the performance of model.ResultsFor mid-axial CT imaging, 6 radiomics parameters were obtained and used for building the models. The logistic regression (LR) algorithm showed the best performance with area under the ROC curves (AUC) of training and validation sets of 0.682 and 0.775. For mid-sagittal CT imaging, 5 parameters were selected, and LR algorithms showed the best performance with AUC of training and validation sets of 0.832 and 0.882. The LR model based on sagittal CT yielded the best performance, with an accuracy of 0.846, sensitivity of 0.846, and specificity of 0.846.ConclusionMachine learning based on CT radiomics features allows for the detection of OVFs, especially the LR model based on the radiomics of sagittal imaging, which indicates it is promising to further combine with deep learning to achieve automatic recognition of OVFs to reduce the associated secondary injury.

Impact of slice thickness on reproducibility of CT radiomic features of lung tumors

Background: Radiomics, a field of research, relies on the theory that quantified characteristics from radiographic images would reflect underlying pathophysiology. Lung cancer continues to stand as one of the prevalent and well-known forms of cancer, causing mortality. The slice thickness (ST) of computed tomography (CT) images would be key concern regarding generalizability of radiomic features (RF) results in oncology. There is scarcity of research that has delved into how ST affects variability of RF in lung tumors. Hence, aim of the study is to evaluate influence of ST on reproducibility of CT-RF for lung tumors. Methods: This is a prospective study, 32 patients with confirmed histopathological diagnosis of lung tumors were included. Contrast Enhanced CT (CECT) thorax was performed using a 128- Incisive CT (Philips Health Care). The image acquisition was performed with 5-mm and 2 mm ST, and was reconstructed retrospectively. RF were extracted from the CECT thorax images of 5-mm and 2-mm ST. We conducted a paired t-test to evaluate the disparity in RF between the two thicknesses. Lin’s Concordance Correlation Coefficient (CCC) was performed to identify the reproducibility of RF between the two thicknesses. Results: Out of 107 RF extracted, 66 (61.6%) exhibited a statistically significant distinction (p&lt;0.05) when comparing two slice thicknesses and while 41 (38.3%) RF did not show significant distinction (p&gt;0.05) between the two ST measurements. 29 features (CCC ≥ 0.90) showed excellent to moderate reproducibility, and 78 features (CCC ≤ 0.90) showed poor reproducibility. Among the 7 RF categories, the shape-based features (57.1%) showed the maximum reproducibility whereas NGTDM-based features showed negligible reproducibility. Conclusions: The slice thickness had a notable impact on the majority of CT-RF of lung tumors. Shape based features (57.1%). First order (44.4%) features showed highest reproducibility compared to other RF categories.

Using CT Radiomics and Machine Learning to Distinguish Radiotherapy vs. Immune Checkpoint Inhibitor-Induced Pneumonitis in Non-Small Cell Lung Cancer Patients

Both radiotherapy (RT) and immune checkpoint inhibitors (ICIs) may induce pneumonitis which is difficult to distinguish clinically. We tested the utility of computed tomography (CT) in identifying the cause of pneumonitis in patients with non-small cell lung cancer (NSCLC) who received RT, ICIs and RT combined with ICIs (RT+ICIs). In this study, we analyzed the clinical characteristics and CT morphology of 130 NSCLC patients who developed pneumonitis after receipt of ICIs only (n = 50), thoracic RT only (n = 50) (ICIs only + RT only, the training cohort, n = 100), and RT+ICIs (the test cohort, n = 30). CT radiomics features were extracted using 3D-Slicer software and selected using LASSO regression for training a random forest (RF) classifier and a linear discriminant analysis (LDA) classifier to discern pneumonitis etiology. Patients who received RT or ICIs only were defined as the training cohort since the cause of their pneumonitis was readily identifiable. Patients who received RT+ICIs were defined as the test cohort and the cause of their pneumonitis were determined by clinicians as the ground truth. In the study, the median time from the initiation of treatment to pneumonitis was 1.9 months in RT+ICIs patients, which was shorter than 2.5 months in RT only and 2.7 months in ICIs only patients. In addition, the patients in RT+ICIs group have more high grade (grade 3-4) pneumonitis compared to patients in ICIs only or RT only group (p<0.05). Pneumonitis after the combined therapy was not a simple superposition mode of RP and CIP, resulting in the distinct characteristics of both RT and ICIs-related pneumonitis. (1) Some cases resembled RP with sharp borders in the ipsilateral lung and others resembled the bilateral distribution of typical CIP, while some patients exhibit an intermediate morphology; (2) Some pneumonitis occurred in the areas of low and medium doses of radiation rather than the high-dose radiation areas similar to RP; (3) Some pneumonitis was initially limited to the ipsilateral lung that had received radiation and then developed in bilateral lungs. The RF classifier showed favorable discrimination between CIP and RP with an area under the receiver operating curve (AUC) of 0.859 (95% CI: 0.7884-0.9292) in the training cohort and 0.851 (95% CI: 0.6998-1) in the test cohort. The simplified LDA classifier consisting of 7 LASSO-selected features achieved an AUC of 0.881 (95% CI: 0.815-0.947) in the training cohort and 0.842 (95% CI: 0.686-0.997) in the test cohort. RT and ICIs-related pneumonitis exhibit distinct clinical characteristics and morphologic features on CT. CT radiomics-based classifiers could provide a noninvasive method to identify the predominant etiology in NSCLC patients who developed pneumonitis after thoracic RT and ICIs.

Association between Contrast-Enhanced Computed Tomography Radiomic Features, Genomic Alterations and Prognosis in Advanced Lung Adenocarcinoma Patients.

Non-invasive methods to assess mutational status, as well as novel prognostic biomarkers, are warranted to foster therapy personalization of patients with advanced non-small cell lung cancer (NSCLC). This study investigated the association of contrast-enhanced Computed Tomography (CT) radiomic features of lung adenocarcinoma lesions, alone or integrated with clinical parameters, with tumor mutational status (EGFR, KRAS, ALK alterations) and Overall Survival (OS). In total, 261 retrospective and 48 prospective patients were enrolled. A Radiomic Score (RS) was created with LASSO-Logistic regression models to predict mutational status. Radiomic, clinical and clinical-radiomic models were trained on retrospective data and tested (Area Under the Curve, AUC) on prospective data. OS prediction models were trained and tested on retrospective data with internal cross-validation (C-index). RS significantly predicted each alteration at training (radiomic and clinical-radiomic AUC 0.95-0.98); validation performance was good for EGFR (AUC 0.86), moderate for KRAS and ALK (AUC 0.61-0.65). RS was also associated with OS at univariate and multivariable analysis, in the latter with stage and type of treatment. The validation C-index was 0.63, 0.79, and 0.80 for clinical, radiomic, and clinical-radiomic models. The study supports the potential role of CT radiomics for non-invasive identification of gene alterations and prognosis prediction in patients with advanced lung adenocarcinoma, to be confirmed with independent studies.

Integration of Multimodal Computed Tomography Radiomic Features of Primary Tumors and the Spleen to Predict Early Recurrence in Patients with Postoperative Adjuvant Transarterial Chemoembolization.

Hepatocellular carcinoma (HCC) is one of the most lethal malignancies in the world. Patients with HCC choose postoperative adjuvant transarterial chemoembolization (PA-TACE) after surgical resection to reduce the risk of recurrence. However, many of them have recurrence within a short period. In this retrospective analysis, a total of 173 patients who underwent PA-TACE between September 2016 and March 2020 were recruited. Radiomic features were derived from the arterial and venous phases of each patient. Early recurrence (ER)-related radiomics features of HCC and the spleen were selected to build two rad-scores using the least absolute shrinkage and selection operator (LASSO) Cox regression analysis. Logistic regression was applied to establish the Radiation (Rad)_score by combining the two regions. We constructed a nomogram containing clinical information and dual-region rad-scores, which was evaluated in terms of discrimination, calibration, and clinical usefulness. All three radiological scores showed good performance for ER prediction. The combined Rad_score performed the best, with an area under the curve (AUC) of 0.853 (95% confidence interval [CI], 0.783-0.908) in the training set and 0.929 (95% CI, 0.789-0.988) in the validation set. Multivariate analysis identified total bilirubin (TBIL) and the combined Rad_score as independent prognostic factors for ER. The nomogram was found to be clinically valuable, as determined by the decision curves (DCA) and clinical impact curves (CIC). A multimodal dual-region radiomics model combining HCC and the spleen is an independent prognostic tool for ER. The combination of dual-region radiomics features and clinicopathological factors has a good clinical application value.

Predictive value of radiomics analysis of enhanced CT for three-tiered microvascular invasion grading in hepatocellular carcinoma.

Microvascular invasion (MVI) is a major risk factor, for recurrence and metastasis of hepatocellular carcinoma (HCC) after radical surgery and liver transplantation. However, its diagnosis depends on the pathological examination of the resected specimen after surgery; therefore, predicting MVI before surgery is necessary to provide reference value for clinical treatment. Meanwhile, predicting only the existence of MVI is not enough, as it ignores the degree, quantity, and distribution of MVI and may lead to MVI-positive patients suffering due to inappropriate treatment. Although some studies have involved M2 (high risk of MVI), majority have adopted the binary classification method or have not included radiomics. To develop three-class classification models for predicting the grade of MVI of HCC by combining enhanced computed tomography radiomics features with clinical risk factors. The data of 166 patients with HCC confirmed by surgery and pathology were analyzed retrospectively. The patients were divided into the training (116 cases) and test (50 cases) groups at a ratio of 7:3. Of them, 69 cases were MVI positive in the training group, including 45 cases in the low-risk group (M1) and 24 cases in the high-risk group (M2), and 47 cases were MVI negative (M0). In the training group, the optimal subset features were obtained through feature selection, and the arterial phase radiomics model, portal venous phase radiomics model, delayed phase radiomics model, three-phase radiomics model, clinical imaging model, and combined model were developed using Linear Support Vector Classification. The test group was used for validation, and the efficacy of each model was evaluated through the receiver operating characteristic curve (ROC). The clinical imaging features of MVI included alpha-fetoprotein, tumor size, tumor margin, peritumoral enhancement, intratumoral artery, and low-density halo. The area under the curve (AUC) of the ROC values of the clinical imaging model for M0, M1, and M2 were 0.831, 0.701, and 0.847, respectively, in the training group and 0.782, 0.534, and 0.785, respectively, in the test group. After combined radiomics analyis, the AUC values for M0, M1, and M2 in the test group were 0.818, 0.688, and 0.867, respectively. The difference between the clinical imaging model and the combined model was statistically significant (p=0.029). The clinical imaging model and radiomics model developed in this study had a specific predictive value for HCC MVI grading, which can provide precise reference value for preoperative clinical diagnosis and treatment. The combined application of the two models had a high predictive efficacy.

Computed tomography radiomic features and clinical factors predicting the response to first transarterial chemoembolization in intermediate-stage hepatocellular carcinoma

BackgroundAccording to clinical practice guidelines, transarterial chemoembolization (TACE) is the standard treatment modality for patients with intermediate-stage hepatocellular carcinoma (HCC). Early prediction of treatment response can help patients choose a reasonable treatment plan. This study aimed to investigate the value of the radiomic-clinical model in predicting the efficacy of the first TACE treatment for HCC to prolong patient survival. MethodsA total of 164 patients with HCC who underwent the first TACE from January 2017 to September 2021 were analyzed. The tumor response was assessed by modified response evaluation criteria in solid tumors (mRECIST), and the response of the first TACE to each session and its correlation with overall survival were evaluated. The radiomic signatures associated with the treatment response were identified by the least absolute shrinkage and selection operator (LASSO), and four machine learning models were built with different types of regions of interest (ROIs) (tumor and corresponding tissues) and the model with the best performance was selected. The predictive performance was assessed with receiver operating characteristic (ROC) curves and calibration curves. ResultsOf all the models, the random forest (RF) model with peritumor (+10 mm) radiomic signatures had the best performance [area under ROC curve (AUC) = 0.964 in the training cohort, AUC = 0.949 in the validation cohort]. The RF model was used to calculate the radiomic score (Rad-score), and the optimal cutoff value (0.34) was calculated according to the Youden's index. Patients were then divided into a high-risk group (Rad-score > 0.34) and a low-risk group (Rad-score ≤ 0.34), and a nomogram model was successfully established to predict treatment response. The predicted treatment response also allowed for significant discrimination of Kaplan-Meier curves. Multivariate Cox regression identified six independent prognostic factors for overall survival, including male [hazard ratio (HR) = 0.500, 95% confidence interval (CI): 0.260–0.962, P = 0.038], alpha-fetoprotein (HR = 1.003, 95% CI: 1.002–1.004, P < 0.001), alanine aminotransferase (HR = 1.003, 95% CI: 1.001–1.005, P = 0.025), performance status (HR = 2.400, 95% CI: 1.200–4.800, P = 0.013), the number of TACE sessions (HR = 0.870, 95% CI: 0.780–0.970, P = 0.012) and Rad-score (HR = 3.480, 95% CI: 1.416–8.552, P = 0.007). ConclusionsThe radiomic signatures and clinical factors can be well-used to predict the response of HCC patients to the first TACE and may help identify the patients most likely to benefit from TACE.

Prediction model based on radiomics and clinical features for preoperative lymphovascular invasion in gastric cancer patients.

Background: We explored whether a model based on contrast-enhanced computed tomography radiomics features and clinicopathological factors can evaluate preoperative lymphovascular invasion (LVI) in patients with gastric cancer (GC) with Lauren classification. Methods: Based on clinical and radiomic characteristics, we established three models: Clinical + Arterial phase_Radcore, Clinical + Venous phase_Radcoreand a combined model. The relationship between Lauren classification and LVI was analyzed using a histogram. Results: We retrospectively analyzed 495 patients with GC. The areas under the curve of the combined model were 0.8629 and 0.8343 in the training and testing datasets, respectively. The combined model showed a superior performance to the other models. Conclusion: CECT-based radiomics models can effectively predict preoperative LVI in GC patients with Lauren classification.

Prognostic value of computed tomography radiomics features in patients with gastric neuroendocrine neoplasm.

The present study aimed to investigate the clinical prognostic significance of radiomics signature (R-signature) in patients with gastric neuroendocrine neoplasm (GNEN). A retrospective study of 182 patients with GNEN who underwent dual-phase enhanced computed tomography (CT) scanning was conducted. LASSO-Cox regression analysis was used to screen the features and establish the arterial, venous and the arteriovenous phase combined R-signature, respectively. The association between the optimal R-signature with the best prognostic performance and overall survival (OS) was assessed in the training cohort and verified in the validation cohort. Univariate and multivariate Cox regression analysis were used to identify the significant factors of clinicopathological characteristics for OS. Furthermore, the performance of a combined radiomics-clinical nomogram integrating the R-signature and independent clinicopathological risk factors was evaluated. The arteriovenous phase combined R-signature had the best performance in predicting OS, and its C-index value was better than the independent arterial and venous phase R-signature (0.803 vs 0.784 and 0.803 vs 0.756, P<0.001, respectively). The optimal R-signature was significantly associated with OS in the training cohort and validation cohort. GNEN patients could be successfully divided into high and low prognostic risk groups with radiomics score median. The combined radiomics-clinical nomogram combining this R-signature and independent clinicopathological risk factors (sex, age, treatment methods, T stage, N stage, M stage, tumor boundary, Ki67, CD56) exhibited significant prognostic superiority over clinical nomogram, R-signature alone, and traditional TNM staging system (C-index, 0.882 vs 0.861, 882 vs 0.803, and 0.882 vs 0.870 respectively, P<0.001). All calibration curves showed remarkable consistency between predicted and actual survival, and decision curve analysis verified the usefulness of the combined radiomics-clinical nomogram for clinical practice. The R-signature could be used to stratify patients with GNEN into high and low risk groups. Furthermore, the combined radiomics-clinical nomogram provided better predictive accuracy than other predictive models and might aid clinicians with therapeutic decision-making and patient counseling.

#4646 CT-DERIVED RADIOMICS ANALYSIS OF DIABETIC NEPHROPATHY BY MACHINE LEARNING MODELS

Abstract Background and Aims Kidney radiomics has been used to develop more accurate diagnostic tools of renal tumor and predict outcomes. However, radiomics studies for diabetic kidney disease (DKD) remain few. In this light, we hypothesized that computed tomography (CT) radiomics features could differentiate DKD from normal kidneys and assess the severity of DKD. Method This retrospective study included 343 subjects with type 2 diabetes mellitus (T2DM) (male 65.5%, mean age 63.6±14.8) and 90 healthy controls (HC) (male 34.4%, mean age 41.9±8.5) who underwent abdominal CT. Whole volumetric CT data of both kidneys were automatically extracted using a deep-learning based model and radiomics features were extracted. T2DM were categorized into three groups according to eGFR (mL/min/1.73 m^2) (group 1, eGFR &amp;gt; 60; group 2, 15 ≤ eGFR ≤ 60; and group 3, eGFR &amp;lt;15). The capability of CT radiomics features to distinguish not only DKD from HC but also various DKD groups based on eGFR, was evaluated using machine learning models. Results A total of 1,723 radiomics features were extracted from the volumetric CT data of both kidneys. A combination of LASSO filter and random forest showed best performance in differentiating between HC and DKD, with an area under curve (AUC) of 0.98 and accuracy of 95.7%. It also showed an excellent performance in differentiating between HC and DKD group 1 with AUC of 0.97 and accuracy of 91.7%, and also it was able to differentiate between DKD groups (AUCs &amp;gt; 0.78). CT radiomics features and eGFR had a moderate degree of correlation (R=-0.50 to R=0.63, all P&amp;lt;.00). Conclusion CT-derived radiomics analysis of kidneys using machine learning models can be useful in differentiating DKD from normal kidneys as well as assessing the severity of DKD.

Limited impact of discretization/interpolation parameters on the predictive power of CT radiomic features in a surgical cohort of pancreatic cancer patients.

To explore the variation of the discriminative power of CT (Computed Tomography) radiomic features (RF) against image discretization/interpolation in predicting early distant relapses (EDR) after upfront surgery. Data of 144 patients with pre-surgical high contrast CT were processed consistently with IBSI (Image Biomarker Standardization Initiative) guidelines. Image interpolation/discretization parameters were intentionally changed, including cubic voxel size (0.21-27 mm3) and binning (32-128 grey levels) in a 15 parameter's sets. After excluding RF with poor inter-observer delineation agreement (ICC < 0.80) and not negligible inter-scanner variability, the variation of 80 RF against discretization/interpolation was first quantified. Then, their ability in classifying patients with early distant relapses (EDR, < 10months, previously assessed at the first quartile value of time-to-relapse) was investigated in terms of AUC (Area Under Curve) variation for those RF significantly associated to EDR. Despite RF variability against discretization/interpolation parameters was large and only 30/80 RF showed %COV < 20 (%COV = 100*STDEV/MEAN), AUC changes were relatively limited: for 30 RF significantly associated with EDR (AUC values around 0.60-0.70), the mean values of SD of AUC variability and AUC range were 0.02 and 0.05 respectively. AUC ranges were between 0.00 and 0.11, with values ≤ 0.05 in 16/30 RF. These variations were further reduced when excluding the extreme values of 32 and 128 for grey levels (Average AUC range 0.04, with values between 0.00 and 0.08). The discriminative power of CT RF in the prediction of EDR after upfront surgery for pancreatic cancer is relatively invariant against image interpolation/discretization within a large range of voxel sizes and binning.