Quantitative computed tomography texture analysis for estimating histological subtypes of thymic epithelial tumors

Comparison of 2D and 3D radiomics features with conventional features based on contrast-enhanced CT images for preoperative prediction the risk of thymic epithelial tumors

BACKGROUND

This study aimed to develop and validate 2-Dimensional (2D) and 3-Dimensional (3D) radiomics signatures based on contrast-enhanced computed tomography (CECT) images for preoperative prediction of the thymic epithelial tumors (TETs) risk and compare the predictive performance with conventional CT features.

PATIENTS AND METHODS

149 TET patients were retrospectively enrolled from January 2016 to December 2018, and divided into high-risk group (B2/B3/TCs, n = 103) and low-risk group (A/AB/B1, n = 46). All patients were randomly assigned into the training (n = 104) and testing (n = 45) set. 14 conventional CT features were collected, and 396 radiomic features were extracted from 2D and 3D CECT images, respectively. Three models including conventional, 2D radiomics and 3D radiomics model were established using multivariate logistic regression analysis. The discriminative performances of the models were demonstrated by receiver operating characteristic (ROC) curves.

RESULTS

In the conventional model, area under the curves (AUCs) in the training and validation sets were 0.863 and 0.853, sensitivity was 78% and 55%, and specificity was 88% and 100%, respectively. The 2D model yielded AUCs of 0.854 and 0.834, sensitivity of 86% and 77%, and specificity of 72% and 86% in the training and validation sets. The 3D model revealed AUC of 0.902 and 0.906, sensitivity of 75% and 68%, and specificity of 94% and 100% in the training and validation sets.

CONCLUSIONS

Radiomics signatures based on 3D images could distinguish high-risk from low-risk TETs and provide complementary diagnostic information.

CT radiomics and human-machine hybrid system for differentiating mediastinal lymphomas from thymic epithelial tumors

BACKGROUND

It is difficult for radiologists, especially junior radiologists with limited experience to make differential diagnoses between mediastinal lymphomas and thymic epithelial tumors (TETs) due to the overlapping imaging features. The purpose of this study was to develop and validate a CT-based clinico-radiomics model for differentiating lymphomas from TETs and to investigate whether a human-machine hybrid system can assist junior radiologists in improving their diagnostic performance.

METHODS

The patients who underwent contrast-enhanced chest CT and pathologically confirmed with lymphoma or TET at two centers from January 2011 to December 2019 and from January 2017 to December 2021 were retrospectively included and split as training/validation set and external test set, respectively. Clinical and radiomic signatures were pre-selected by elastic-net, and the models were established with the selected signatures using ensemble learning. Three radiologists independently reviewed CT images and assessed each case of the external test set with knowledge of the relevant clinical information. The diagnoses of reader 1, reader 2, and reader 3 were compared with those of the models in the external test set and further separately input to the model's ensemble process as a human-machine system to make final decisions in the external test set. The improvement of diagnostic performance of radiologists by human-machine system was evaluated by the area under the receiver operating characteristic curve and increase rate.

RESULTS

A total of 95 patients (51 with lymphomas and 44 with TETs) at Center 1 and 94 (52 with lymphomas and 42 with TETs) at Center 2 were enrolled and divided into training/validation sets and external test set, respectively. The diagnostic performance of the clinico-radiomics model has outperformed the junior radiologists and senior radiologist in AUC (clinico-radiomics model: 0.85 (0.76,0.92); reader 2: 0.70 (0.60,0.80); reader 3: 0.60 (0.49,0.71), reader 1: 0.76 (0.66,0.86), respectively) in the external test set. The human-machine hybrid system demonstrated significant increases in AUC (reader 1 + model: 0.87 (0.79,0.94), an increase of 14%; reader 2 + model: 0.86 (0.77,0.93), an increase of 23%; reader 3 + model: 0.84 (0.76,0.91), an increase of 40%), compared to the human performance alone.

CONCLUSIONS

The clinico-radiomics model outperformed three radiologists in differentiating lymphomas from TETs on CT. The use of the human-machine hybrid system significantly improved the performance of radiologists, especially junior radiologists. It provides a real-time decision tool to reduce bias and mistakes in radiologist diagnosis and enhances the diagnostic confidence of junior radiologists. This attempt may lead to more human-machine hybrid systems being explored in the diagnosis of different diseases to drive future clinical applications.

Differentiating low-risk thymomas from high-risk thymomas: preoperative radiomics nomogram based on contrast enhanced CT to minimize unnecessary invasive thoracotomy

BACKGROUND

This study was designed to develop a combined radiomics nomogram to preoperatively predict the risk categorization of thymomas based on contrast-enhanced computed tomography (CE-CT) images.

MATERIALS

The clinical and CT data of 178 patients with thymoma (100 patients with low-risk thymomas and 78 patients with high-risk thymomas) collected in our hospital from March 2018 to July 2023 were retrospectively analyzed. The patients were randomly divided into a training set (n = 125) and a validation set (n = 53) in a 7:3 ratio. Qualitative radiological features were recorded, including (a) tumor diameter, (b) location, (c) shape, (d) capsule integrity, (e) calcification, (f) necrosis, (g) fatty infiltration, (h) lymphadenopathy, and (i) enhanced CT value. Radiomics features were extracted from each CE-CT volume of interest (VOI), and the least absolute shrinkage and selection operator (LASSO) algorithm was performed to select the optimal discriminative ones. A combined radiomics nomogram was further established based on the clinical factors and radiomics scores. The differentiating efficacy was determined using receiver operating characteristic (ROC) analysis.

RESULTS

Only one clinical factor (incomplete capsule) and seven radiomics features were found to be independent predictors and were used to establish the radiomics nomogram. In differentiating low-risk thymomas (types A, AB, and B1) from high-risk ones (types B2 and B3), the nomogram demonstrated better diagnostic efficacy than any single model, with the respective area under the curve (AUC), accuracy, sensitivity, and specificity of 0.974, 0.921, 0.962 and 0.900 in the training cohort, 0.960, 0.892, 0923 and 0.897 in the validation cohort, respectively. The calibration curve showed good agreement between the prediction probability and actual clinical findings.

CONCLUSIONS

The nomogram incorporating clinical factors and radiomics features provides additional value in differentiating the risk categorization of thymomas, which could potentially be useful in clinical practice for planning personalized treatment strategies.

Computed tomography radiomic feature analysis of thymic epithelial tumors: Differentiation of thymic epithelial tumors from thymic cysts and prediction of histological subtypes

PURPOSE

To investigate the value of computed tomography (CT) radiomic feature analysis for the differential diagnosis between thymic epithelial tumors (TETs) and thymic cysts, and prediction of histological subtypes of TETs.

MATERIALS AND METHODS

Twenty-four patients with TETs (13 low-risk and 9 high-risk thymomas, and 2 thymic carcinomas) and 12 with thymic cysts were included in this study. For each lesion, the radiomic features of a volume of interest covering the lesion were extracted from non-contrast enhanced CT images. The Least Absolute Shrinkage and Selection Operator (Lasso) method was used for the feature selection. Predictive models for differentiating TETs from thymic cysts (model A), and high risk thymomas + thymic carcinomas from low risk thymomas (model B) were created from the selected features. The receiver operating characteristic curve was used to evaluate the effectiveness of radiomic feature analysis for differentiating among these tumors.

RESULTS

In model A, the selected 5 radiomic features for the model A were NGLDM_Contrast, GLCM_Correlation, GLZLM_SZLGE, DISCRETIZED_HISTO_Entropy_log2, and DISCRETIZED_HUmin. In model B, sphericity was the only selected feature. The area under the curve, sensitivity, and specificity of radiomic feature analysis were 1 (95% confidence interval [CI]: 1-1), 100%, and 100%, respectively, for differentiating TETs from thymic cysts (model A), and 0.76 (95%CI: 0.53-0.99), 64%, and 100% respectively, for differentiating high-risk thymomas + thymic carcinomas from low-risk thymomas (model B).

CONCLUSION

CT radiomic analysis could be utilized as a non-invasive imaging technique for differentiating TETs from thymic cysts, and high-risk thymomas + thymic carcinomas from low-risk thymomas.

Quantitative CT parameters combined with preoperative systemic inflammatory markers for differentiating risk subgroups of thymic epithelial tumors

BACKGROUND

Thymic epithelial tumors (TETs) are the most common primary neoplasms of the anterior mediastinum. Different risk subgroups of TETs have different prognosis and therapeutic strategies, therefore, preoperative identification of different risk subgroups is of high clinical significance. This study aims to explore the diagnostic efficiency of quantitative computed tomography (CT) parameters combined with preoperative systemic inflammatory markers in differentiating low-risk thymic epithelial tumors (LTETs) from high-risk thymic epithelial tumors (HTETs).

METHODS

74 Asian patients with TETs confirmed by biopsy or postoperative pathology between January 2013 and October 2022 were collected retrospectively and divided into two risk subgroups: LTET group (type A, AB and B1 thymomas) and HTET group (type B2, B3 thymomas and thymic carcinoma). Statistical analysis were performed between the two groups in terms of quantitative CT parameters and preoperative systemic inflammatory markers. Multivariate logistic regression analysis was used to determine the independent predictors of risk subgroups of TETs. The area under curve (AUC) and optimal cut-off values were calculated by receiver operating characteristic (ROC) curves.

RESULTS

47 TETs were in LTET group, while 27 TETs were in HTET group. In addition to tumor size and CT value of the tumor on plain scan, there were statistical significance comparing in CT value of the tumor on arterial phase (CTv-AP) and venous phase (CTv-VP), and maximum enhanced CT value (CEmax) of the tumor between the two groups (for all, P < 0.05). For systemic inflammatory markers, HTET group was significantly higher than LTET group (for all, P < 0.05), including platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR) and systemic immune-inflammation index (SII). Multivariate logistic regression analysis showed that NLR (odds ratio [OR] = 2.511, 95% confidence interval [CI]: 1.322-4.772, P = 0.005), CTv-AP (OR = 0.939, 95%CI: 0.888-0.994, P = 0.031) and CTv-VP (OR = 0.923, 95%CI: 0.871-0.979, P = 0.008) were the independent predictors of risk subgroups of TETs. The AUC value of 0.887 for the combined model was significantly higher than NLR (0.698), CTv-AP (0.800) or CTv-VP (0.811) alone. The optimal cut-off values for NLR, CTv-AP and CTv-VP were 2.523, 63.44 Hounsfeld Unit (HU) and 88.29HU, respectively.

CONCLUSIONS

Quantitative CT parameters and preoperative systemic inflammatory markers can differentiate LTETs from HTETs, and the combined model has the potential to improve diagnostic efficiency and to help the patient management.

Development and validation of a deep learning radiomics nomogram for preoperatively differentiating thymic epithelial tumor histologic subtypes

OBJECTIVES

Using contrast-enhanced computed tomography (CECT) and deep learning technology to develop a deep learning radiomics nomogram (DLRN) to preoperative predict risk status of patients with thymic epithelial tumors (TETs).

METHODS

Between October 2008 and May 2020, 257 consecutive patients with surgically and pathologically confirmed TETs were enrolled from three medical centers. We extracted deep learning features from all lesions using a transformer-based convolutional neural network and created a deep learning signature (DLS) using selector operator regression and least absolute shrinkage. The predictive capability of a DLRN incorporating clinical characteristics, subjective CT findings and DLS was evaluated by the area under the curve (AUC) of a receiver operating characteristic curve.

RESULTS

To construct a DLS, 25 deep learning features with non-zero coefficients were selected from 116 low-risk TETs (subtypes A, AB, and B1) and 141 high-risk TETs (subtypes B2, B3, and C). The combination of subjective CT features such as infiltration and DLS demonstrated the best performance in differentiating TETs risk status. The AUCs in the training, internal validation, external validation 1 and 2 cohorts were 0.959 (95% confidence interval [CI]: 0.924-0.993), 0.868 (95% CI: 0.765-0.970), 0.846 (95% CI: 0.750-0.942), and 0.846 (95% CI: 0.735-0.957), respectively. The DeLong test and decision in curve analysis revealed that the DLRN was the most predictive and clinically useful model.

CONCLUSIONS

The DLRN comprised of CECT-derived DLS and subjective CT findings showed a high performance in predicting risk status of patients with TETs.

CLINICAL RELEVANCE STATEMENT

Accurate risk status assessment of thymic epithelial tumors (TETs) may aid in determining whether preoperative neoadjuvant treatment is necessary. A deep learning radiomics nomogram incorporating enhancement CT-based deep learning features, clinical characteristics, and subjective CT findings has the potential to predict the histologic subtypes of TETs, which can facilitate decision-making and personalized therapy in clinical practice.

KEY POINTS

• A non-invasive diagnostic method that can predict the pathological risk status may be useful for pretreatment stratification and prognostic evaluation in TET patients. • DLRN demonstrated superior performance in differentiating the risk status of TETs when compared to the deep learning signature, radiomics signature, or clinical model. • The DeLong test and decision in curve analysis revealed that the DLRN was the most predictive and clinically useful in differentiating the risk status of TETs.

Development and Validation of Multi-Omics Thymoma Risk Classification Model Based on Transfer Learning

The paper aims to develop prediction model that integrates clinical, radiomics, and deep features using transfer learning to stratifying between high and low risk of thymoma. Our study enrolled 150 patients with thymoma (76 low-risk and 74 high-risk) who underwent surgical resection and pathologically confirmed in Shengjing Hospital of China Medical University from January 2018 to December 2020. The training cohort consisted of 120 patients (80%) and the test cohort consisted of 30 patients (20%). The 2590 radiomics and 192 deep features from non-enhanced, arterial, and venous phase CT images were extracted and ANOVA, Pearson correlation coefficient, PCA, and LASSO were used to select the most significant features. A fusion model that integrated clinical, radiomics, and deep features was developed with SVM classifiers to predict the risk level of thymoma, and accuracy, sensitivity, specificity, ROC curves, and AUC were applied to evaluate the classification model. In both the training and test cohorts, the fusion model demonstrated better performance in stratifying high and low risk of thymoma. It had AUCs of 0.99 and 0.95, and an accuracy of 0.93 and 0.83, respectively. This was compared to the clinical model (AUCs of 0.70 and 0.51, accuracy of 0.68 and 0.47), the radiomics model (AUCs of 0.97 and 0.82, accuracy of 0.93 and 0.80), and the deep model (AUCs of 0.94 and 0.85, accuracy of 0.88 and 0.80). The fusion model integrating clinical, radiomics and deep features based on transfer learning was efficient for noninvasively stratifying high risk and low risk of thymoma. The models could help to determine surgery strategy for thymoma cancer.

Conventional and radiomic features to predict pathology in the preoperative assessment of anterior mediastinal masses

OBJECTIVES

The aim of this study was to differentiate benign from malignant tumors in the anterior mediastinum based on computed tomography (CT) imaging characteristics, which could be useful in preoperative planning. Additionally, our secondary aim was to differentiate thymoma from thymic carcinoma, which could guide the use of neoadjuvant therapy.

MATERIALS AND METHODS

Patients referred for thymectomy were retrospectively selected from our database. Twenty-five conventional characteristics were evaluated by visual analysis, and 101 radiomic features were extracted from each CT. In the step of model training, we applied support vector machines to train classification models. Model performance was assessed using the area under the receiver operating curves (AUC).

RESULTS

Our final study sample comprised 239 patients, 59 (24.7 %) with benign mediastinal lesions and 180 (75.3 %) with malignant thymic tumors. Among the malignant masses, there were 140 (58.6 %) thymomas, 23 (9.6 %) thymic carcinomas, and 17 (7.1 %) non-thymic lesions. For the benign versus malignant differentiation, the model that integrated both conventional and radiomic features achieved the highest diagnostic performance (AUC = 0.715), in comparison to the conventional (AUC = 0.605) and radiomic-only (AUC = 0.678) models. Similarly, regarding thymoma versus thymic carcinoma differentiation, the model that integrated both conventional and radiomic features also achieved the highest diagnostic performance (AUC = 0.810), in comparison to the conventional (AUC = 0.558) and radiomic-only (AUC = 0.774) models.

CONCLUSION

CT-based conventional and radiomic features with machine learning analysis could be useful for predicting pathologic diagnoses of anterior mediastinal masses. The diagnostic performance was moderate for differentiating benign from malignant lesions and good for differentiating thymomas from thymic carcinomas. The best diagnostic performance was achieved when both conventional and radiomic features were integrated in the machine learning algorithms.

Predicting the Risk of Thymic Tumors Using Texture Analysis of Contrast-Enhanced Chest Computed Tomography

OBJECTIVE

This study aimed to explore the value of contrast-enhanced computed tomography texture features for predicting the risk of malignant thymic epithelial tumor.

METHODS

Data of 97 patients with pathologically confirmed thymic epithelial tumors treated at in our hospital from March 2015 to October 2021 were retrospectively analyzed. Based on the World Health Organization classification of thymic epithelial tumors, patients were divided into a high-risk group (types B2, B3, and C; n = 45) and a low-risk group (types A, AB, and B1; n = 52). Texture analysis was performed using a first-order, gray-level histogram method. Six features were evaluated: mean, variance, skewness, kurtosis, energy, and entropy. The association between contrast-enhanced computed tomography texture features and the risk of malignancy in thymic epithelial tumors was analyzed. The predictive thresholds of predictive texture features were determined by receiver operating characteristics analysis.

RESULTS

The mean, skewness, and entropy were significantly greater in the high-risk group than in the low-risk group (P < 0.05); however, variance, kurtosis, and energy were comparable in the two groups (P > 0.05). The area under curve of mean, skewness, and entropy was 0.670, 0.760, and 0.880, respectively. The optimal cutoff value of entropy for predicting risk of malignancy was 7.74, with sensitivity, specificity, and accuracy of 80.0%, 80.0%, and 75%, respectively.

CONCLUSIONS

Contrast-enhanced computed tomography texture features, especially entropy, may be a useful tool to predict the risk of malignancy in thymic epithelial tumors.

Radiomics Analysis of Multiphasic Computed Tomography Images for Distinguishing High-Risk Thymic Epithelial Tumors From Low-Risk Thymic Epithelial Tumors

OBJECTIVES

The objective of this study is to preoperatively investigate the value of multiphasic contrast-enhanced computed tomography (CT)-based radiomics signatures for distinguishing high-risk thymic epithelial tumors (HTET) from low-risk thymic epithelial tumors (LTET) compared with conventional CT signatures.

MATERIALS AND METHODS

Pathologically confirmed 305 thymic epithelial tumors (TETs), including 147 LTET (Type A/AB/B1) and 158 HTET (Type B2/B3/C), were retrospectively analyzed, and were randomly divided into training (n = 214) and validation cohorts (n = 91). All patients underwent nonenhanced, arterial contrast-enhanced, and venous contrast-enhanced CT analysis. The least absolute shrinkage and selection operator regression with 10-fold cross-validation was performed for radiomic models building, and multivariate logistic regression analysis was performed for radiological and combined models building. The performance of the model was evaluated by the area under the receiver operating characteristic curve (AUC of ROC), and the AUCs were compared using the Delong test. Decision curve analysis was used to evaluate the clinical value of each model. Nomogram and calibration curves were plotted for the combined model.

RESULTS

The AUCs for radiological model in the training and validation cohorts were 0.756 and 0.733, respectively. For nonenhanced, arterial contrast-enhanced, venous contrast-enhanced CT and 3-phase images combined radiomics models, the AUCs were 0.940, 0.946, 0.960, and 0.986, respectively, in the training cohort, whereas 0.859, 0.876, 0.930, and 0.923, respectively, in the validation cohort. The combined model, including CT morphology and radiomics signature, showed AUCs of 0.990 and 0.943 in the training and validation cohorts, respectively. Delong test and decision curve analysis showed that the predictive performance and clinical value of the 4 radiomics models and combined model were greater than the radiological model ( P < 0.05).

CONCLUSIONS

The combined model, including CT morphology and radiomics signature, greatly improved the predictive performance for distinguishing HTET from LTET. Radiomics texture analysis can be used as a noninvasive method for preoperative prediction of the pathological subtypes of TET.

Clinical radiomics-based machine learning versus three-dimension convolutional neural network analysis for differentiation of thymic epithelial tumors from other prevascular mediastinal tumors on chest computed tomography scan

Purpose

To compare the diagnostic performance of radiomic analysis with machine learning (ML) model with a convolutional neural network (CNN) in differentiating thymic epithelial tumors (TETs) from other prevascular mediastinal tumors (PMTs).

Methods

A retrospective study was performed in patients with PMTs and undergoing surgical resection or biopsy in National Cheng Kung University Hospital, Tainan, Taiwan, E-Da Hospital, Kaohsiung, Taiwan, and Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan between January 2010 and December 2019. Clinical data including age, sex, myasthenia gravis (MG) symptoms and pathologic diagnosis were collected. The datasets were divided into UECT (unenhanced computed tomography) and CECT (enhanced computed tomography) for analysis and modelling. Radiomics model and 3D CNN model were used to differentiate TETs from non-TET PMTs (including cyst, malignant germ cell tumor, lymphoma and teratoma). The macro F1-score and receiver operating characteristic (ROC) analysis were performed to evaluate the prediction models.

Result

In the UECT dataset, there were 297 patients with TETs and 79 patients with other PMTs. The performance of radiomic analysis with machine learning model using LightGBM with Extra Tree (macro F1-Score = 83.95%, ROC-AUC = 0.9117) had better performance than the 3D CNN model (macro F1-score = 75.54%, ROC-AUC = 0.9015). In the CECT dataset, there were 296 patients with TETs and 77 patients with other PMTs. The performance of radiomic analysis with machine learning model using LightGBM with Extra Tree (macro F1-Score = 85.65%, ROC-AUC = 0.9464) had better performance than the 3D CNN model (macro F1-score = 81.01%, ROC-AUC = 0.9275).

Conclusion

Our study revealed that the individualized prediction model integrating clinical information and radiomic features using machine learning demonstrated better predictive performance in the differentiation of TETs from other PMTs at chest CT scan than 3D CNN model.

Machine-learning classifiers based on non-enhanced computed tomography radiomics to differentiate anterior mediastinal cysts from thymomas and low-risk from high-risk thymomas: A multi-center study

Background

This study aimed to investigate the diagnostic value of machine-learning (ML) models with multiple classifiers based on non-enhanced CT Radiomics features for differentiating anterior mediastinal cysts (AMCs) from thymomas, and high-risk from low risk thymomas.

Methods

In total, 201 patients with AMCs and thymomas from three centers were included and divided into two groups: AMCs vs. thymomas, and high-risk vs low-risk thymomas. A radiomics model (RM) was built with 73 radiomics features that were extracted from the three-dimensional images of each patient. A combined model (CM) was built with clinical features and subjective CT finding features combined with radiomics features. For the RM and CM in each group, five selection methods were adopted to select suitable features for the classifier, and seven ML classifiers were employed to build discriminative models. Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic performance of each combination.

Results

Several classifiers combined with suitable selection methods demonstrated good diagnostic performance with areas under the curves (AUCs) of 0.876 and 0.922 for the RM and CM in group 1 and 0.747 and 0.783 for the RM and CM in group 2, respectively. The combination of support vector machine (SVM) as the feature-selection method and Gradient Boosting Decision Tree (GBDT) as the classification algorithm represented the best comprehensive discriminative ability in both group. Comparatively, assessments by radiologists achieved a middle AUCs of 0.656 and 0.626 in the two groups, which were lower than the AUCs of the RM and CM. Most CMs exhibited higher AUC value compared to RMs in both groups, among them only a few CMs demonstrated better performance with significant difference in group 1.

Conclusion

Our ML models demonstrated good performance for differentiation of AMCs from thymomas and low-risk from high-risk thymomas. ML based on non-enhanced CT radiomics may serve as a novel preoperative tool.

CT Radiomic Features for Predicting Resectability and TNM Staging in Thymic Epithelial Tumors

BACKGROUND

To explore the performance of a computed tomography based radiomics model in the preoperative prediction of resectability status and TNM staging in thymic epithelial tumors.

METHODS

We reviewed the last preoperative computed tomography scan of patients with thymic epithelial tumors prior to resection and pathology evaluation at our institution between February 2008 and June 2019. A total of 101 quantitative features were extracted and a radiomics model was trained using elastic net penalized logistic regressions for each aim. In the set-aside testing sets, discriminating performance of each model was assessed with area under receiver operating characteristic curve.

RESULTS

Our final population consisted of 243 patients with: 153 (87%) thymomas, 23 (9%) thymic carcinomas, and 9 (4%) thymic carcinoids. Incomplete resections (R1 or R2) occurred in 38 (16%) patients, and 67 (28%) patients had more advanced stage tumors (stage III or IV). In the set-aside testing sets, the radiomics model achieved good performance in preoperatively predicting incomplete resections (area under receiver operating characteristic curve: 0.80) and advanced stage tumors (area under receiver operating characteristic curve: 0.70).

CONCLUSIONS

Our computed tomography radiomics model achieved good performance to predict resectability status and staging in thymic epithelial tumors, suggesting a potential value for the evaluation of radiomic features in the preoperative prediction of surgical outcomes in thymic malignancies.

Computed tomography radiomics for the prediction of thymic epithelial tumor histology, TNM stage and myasthenia gravis

Objectives

To evaluate CT-derived radiomics for machine learning-based classification of thymic epithelial tumor (TET) stage (TNM classification), histology (WHO classification) and the presence of myasthenia gravis (MG).

Methods

Patients with histologically confirmed TET in the years 2000–2018 were retrospectively included, excluding patients with incompatible imaging or other tumors. CT scans were reformatted uniformly, gray values were normalized and discretized. Tumors were segmented manually; 15 scans were re-segmented after 2 weeks by two readers. 1316 radiomic features were calculated (pyRadiomics). Features with low intra-/inter-reader agreement (ICC<0.75) were excluded. Repeated nested cross-validation was used for feature selection (Boruta algorithm), model training, and evaluation (out-of-fold predictions). Shapley additive explanation (SHAP) values were calculated to assess feature importance.

Results

105 patients undergoing surgery for TET were identified. After applying exclusion criteria, 62 patients (28 female; mean age, 57±14 years; range, 22–82 years) with 34 low-risk TET (LRT; WHO types A/AB/B1), 28 high-risk TET (HRT; WHO B2/B3/C) in early stage (49, TNM stage I-II) or advanced stage (13, TNM III-IV) were included. 14(23%) of the patients had MG. 334(25%) features were excluded after intra-/inter-reader analysis. Discriminatory performance of the random forest classifiers was good for histology(AUC, 87.6%; 95% confidence interval, 76.3–94.3) and TNM stage(AUC, 83.8%; 95%CI, 66.9–93.4) but poor for the prediction of MG (AUC, 63.9%; 95%CI, 44.8–79.5).

Conclusions

CT-derived radiomic features may be a useful imaging biomarker for TET histology and TNM stage.

CT texture analysis in evaluation of thymic tumors and thymic hyperplasia: correlation with the international thymic malignancy interest group (ITMIG) stage and WHO grade

OBJECTIVES

To evaluate the effectiveness of CT texture analysis (CTTA) in (1) differentiating Thymoma (THY) from thymic hyperplasia (TH) (2) low from high WHO grade, and (3) low from high Masaoka Koga (MK)/International Thymic Malignancy Interest Group (ITMIG) stages.

METHODS

After institute ethical clearance, this cross-sectional study analyzed 26 patients (THY-18, TH-8) who underwent dual energy CT (DECT) and surgery between January 2016 and December 2018. CTTA was performed using TexRad (Feedback Medical Ltd., Cambridge, UK- www.fbkmed.com) by a single observer. Free hand regions of interest (ROIs) were placed over axial sections where there was maximum enhancement and homogeneity. Filtration histogram was used to generate six first-order texture parameters [mean, standard deviation (SD), mean of positive pixels (MPP), entropy, skewness, and kurtosis] at six spatial scaling factors "SSF 0, 2, 3, 4, 5, and 6". Mann-Whitney test was applied among various categories and p value < 0.05 was considered significant. Three-step feature selection was performed to determine the best parameters among each category.

RESULTS

The best performing parameters were (1) THY vs TH- Mean at "SSF 0" (AUC: 0.8889) and MPP at "SSF 0" (AUC: 0.8889), (2) Low vs high WHO grade - no parameter showed statistical significance with good AUC, and (3) Low vs high MK/ITMIG stage- SD at "SSF 6" (AUC: 0.8052 and 0.8333 respectively]).

CONCLUSION

CTTA revealed several parameters with excellent diagnostic performance in differentiating thymoma from thymic hyperplasia and MK/ITMIG high vs low stages. CTTA could potentially serve as a non-invasive tool for this stratification.

ADVANCES IN KNOWLEDGE

This study has employed texture analysis, a novel radiomics method on DECT scans to determine the best performing parameter and their corresponding cut-off values to differentiate among the above-mentioned categories. These new parameters may help add another layer of confidence to non-invasively stratify and prognosticate patients accurately which was only previously possible with a biopsy.

Quantification of Cancer-Developing Idiopathic Pulmonary Fibrosis Using Whole-Lung Texture Analysis of HRCT Images

Simple Summary

Idiopathic pulmonary fibrosis (IPF) patients have a significantly higher risk of developing lung cancer. Traditional risk factors including age, male gender, smoking status, and emphysema have been reported. However, there are only limited data on radiomics features from HRCT images useful for risk stratification of IPF patients for lung cancer. In this study, we found that texture-based radiomics features can be differentiated between IPF patients with and without cancer development, and their diagnostic accuracy is not inferior to that of traditional risk factors. By combining radiomics features and traditional risk factors, the diagnostic accuracy can be improved.

Abstract

Idiopathic pulmonary fibrosis (IPF) patients have a significantly higher risk of developing lung cancer (LC). There is only limited evidence of the use of texture-based radiomics features from high-resolution computed tomography (HRCT) images for risk stratification of IPF patients for LC. We retrospectively enrolled subjects who suffered from IPF in this study. Clinical data including age, gender, smoking status, and pulmonary function were recorded. Non-contrast chest CT for fibrotic score calculation and determination of three dimensional measures of whole-lung texture and emphysema were performed using a promising deep learning imaging platform. The results revealed that among 116 subjects with IPF (90 non-cancer and 26 lung cancer cases), the radiomics features showed significant differences between non-cancer and cancer patients. In the training cohort, the diagnostic accuracy using selected radiomics features with AUC of 0.66–0.73 (sensitivity of 80.0–85.0% and specificity of 54.2–59.7%) was not inferior to that obtained using traditional risk factors, such as gender, smoking status, and emphysema (%). In the validation cohort, the combination of radiomics features and traditional risk factors produced a diagnostic accuracy of 0.87 AUC and an accuracy of 75.0%. In this study, we found that whole-lung CT texture analysis is a promising tool for LC risk stratification of IPF patients.

Quantitative Histogram Analysis of T2-Weighted and Diffusion-Weighted Magnetic Resonance Images for Prediction of Malignant Thymic Epithelial Tumors

PURPOSE

To assess the value of histogram analysis for differentiating a high-risk thymic epithelial tumor (TET) from a low-risk TET using T2-weighted images and the apparent diffusion coefficient (ADC).

METHODS

Forty-nine patients with histopathologically proven TET after thymectomy were enrolled in this study and retrospectively classified as having low-risk TET (low-risk thymoma) or high-risk TET (high-risk thymoma or thymic carcinoma). Twelve parameters were obtained from the quantitative histogram analysis. The histogram parameters were compared using the Mann-Whitney U test. Diagnostic efficacy was estimated by receiver-operating characteristic curve analysis.

RESULTS

Twenty-five patients were classified as having low-risk TET and 24 as having high-risk TET. The mean ADC value showed diagnostic efficacy for differentiating high-risk TET from low-risk TET, with an area under the curve of 0.7, and was better than when using conventional methods alone.

CONCLUSION

The ADC-based histogram analysis could help to differentiate between high-risk and low-risk TETs.

CT imaging-based machine learning model: a potential modality for predicting low-risk and high-risk groups of thymoma: "Impact of surgical modality choice"

INTRODUCTION

Radiomics methods are used to analyze various medical images, including computed tomography (CT), magnetic resonance, and positron emission tomography to provide information regarding the diagnosis, patient outcome, tumor phenotype, and the gene-protein signatures of various diseases. In low-risk group, complete surgical resection is typically sufficient, whereas in high-risk thymoma, adjuvant therapy is usually required. Therefore, it is important to distinguish between both. This study evaluated the CT radiomics features of thymomas to discriminate between low- and high-risk thymoma groups.

MATERIALS AND METHODS

In total, 83 patients with thymoma were included in this study between 2004 and 2019. We used the Radcloud platform (Huiying Medical Technology Co., Ltd.) to manage the imaging and clinical data and perform the radiomics statistical analysis. The training and validation datasets were separated by a random method with a ratio of 2:8 and 502 random seeds. The histopathological diagnosis was noted from the pathology report.

RESULTS

Four machine-learning radiomics features were identified to differentiate a low-risk thymoma group from a high-risk thymoma group. The radiomics feature names were Energy, Zone Entropy, Long Run Low Gray Level Emphasis, and Large Dependence Low Gray Level Emphasis.

CONCLUSIONS

The results demonstrated that a machine-learning model and a multilayer perceptron classifier analysis can be used on CT images to predict low- and high-risk thymomas. This combination could be a useful preoperative method to determine the surgical approach for thymoma.

CT-Based Radiomics Signatures for Predicting the Risk Categorization of Thymic Epithelial Tumors

Objectives

This study aims to assess the performance of radiomics approaches based on 3D computed tomography (CT), clinical and semantic features in predicting the pathological classification of thymic epithelial tumors (TETs).

Methods

A total of 190 patients who underwent surgical resection and had pathologically confirmed TETs were enrolled in this retrospective study. All patients underwent non-contrast-enhanced CT (NECT) scans and contrast-enhanced CT (CECT) scans before treatment. A total of 396 hand-crafted radiomics features of each patient were extracted from the volume of interest in NECT and CECT images. We compared three clinical features and six semantic features (observed radiological traits) between patients with TETs. Two triple-classification radiomics models (RMs), two corresponding clinical RMs, and two corresponding clinical-semantic RMs were built to identify the types of the TETs. The area under the receiver operating characteristic curve (AUC) and accuracy (ACC) were useful to evaluate the different models.

Results

Of the 190 patients, 83 had low-risk thymoma, 58 had high-risk thymoma, and 49 had thymic carcinoma. Clinical features (Age) and semantic features (mediastinal fat infiltration, mediastinal lymph node enlargement, and pleural effusion) were significantly different among the groups(P < 0.001). In the validation set, the NECT-based clinical RM (AUC = 0.770 for low-risk thymoma, 0.689 for high-risk thymoma, and 0.783 for thymic carcinoma; ACC = 0.569) performed better than the CECT-based clinical-semantic RM (AUC = 0.785 for low-risk thymoma, 0.576 for high-risk thymoma, and 0.774 for thymic carcinoma; ACC = 0.483).

Conclusions

NECT-based and CECT-based RMs may provide a non-invasive method to distinguish low-risk thymoma, high-risk thymoma, and thymic carcinoma, and NECT-based RMs performed better.

Advances in Knowledge

Radiomics models may be used for the preoperative prediction of the pathological classification of TETs.

Development and validation of a CT-texture analysis nomogram for preoperatively differentiating thymic epithelial tumor histologic subtypes

BACKGROUND

Thymic epithelial tumors (TETs) are the most common primary tumors in the anterior mediastinum, which have considerable histologic heterogeneity. This study aimed to develop and validate a nomogram based on computed tomography (CT) and texture analysis (TA) for preoperatively predicting the pathological classifications for TET patients.

METHODS

Totally TET 172 patients confirmed by postoperative pathology between January 2011 to April 2019 were retrospectively analyzed and randomly divided into training (n = 120) and validation (n = 52) cohorts. Preoperative clinical factors, CT signs and texture features of each patient were analyzed, and prediction models were developed using the least absolute shrinkage and selection operator (LASSO) regression. The performance of the models was evaluated and compared by the area under receiver-operator characteristic (ROC) curve (AUC) and the DeLong test. The clinical application value of the models was determined via the decision curve analysis (DCA). Then, a nomogram was developed based on the model with the best predictive efficiency and clinical utility and validated using the calibration plots.

RESULTS

Totally 87 patients with low-risk TET (LTET) (types A, AB, B1) and 85 patients with high-risk TET (HTET) (types B2, B3, C) were enrolled in this study. We separately constructed 4 prediction models for differentiating LTET from HTET using clinical, CT, texture features, and their combination. These 4 prediction models achieved AUCs of 0.66, 0.79, 0.82, 0.88 in the training cohort and 0.64, 0.82, 0.86, 0.94 in the validation cohort, respectively. The DeLong test and DCA showed that the Combined model, consisting of 2 CT signs and 2 texture parameters, held the highest predictive efficiency and clinical utility (p < 0.05). A prediction nomogram was subsequently developed using the 4 independently risk factors from the Combined model. The calibration curves indicated a good consistency between the actual observations and nomogram predictions for differentiating TET classifications.

CONCLUSION

A prediction nomogram incorporating both the CT and texture parameters was constructed and validated in our study, which can be conveniently used for the preoperative individualized prediction of the simplified histologic subtypes in TET patients.

Can computed tomography-based radiomics potentially discriminate between anterior mediastinal cysts and type B1 and B2 thymomas?

Background

Anterior mediastinal cysts (AMC) are often misdiagnosed as thymomas and undergo surgical resection, which caused unnecessary treatment and medical resource waste. The purpose of this study is to explore potential possibility of computed tomography (CT)-based radiomics for the diagnosis of AMC and type B1 and B2 thymomas.

Methods

A group of 188 patients with pathologically confirmed AMC (106 cases misdiagnosed as thymomas in CT) and thymomas (82 cases) and underwent routine chest CT from January 2010 to December 2018 were retrospectively analyzed. The lesions were manually delineated using ITK-SNAP software, and radiomics features were performed using the artificial intelligence kit (AK) software. A total of 180 tumour texture features were extracted from enhanced CT and unenhanced CT, respectively. The general test, correlation analysis, and LASSO were used to features selection and then the radiomics signature (radscore) was obtained. The combined model including radscore and independent clinical factors was developed. The model performances were evaluated on discrimination, calibration curve.

Results

Two radscore models were constructed from the unenhanced and enhanced phases based on the selected four and three features, respectively. The AUC, sensitivity, and specificity of the enhanced radscore model were 0.928, 89.3%, and 83.8% in the training dataset and 0.899, 84.6%, and 87.5% in the test dataset (higher than the unenhanced radscore model). The combined model of enhanced CT including radiomics features and independent clinical factors yielded an AUC, sensitivity and specificity of 0.941, 82.1%, and 94.6% in the training dataset and 0.938, 92.3%, and 87.5% in the test dataset (higher than the unenhanced combined model and enhanced radscore model).

Conclusions

The study suggested the possibility that the combined model in enhanced CT provided a potential tool to facilitate the differential diagnosis of AMC and type B1 and B2 thymomas.

Applicability of radiomics in interstitial lung disease associated with systemic sclerosis: proof of concept

Objective

To retrospectively evaluate if texture-based radiomics features are able to detect interstitial lung disease (ILD) and to distinguish between the different disease stages in patients with systemic sclerosis (SSc) in comparison with mere visual analysis of high-resolution computed tomography (HRCT).

Methods

Sixty patients (46 females, median age 56 years) with SSc who underwent HRCT of the thorax were retrospectively analyzed. Visual analysis was performed by two radiologists for the presence of ILD features. Gender, age, and pulmonary function (GAP) stage was calculated from clinical data (gender, age, pulmonary function test). Data augmentation was performed and the balanced dataset was split into a training (70%) and a testing dataset (30%). For selecting variables that allow classification of the GAP stage, single and multiple logistic regression models were fitted and compared by using the Akaike information criterion (AIC). Diagnostic accuracy was evaluated from the area under the curve (AUC) from receiver operating characteristic (ROC) analyses, and diagnostic sensitivity and specificity were calculated.

Results

Values for some radiomics features were significantly lower (p < 0.05) and those of other radiomics features were significantly higher (p = 0.001) in patients with GAP2 compared with those in patients with GAP1. The combination of two specific radiomics features in a multivariable model resulted in the lowest AIC of 10.73 with an AUC of 0.96, 84% sensitivity, and 99% specificity. Visual assessment of fibrosis was inferior in predicting individual GAP stages (AUC 0.86; 83% sensitivity; 74% specificity).

Conclusion

The correlation of radiomics with GAP stage, but not with the visually defined features of ILD-HRCT, implies that radiomics might capture features indicating severity of SSc-ILD on HRCT, which are not recognized by visual analysis.

Key Points

• Radiomics features can predict GAP stage with a sensitivity of 84% and a specificity of almost 100%.

• Extent of fibrosis on HRCT and a combined model of different visual HRCT-ILD features perform worse in predicting GAP stage.

• The correlation of radiomics with GAP stage, but not with the visually defined features of ILD-HRCT, implies that radiomics might capture features on HRCT, which are not recognized by visual analysis.

Electronic supplementary material

The online version of this article (10.1007/s00330-020-07293-8) contains supplementary material, which is available to authorized users.

Computed tomography (CT)-derived radiomic features differentiate prevascular mediastinum masses as thymic neoplasms versus lymphomas

OBJECTIVES

We aimed to assess the ability of radiomics, applied to not-enhanced computed tomography (CT), to differentiate mediastinal masses as thymic neoplasms vs lymphomas.

METHODS

The present study was an observational retrospective trial. Inclusion criteria were pathology-proven thymic neoplasia or lymphoma with mediastinal localization, availability of CT. Exclusion criteria were age < 16 years and mediastinal lymphoma lesion < 4 cm. We selected 108 patients (M:F = 47:61, median age 48 years, range 17-79) and divided them into a training and a validation group. Radiomic features were used as predictors in linear discriminant analysis. We built different radiomic models considering segmentation software and resampling setting. Clinical variables were used as predictors to build a clinical model. Scoring metrics included sensitivity, specificity, accuracy and area under the curve (AUC). Wilcoxon paired test was used to compare the AUCs.

RESULTS

Fifty-five patients were affected by thymic neoplasia and 53 by lymphoma. In the validation analysis, the best radiomics model sensitivity, specificity, accuracy and AUC resulted 76.2 ± 7.0, 77.8 ± 5.5, 76.9 ± 6.0 and 0.84 ± 0.06, respectively. In the validation analysis of the clinical model, the same metrics resulted 95.2 ± 7.0, 88.9 ± 8.9, 92.3 ± 8.5 and 0.98 ± 0.07, respectively. The AUCs of the best radiomic and the clinical model not differed.

CONCLUSIONS

We developed and validated a CT-based radiomic model able to differentiate mediastinal masses on non-contrast-enhanced images, as thymic neoplasms or lymphoma. The proposed method was not affected by image postprocessing. Therefore, the present image-derived method has the potential to noninvasively support diagnosis in patients with prevascular mediastinal masses with major impact on management of asymptomatic cases.

Machine-learning-based computed tomography radiomic analysis for histologic subtype classification of thymic epithelial tumours

PURPOSE

To evaluate the performance of machine-learning-based computed tomography (CT) radiomic analysis to differentiate high-risk thymic epithelial tumours (TETs) from low-risk TETs according to the WHO classification.

METHOD

This retrospective study included 155 patients with a histologic diagnosis of high-risk TET (n = 72) and low-risk TET (n = 83) who underwent unenhanced CT (UECT) and contrast-enhanced CT (CECT). The radiomic features were extracted from the UECT and CECT of each patient at the largest cross-section of the lesion. The classification performance was evaluated with a nested leave-one-out cross-validation approach combining the least absolute shrinkage and selection operator feature selection and four classifiers: generalised linear model (GLM), k-nearest neighbor (KNN), support vector machine (SVM) and random forest (RF). The receiver-operating characteristic curve (ROC) and the area under the curve (AUC) were used to evaluate the performance of the classifiers.

RESULTS

The combination of UECT and CECT radiomic features demonstrated the best performance to differentiate high-risk TETs from low-risk TETs for all four classifiers. Among these classifiers, the RF had the highest AUC of 0.87, followed by GLM (AUC = 0.86), KNN (AUC = 0.86) and SVM (AUC = 0.84).

CONCLUSIONS

Machine learning-based CT radiomic analysis allows for the differentiation of high-risk TETs and low-risk TETs with excellent performance, representing a promising tool to assist clinical decision making in patients with TETs.

Application of CT texture analysis to assess the localization of primary aldosteronism

We performed present study to investigate whether the localization of primary aldosteronism (PA) can be predicted using quantitative texture analysis on unenhanced computed tomography (CT). Plain CT data of 82 PA patients (54 unilateral (right-sided:left-sided = 24:30), 28 bilateral) were analyzed retrospectively. After semi-automatically setting the region of interest to include the whole adrenal gland, texture analyses were performed with or without a Laplacian of Gaussian filter with various spatial scaling factors (SSFs). Logistic regression analysis was performed using the extracted histogram-based texture features to identify parameters capable of predicting excessive aldosterone production. The result of adrenal venous sampling served as gold standard in present study. As a result, logistic regression analysis indicated that the mean gray level intensity (p = 0.026), the mean value of the positive pixels (p = 0.003) in the unfiltered image, and entropy (p = 0.027) in the filtered image (SSF: 2 mm) were significant parameters. Using the model constructed by logistic regression analysis and the optimum cutoff value, the localization of PA (three multiple choices of left, right or bilateral) was determined with an accuracy of 67.1% (55/82). CT texture analysis may provide a potential avenue for less invasive prediction of the localization of PA.

Radiomics Signatures of Computed Tomography Imaging for Predicting Risk Categorization and Clinical Stage of Thymomas

Purpose

The aim of this study is to develop and compare performance of radiomics signatures using texture features extracted from noncontrast enhanced CT (NECT) and contrast enhanced CT (CECT) images for preoperative predicting risk categorization and clinical stage of thymomas.

Materials and Methods

Between January 2010 and October 2018, 199 patients with surgical resection and histopathologically confirmed thymoma were enrolled in this retrospective study. We extracted 841 radiomics features separately from volume of interest (VOI) in NECT and CECT images. The features with poor reproducibility and highly redundancy were removed. Then a least absolute shrinkage and selection operator method (LASSO) logistic regression model with 10-fold cross validation was used for further feature selection and radiomics signatures build. The predictive performances of radiomics signatures were assessed by receiver operating characteristic (ROC) analysis. The areas under the receiver operating characteristic curve (AUC) between radiomics signatures were compared by using Delong test.

Result

In differentiating high risk thymomas from low risk thymomas, the AUC, sensitivity, and specificity were 0.801(95% CI 0.740–0.863), 0.752 and 0.767 for radiomics signature based on NECT images, and 0.827 (95% CI 0.771 -0.884), 0.798, and 0.722 for radiomics signature based on CECT images. But there was no significant difference (p=0.365) between them. In differentiating advanced stage thymomas from early stage thymomas, the AUC, sensitivity, and specificity were 0.829 (95%CI 0.757-0.900), 0.712, and 0.806 for radiomics signature based on NECT images and 0.860 (95%CI 0.803-0.917), 0.699, and 0.889 for radiomics signature based on CECT images. There was no significant difference (p=0.069) between them. The accuracy was 0.819 for radiomics signature based on NECT images, 0.869 for radiomics signature based on CECT images, and 0.779 for radiologists. Both radiomics signatures had a better performance than radiologists. But there was significant difference (p = 0.025) only between CECT radiomics signature and radiologists.

Conclusion

Radiomics signatures based on texture analysis from NECT and CECT images could be utilized as noninvasive biomarkers for differentiating high risk thymomas from low risk thymomas and advanced stage thymomas from early stage thymoma. As a quantitative method, radiomics signature can provide complementary diagnostic information and help to plan personalized treatment for patients with thymomas.

CT-based radiomics features analysis for predicting the risk of anterior mediastinal lesions

Background

To retrospectively validate CT-based radiomics features for predicting the risk of anterior mediastinal lesions.

Methods

A retrospective study was performed through February 2013 to March 2018 on 298 patients who had pathologically confirmed anterior mediastinal lesions. The patients all underwent CT scans before their treatment, including 130 unenhanced computed tomography (UECT) and 168 contrast-enhanced CT (CECT) scans. The lesion areas were delineated, and a total of 1,029 radiomics features were extracted. The least absolute shrinkage and selection operator (Lasso) algorithm method was used to select the radiomics features significantly associated with discrimination of high-risk from low-risk lesions in the anterior mediastinum. Then, 8-fold and 3-fold cross-validation logistic regression (LR) models were taken as the feature selection classifiers to build the radiomics models for UECT and CECT scan respectively. The predictive performance of the radiomics features was evaluated based on the receiver operating characteristics (ROC) curve.

Results

Each of the two radiomics classifiers included the optimal 12 radiomic features. In terms of the area under ROC curve, using the radiomics model in discriminating high-risk lesions from the low-risks, CECT images accounted for 74.1% with a sensitivity of 66.67% and specificity of 64.81%. Meanwhile, UECT images were 84.2% with a sensitivity of 71.43% and specificity of 74.07%.

Conclusions

The association of the two proposed CT-based radiomics features with the discrimination of high and low-risk lesions in anterior mediastinum was confirmed, and the radiomics features of the UECT scan were proven to have better prediction performance than the CECT's in risk grading.

Texture analysis of myocardial infarction in CT: Comparison with visual analysis and impact of iterative reconstruction

OBJECTIVES

To compare texture analysis (TA) with subjective visual diagnosis of myocardial infarction (MI) in cardiac computed tomography (CT) and to evaluate the impact of iterative reconstruction (IR).

METHODS

Ten patients (4 women, mean age 68 ± 11 years) with confirmed chronic MI and 20 controls (8 women, mean age 52 ± 11 years) with no cardiac abnormality underwent contrast-enhanced cardiac CT with the same protocol. Images were reconstructed with filtered back projection (FBP) and with advanced modeled IR at strength levels 3-5. Subjective diagnosis of MI was made by three independent, blinded readers with different experience levels. Classification of MI was performed using machine learning-based decision tree models for the entire data set and after splitting into training and test data to avoid overfitting.

RESULTS

Subjective visual analysis for diagnosis of MI showed excellent intrareader (kappa: 0.93) but poor interreader agreement (kappa: 0.3), with variable performance at different image reconstructions. TA showed high performance for all image reconstructions (correct classifications: 94%-97%, areas under the curve: 0.94-0.99). After splitting into training and test data, overall lower performances were observed, with best results for IR at level 5 (correct classifications: 73%, area under the curve: 0.65).

CONCLUSIONS

As compared with subjective, nonreliable visual analysis of inexperienced readers, TA enables objective and reproducible diagnosis of chronic MI in cardiac CT with higher accuracy. IR has a considerable impact on both subjective and objective image analysis.

CT Enhancement and 3D Texture Analysis of Pancreatic Neuroendocrine Neoplasms

To evaluate pancreatic neuroendocrine neoplasms (panNENs) grade prediction by means of qualitative and quantitative CT evaluation, and 3D CT-texture analysis. Patients with histopathologically-proven panNEN, availability of Ki67% values and pre-treatment CT were included. CT images were retrospectively reviewed, and qualitative and quantitative images analysis were done; for quantitative analysis four enhancement-ratios and three permeability-ratios were created. 3D CT-texture imaging analysis was done (Mean Value; Variance; Skewness; Kurtosis; Entropy). Subsequently, these features were compared among the three grading (G) groups. 304 patients affected by panNENs were considered, and 100 patients were included. At qualitative evaluation, frequency of irregular margins was significantly different between tumor G groups. At quantitative evaluation, for all ratios, comparisons resulted statistical significant different between G1 and G3 groups and between G2 and G3 groups. At 3D CT-texture analysis, Kurtosis resulted statistical significant different among three G groups and Entropy resulted statistical significant different between G1 and G3 and between G2 and G3 groups. Quantitative CT evaluation of panNENs can predict tumor grade, discerning G1 from G3 and G2 from G3 tumors. CT-texture analysis can predict panNENs tumor grade, distinguishing G1 from G3 and G2 from G3, and G1 from G2 tumors.

Deep learning with convolutional neural network in radiology

Deep learning with a convolutional neural network (CNN) is gaining attention recently for its high performance in image recognition. Images themselves can be utilized in a learning process with this technique, and feature extraction in advance of the learning process is not required. Important features can be automatically learned. Thanks to the development of hardware and software in addition to techniques regarding deep learning, application of this technique to radiological images for predicting clinically useful information, such as the detection and the evaluation of lesions, etc., are beginning to be investigated. This article illustrates basic technical knowledge regarding deep learning with CNNs along the actual course (collecting data, implementing CNNs, and training and testing phases). Pitfalls regarding this technique and how to manage them are also illustrated. We also described some advanced topics of deep learning, results of recent clinical studies, and the future directions of clinical application of deep learning techniques.

Imaging prediction of nonalcoholic steatohepatitis using computed tomography texture analysis

OBJECTIVES

To determine if texture analysis of non-contrast-enhanced CT (NECT) images is able to predict nonalcoholic steatohepatitis (NASH).

METHODS

NECT images from 88 patients who underwent a liver biopsy for the diagnosis of suspected NASH were assessed and texture feature parameters were obtained without and with filtration. The patient population was divided into a predictive learning dataset and a validation dataset, and further divided into groups according to the prediction of liver fibrosis as assessed by hyaluronic acid levels. The reference standard was the histological result of a liver biopsy. A predictive model for NASH was developed using parameters derived from the learning dataset that demonstrated areas under the receiver operating characteristic curve (AUC) of >0.65. The resulting model was then applied to the validation dataset.

RESULTS

In patients without suspected fibrosis, the texture parameter mean without filter and skewness with a 2-mm filter were selected for the NASH prediction model. The AUC of the predictive model for the validation dataset was 0.94 and the accuracy was 94%. In patients with suspicion of fibrosis, the mean without filtration and kurtosis with a 4-mm filter were selected for the NASH prediction model. The AUC for the validation dataset was 0.60 and the accuracy was 42%.

CONCLUSIONS

In patients without suspicion of fibrosis, NECT texture analysis effectively predicted NASH.

KEY POINTS

• In patients without suspicion of fibrosis, NECT texture analysis effectively predicted NASH. • The mean without filtration and skewness with a 2-mm filter were modest predictors of NASH in patients without suspicion of liver fibrosis. • Hepatic fibrosis masks the characteristic texture features of NASH.

Quantitative computed tomography texture analyses for anterior mediastinal masses: Differentiation between solid masses and cysts

OBJECTIVES

To investigate whether solid anterior mediastinal masses could be differentiated from cysts using quantitative computed tomography (CT) texture analyses in unenhanced CT (UECT) or contrast enhanced CT (CECT).

MATERIALS AND METHODS

This clinical retrospective study included 76 UECT images (40 men and 36 women, 28 cystic (mean diameter, 29.5 mm) and 48 solid (mean diameter, 48.2 mm)) and 84 CECT images (45 men and 39 women, 26 cystic (mean diameter, 31.4 mm) and 58 solid (mean diameter, 51.4 mm)) of anterior mediastinal masses, which were diagnosed histopathologically or using imaging criteria. Polygonal regions of interest were placed on these masses. CT histogram analyses for images of masses with or without filtration (Laplacian of Gaussian filters with various spatial scaling factors) were performed. DeLong's test was performed to compare areas under the curve (AUC) with receiver operating characteristic analyses.

RESULTS

From logistic regression analyses with a stepwise procedure, a combination of the mean in unfiltered images (mean0; i.e., CT attenuation) and mean in filtered images featuring coarse texture for UECT (AUC = 0.869) and the combination of mean0 and entropy in filtered images featuring fine texture for CECT (AUC = 0.997) were found to predict better the internal characteristics of anterior mediastinal masses. In UECT and CECT, diagnostic performance using these combinations tended to be high compared to mean0 alone (AUC = 0.780 [p = 0.033] and AUC = 0.983 [p = 0.130], respectively).

CONCLUSION

Solid anterior mediastinal masses can be differentiated from cysts using quantitative CT texture analyses with moderate and high diagnostic performance in UECT and CECT, respectively.

Liver Fibrosis: Deep Convolutional Neural Network for Staging by Using Gadoxetic Acid-enhanced Hepatobiliary Phase MR Images

Purpose To investigate the performance of a deep convolutional neural network (DCNN) model in the staging of liver fibrosis using gadoxetic acid-enhanced hepatobiliary phase magnetic resonance (MR) imaging. Materials and Methods This retrospective study included patients for whom input data (hepatobiliary phase MR images, static magnetic field of the imaging unit, and hepatitis B and C virus testing results available, either positive or negative) and reference standard data (liver fibrosis stage evaluated from biopsy or surgical specimens obtained within 6 months of the MR examinations) were available were assigned to the training (534 patients) or the test (100 patients) group. For the training group (54, 53, 81, 113, and 233 patients with fibrosis stages F0, F1, F2, F3, and F4, respectively; mean patient age, 67.4 ± 9.7 years; 388 men and 146 women), MR images with three different section levels were augmented 90-fold (rotated, parallel-shifted, brightness-changed and contrast-changed images were generated; a total of 144 180 images). Supervised training was performed by using the DCNN model to minimize the difference between the output data (fibrosis score obtained through deep learning [F_DL score]) and liver fibrosis stage. The performance of the DCNN model was evaluated in the test group (10, 10, 15, 20, and 45 patients with fibrosis stages F0, F1, F2, F3, and F4, respectively; mean patient age, 66.8 years ± 10.7; 71 male patients and 29 female patients) with receiver operating characteristic (ROC) analyses. Results The F_DL score was correlated significantly with fibrosis stage (Spearman rank correlation coefficient: 0.63; P < .001). Fibrosis stages F4, F3, and F2 were diagnosed with areas under the ROC curve of 0.84, 0.84, and 0.85, respectively. Conclusion The DCNN model exhibited a high diagnostic performance in the staging of liver fibrosis. ^© RSNA, 2017 Online supplemental material is available for this article.