Diagnostic performance of EfficientNetV2-S method for staging liver fibrosis based on multiparametric MRI.

Problem: Previous studies had confirmed that some deep learning models had high diagnostic performance in staging liver fibrosis. However, training efficiency of models predicting liver fibrosis need to be improved to achieve rapid diagnosis and precision medicine. Aim: The deep learning framework of EfficientNetV2-S was noted because of its faster training speed and better parameter efficiency compared with other models. Our study sought to develop noninvasive predictive models based on EfficientNetV2-S framework for staging liver fibrosis. Methods: Patients with chronic liver disease who underwent multi-parametric abdominal MRI were included in the retrospective study. Data augmentation methods including horizontal flip, vertical flip, perspective transformation and edge enhancement were applied to multi-parametric MR images to solve the data imbalance between different liver fibrosis groups. The EfficientNetV2-S models were used for the prediction of liver fibrosis stages F1-2, F1-3, F3, F4 and F3-4. We evaluated the diagnostic performance of our models in training, validation, and test sets by using receiver operating characteristic curve (ROC) analysis. Results: The total training time of EfficientNetV2-S was about 6 h. For differentiating of F1-2 vs F3, the accuracy, sensitivity and specificity of EfficientNetV2-S model were 96.2 %, 96.4 % and 96.0 % in the test set. The AUC in test set was 0.559. The accuracy, sensitivity and specificity were 82.1 %, 74.5 % and 89.6 % in the test set by using EfficientNetV2-S model to differentiate F1-2 vs F3-4, and the AUC in test set were 0.763. For differentiating F1-3 vs F4, the accuracy, sensitivity and specificity of EfficientNetV2-S model were 71.5 %, 73.4 % and 69.5 % in the test set. The AUC was 0.553 in test set. For differentiating F1-2 vs F4, the accuracy, sensitivity and specificity of our model were 84.3 %, 80.2 % and 88.3 % in the test set, and the AUC was 0.715, respectively. For differentiating F3 vs F4, the accuracy, sensitivity and specificity of EfficientNetV2-S model were 92.5 %, 89.1 % and 95.6 % in the test set, and the AUC was 0.696 in the test set. Conclusions: The EfficientNetV2-S models based on multi-parametric MRI had the feasibility for staging of liver fibrosis because they showed high training speed and diagnostic performance in our study.

Staging liver fibrosis: comparison of radiomics model and fusion model based on multiparametric MRI in patients with chronic liver disease.

OBJECTIVES: To develop and compare radiomics model and fusion model based on multiple MR parameters for staging liver fibrosis in patients with chronic liver disease. MATERIALS AND METHODS: Patients with chronic liver disease who underwent multiparametric abdominal MRI were included in this retrospective study. Multiparametric MR images were imported into 3D-Slicer software for drawing bounding boxes on MR images. By using a 3D-Slicer extension of SlicerRadiomics, radiomics features were extracted from these MR images. The z-score normalization method was used for post-processing radiomics features. The least absolute shrinkage and selection operator method (LASSO) was performed for selecting significant radiomics features. The logistic regression analysis was used for building the radiomics model. A fusion model was built by integrating serum fibrosis biomarkers of aspartate transaminase-to-platelet ratio index (APRI) and the fibrosis-4 index (FIB-4) with radiomics signatures. RESULTS: In the training cohort, AUCs of radiomics and fusion model were 0.707-0.842 and 0.718-0.854 for differentiating different groups. In the testing cohort, AUCs were 0.514-0.724 and 0.609-0.728. For the training cohort, there was no significant difference of AUCs between radiomics and fusion model (p > 0.05). For the testing cohort, AUCs of fusion model were higher than those of radiomics model in differentiating F1-3 vs. F4 and F1-2 vs. F4 (p = 0.011 & 0.042). CONCLUSIONS: Radiomics model and fusion model based on multiparametric MRI exhibited the feasibility for staging liver fibrosis in patients with CLD, and APRI and FIB-4 could improve the diagnostic performance of radiomics model in differentiating F1-3 vs. F4 and F1-2 vs. F4.

Automatic liver segmentation and assessment of liver fibrosis using deep learning with MR T1-weighted images in rats.

OBJECTIVES: To validate the performance of nnU-Net in segmentation and CNN in classification for liver fibrosis using T1-weighted images. MATERIALS AND METHODS: In this prospective study, animal models of liver fibrosis were induced by injecting subcutaneously a mixture of Carbon tetrachloride and olive oil. A total of 99 male Wistar rats were successfully induced and underwent MR scanning with no contrast agent to get T1-weighted images. The regions of interest (ROIs) of the whole liver were delineated layer by layer along the liver edge by 3D Slicer. For segmentation task, all T1-weighted images were randomly divided into training and test cohorts in a ratio of 7:3. For classification, images containing the hepatic maximum diameter of every rat were selected and 80% images of no liver fibrosis (NLF), early liver fibrosis (ELF) and progressive liver fibrosis (PLF) stages were randomly selected for training, while the rest were used for testing. Liver segmentation was performed by the nnU-Net model. The convolutional neural network (CNN) was used for classification task of liver fibrosis stages. The Dice similarity coefficient was used to evaluate the segmentation performance of nnU-Net. Confusion matrix, ROC curve and accuracy were used to show the classification performance of CNN. RESULTS: A total of 2628 images were obtained from 99 Wistar rats by MR scanning. For liver segmentation by nnU-Net, the Dice similarity coefficient in the test set was 0.8477. The accuracies of CNN in staging NLF, ELF and PLF were 0.73, 0.89 and 0.84, respectively. The AUCs were 0.76, 0.88 and 0.79, respectively. CONCLUSION: The nnU-Net architecture is of high accuracy for liver segmentation and CNN for assessment of liver fibrosis with T1-weighted images.

A computational model of learning flexible navigation in a maze by layout-conforming replay of place cells.

Recent experimental observations have shown that the reactivation of hippocampal place cells (PC) during sleep or wakeful immobility depicts trajectories that can go around barriers and can flexibly adapt to a changing maze layout. However, existing computational models of replay fall short of generating such layout-conforming replay, restricting their usage to simple environments, like linear tracks or open fields. In this paper, we propose a computational model that generates layout-conforming replay and explains how such replay drives the learning of flexible navigation in a maze. First, we propose a Hebbian-like rule to learn the inter-PC synaptic strength during exploration. Then we use a continuous attractor network (CAN) with feedback inhibition to model the interaction among place cells and hippocampal interneurons. The activity bump of place cells drifts along paths in the maze, which models layout-conforming replay. During replay in sleep, the synaptic strengths from place cells to striatal medium spiny neurons (MSN) are learned by a novel dopamine-modulated three-factor rule to store place-reward associations. During goal-directed navigation, the CAN periodically generates replay trajectories from the animal's location for path planning, and the trajectory leading to a maximal MSN activity is followed by the animal. We have implemented our model into a high-fidelity virtual rat in the MuJoCo physics simulator. Extensive experiments have demonstrated that its superior flexibility during navigation in a maze is due to a continuous re-learning of inter-PC and PC-MSN synaptic strength.

Radiomics nomogram for prediction of microvascular invasion in hepatocellular carcinoma based on MR imaging with Gd-EOB-DTPA.

Objective: To develop a radiomics nomogram for predicting microvascular invasion (MVI) before surgery in hepatocellular carcinoma (HCC) patients. Materials and Methods: The data from a total of 189 HCC patients (training cohort: n = 141; validation cohort: n = 48) were collected, involving the clinical data and imaging characteristics. Radiomics features of all patients were extracted from hepatobiliary phase (HBP) in 15 min. Least absolute shrinkage selection operator (LASSO) regression and logistic regression were utilized to reduce data dimensions, feature selection, and to construct a radiomics signature. Clinicoradiological factors were identified according to the univariate and multivariate analyses, which were incorporated into the final predicted nomogram. A nomogram was developed to predict MVI of HCC by combining radiomics signatures and clinicoradiological factors. Radiomics nomograms were evaluated for their discrimination capability, calibration, and clinical usefulness. Results: In the clinicoradiological factors, gender, alpha-fetoprotein (AFP) level, tumor shape and halo sign served as the independent risk factors of MVI, with which the area under the curve (AUC) is 0.802. Radiomics signatures covering 14 features at HBP 15 min can effectively predict MVI in HCC, to construct radiomics signature model, with the AUC of 0.732. In the final nomogram model the clinicoradiological factors and radiomics signatures were integrated, outperforming the clinicoradiological model (AUC 0.884 vs. 0.802; p <0.001) and radiomics signatures model (AUC 0.884 vs. 0.732; p < 0.001) according to Delong test results. A robust calibration and discrimination were demonstrated in the nomogram model. The results of decision curve analysis (DCA) showed more significantly clinical efficiency of the nomogram model in comparison to the clinicoradiological model and the radiomic signature model. Conclusions: Depending on the clinicoradiological factors and radiological features on HBP 15 min images, nomograms can effectively predict MVI status in HCC patients.

Machine learning-based multiparametric MRI radiomics for predicting poor responders after neoadjuvant chemoradiotherapy in rectal Cancer patients.

BACKGROUND: The purpose of this study was to investigate and validate multiparametric magnetic resonance imaging (MRI)-based machine learning classifiers for early identification of poor responders after neoadjuvant chemoradiotherapy (nCRT) in patients with locally advanced rectal cancer (LARC). METHODS: Patients with LARC who underwent nCRT were included in this retrospective study (207 patients). After preprocessing of multiparametric MRI, radiomics features were extracted and four feature selection methods were used to select robust features. The selected features were used to build five machine learning classifiers, and 20 (four feature selection methods × five machine learning classifiers) predictive models for the screening of poor responders were constructed. The predictive models were evaluated according to the area under the curve (AUC), F1 score, accuracy, sensitivity, and specificity. RESULTS: Eighty percent of all predictive models constructed achieved an AUC of more than 0.70. A predictive model using a support vector machine classifier with the minimum redundancy maximum relevance (mRMR) selection method followed by the least absolute shrinkage and selection operator (LASSO) selection method showed superior prediction performance, with an AUC of 0.923, an F1 score of 88.14%, and accuracy of 91.03%. The predictive performance of the constructed models was not improved by ComBat compensation. CONCLUSIONS: In rectal cancer patients who underwent neoadjuvant chemoradiotherapy, machine learning classifiers with radiomics features extracted from multiparametric MRI were able to accurately discriminate poor responders from good responders. The techniques should provide additional information to guide patient-tailored treatment.

Radiomics Analysis of MR Imaging with Gd-EOB-DTPA for Preoperative Prediction of Microvascular Invasion in Hepatocellular Carcinoma: Investigation and Comparison of Different Hepatobiliary Phase Delay Times.

PURPOSE: To investigate whether the radiomics analysis of MR imaging in the hepatobiliary phase (HBP) can be used to predict microvascular invasion (MVI) in patients with hepatocellular carcinoma (HCC). METHOD: A total of 130 patients with HCC, including 80 MVI-positive patients and 50 MVI-negative patients, who underwent MR imaging with Gd-EOB-DTPA were enrolled. Least absolute shrinkage and selection operator (LASSO) regression was applied to select radiomics parameters derived from MR images obtained in the HBP 5 min, 10 min, and 15 min images. The selected features at each phase were adopted into support vector machine (SVM) classifiers to establish models. Multiple comparisons of the AUCs at each phase were performed by the Delong test. The decision curve analysis (DCA) was used to analyze the classification of MVI-positive and MVI-negative patients. RESULTS: The most predictive features between MVI-positive and MVI-negative patients included 9, 8, and 14 radiomics parameters on HBP 5 min, 10 min, and 15 min images, respectively. A model incorporating the selected features produced an AUC of 0.685, 0.718, and 0.795 on HBP 5 min, 10 min, and 15 min images, respectively. The predictive model for HBP 5 min, 10 min and 15 min showed no significant difference by the Delong test. DCA indicated that the predictive model for HBP 15 min outperformed the models for HBP 5 min and 10 min. CONCLUSIONS: Radiomics parameters in the HBP can be used to predict MVI, with the HBP 15 min model having the best differential diagnosis ability.

Development and validation of an MRI-based radiomic nomogram to distinguish between good and poor responders in patients with locally advanced rectal cancer undergoing neoadjuvant chemoradiotherapy.

PURPOSE: In the clinical management of patients with locally advanced rectal cancer (LARC), the early identification of poor and good responders after neoadjuvant chemoradiotherapy (N-CRT) is essential. Therefore, we developed and validated predictive models including MRI findings from the structured report template, clinical and radiomics parameters to differentiate between poor and good responders in patients with locally advanced rectal cancer who underwent neoadjuvant chemoradiotherapy. METHODS: Preoperative multiparametric MRI from 183 patients with locally advanced rectal cancer (122 in the training cohort, 61 in the validation cohort) was included in this retrospective study. After preprocessing, radiomic features were extracted and two methods of feature selection was applied to reduce the number of radiomics features. Logistic regression (LR) and random forest (RF) machine learning classifiers were trained to identify good responders from poor responders. Multivariable logistic regression analysis was used to incorporate the radiomic signature and clinical risk factors into a nomogram. Classifier performance was evaluated using the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). RESULTS: For the differentiation of poor and good responders, the radiomics model with an LR classifier achieved AUCs of 0.869 and 0.842 for the training and validation cohorts, respectively. The nomogram showed the highest reproducibility and prognostic ability in the training and validation cohorts, with AUCs of 0.923 (95% confidence interval, 0.872-0.975) and 0.898 (0.819-0.978), respectively. Additionally, the nomogram achieved significant risk stratification of patients in respect to progression free survival (PFS). CONCLUSIONS: The nomogram accurately differentiated good and poor responders in patients with LARC undergoing N-CRT, and showed significant performance for predicting PFS.

Prediction of the clinicopathological subtypes of breast cancer using a fisher discriminant analysis model based on radiomic features of diffusion-weighted MRI.

BACKGROUND: The clinicopathological classification of breast cancer is proposed according to therapeutic purposes. It is simplified and can be conducted easily in clinical practice, and this subtyping undoubtedly contributes to the treatment selection of breast cancer. This study aims to investigate the feasibility of using a Fisher discriminant analysis model based on radiomic features of diffusion-weighted MRI for predicting the clinicopathological subtypes of breast cancer. METHODS: Patients who underwent breast magnetic resonance imaging were confirmed by retrieving data from our institutional picture archiving and communication system (PACS) between March 2013 and September 2017. Five clinicopathological subtypes were determined based on the status of ER, PR, HER2 and Ki-67 from the immunohistochemical test. The radiomic features of diffusion-weighted imaging were derived from the volume of interest (VOI) of each tumour. Fisher discriminant analysis was performed for clinicopathological subtyping by using a backward selection method. To evaluate the diagnostic performance of the radiomic features, ROC analyses were performed to differentiate between immunohistochemical biomarker-positive and -negative groups. RESULTS: A total of 84 radiomic features of four statistical methods were included after preprocessing. The overall accuracy for predicting the clinicopathological subtypes was 96.4% by Fisher discriminant analysis, and the weighted accuracy was 96.6%. For predicting diverse clinicopathological subtypes, the prediction accuracies ranged from 92 to 100%. According to the cross-validation, the overall accuracy of the model was 82.1%, and the accuracies of the model for predicting the luminal A, luminal BHER2-, luminal BHER2+, HER2 positive and triple negative subtypes were 79, 77, 88, 92 and 73%, respectively. According to the ROC analysis, the radiomic features had excellent performance in differentiating between different statuses of ER, PR, HER2 and Ki-67. CONCLUSIONS: The Fisher discriminant analysis model based on radiomic features of diffusion-weighted MRI is a reliable method for the prediction of clinicopathological breast cancer subtypes.

Gut Microbiota and Liver Fibrosis: One Potential Biomarker for Predicting Liver Fibrosis.

PURPOSE: To investigate the relationship between gut microbiota and liver fibrosis and establish a microbiota biomarker for detecting and staging liver fibrosis. METHODS: 131 Wistar rats were used in our study, and liver fibrosis was induced by carbon tetrachloride. Stool samples were collected within 72 hours after the last administration. The V4 regions of 16S rRNA gene were amplified. The sequencing data was processed using the Quantitative Insights Into Microbial Ecology (QIIME version 1.9). The diversity, principal coordinate analysis (PCoA), nonmetric multidimensional scaling (NMDS), and linear discriminant analysis (LDA) effect size (LEfSe) were performed. Random-Forest classification was performed for discriminating the samples from different groups. Microbial function was assessed using the PICRUST. RESULTS: The Simpson in the control group was lower than that in the liver fibrosis group (p = 0.048) and differed significantly among different fibrosis stages (p = 0.047). The Chao1 index in the control group was higher than that in the liver fibrosis group (p < 0.001). NMDS analysis showed a marked difference between the control and liver fibrosis groups (p < 0.001). PCoA analysis indicated the different community composition between the control and liver fibrosis groups with variances of PC1 13.76% and PC2 5.89% and between different liver fibrosis stages with variances of PC1 10.51% and PC2 7.78%. LEfSe analysis showed alteration of gut microbiota in the liver fibrosis group. Biomarkers obtained from Random-Forest classification showed excellent diagnostic accuracy in prediction of liver fibrosis with AUROCs of 0.99. The AUROCs were 0.77~0.84 in prediction of stage F4. There were six increased and 17 decreased metabolic functions in the liver fibrosis group and 6 metabolic functions significantly differed among four liver fibrosis stages. CONCLUSION: Gut microbiota is a potential biomarker for detecting and staging liver fibrosis with high diagnostic accuracies.

Quantitative Analysis of Enhanced Computed Tomography in Differentiating Cystitis Glandularis and Bladder Cancer.

OBJECTIVE: This study was performed to assess the value of quantitative analysis of enhanced computed tomography (CT) values in the differential diagnosis of bladder cancer and cystitis glandularis (CG). METHODS: Eighty patients with bladder masses (39 with CG and 41 with bladder cancer) who underwent enhanced CT were retrospectively reviewed. The CT enhancement values of the lesion and normal bladder wall in the arterial phase, venous phase, and delayed phase were measured. The relative enhancement CT values (relative enhancement CT value = enhancement CT value of lesion - enhancement CT value of normal bladder) in the arterial phase, venous phase, and delayed phase were also calculated. The pathological results were used as the gold standard, and the area under the curve (AUC), sensitivity, and specificity were calculated for the six groups of quantitative indicators (enhanced CT values and relative enhanced CT values of CG and bladder cancer in the arterial, venous, and delayed phases). We performed the leave-group-out cross-validation method to validate the accuracy, AUC, sensitivity, and specificity. The differences in accuracy, AUC, sensitivity, and specificity among the six groups of quantitative indicators were compared by the t-test. RESULTS: In a combined analysis of the AUC, sensitivity, and specificity performance, the best indicator was the arterial-phase relative enhancement CT value with a cut-off of 25.85 HU (AUC, 0.966; sensitivity, 95.1%; specificity, 92.3%). We used the 100-times leave-group-out cross-validation method to validate the accuracy, AUC, sensitivity, and specificity. Arterial-phase relative enhancement CT values showed the highest AUC and accuracy among the six groups, with statistical significance (P < 0.05). CONCLUSION: Quantitative analysis of enhanced CT is of great clinical value in the differential diagnosis of CG and bladder cancer.

Radiomics models for diagnosing microvascular invasion in hepatocellular carcinoma: which model is the best model?

OBJECTIVES: To explore the feasibility of diagnosing microvascular invasion (MVI) with radiomics, to compare the diagnostic performance of different models established by each method, and to determine the best diagnostic model based on radiomics. METHODS: A retrospective analysis was conducted with 206 cases of hepatocellular carcinoma (HCC) confirmed through surgery and pathology in our hospital from June 2015 to September 2018. Among the samples, 88 were MVI-positive, and 118 were MVI-negative. The radiomics analysis process included tumor segmentation, feature extraction, data preprocessing, dimensionality reduction, modeling and model evaluation. RESULTS: A total of 1044 sets of texture feature parameters were extracted, and 21 methods were used for the radiomics analysis. All research methods could be used to diagnose MVI. Of all the methods, the LASSO+GBDT method had the highest accuracy, the LASSO+RF method had the highest sensitivity, the LASSO+BPNet method had the highest specificity, and the LASSO+GBDT method had the highest AUC. Through Z-tests of the AUCs, LASSO+GBDT, LASSO+K-NN, LASSO+RF, PCA + DT, and PCA + RF had Z-values greater than 1.96 (p<0.05). The DCA results showed that the LASSO + GBDT method was better than the other methods when the threshold probability was greater than 0.22. CONCLUSIONS: Radiomics can be used for the preoperative, noninvasive diagnosis of MVI, but different dimensionality reduction and modeling methods will affect the diagnostic performance of the final model. The model established with the LASSO+GBDT method had the optimal diagnostic performance and the greatest diagnostic value for MVI.

Identification of Metastatic Lymph Nodes in MR Imaging with Faster Region-Based Convolutional Neural Networks.

MRI is the gold standard for confirming a pelvic lymph node metastasis diagnosis. Traditionally, medical radiologists have analyzed MRI image features of regional lymph nodes to make diagnostic decisions based on their subjective experience; this diagnosis lacks objectivity and accuracy. This study trained a faster region-based convolutional neural network (Faster R-CNN) with 28,080 MRI images of lymph node metastasis, allowing the Faster R-CNN to read those images and to make diagnoses. For clinical verification, 414 cases of rectal cancer at various medical centers were collected, and Faster R-CNN-based diagnoses were compared with radiologist diagnoses using receiver operating characteristic curves (ROC). The area under the Faster R-CNN ROC was 0.912, indicating a more effective and objective diagnosis. The Faster R-CNN diagnosis time was 20 s/case, which was much shorter than the average time (600 s/case) of the radiologist diagnoses.Significance: Faster R-CNN enables accurate and efficient diagnosis of lymph node metastases. Cancer Res; 78(17); 5135-43. ©2018 AACR.

Diagnostic Performance of Mammographic Texture Analysis in the Differential Diagnosis of Benign and Malignant Breast Tumors.

BACKGROUND: The purpose of this study was to investigate the diagnostic performance of mammographic texture analysis in the differential diagnosis of benign and malignant breast tumors. PATIENTS AND METHODS: Digital mammography images were obtained from the Picture Archiving and Communication System at our institute. Texture features of mammographic images were calculated. Mann-Whitney U test was used to identify differences between the benign and malignant group. The receiver operating characteristic (ROC) curve analysis was used to assess the diagnostic performance of texture features. RESULTS: Significant differences of texture features of histogram, gray-level co-occurrence matrix (GLCM) and run length matrix (RLM) were found between the benign and malignant breast group (P < .05). The area under the ROC (AUROC) of histogram, GLCM, and RLM were 0.800, 0.787, and 0.761, with no differences between them (P > .05). The AUROCs of imaging-based diagnosis, texture analysis, and imaging-based diagnosis combined with texture analysis were 0.873, 0.863, and 0.961, respectively. When imaging-based diagnosis was combined with texture analysis, the AUROC was higher than that of imaging-based diagnosis or texture analysis (P < .05). CONCLUSION: Mammographic texture analysis is a reliable technique for differential diagnosis of benign and malignant breast tumors. Furthermore, the combination of imaging-based diagnosis and texture analysis can significantly improve diagnostic performance.