Combined deep learning and radiomics in pretreatment radiation esophagitis prediction for patients with esophageal cancer underwent volumetric modulated arc therapy.

PURPOSE: To develop a combined radiomics and deep learning (DL) model in predicting radiation esophagitis (RE) of a grade ≥ 2 for patients with esophageal cancer (EC) underwent volumetric modulated arc therapy (VMAT) based on computed tomography (CT) and radiation dose (RD) distribution images. MATERIALS AND METHODS: A total of 273 EC patients underwent VMAT were retrospectively reviewed and enrolled from two centers and divided into training (n = 152), internal validation (n = 66), and external validation (n = 55) cohorts, respectively. Radiomic and dosiomic features along with DL features using convolutional neural networks were extracted and screened from CT and RD images to predict RE. The performance of these models was evaluated and compared using the area under curve (AUC) of the receiver operating characteristic curves (ROC). RESULTS: There were 5 and 10 radiomic and dosiomic features were screened, respectively. XGBoost achieved a best AUC of 0.703, 0.694 and 0.801, 0.729 with radiomic and dosiomic features in the internal and external validation cohorts, respectively. ResNet34 achieved a best prediction AUC of 0.642, 0.657 and 0.762, 0.737 for radiomics based DL model (DLR) and RD based DL model (DLD) in the internal and external validation cohorts, respectively. Combined model of DLD + Dosiomics + clinical factors achieved a best AUC of 0.913, 0.821 and 0.805 in the training, internal, and external validation cohorts, respectively. CONCLUSION: Although the dose was not responsible for the prediction accuracy, the combination of various feature extraction methods was a factor in improving the RE prediction accuracy. Combining DLD with dosiomic features was promising in the pretreatment prediction of RE for EC patients underwent VMAT.

Radiation pneumonitis prediction with dual-radiomics for esophageal cancer underwent radiotherapy.

BACKGROUND: To integrate radiomics and dosiomics features from multiple regions in the radiation pneumonia (RP grade ≥ 2) prediction for esophageal cancer (EC) patients underwent radiotherapy (RT). METHODS: Total of 143 EC patients in the authors' hospital (training and internal validation: 70%:30%) and 32 EC patients from another hospital (external validation) underwent RT from 2015 to 2022 were retrospectively reviewed and analyzed. Patients were dichotomized as positive (RP+) or negative (RP-) according to CTCAE V5.0. Models with radiomics and dosiomics features extracted from single region of interest (ROI), multiple ROIs and combined models were constructed and evaluated. A nomogram integrating radiomics score (Rad_score), dosiomics score (Dos_score), clinical factors, dose-volume histogram (DVH) factors, and mean lung dose (MLD) was also constructed and validated. RESULTS: Models with Rad_score_Lung&Overlap and Dos_score_Lung&Overlap achieved a better area under curve (AUC) of 0.818 and 0.844 in the external validation in comparison with radiomics and dosiomics models with features extracted from single ROI. Combining four radiomics and dosiomics models using support vector machine (SVM) improved the AUC to 0.854 in the external validation. Nomogram integrating Rad_score, and Dos_score with clinical factors, DVH factors, and MLD further improved the RP prediction AUC to 0.937 and 0.912 in the internal and external validation, respectively. CONCLUSION: CT-based RP prediction model integrating radiomics and dosiomics features from multiple ROIs outperformed those with features from a single ROI with increased reliability for EC patients who underwent RT.

Multi-omics deep learning for radiation pneumonitis prediction in lung cancer patients underwent volumetric modulated arc therapy.

BACKGROUND AND OBJECTIVE: To evaluate the feasibility and accuracy of radiomics, dosiomics, and deep learning (DL) in predicting Radiation Pneumonitis (RP) in lung cancer patients underwent volumetric modulated arc therapy (VMAT) to improve radiotherapy safety and management. METHODS: Total of 318 and 31 lung cancer patients underwent VMAT from First Affiliated Hospital of Wenzhou Medical University (WMU) and Quzhou Affiliated Hospital of WMU were enrolled for training and external validation, respectively. Models based on radiomics (R), dosiomics (D), and combined radiomics and dosiomics features (R+D) were constructed and validated using three machine learning (ML) methods. DL models trained with CT (DLR), dose distribution (DLD), and combined CT and dose distribution (DL(R+D)) images were constructed. DL features were then extracted from the fully connected layers of the best-performing DL model to combine with features of the ML model with the best performance to construct models of R+DLR, D+DLD, R+D+DL(R+D)) for RP prediction. RESULTS: The R+D model achieved a best area under curve (AUC) of 0.84, 0.73, and 0.73 in the internal validation cohorts with Support Vector Machine (SVM), XGBoost, and Logistic Regression (LR), respectively. The DL(R+D) model achieved a best AUC of 0.89 and 0.86 using ResNet-34 in training and internal validation cohorts, respectively. The R+D+DL(R+D) model achieved a best performance in the external validation cohorts with an AUC, accuracy, sensitivity, and specificity of 0.81(0.62-0.99), 0.81, 0.84, and 0.67, respectively. CONCLUSIONS: The integration of radiomics, dosiomics, and DL features is feasible and accurate for the RP prediction to improve the management of lung cancer patients underwent VMAT.

MRI radiomics nomogram integrating postoperative adjuvant treatments in recurrence risk prediction for patients with early-stage cervical cancer.

BACKGROUND AND PURPOSE: Adjuvant treatments are valuable to decrease the recurrence rate and improve survival for early-stage cervical cancer patients (ESCC), Therefore, recurrence risk evaluation is critical for the choice of postoperative treatment. A magnetic resonance imaging (MRI) based radiomics nomogram integrating postoperative adjuvant treatments was constructed and validated externally to improve the recurrence risk prediction for ESCC. MATERIAL AND METHODS: 212 ESCC patients underwent surgery and adjuvant treatments from three centers were enrolled and divided into the training, internal validation, and external validation cohorts. Their clinical data, pretreatment T2-weighted images (T2WI) were retrieved and analyzed. Radiomics models were constructed using machine learning methods with features extracted and screen from sagittal and axial T2WI. A nomogram for recurrence prediction was build and evaluated using multivariable logistic regression analysis integrating radiomic signature and adjuvant treatments. RESULTS: A total of 8 radiomic features were screened out of 1020 extracted features. The extreme gradient boosting (XGboost) model based on MRI radiomic features performed best in recurrence prediction with an area under curve (AUC) of 0.833, 0.822 in the internal and external validation cohorts, respectively. The nomogram integrating radiomic signature and clinical factors achieved an AUC of 0.806, 0.718 in the internal and external validation cohorts, respectively, for recurrence risk prediction for ESCC. CONCLUSION: In this study, the nomogram integrating T2WI radiomic signature and clinical factors is valuable to predict the recurrence risk, thereby allowing timely planning for effective treatments for ESCC with high risk of recurrence.

Automatic dose prediction using deep learning and plan optimization with finite-element control for intensity modulated radiation therapy.

BACKGROUND: Automatic solutions for generating radiotherapy treatment plans using deep learning (DL) have been investigated by mimicking the voxel's dose. However, plan optimization using voxel-dose features has not been extensively studied. PURPOSE: This study aims to investigate the efficiency of a direct optimization strategy with finite elements (FEs) after DL dose prediction for automatic intensity-modulated radiation therapy (IMRT) treatment planning. METHODS: A double-UNet DL model was adapted for 220 cervical cancer patients (200 for training and 20 for testing), who underwent IMRT between 2016 and 2020 at our clinic. The model inputs were computed tomography (CT) slices, organs at risk (OARs), and planning target volumes (PTVs), and the outputs were dose distributions of uniformly generated high-dose region-controlled plans. The FEs were discretized into equal intervals of the dose prediction value within the [OARs avoid PTV(O-P)] and [body avoids OARs & PTV(B-OP)] regions in the test cohort and used to define the objectives for IMRT plan optimization. The plans were optimized using a two-step process. In the beginning, the plans of two extra cases with and without low-dose region control were compared to pursue robust and optimal dose adjustment degree pattern of FEs. In the first step, the mean dose of O-P FEs were constrained to differing degrees according to the pattern. The further the FEs from the PTV, the tighter the constraints. In the second step, the mean dose of O-P FEs from first step were constrained again but weakly and the dose of the B-OP FEs from dose prediction and PTV were tightly regulated. The dosimetric parameters of the OARs and PTV were evaluated and compared using an interstep approach. In another 10 cases, the plans optimized via the aforementioned steps (method 1) were compared with those directly generated by the double-UNet dose prediction model trained by low and high region-controlled plans (method 2). RESULTS: The mean differences in dose metrics between the UNet-predicted dose and the clinical plans were: 0.47 Gy for bladder D50% ; 0.62 Gy for rectum D50% ; 0% for small intestine V30Gy ; 1% for small intestine V40Gy ; 4% for left femoral head V30Gy ; and 6% for right femoral head V30Gy . The reductions in mean dose (p < 0.001) after FE-based optimization were: 4.0, 1.9, 2.8, 5.9, and 5.7 Gy for the bladder, rectum, small intestine, left femoral head, and right femoral head, respectively, with flat PTV homogeneity and conformity. Method 1 plans produced lower mean doses than those of method 2 for the bladder (0.7 Gy), rectum (1.0 Gy), and small intestine (0.6 Gy), while maintaining  PTV homogeneity and conformity. CONCLUSION: FE-based direct optimization produced lower OAR doses and adequate PTV doses after DL prediction. This solution offers rapid and automatic plan optimization without manual adjustment, particularly in low-dose regions.

Preoperative prediction of clinical and pathological stages for patients with esophageal cancer using PET/CT radiomics.

BACKGROUND: Preoperative stratification is critical for the management of patients with esophageal cancer (EC). To investigate the feasibility and accuracy of PET-CT-based radiomics in preoperative prediction of clinical and pathological stages for patients with EC. METHODS: Histologically confirmed 100 EC patients with preoperative PET-CT images were enrolled retrospectively and randomly divided into training and validation cohorts at a ratio of 7:3. The maximum relevance minimum redundancy (mRMR) was applied to select optimal radiomics features from PET, CT, and fused PET-CT images, respectively. Logistic regression (LR) was applied to classify the T stage (T1,2 vs. T3,4), lymph node metastasis (LNM) (LNM(-) vs. LNM(+)), and pathological state (pstage) (I-II vs. III-IV) with features from CT (CT_LR_Score), PET (PET_LR_Score), fused PET/CT (Fused_LR_Score), and combined CT and PET features (CT + PET_LR_Score), respectively. RESULTS: Seven, 10, and 7 CT features; 7, 8, and 7 PET features; and 3, 6, and 3 fused PET/CT features were selected using mRMR for the prediction of T stage, LNM, and pstage, respectively. The area under curves (AUCs) for T stage, LNM, and pstage prediction in the validation cohorts were 0.846, 0.756, 0.665, and 0.815; 0.769, 0.760, 0.665, and 0.824; and 0.727, 0.785, 0.689, and 0.837 for models of CT_LR_Score, PET_ LR_Score, Fused_ LR_Score, and CT + PET_ LR_Score, respectively. CONCLUSIONS: Accurate prediction ability was observed with combined PET and CT radiomics in the prediction of T stage, LNM, and pstage for EC patients. CRITICAL RELEVANCE STATEMENT: PET/CT radiomics is feasible and promising to stratify stages for esophageal cancer preoperatively. KEY POINTS: • PET-CT radiomics achieved the best performance for Node and pathological stage prediction. • CT radiomics achieved the best AUC for T stage prediction. • PET-CT radiomics is feasible and promising to stratify stages for EC preoperatively.

Integrating plan complexity and dosiomics features with deep learning in patient-specific quality assurance for volumetric modulated arc therapy.

PURPOSE: To investigate the feasibility and performance of deep learning (DL) models combined with plan complexity (PC) and dosiomics features in the patient-specific quality assurance (PSQA) for patients underwent volumetric modulated arc therapy (VMAT). METHODS: Total of 201 VMAT plans with measured PSQA results were retrospectively enrolled and divided into training and testing sets randomly at 7:3. PC metrics were calculated using house-built algorithm based on Matlab. Dosiomics features were extracted and selected using Random Forest (RF) from planning target volume (PTV) and overlap regions with 3D dose distributions. The top 50 dosiomics and 5 PC features were selected based on feature importance screening. A DL DenseNet was adapted and trained for the PSQA prediction. RESULTS: The measured average gamma passing rate (GPR) of these VMAT plans was 97.94% ± 1.87%, 94.33% ± 3.22%, and 87.27% ± 4.81% at the criteria of 3%/3 mm, 3%/2 mm, and 2%/2 mm, respectively. Models with PC features alone demonstrated the lowest area under curve (AUC). The AUC and sensitivity of PC and dosiomics (D) combined model at 2%/2 mm were 0.915 and 0.833, respectively. The AUCs of DL models were improved from 0.943, 0.849, 0.841 to 0.948, 0.890, 0.942 in the combined models (PC + D + DL) at 3%/3 mm, 3%/2 mm and 2%/2 mm, respectively. A best AUC of 0.942 with a sensitivity, specificity and accuracy of 100%, 81.8%, and 83.6% was achieved with combined model (PC + D + DL) at 2%/2 mm. CONCLUSIONS: Integrating DL with dosiomics and PC metrics is promising in the prediction of GPRs in PSQA for patients underwent VMAT.

Evaluation the benefits of additional radiotherapy for gastric cancer patients after D2 resection using CT based radiomics.

OBJECTIVES: The value of adding radiotherapy (RT) is still unclear for patients with gastric cancer (GC) after D2 lymphadenectomy. The purpose of this study is to predict and compare the overall survival (OS) and disease-free survival (DFS) of GC patients treated by chemotherapy and chemoradiation based on contrast-enhanced CT (CECT) radiomics. METHODS: A total of 154 patients treated by chemotherapy and chemoradiation in authors' hospital were retrospectively reviewed and randomly divided into the training and testing cohorts (7:3). Radiomics features were extracted from contoured tumor volumes in CECT using the pyradiomics software. Radiomics score and nomogram with integrated clinical factors were developed to predict the OS and DFS and evaluated with Harrell's Consistency Index (C-index). RESULTS: Radiomics score achieved a C index of 0.721(95%CI: 0.681-0.761) and 0.774 (95%CI: 0.738-0.810) in the prediction of DFS and OS for GC patients treated by chemotherapy and chemoradiation, respectively. The benefits of additional RT only demonstrated in subgroup of GC patients with Lauren intestinal type and perineural invasion (PNI). Integrating clinical factors further improved the prediction ability of radiomics models with a C-index of 0.773 (95%CI: 0.736-0.810) and 0.802 (95%CI: 0.765-0.839) for DFS and OS, respectively. CONCLUSIONS: CECT based radiomics is feasible to predict the OS and DFS for GC patients underwent chemotherapy and chemoradiation after D2 resection. The benefits of additional RT only observed in GC patients with intestinal cancer and PNI.

Intra- and peri-tumoral MRI radiomics features for preoperative lymph node metastasis prediction in early-stage cervical cancer.

BACKGROUND: Noninvasive and accurate prediction of lymph node metastasis (LNM) is very important for patients with early-stage cervical cancer (ECC). Our study aimed to investigate the accuracy and sensitivity of radiomics models with features extracted from both intra- and peritumoral regions in magnetic resonance imaging (MRI) with T2 weighted imaging (T2WI) and diffusion weighted imaging (DWI) for predicting LNM. METHODS: A total of 247 ECC patients with confirmed lymph node status were enrolled retrospectively and randomly divided into training (n = 172) and testing sets (n = 75). Radiomics features were extracted from both intra- and peritumoral regions with different expansion dimensions (3, 5, and 7 mm) in T2WI and DWI. Radiomics signature and combined radiomics models were constructed with selected features. A nomogram was also constructed by combining radiomics model with clinical factors for predicting LNM. RESULTS: The area under curves (AUCs) of radiomics signature with features from tumors in T2WI and DWI were 0.841 vs. 0.791 and 0.820 vs. 0.771 in the training and testing sets, respectively. Combining radiomics features from tumors in the T2WI, DWI and peritumoral 3 mm expansion in T2WI achieved the best performance with an AUC of 0.868 and 0.846 in the training and testing sets, respectively. A nomogram combining age and maximum tumor diameter (MTD) with radiomics signature achieved a C-index of 0.884 in the prediction of LNM for ECC. CONCLUSIONS:  Radiomics features extracted from both intra- and peritumoral regions in T2WI and DWI are feasible and promising for the preoperative prediction of LNM for patients with ECC.

Direct Dose Prediction With Deep Learning for Postoperative Cervical Cancer Underwent Volumetric Modulated Arc Therapy.

PURPOSE: To predict the voxel-based dose distribution for postoperative cervical cancer patients underwent volumetric modulated arc therapy using deep learning models. METHOD: A total of 254 patients with cervical cancer received volumetric modulated arc therapy in authors' hospital from January 2018 to September 2021 were enrolled in this retrospective study. Two deep learning networks (3D deep residual neural network and 3DUnet) were adapted to train (203 cases) and test (51 cases) the feasibility and effectiveness of the prediction method. The performance of deep learning models was evaluated by comparing the results with those of treatment planning system according to metrics of dose-volume histogram of target volumes and organs at risk. RESULTS: The dose distributions predicted by deep learning models were clinically acceptable. The automatic dose prediction time was around 5 to 10 min, which was about one-eighth to one-tenth of the manual optimization time. The maximum dose difference was observed in D98 of rectum with a | δD| of 5.00 ± 3.40% and 4.88 ± 3.99% for Unet3D and ResUnet3D, respectively. The minimum difference was observed in the D2 of clinical target volume with a |δD| of 0.53 ± 0.45% and 0.83 ± 0.45% for ResUnet3D and Unet3D, respectively. CONCLUSION: The 2 deep learning models adapted in the study showed the feasibility and reasonable accuracy in the voxel-based dose prediction for postoperative cervical cancer underwent volumetric modulated arc therapy. Automatic dose distribution prediction of volumetric modulated arc therapy with deep learning models is of clinical significance for the postoperative management of patients with cervical cancer.

Models of ultrasonic radiomics and clinical characters for lymph node metastasis assessment in thyroid cancer: a retrospective study.

Background: Preoperative prediction of cervical lymph node metastasis in papillary thyroid carcinoma provided a basis for tumor staging and treatment decision. This study aimed to investigate the utility of machine learning and develop different models to preoperatively predict cervical lymph node metastasis based on ultrasonic radiomic features and clinical characteristics in papillary thyroid carcinoma nodules. Methods: Data from 400 papillary thyroid carcinoma nodules were included and divided into training and validation group. With the help of machine learning, clinical characteristics and ultrasonic radiomic features were extracted and selected using randomforest and least absolute shrinkage and selection operator regression before classified by five classifiers. Finally, 10 models were built and their area under the receiver operating characteristic curve, accuracy, sensitivity, specificity, positive predictive value and negative predictive value were measured. Results: Among the 10 models, RF-RF model revealed the highest area under curve (0.812) and accuracy (0.7542) in validation group. The top 10 variables of it included age, seven textural features, one shape feature and one first-order feature, in which eight were high-dimensional features. Conclusions: RF-RF model showed the best predictive performance for cervical lymph node metastasis. And the importance features selected by it highlighted the unique role of higher-dimensional statistical methods for radiomics analysis.

Multi-modality radiomics nomogram based on DCE-MRI and ultrasound images for benign and malignant breast lesion classification.

Objective: To develop a multi-modality radiomics nomogram based on DCE-MRI, B-mode ultrasound (BMUS) and strain elastography (SE) images for classifying benign and malignant breast lesions. Material and Methods: In this retrospective study, 345 breast lesions from 305 patients who underwent DCE-MRI, BMUS and SE examinations were randomly divided into training (n = 241) and testing (n = 104) datasets. Radiomics features were extracted from manually contoured images. The inter-class correlation coefficient (ICC), Mann-Whitney U test and the least absolute shrinkage and selection operator (LASSO) regression were applied for feature selection and radiomics signature building. Multivariable logistic regression was used to develop a radiomics nomogram incorporating radiomics signature and clinical factors. The performance of the radiomics nomogram was evaluated by its discrimination, calibration, and clinical usefulness and was compared with BI-RADS classification evaluated by a senior breast radiologist. Results: The All-Combination radiomics signature derived from the combination of DCE-MRI, BMUS and SE images showed better diagnostic performance than signatures derived from single modality alone, with area under the curves (AUCs) of 0.953 and 0.941 in training and testing datasets, respectively. The multi-modality radiomics nomogram incorporating the All-Combination radiomics signature and age showed excellent discrimination with the highest AUCs of 0.964 and 0.951 in two datasets, respectively, which outperformed all single modality radiomics signatures and BI-RADS classification. Furthermore, the specificity of radiomics nomogram was significantly higher than BI-RADS classification (both p < 0.04) with the same sensitivity in both datasets. Conclusion: The proposed multi-modality radiomics nomogram based on DCE-MRI and ultrasound images has the potential to serve as a non-invasive tool for classifying benign and malignant breast lesions and reduce unnecessary biopsy.

Differentiating Primary Tumors for Brain Metastasis with Integrated Radiomics from Multiple Imaging Modalities.

Objectives: To differentiate the primary site of brain metastases (BMs) is of high clinical value for the successful management of patients with BM. The purpose of this study is to investigate a combined radiomics model with computer tomography (CT) and magnetic resonance imaging (MRI) images in differentiating BMs originated from lung and breast cancer. Methods: Pretreatment cerebral contrast enhanced CT and T1-weighted MRI images of 78 patients with 179 BMs from primary lung and breast cancer were retrospectively analyzed. Radiomic features were extracted from contoured BM lesions and selected using the Mann-Whitney U test and the least absolute shrinkage and selection operator (LASSO) logistic regression. Binary logistic regression (BLR) and support vector machine (SVM) models were built and evaluated based on selected radiomic features from CT alone, MRI alone, and combined images to differentiate BMs originated from lung and breast cancer. Results: A total of 10 and 6 optimal radiomic features were screened out of 1288 CT and 1197 MRI features, respectively. The mean area under the curves (AUCs) of the BLR and SVM models using fivefolds cross-validation were 0.703 vs. 0.751, 0.718 vs. 0.754, and 0.781 vs. 0.803 in the training dataset and 0.708 vs. 0.763, 0.715 vs. 0.717, and 0.771 vs. 0.805 in the testing dataset for models with CT alone, MRI alone, and combined CT and MRI radiomic features, respectively. Conclusions: Radiomics model based on combined CT and MRI features is feasible and accurate in the differentiation of the primary site of BMs from lung and breast cancer.

Efficient dose-volume histogram-based pretreatment patient-specific quality assurance methodology with combined deep learning and machine learning models for volumetric modulated arc radiotherapy.

BACKGROUND: Weak correlation between gamma passing rates and dose differences in target volumes and organs at risk (OARs) has been reported in several studies. Evaluation on the differences between planned dose-volume histogram (DVH) and reconstructed DVH from measurement was adopted and incorporated into patient-specific quality assurance (PSQA). However, it is difficult to develop a methodology allowing the evaluation of errors on DVHs accurately and quickly. PURPOSE: To develop a DVH-based pretreatment PSQA for volumetric modulated arc therapy (VMAT) with combined deep learning (DL) and machine learning models to overcome the limitation of conventional gamma index (GI) and improve the efficiency of DVH-based PSQA. METHODS: A DL model with a three-dimensional squeeze-and-excitation residual blocks incorporated into a modified U-net was developed to predict the measured PSQA DVHs of 208 head-and-neck (H&N) cancer patients underwent VMAT between 2018 and 2021 from two hospitals, in which 162 cases was randomly selected for training, 18 for validation, and 28 for testing. After evaluating the differences between treatment planning system (TPS) and PSQA DVHs predicted by DL model with multiple metrics, a pass or fail (PoF) classification model was developed using XGBoost algorithm. Evaluation of domain experts on dose errors between TPS and reconstructed PSQA DVHs was taken as ground truth for PoF classification model training. RESULTS: The prediction model was able to achieve a good agreement between predicted, measured, and TPS doses. Quantitative evaluation demonstrated no significant difference between predicted PSQA dose and measured dose for target and OARs, except for Dmean of PTV6900 (p = 0.001), D50 of PTV6000 (p = 0.014), D2 of PTV5400 (p = 0.009), D50 of left parotid (p = 0.015), and Dmax of left inner ear (p = 0.007). The XGBoost model achieved an area under curves, accuracy, sensitivity, and specificity of 0.89 versus 0.88, 0.89 versus 0.86, 0. 71 versus 0.71, and 0.95 versus 0.91 with measured and predicted PSQA doses, respectively. The agreement between domain experts and the classification model was 86% for 28 test cases. CONCLUSIONS: The successful prediction of PSQA doses and classification of PoF for H&N VMAT PSQA indicating that this DVH-based PSQA method is promising to overcome the limitations of GI and to improve the efficiency and accuracy of VMAT delivery.

The Influence of Different Ultrasonic Machines on Radiomics Models in Prediction Lymph Node Metastasis for Patients with Cervical Cancer.

Objective To investigate the effects of different ultrasonic machines on the performance of radiomics models using ultrasound (US) images in the prediction of lymph node metastasis (LNM) for patients with cervical cancer (CC) preoperatively. Methods A total of 536 CC patients with confirmed histological characteristics and lymph node status after radical hysterectomy and pelvic lymphadenectomy were enrolled. Radiomics features were extracted and selected with US images acquired with ATL HDI5000, Voluson E8, MyLab classC, ACUSON S2000, and HI VISION Preirus to build radiomics models for LNM prediction using support vector machine (SVM) and logistic regression, respectively. Results There were 148 patients (training vs validation: 102:46) scanned in machine HDI5000, 75 patients (53:22) in machine Voluson E8, 100 patients (69:31) in machine MyLab classC, 110 patients (76:34) in machine ACUSON S2000, and 103 patients (73:30) in machine HI VISION Preirus, respectively. Few radiomics features were reproducible among different machines. The area under the curves (AUCs) ranged from 0.75 to 0.86, 0.73 to 0.86 in the training cohorts, and from 0.71 to 0.82, 0.70 to 0.80 in the validation cohorts for SVM and logistic regression models, respectively. The highest difference in AUCs for different machines reaches 17.8% and 15.5% in the training and validation cohorts, respectively. Conclusions The performance of radiomics model is dependent on the type of scanner. The problem of scanner dependency on radiomics features should be considered, and their effects should be minimized in future studies for US images.

Contrast-enhanced CT based radiomics in the preoperative prediction of perineural invasion for patients with gastric cancer.

PURPOSE: To investigate the feasibility and accuracy of radiomics models based on contrast-enhanced CT (CECT) in the prediction of perineural invasion (PNI), so as to stratify high-risk recurrence and improve the management of patients with gastric cancer (GC) preoperatively. METHODS: Total of 154 GC patients underwent D2 lymph node dissection with pathologically confirmed GC and preoperative CECT from an open-label, investigator-sponsored trial (NCT01711242) were enrolled. Radiomics features were extracted from contoured images and selected using Mann-Whitney U test and the least absolute shrinkage and selection operator (LASSO) after inter-class correlation coefficient (ICC) analysis. Models based on radiomics features (R), clinical factors (C) and combined parameters (R + C) were built and evaluated using Support Vector Machine (SVM) and logistic regression to predict the PNI for patients with GC preoperatively. RESULTS: Total of 11 radiomics features were selected for final analysis, along with two clinical factors. The area under curve (AUC) of models based on R, C, and R + C with logistic regression and SVM were 0.77 vs. 0.83, 0.71 vs.0.70, 0.86 vs. 0.90, and 0.73 vs.0.80, 0.62 vs. 0.64, 0.77 vs. 0.82 in the training and testing cohorts, respectively. SVM(R + C) achieved a best AUC of 0.82(0.69-0.94) in the test cohorts with a sensitivity, specificity and accuracy of 0.63, 0.91, and 0.77, respectively. CONCLUSIONS: The performance of these models indicates that radiomics features alone or combined with clinical factors provide a feasible way to classify patients preoperatively and improve the management of patients with GC.

RefineNet-based 2D and 3D automatic segmentations for clinical target volume and organs at risks for patients with cervical cancer in postoperative radiotherapy.

PURPOSE: An accurate and reliable target volume delineation is critical for the safe and successful radiotherapy. The purpose of this study is to develop new 2D and 3D automatic segmentation models based on RefineNet for clinical target volume (CTV) and organs at risk (OARs) for postoperative cervical cancer based on computed tomography (CT) images. METHODS: A 2D RefineNet and 3D RefineNetPlus3D were adapted and built to automatically segment CTVs and OARs on a total of 44 222 CT slices of 313 patients with stage I-III cervical cancer. Fully convolutional networks (FCNs), U-Net, context encoder network (CE-Net), UNet3D, and ResUNet3D were also trained and tested with randomly divided training and validation sets, respectively. The performances of these automatic segmentation models were evaluated by Dice similarity coefficient (DSC), Jaccard similarity coefficient, and average symmetric surface distance when comparing them with manual segmentations with the test data. RESULTS: The DSC for RefineNet, FCN, U-Net, CE-Net, UNet3D, ResUNet3D, and RefineNet3D were 0.82, 0.80, 0.82, 0.81, 0.80, 0.81, and 0.82 with a mean contouring time of 3.2, 3.4, 8.2, 3.9, 9.8, 11.4, and 6.4 s, respectively. The generated RefineNetPlus3D demonstrated a good performance in the automatic segmentation of bladder, small intestine, rectum, right and left femoral heads with a DSC of 0.97, 0.95, 091, 0.98, and 0.98, respectively, with a mean computation time of 6.6 s. CONCLUSIONS: The newly adapted RefineNet and developed RefineNetPlus3D were promising automatic segmentation models with accurate and clinically acceptable CTV and OARs for cervical cancer patients in postoperative radiotherapy.

The Effects of Automatic Segmentations on Preoperative Lymph Node Status Prediction Models With Ultrasound Radiomics for Patients With Early Stage Cervical Cancer.

Introduction: The purpose of this study is to investigate the effects of automatic segmentation algorithms on the performance of ultrasound (US) radiomics models in predicting the status of lymph node metastasis (LNM) for patients with early stage cervical cancer preoperatively. Methods: US images of 148 cervical cancer patients were collected and manually contoured by two senior radiologists. The four deep learning-based automatic segmentation models, namely U-net, context encoder network (CE-net), Resnet, and attention U-net were constructed to segment the tumor volumes automatically. Radiomics features were extracted and selected from manual and automatically segmented regions of interest (ROIs) to predict the LNM of these cervical cancer patients preoperatively. The reliability and reproducibility of radiomics features and the performances of prediction models were evaluated. Results: A total of 449 radiomics features were extracted from manual and automatic segmented ROIs with Pyradiomics. Features with an intraclass coefficient (ICC) > 0.9 were all 257 (57.2%) from manual and automatic segmented contours. The area under the curve (AUCs) of validation models with radiomics features extracted from manual, attention U-net, CE-net, Resnet, and U-net were 0.692, 0.755, 0.696, 0.689, and 0.710, respectively. Attention U-net showed best performance in the LNM prediction model with a lowest discrepancy between training and validation. The AUCs of models with automatic segmentation features from attention U-net, CE-net, Resnet, and U-net were 9.11%, 0.58%, -0.44%, and 2.61% higher than AUC of model with manual contoured features, respectively. Conclusion: The reliability and reproducibility of radiomics features, as well as the performance of radiomics models, were affected by manual segmentation and automatic segmentations.

Differentiate Thyroid Follicular Adenoma from Carcinoma with Combined Ultrasound Radiomics Features and Clinical Ultrasound Features.

Noninvasive differentiating thyroid follicular adenoma from carcinoma preoperatively is of great clinical value to decrease the risks resulted from excessive surgery for patients with follicular neoplasm. The purpose of this study is to investigate the accuracy of ultrasound radiomics features integrating with ultrasound features in the differentiation between thyroid follicular carcinoma and adenoma. A total of 129 patients diagnosed as thyroid follicular neoplasm with pathologically confirmed follicular adenoma and carcinoma were enrolled and analyzed retrospectively. Radiomics features were extracted from preoperative ultrasound images with manually contoured targets. Ultrasound features and clinical parameters were also obtained from electronic medical records. Radiomics signature, combined model integrating radiomics features, ultrasound features, and clinical parameters were constructed and validated to differentiate the follicular carcinoma from adenoma. A total of 23 optimal features were selected from 449 extracted radiomics features. Clinical and ultrasound parameters of sex (p = 0.003), interior structure (p = 0.035), edge (p = 0.02), platelets (p = 0.007), and creatinine (p = 0.001) were associated with the differentiation between benign and malignant follicular neoplasm. The values of area under curves (AUCs) of the radiomics signature, clinical model, and combined model were 0.772 (95% CI: 0.707-0.838), 0.792 (95% CI: 0.715-0.869), and 0.861 (95% CI: 0.775-0.909), respectively. A final corrected AUC of 0.844 was achieved for the combined model after internal validation. Radiomics features from ultrasound images combined with ultrasound features and clinical factors are feasible to differentiate thyroid follicular carcinoma from adenoma noninvasive before operation to decrease the unnecessary of diagnostic thyroidectomy for patients with benign follicular adenoma.

The Accuracy and Radiomics Feature Effects of Multiple U-net-Based Automatic Segmentation Models for Transvaginal Ultrasound Images of Cervical Cancer.

Ultrasound (US) imaging has been recognized and widely used as a screening and diagnostic imaging modality for cervical cancer all over the world. However, few studies have investigated the U-net-based automatic segmentation models for cervical cancer on US images and investigated the effects of automatic segmentations on radiomics features. A total of 1102 transvaginal US images from 796 cervical cancer patients were collected and randomly divided into training (800), validation (100) and test sets (202), respectively, in this study. Four U-net models (U-net, U-net with ResNet, context encoder network (CE-net), and Attention U-net) were adapted to segment the target of cervical cancer automatically on these US images. Radiomics features were extracted and evaluated from both manually and automatically segmented area. The mean Dice similarity coefficient (DSC) of U-net, Attention U-net, CE-net, and U-net with ResNet were 0.88, 0.89, 0.88, and 0.90, respectively. The average Pearson coefficients for the evaluation of the reliability of US image-based radiomics were 0.94, 0.96, 0.94, and 0.95 for U-net, U-net with ResNet, Attention U-net, and CE-net, respectively, in their comparison with manual segmentation. The reproducibility of the radiomics parameters evaluated by intraclass correlation coefficients (ICC) showed robustness of automatic segmentation with an average ICC coefficient of 0.99. In conclusion, high accuracy of U-net-based automatic segmentations was achieved in delineating the target area of cervical cancer US images. It is feasible and reliable for further radiomics studies with features extracted from automatic segmented target areas.