The value of the computer-aided diagnosis system for thyroid lesions based on computed tomography images

The application of computer-aided diagnosis in Breast Imaging Reporting and Data System ultrasound training for residents-a randomized controlled study

Background

The consistency of Breast Imaging Reporting and Data System (BI-RADS) classification among experienced radiologists is different, which is difficult for inexperienced radiologists to master. This study aims to explore the value of computer-aided diagnosis (CAD) (AI-SONIC breast automatic detection system) in the BI-RADS training for residents.

Methods

A total of 12 residents who participated in the first year and the second year of standardized resident training in Ningbo No. 2 Hospital from May 2020 to May 2021 were randomly divided into 3 groups (Group 1, Group 2, Group 3) for BI-RADS training. They were asked to complete 2 tests and questionnaires at the beginning and end of the training. After the first test, the educational materials were given to the residents and reviewed during the breast imaging training month. Group 1 studied independently, Group 2 studied with CAD, and Group 3 was taught face-to-face by experts. The test scores and ultrasonographic descriptors of the residents were evaluated and compared with those of the radiology specialists. The trainees' confidence and recognition degree of CAD were investigated by questionnaire.

Results

There was no statistical significance in the scores of residents in the first test among the 3 groups (P=0.637). After training and learning, the scores of all 3 groups of residents were improved in the second test (P=0.006). Group 2 (52±7.30) and Group 3 (54±5.16) scored significantly higher than Group 1 (38±3.65). The consistency of ultrasonographic descriptors and final assessments between the residents and senior radiologists were improved (κ3 > κ2 > κ1), with κ2 and κ3 >0.4 (moderately consistent with experts), and κ1 =0.225 (fairly agreed with experts). The results of the questionnaire showed that the trainees had increased confidence in BI-RADS classification, especially Group 2 (1.5 to 3.5) and Group 3 (1.25 to 3.75). All trainees agreed that CAD was helpful for BI-RADS learning (Likert scale score: 4.75 out of 5) and were willing to use CAD as an aid (4.5, max. 5).

Conclusions

The AI-SONIC breast automatic detection system can help residents to quickly master BI-RADS, improve the consistency between residents and experts, and help to improve the confidence of residents in the classification of BI-RADS, which may have potential value in the BI-RADS training for radiology residents.

Trial Registration

Chinese Clinical Trial Registry (ChiCTR2400081672).

Computer-aided detection of prostate cancer in early stages using multi-parameter MRI: A promising approach for early diagnosis

BACKGROUND

Transrectal ultrasound-guided prostate biopsy is the gold standard diagnostic test for prostate cancer, but it is an invasive examination of non-targeted puncture and has a high false-negative rate.

OBJECTIVE

In this study, we aimed to develop a computer-assisted prostate cancer diagnosis method based on multiparametric MRI (mpMRI) images.

METHODS

We retrospectively collected 106 patients who underwent radical prostatectomy after diagnosis with prostate biopsy. mpMRI images, including T2 weighted imaging (T2WI), diffusion weighted imaging (DWI), and dynamic-contrast enhanced (DCE), and were accordingly analyzed. We extracted the region of interest (ROI) about the tumor and benign area on the three sequential MRI axial images at the same level. The ROI data of 433 mpMRI images were obtained, of which 202 were benign and 231 were malignant. Of those, 50 benign and 50 malignant images were used for training, and the 333 images were used for verification. Five main feature groups, including histogram, GLCM, GLGCM, wavelet-based multi-fractional Brownian motion features and Minkowski function features, were extracted from the mpMRI images. The selected characteristic parameters were analyzed by MATLAB software, and three analysis methods with higher accuracy were selected.

RESULTS

Through prostate cancer identification based on mpMRI images, we found that the system uses 58 texture features and 3 classification algorithms, including Support Vector Machine (SVM), K-nearest Neighbor (KNN), and Ensemble Learning (EL), performed well. In the T2WI-based classification results, the SVM achieved the optimal accuracy and AUC values of 64.3% and 0.67. In the DCE-based classification results, the SVM achieved the optimal accuracy and AUC values of 72.2% and 0.77. In the DWI-based classification results, the ensemble learning achieved optimal accuracy as well as AUC values of 75.1% and 0.82. In the classification results based on all data combinations, the SVM achieved the optimal accuracy and AUC values of 66.4% and 0.73.

CONCLUSION

The proposed computer-aided diagnosis system provides a good assessment of the diagnosis of the prostate cancer, which may reduce the burden of radiologists and improve the early diagnosis of prostate cancer.

Comparison of the cox regression to machine learning in predicting the survival of anaplastic thyroid carcinoma

BACKGROUND

To compare the ability of the Cox regression and machine learning algorithms to predict the survival of patients with Anaplastic thyroid carcinoma (ATC).

METHODS

Patients diagnosed with ATC were extracted from the Surveillance, Epidemiology, and End Results database. The outcomes were overall survival (OS) and cancer-specific survival (CSS), divided into: (1) binary data: survival or not at 6 months and 1 year; (2): time-to-event data. The Cox regression method and machine learnings were used to construct models. Model performance was evaluated using the concordance index (C-index), brier score and calibration curves. The SHapley Additive exPlanations (SHAP) method was deployed to interpret the results of machine learning models.

RESULTS

For binary outcomes, the Logistic algorithm performed best in the prediction of 6-month OS, 12-month OS, 6-month CSS, and 12-month CSS (C-index = 0.790, 0.811, 0.775, 0.768). For time-event outcomes, traditional Cox regression exhibited good performances (OS: C-index = 0.713; CSS: C-index = 0.712). The DeepSurv algorithm performed the best in the training set (OS: C-index = 0.945; CSS: C-index = 0.834) but performs poorly in the verification set (OS: C-index = 0.658; CSS: C-index = 0.676). The brier score and calibration curve showed favorable consistency between the predicted and actual survival. The SHAP values was deployed to explain the best machine learning prediction model.

CONCLUSIONS

Cox regression and machine learning models combined with the SHAP method can predict the prognosis of ATC patients in clinical practice. However, due to the small sample size and lack of external validation, our findings should be interpreted with caution.

Clinical diagnostic value of American College of Radiology thyroid imaging report and data system in different kinds of thyroid nodules

BACKGROUND

To evaluate the diagnostic value of American College of Radiology (ACR) score and ACR Thyroid Imaging Report and Data System (TI-RADS) for benign nodules, medullary thyroid carcinoma (MTC) and papillary thyroid carcinoma (PTC) through comparing with Kwak TI-RADS.

METHODS

Five hundred nine patients diagnosed with PTC, MTC or benign thyroid nodules were included and classified into the benign thyroid nodules group (n = 264), the PTC group (n = 189) and the MTC group (n = 56). The area under the curve (AUC) values were analyzed and the receiver operator characteristic (ROC) curves were drawn to compare the diagnostic efficiencies of ACR score, ACR TI-RADS and KWAK TI-RADS on benign thyroid nodules, MTC and PTC.

RESULTS

The AUC values of ACR score, ACR TI-RADS and Kwak TI-RADS for distinguishing malignant nodules from benign nodules were 0.914 (95%CI: 0.886-0.937), 0.871 (95%CI: 0.839-0.899) and 0.885 (95%CI: 0.854-0.911), respectively. In distinguishing of patients with MTC from PTC, the AUC values of ACR score, ACR TI-RADS and Kwak TI-RADS were 0.650 (95%CI: 0.565-0.734), 0.596 (95%CI: 0.527-0.664), and 0.613 (95%CI: 0.545-0.681), respectively. The AUC values of ACR score, ACR TI-RADS and Kwak TI-RADS for the discrimination of patients with MTC, PTC or benign nodules from patients without MTC, PTC or benign nodules were 0.899 (95%CI: 0.882-0.915), 0.865 (95%CI: 0.846-0.885), and 0.873 (95%CI: 0.854-0.893), respectively.

CONCLUSION

The ACR score performed the best, followed ex aequo by the ACR and Kwak TI-RADS in discriminating patients with malignant nodules from benign nodules and patients with MTC from PTC.

Incidental Thyroid Nodule on Chest Computed Tomography: Application of Computed Tomography Texture Analysis in Prediction of Ultrasound Classification

OBJECTIVE

The aim of the study was to evaluate the value of computed tomography (CT) texture analysis (CTTA) in predicting ultrasound (US) classification of incidentally detected thyroid nodule (ITN) on chest CT.

METHODS

A total of 117 ITNs (≥1 cm in the longest diameter) on chest CT scan of 107 patients was divided into 4 categories according to the Korean Thyroid Imaging Reporting and Data System (K-TIRADS) classification on recent thyroid US within 3 months. Computed tomography texture features were extracted with or without filtration using commercial software. The texture features were compared between the benign (K-TIRADS 2; n = 21) and the suspicious (K-TIRADS 3, 4, 5; n = 96) nodules. Multivariate regression and area under the receiver operating characteristic curve analysis were performed to determine significant prediction factors of the suspicious nodules.

RESULTS

The mean value of positive pixels was significantly higher in the suspicious nodules except the unfiltered image (P < 0.05). Entropy of the suspicious nodules was significantly higher with unfiltered and fine filters (P < 0.05), and kurtosis of the suspicious nodules was significantly higher with medium and coarse filters (P < 0.05). A logistic regression model incorporating mean value of positive pixels and kurtosis with a medium filter using volumetric analysis demonstrated the best performance to predict the suspicious nodules with an area under the receiver operating characteristic curve of 0.842 (P < 0.001, sensitivity 82.3%, and specificity 81.0%).

CONCLUSIONS

Computed tomography texture analysis for ITN larger than 1 cm showed significant correlation with systematic thyroid US classification and presented excellent performance to predict the suspicious nodules.

A new prediction model for lateral cervical lymph node metastasis in patients with papillary thyroid carcinoma: Based on dual-energy CT

PURPOSE

The current study aimed to develop and validate a prediction model to estimate the independent risk factors for lateral cervical lymph node metastasis (LLNM) in papillary thyroid carcinoma (PTC) patients based on dual-energy computed tomography (DECT).

METHOD

This study retrospectively conducted 406 consecutive patients from July 2015 to June 2019 to form the derivation cohorts and performed internal validation. 101 consecutive patients from July 2019 to June 2020 were included to create the external validation cohort. Univariable and multivariable logistic regression analyses were used to evaluate independent risk factors for LLNM. A prediction model based on DECT parameters was built and presented on a nomogram. The internal and external validations were performed.

RESULTS

Iodine concentration (IC) in the arterial phase (OR 2.761, 95% CI 1.028-7.415, P 0.044), IC in venous phase (OR 3.820, 95% CI 1.430-10.209, P 0.008), located in the superior pole (OR 4.181, 95% CI 2.645-6.609, P 0.000), and extrathyroidal extension (OR 4.392, 95% CI 2.142-9.004, P 0.000) were independently associated with LLNM in the derivation cohort. These four predictors were incorporated into the nomogram. The model showed good discrimination in the derivation (AUC, 0.899), internal (AUC, 0.905), and external validation (AUC, 0.912) cohorts. The decision curve revealed that more advantages would be added using the nomogram to estimate LLNM, which implied that the lateral lymph node dissection was recommended.

CONCLUSIONS

DECT parameters could provide independent indicators of LLNM in PTC patients, and the nomogram based on them may be helpful in treatment decision-making.

Automatic Recognition and Classification System of Thyroid Nodules in CT Images Based on CNN

Thyroid nodule lesions are one of the most common lesions of the thyroid; the incidence rate has been the highest in the past thirty years. X-ray computed tomography (CT) plays an increasingly important role in the diagnosis of thyroid diseases. Nonetheless, as a result of the artifact and high complexity of thyroid CT image, the traditional machine learning method cannot be applied to CT image processing. In this paper, an end-to-end thyroid nodule automatic recognition and classification system is designed based on CNN. An improved Eff-Unet segmentation network is used to segment thyroid nodules as ROI. The image processing algorithm optimizes the ROI region and divides the nodules. A low-level and high-level feature fusion classification network CNN-F is proposed to classify the benign and malignant nodules. After each module is connected in series with the algorithm, the automatic classification of each nodule can be realized. Experimental results demonstrate that the proposed end-to-end thyroid nodule automatic recognition and classification system has excellent performance in diagnosing thyroid diseases. In the test set, the segmentation IOU reaches 0.855, and the classification output accuracy reaches 85.92%.

An efficient deep convolutional neural network model for visual localization and automatic diagnosis of thyroid nodules on ultrasound images

Background

The aim of this study was to construct a deep convolutional neural network (CNN) model for localization and diagnosis of thyroid nodules on ultrasound and evaluate its diagnostic performance.

Methods

We developed and trained a deep CNN model called the Brief Efficient Thyroid Network (BETNET) using 16,401 ultrasound images. According to the parameters of the model, we developed a computer-aided diagnosis (CAD) system to localize and differentiate thyroid nodules. The validation dataset (1,000 images) was used to compare the diagnostic performance of the model using three state-of-the-art algorithms. We used an internal test set (300 images) to evaluate the BETNET model by comparing it with diagnoses from five radiologists with varying degrees of experience in thyroid nodule diagnosis. Lastly, we demonstrated the general applicability of our artificial intelligence (AI) system for diagnosing thyroid cancer in an external test set (1,032 images).

Results

The BETNET model accurately detected thyroid nodules in visualization experiments. The model demonstrated higher values for area under the receiver operating characteristic (AUC-ROC) curve [0.983, 95% confidence interval (CI): 0.973-0.990], sensitivity (99.19%), accuracy (98.30%), and Youden index (0.9663) than the three state-of-the-art algorithms (P<0.05). In the internal test dataset, the diagnostic accuracy of the BETNET model was 91.33%, which was markedly higher than the accuracy of one experienced (85.67%) and two less experienced radiologists (77.67% and 69.33%). The area under the ROC curve of the BETNET model (0.951) was similar to that of the two highly skilled radiologists (0.940 and 0.953) and significantly higher than that of one experienced and two less experienced radiologists (P<0.01). The kappa coefficient of the BETNET model and the pathology results showed good agreement (0.769). In addition, the BETNET model achieved an excellent diagnostic performance (AUC =0.970, 95% CI: 0.958-0.980) when applied to ultrasound images from another independent hospital.

Conclusions

We developed a deep learning model which could accurately locate and automatically diagnose thyroid nodules on ultrasound images. The BETNET model exhibited better diagnostic performance than three state-of-the-art algorithms, which in turn performed similarly in diagnosis as the experienced radiologists. The BETNET model has the potential to be applied to ultrasound images from other hospitals.

Dual-energy computed tomography could reliably differentiate metastatic from non-metastatic lymph nodes of less than 0.5 cm in patients with papillary thyroid carcinoma

Background

Dual-energy computed tomography (DECT) has been widely applied to detect lymph node (LN) and lymph node metastasis (LNM) in various cancers, including papillary thyroid carcinoma (PTC). This study aimed to quantitatively evaluate metastatic cervical lymph nodes (LNs) <0.5 cm in patients with PTC using DECT, which has not been done in previous studies.

Methods

Preoperative DECT data of patients with pathologically confirmed PTC were retrospectively collected and analyzed between May 2016 and June 2018. A total of 359 LNs from 52 patients were included. Diameter, iodine concentration (IC), normalized iodine concentration (NIC), and the slope of the energy spectrum curve (λ_HU) of LNs in the arterial and the venous phases were compared between metastatic and non-metastatic LNs. The optimal parameters were obtained from the receiver operating characteristic (ROC) curves. The generalized estimation equation (GEE) model was used to evaluate independent diagnostic factors for LNM.

Results

A total of 139 metastatic and 220 non-metastatic LNs were analyzed. There were statistical differences of quantitative parameters between the two groups (P value 0.000-0.007). The optimal parameter for diagnosing LNM was IC in the arterial phase, and its area under the curve (AUC), sensitivity, and specificity were 0.775, 71.9%, and 73.6%, respectively. When the three parameters of diameter, IC in the arterial phase, and NIC in the venous phase were combined, the prediction efficiency was better, and the AUC was 0.819. The GEE results showed that LNs located in level VIa [odds ratio (OR) 2.030, 95% confidence interval (CI): 1.134-3.634, P=0.017], VIb (OR 2.836, 95% CI: 1.597-5.038, P=0.000), diameter (OR 2.023, 95% CI: 1.158-3.532, P=0.013), IC in the arterial phase (OR 4.444, 95% CI: 2.808-7.035, P=0.000), and IC in the venous phase (OR 5.387, 95% CI: 3.449-8.413, P=0.000) were independent risk factors for LNM in patients with PTC.

Conclusions

DECT had good diagnostic performance in the differentiation of cervical metastatic LNs <0.5 cm in patients with PTC.

Differential Diagnosis of Malignant Thyroid Calcification Nodule Based on Computed Tomography Image Texture

Purpose: Calcification nodules in thyroid can be found in thyroid disease. Current clinical computed tomography systems can be used to detect calcification nodules. Our aim is to identify the nature of thyroid calcification nodule based on plain CT images. Method: Sixty-three patients (36 benign and 27 malignant nodules) found thyroid calcification nodules were retrospectively analyzed, together with computed tomography images and pathology finding. The regions of interest (ROI) of 6464 pixels containing calcification nodules were manually delineated by radiologists in CT plain images. We extracted thirty-one texture features from each ROI. And nineteen texture features were picked up after feature optimization by logistic regression analysis. All the texture features were normalized to [0, 1]. Four classification algorithms, including ensemble learning, support vector machine, K-nearest neighbor, decision tree, were used as classification algorithms to identity the benign and malignant nodule. Accuracy, PPV, NPV, SEN, and AUC were calculated to evaluate the performance of different classifiers. Results: Nineteen texture features were selected after feature optimization by logistic regression analysis (P <0.05). Both Ensemble Learning and Support Vector Machine achieved the highest accuracy of 97.1%. The PPV, NPV, SEN, and SPC are 96.9%, 97.4%, 98.4%, and 95.0%, respectively. The AUC was 1. Conclusion: Texture features extracted from calcification nodules could be used as biomarkers to identify benign or malignant thyroid calcification.

Can contrast-enhancement computed tomography texture and histogram analyses help to differentiate malignant from benign thyroid nodules?

PURPOSE

We aimed to determine the ability of contrast-enhanced computed tomography (CECT) texture and histogram analyses to differentiate between benign and malignant thyroid nodules.

MATERIALS AND METHODS

The clinical data from 49 patients with 60 thyroid nodules were retrospectively analyzed. Nodules were classified as malignant or benign based on their histological results. Five texture and histogram parameters of thyroid nodules from CECT images, including entropy, mean, standard deviation, skewness, and kurtosis, were compared and analyzed between the two groups. Regions of interest in axial CECT images were delineated manually by two radiologists. Interobserver agreement in texture and histogram parameters between the two radiologists was assessed using the intraclass correlation coefficient (ICC). The Mann-Whitney U test and receiver operating characteristic curve analysis were conducted to estimate the diagnostic capability of texture parameters.

RESULTS

Interobserver reproducibility (ICC = 0.919-0.969) was excellent. Among the 60 nodules, 36 were malignant and 24 were benign. Entropy of malignant thyroid nodules was significantly higher compared with benign thyroid nodules (P = 0.005). A trend toward a higher kurtosis value was observed in malignant thyroid nodules (P = 0.062). When an entropy value of 6.55 was used as a cutoff for differentiating benign from malignant thyroid nodules, the optimal area under the curve, sensitivity, and specificity were 0.716 (0.585-0.847, 95% confidence interval, P = 0.005), 75.0%, and 62.5%, respectively.

CONCLUSIONS

CECT texture and histogram analyses can be used to differentiate benign from malignant thyroid nodules.

Diagnostic accuracy of single-source dual-energy computed tomography and ultrasonography for detection of lateral cervical lymph node metastases of papillary thyroid carcinoma

Background

Dual-energy computed tomography (DECT) imaging can generate iodine-based material decomposition (MD) images and spectral HU curve. This study aimed to investigate the diagnostic accuracy of single-source dual-energy CT (DECT) and ultrasonography (US) for detecting lateral cervical lymph node metastases of papillary thyroid carcinoma (PTC).

Methods

Thirty patients with PTC were enrolled in the study and underwent DECT and US examination before thyroidectomy and cervical lymph node dissection. The spectral parameters included iodine concentration (IC), normalized iodine concentration (NIC) and slope (λ_HU) of lymph nodes. The CT morphological parameters included maximal short diameter, shape, margin, calcification and cystic degeneration of lymph nodes. The US morphological parameters included maximal short diameter, calcification and cystic degeneration of lymph nodes. The diagnostic value of every single spectral parameter, combined gemstone spectral image (GSI) parameters, CT morphological parameters and US morphological parameters between metastatic and non-metastatic lymph nodes were statistically compared. Receiver operating characteristic (ROC) curves, sensitivity, and specificity were used to determine the diagnostic value.

Results

Ninety-nine lymph nodes from thirty patients were pathologically confirmed. Among them, 70 (70.7%) were metastatic. For single GSI parameters, ROC analysis showed that the area under the curve (AUC) for IC was the highest (AUC =0.937) but the difference was not statistically significant when compared with NIC or slope (λ_HU) (P>0.05). The optimal diagnostic threshold for IC was 2.56 mg/mL, with a sensitivity, specificity and accuracy of 87.1%, 93.1%, and 88.9%, respectively. The AUC for combined GSI parameter (AUC =0.942) was higher compared with the US morphological parameters (AUC =0.771, P<0.001), with a sensitivity, specificity, and accuracy of 92.9%, 86.2%, and 90.9%, respectively. However AUC did not differ significantly among combined GSI parameters, combined CT morphological parameters and a single application for spectral CT parameters IC (P>0.05).

Conclusions

Combined GSI parameters showed better diagnostic accuracy in lateral cervical lymph node metastasis of PTC compared with that of combined US morphological parameters. IC alone showed excellent diagnostic stability and could be performed easily.

Cascade and Fusion of Multitask Convolutional Neural Networks for Detection of Thyroid Nodules in Contrast-Enhanced CT

With the development of computed tomography (CT), the contrast-enhanced CT scan is widely used in the diagnosis of thyroid nodules. However, due to the artifacts and high complexity of thyroid CT images, traditional machine learning has difficulty in detecting thyroid nodules in contrast-enhanced CT. A fully automated detection algorithm for thyroid nodules using contrast-enhanced CT images is developed. A modified U-Net architecture of fully convolutional networks is employed to segment the thyroid region of interest (ROI), and a fusion of convolutional neural networks (CNN-Fs) is proposed to detect benign and malignant thyroid nodules from the ROI images and original contrast-enhanced CT images. Experimental results demonstrate that the proposed cascade and fusion method of multitask convolutional neural networks (CNNs) is efficient in diagnosing thyroid diseases with contrast-enhanced CT images and has superior performance compared with other CNN methods.