Texture analysis and machine learning to characterize suspected thyroid nodules and differentiated thyroid cancer: Where do we stand?

DEL-Thyroid: deep ensemble learning framework for detection of thyroid cancer progression through genomic mutation

Genes, expressed as sequences of nucleotides, are susceptible to mutations, some of which can lead to cancer. Machine learning and deep learning methods have emerged as vital tools in identifying mutations associated with cancer. Thyroid cancer ranks as the 5th most prevalent cancer in the USA, with thousands diagnosed annually. This paper presents an ensemble learning model leveraging deep learning techniques such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs), and Bi-directional LSTM (Bi-LSTM) to detect thyroid cancer mutations early. The model is trained on a dataset sourced from asia.ensembl.org and IntOGen.org, consisting of 633 samples with 969 mutations across 41 genes, collected from individuals of various demographics. Feature extraction encompasses techniques including Hahn moments, central moments, raw moments, and various matrix-based methods. Evaluation employs three testing methods: self-consistency test (SCT), independent set test (IST), and 10-fold cross-validation test (10-FCVT). The proposed ensemble learning model demonstrates promising performance, achieving 96% accuracy in the independent set test (IST). Statistical measures such as training accuracy, testing accuracy, recall, sensitivity, specificity, Mathew's Correlation Coefficient (MCC), loss, training accuracy, F1 Score, and Cohen's kappa are utilized for comprehensive evaluation.

Differential diagnosis of thyroid nodules using heterogeneity quantification software on ultrasound images: correlation with the Bethesda system and surgical pathology

Ultrasonography (US)-guided fine-needle aspiration cytology (FNAC) is the primary modality for evaluating thyroid nodules. However, in cases of atypia of undetermined significance (AUS) or follicular lesion of undetermined significance (FLUS), supplemental tests are necessary for a definitive diagnosis. Accordingly, we aimed to develop a non-invasive quantification software using the heterogeneity scores of thyroid nodules. This cross-sectional study retrospectively enrolled 188 patients who were categorized into four groups according to their diagnostic classification in the Bethesda system and surgical pathology [II-benign (B) (n = 24); III-B (n = 52); III-malignant (M) (n = 54); V/VI-M (n = 58)]. Heterogeneity scores were derived using an image pixel-based heterogeneity index, utilized as a coefficient of variation (CV) value, and analyzed across all US images. Differences in heterogeneity scores were compared using one-way analysis of variance with Tukey's test. Diagnostic accuracy was determined by calculating the area under the receiver operating characteristic (AUROC) curve. The results of this study indicated significant differences in mean heterogeneity scores between benign and malignant thyroid nodules, except in the comparison between III-M and V/VI-M nodules. Among malignant nodules, the Bethesda classification was not observed to be associated with mean heterogeneity scores. Moreover, there was a positive correlation between heterogeneity scores and the combined diagnostic category, which was based on the Bethesda system and surgical cytology grades (R = 0.639, p < 0.001). AUROC for heterogeneity scores showed the highest diagnostic performance (0.818; cut-off: 30.22% CV value) for differentiating the benign group (normal/II-B/III-B) from the malignant group (III-M/V&VI-M), with a diagnostic accuracy of 72.5% (161/122). Quantitative heterogeneity measurement of US images is a valuable non-invasive diagnostic tool for predicting the likelihood of malignancy in thyroid nodules, including AUS or FLUS.

A Novel Interpretable Radiomics Model to Distinguish Nodular Goiter From Malignant Thyroid Nodules

OBJECTIVES

The purpose of this study is to inquire about the potential association between radiomics features and the pathological nature of thyroid nodules (TNs), and to propose an interpretable radiomics-based model for predicting the risk of malignant TN.

METHODS

In this retrospective study, computed tomography (CT) imaging and pathological data from 141 patients with TN were collected. The data were randomly stratified into a training group (n = 112) and a validation group (n = 29) at a ratio of 4:1. A total of 1316 radiomics features were extracted by using the pyradiomics tool. The redundant features were removed through correlation testing, and the least absolute shrinkage and selection operator (LASSO) or the minimum redundancy maximum relevance standard was used to select features. Finally, 4 different machine learning models (RF Hybrid Feature, SVM Hybrid Feature, RF, and LASSO) were constructed. The performance of the 4 models was evaluated using the receiver operating characteristic curve. The calibration curve, decision curve analysis, and SHapley Additive exPlanations method were used to evaluate or explain the best radiomics machine learning model.

RESULTS

The optimal radiomics model (RF Hybrid Feature model) demonstrated a relatively high degree of discrimination with an area under the receiver operating characteristic curve (AUC) of 0.87 (95% CI, 0.70-0.97; P < 0.001) for the validation cohort. Compared with the commonly used LASSO model (AUC, 0.78; 95% CI, 0.60-0.91; P < 0.01), there is a significant improvement in AUC in the validation set, net reclassification improvement, 0.79 (95% CI, 0.13-1.46; P < 0.05), and integrated discrimination improvement, 0. 20 (95% CI, 0.10-0.30; P < 0.001).

CONCLUSION

The interpretable radiomics model based on CT performs well in predicting benign and malignant TNs by using quantitative radiomics features of the unilateral total thyroid. In addition, the data preprocessing method incorporating different layers of features has achieved excellent experimental results.

CLINICAL RELEVANCE STATEMENT

As the detection rate of TNs continues to increase, so does the diagnostic burden on radiologists. This study establishes a noninvasive, interpretable and accurate machine learning model to rapidly identify the nature of TN found in CT.

Application of radiomics and machine learning to thyroid diseases in nuclear medicine: a systematic review

BACKGROUND

In the last years growing evidences on the role of radiomics and machine learning (ML) applied to different nuclear medicine imaging modalities for the assessment of thyroid diseases are starting to emerge. The aim of this systematic review was therefore to analyze the diagnostic performances of these technologies in this setting.

METHODS

A wide literature search of the PubMed/MEDLINE, Scopus and Web of Science databases was made in order to find relevant published articles about the role of radiomics or ML on nuclear medicine imaging for the evaluation of different thyroid diseases.

RESULTS

Seventeen studies were included in the systematic review. Radiomics and ML were applied for assessment of thyroid incidentalomas at 18 F-FDG PET, evaluation of cytologically indeterminate thyroid nodules, assessment of thyroid cancer and classification of thyroid diseases using nuclear medicine techniques.

CONCLUSION

Despite some intrinsic limitations of radiomics and ML may have affect the results of this review, these technologies seem to have a promising role in the assessment of thyroid diseases. Validation of preliminary findings in multicentric studies is needed to translate radiomics and ML approaches in the clinical setting.

Predicting BRAFV600E mutations in papillary thyroid carcinoma using six machine learning algorithms based on ultrasound elastography

The most common BRAF mutation is thymine (T) to adenine (A) missense mutation in nucleotide 1796 (T1796A, V600E). The BRAFV600E gene encodes a protein-dependent kinase (PDK), which is a key component of the mitogen-activated protein kinase pathway and essential for controlling cell proliferation, differentiation, and death. The BRAFV600E mutation causes PDK to be activated improperly and continuously, resulting in abnormal proliferation and differentiation in PTC. Based on elastography ultrasound (US) radiomic features, this study seeks to create and validate six distinct machine learning algorithms to predict BRAFV6OOE mutation in PTC patients prior to surgery. This study employed routine US strain elastography image data from 138 PTC patients. The patients were separated into two groups: those who did not have the BRAFV600E mutation (n = 75) and those who did have the mutation (n = 63). The patients were randomly assigned to one of two data sets: training (70%), or validation (30%). From strain elastography US images, a total of 479 radiomic features were retrieved. Pearson's Correlation Coefficient (PCC) and Recursive Feature Elimination (RFE) with stratified tenfold cross-validation were used to decrease the features. Based on selected radiomic features, six machine learning algorithms including support vector machine with the linear kernel (SVM_L), support vector machine with radial basis function kernel (SVM_RBF), logistic regression (LR), Naïve Bayes (NB), K-nearest neighbors (KNN), and linear discriminant analysis (LDA) were compared to predict the possibility of BRAFV600E. The accuracy (ACC), the area under the curve (AUC), sensitivity (SEN), specificity (SPEC), positive predictive value (PPV), negative predictive value (NPV), decision curve analysis (DCA), and calibration curves of the machine learning algorithms were used to evaluate their performance. ① The machine learning algorithms' diagnostic performance depended on 27 radiomic features. ② AUCs for NB, KNN, LDA, LR, SVM_L, and SVM_RBF were 0.80 (95% confidence interval [CI]: 0.65-0.91), 0.87 (95% CI 0.73-0.95), 0.91(95% CI 0.79-0.98), 0.92 (95% CI 0.80-0.98), 0.93 (95% CI 0.80-0.98), and 0.98 (95% CI 0.88-1.00), respectively. ③ There was a significant difference in echogenicity,vertical and horizontal diameter ratios, and elasticity between PTC patients with BRAFV600E and PTC patients without BRAFV600E. Machine learning algorithms based on US elastography radiomic features are capable of predicting the likelihood of BRAFV600E in PTC patients, which can assist physicians in identifying the risk of BRAFV600E in PTC patients. Among the six machine learning algorithms, the support vector machine with radial basis function (SVM_RBF) achieved the best ACC (0.93), AUC (0.98), SEN (0.95), SPEC (0.90), PPV (0.91), and NPV (0.95).

A Novel Deep-Learning-Based CADx Architecture for Classification of Thyroid Nodules Using Ultrasound Images

Nodules of thyroid cancer occur in the cells of the thyroid as benign or malign types. Thyroid sonographic images are mostly used for diagnosis of thyroid cancer. The aim of this study is to introduce a computer-aided diagnosis system that can classify the thyroid nodules with high accuracy using the data gathered from ultrasound images. Acquisition and labeling of sub-images were performed by a specialist physician. Then the number of these sub-images were increased using data augmentation methods. Deep features were obtained from the images using a pre-trained deep neural network. The dimensions of the features were reduced and features were improved. The improved features were combined with morphological and texture features. This feature group was rated by a value called similarity coefficient value which was obtained from a similarity coefficient generator module. The nodules were classified as benign or malignant using a multi-layer deep neural network with a pre-weighting layer designed with a novel approach. In this study, a novel multi-layer computer-aided diagnosis system was proposed for thyroid cancer detection. In the first layer of the system, a novel feature extraction method based on the class similarity of images was developed. In the second layer, a novel pre-weighting layer was proposed by modifying the genetic algorithm. The proposed system showed superior performance in different metrics compared to the literature.

Ovarian imaging radiomics quality score assessment: an EuSoMII radiomics auditing group initiative

OBJECTIVE

To evaluate the methodological rigor of radiomics-based studies using noninvasive imaging in ovarian setting.

METHODS

Multiple medical literature archives (PubMed, Web of Science, and Scopus) were searched to retrieve original studies focused on computed tomography (CT), magnetic resonance imaging (MRI), ultrasound (US), or positron emission tomography (PET) radiomics for ovarian disorders' assessment. Two researchers in consensus evaluated each investigation using the radiomics quality score (RQS). Subgroup analyses were performed to assess whether the total RQS varied according to first author category, study aim and topic, imaging modality, and journal quartile.

RESULTS

From a total of 531 items, 63 investigations were finally included in the analysis. The studies were greatly focused (94%) on the field of oncology, with CT representing the most used imaging technique (41%). Overall, the papers achieved a median total RQS 6 (IQR, -0.5 to 11), corresponding to a percentage of 16.7% of the maximum score (IQR, 0-30.6%). The scoring was low especially due to the lack of prospective design and formal validation of the results. At subgroup analysis, the 4 studies not focused on oncological topic showed significantly lower quality scores than the others.

CONCLUSIONS

The overall methodological rigor of radiomics studies in the ovarian field is still not ideal, limiting the reproducibility of results and potential translation to clinical setting. More efforts towards a standardized methodology in the workflow are needed to allow radiomics to become a viable tool for clinical decision-making.

KEY POINTS

• The 63 included studies using noninvasive imaging for ovarian applications were mostly focused on oncologic topic (94%). • The included investigations achieved a median total RQS 6 (IQR, -0.5 to 11), indicating poor methodological rigor. • The RQS was low especially due to the lack of prospective design and formal validation of the results.

The progress of radiomics in thyroid nodules

Due to the development of Artificial Intelligence (AI), Machine Learning (ML), and the improvement of medical imaging equipment, radiomics has become a popular research in recent years. Radiomics can obtain various quantitative features from medical images, highlighting the invisible image traits and significantly enhancing the ability of medical imaging identification and prediction. The literature indicates that radiomics has a high potential in identifying and predicting thyroid nodules. So in this article, we explain the development, definition, and workflow of radiomics. And then, we summarize the applications of various imaging techniques in identifying benign and malignant thyroid nodules, predicting invasiveness and metastasis of thyroid lymph nodes, forecasting the prognosis of thyroid malignancies, and some new advances in molecular level and deep learning. The shortcomings of this technique are also summarized, and future development prospects are provided.

Differential diagnosis and feature visualization for thyroid nodules using computer-aided ultrasonic diagnosis system: initial clinical assessment

BACKGROUND

To assess the diagnostic efficacy of the computer-aided ultrasonic diagnosis system (CAD system) in differentiating benign and malignant thyroid nodules.

METHODS

The images of 296 thyroid nodules were included in validation sets. The diagnostic efficacy of the CAD system was compared with that of junior physicians and senior physicians, as well as that of the combination diagnosis of the CAD system with junior physicians. The diagnostic efficacy of the CAD system for different sizes of thyroid nodules was compared.

RESULTS

The diagnostic sensitivity and accuracy of the CAD system were higher than those of junior physicians (83.4% vs. 72.2%, 73.0% vs. 69.6%), but the diagnostic specificity of the CAD system was lower than that of junior physicians (62.1% vs. 66.9%). The diagnostic accuracy of the CAD system was lower than that of senior physicians (73.0% vs. 83.8%). However, the combination diagnosis of the CAD system with junior physicians had higher accuracy (81.8%) and AUC (0.842) than those of either the CAD system or junior physicians alone, and comparable diagnostic performance with those of senior physicians. The Kappa was 0.635 in the combination diagnosis of the CAD system with junior physicians, showing good consistency with the pathological results. The accuracy (76.4%) of the CAD system was the highest for nodules of 1-2 cm.

CONCLUSION

The CAD system can effectively assist physicians to identify malignant and benign thyroid nodules, reduce the overdiagnosis and overtreatment of thyroid nodules, avoid unnecessary invasive fine needle aspiration, and improve the diagnostic accuracy of junior physicians.

Machine Learning-Assisted Diagnostic System for Indeterminate Thyroid Nodules

To develop an ultrasound-based machine learning classifier to diagnose benignity within indeterminate thyroid nodules (ITNs) by fine-needle aspiration, 180 patients with 194 ITNs (Bethesda classes III, IV and V) undergoing surgery over a 5-y study period were analyzed. The data set was randomly divided into training and testing data sets with 155 and 39 ITNs, respectively. All nodules were evaluated by ultrasound using the American College of Radiology Thyroid Imaging Reporting and Data System by manually scoring composition, echogenicity, shape, margin and echogenic foci. Nodule size, participant age and patient sex were recorded. A support vector machine (SVM) model with a cost-sensitive approach was developed using the aforementioned eight parameters with surgical histopathology as the reference standard. Surgical pathology determined 90 (46.4%) ITNs were malignant and 104 (53.6%) were benign. The SVM model classified 14 nodules as benign in the testing data set, of which 13 were correct (sensitivity = 93.8%, specificity = 56.5%). Considering malignancy prevalence by Bethesda group, the negative predictive values of this model for Bethesda III and IV categories were 93.9% and 93. 8%, respectively. The high negative predictive value of the SVM ultrasound-based model suggests a pathway by which surgical excision of Bethesda III and IV ITNs classified as benign may be avoided.

Artificial Intelligence for Thyroid Nodule Characterization: Where Are We Standing?

Simple Summary

In the present review, an up-to-date summary of the state of the art of artificial intelligence (AI) implementation for thyroid nodule characterization and cancer is provided. The opinion on the real effectiveness of AI systems remains controversial. Taking into consideration the largest and most scientifically valid studies, it is possible to state that AI provides results that are comparable or inferior to expert ultrasound specialists and radiologists. Promising data approve AI as a support tool and simultaneously highlight the need for a radiologist supervisory framework for AI provided results. Therefore, current solutions might be more suitable for educational purposes.

Abstract

Machine learning (ML) is an interdisciplinary sector in the subset of artificial intelligence (AI) that creates systems to set up logical connections using algorithms, and thus offers predictions for complex data analysis. In the present review, an up-to-date summary of the current state of the art regarding ML and AI implementation for thyroid nodule ultrasound characterization and cancer is provided, highlighting controversies over AI application as well as possible benefits of ML, such as, for example, training purposes. There is evidence that AI increases diagnostic accuracy and significantly limits inter-observer variability by using standardized mathematical algorithms. It could also be of aid in practice settings with limited sub-specialty expertise, offering a second opinion by means of radiomics and computer-assisted diagnosis. The introduction of AI represents a revolutionary event in thyroid nodule evaluation, but key issues for further implementation include integration with radiologist expertise, impact on workflow and efficiency, and performance monitoring.

The European Institute of Oncology Thyroid Imaging Reporting and Data System for Classification of Thyroid Nodules: A Prospective Study

Background: To evaluate the performance, quality and effectiveness of “IEO-TIRADS” in assigning a TI-RADS score to thyroid nodules (TN) when compared with “EU-TIRADS” and the US risk score calculated with the S-Detect software (“S-Detect”). The primary objective is the evaluation of diagnostic accuracy (DA) by “IEO-TIRADS”, “S-Detect” and “EU-TIRADS”, and the secondary objective is to evaluate the diagnostic performances of the scores, using the histological report as the gold standard. Methods: A radiologist collected all three scores of the TNs detected and determined the risk of malignancy. The results of all the scores were compared with the histological specimens. The sensitivity (SE), specificity (SP), and diagnostic accuracy (DA), with their 95% confidence interval (95% CI), were calculated for each method. Results: 140 TNs were observed in 93 patients and classified according to all three scores. “IEO-TIRADS” has an SE of 73.6%, an SP of 59.2% and a DA of 68.6%. “EU-TIRADS” has an SE of 90.1%, an SP of 32.7% and a DA of 70.0%. “S-Detect” has an SE of 67.0%, an SP of 69.4% and a DA of 67.9%. Conclusion: “IEO-TIRADS” has a similar diagnostic performance to “S-Detect” and “EU-TIRADS”. Providing a comparable DA with other reporting systems, IEO-TIRADS holds the potential of being applied in clinical practice.

Value of Whole-Thyroid CT-Based Radiomics in Predicting Benign and Malignant Thyroid Nodules

The objective of this research is to explore the value of whole-thyroid CT-based radiomics in predicting benign (noncancerous) and malignant thyroid nodules. The imaging and clinical data of 161 patients with thyroid nodules that were confirmed by pathology were retrospectively analyzed. The entire thyroid regions of interest (ROIs) were manually sketched for all 161 cases. After extracting CT radiomic features, the patients were divided into a training group (128 cases) and a test group (33 cases) according to the 4:1 ratio with stratified random sampling (fivefold cross validation). All the data were normalized by the maximum absolute value and screened using selection operator regression analysis and K best. The data generation model was trained by logistic regression. The effectiveness of the model in differentiating between benign and malignant thyroid nodules was validated by a receiver operating characteristic (ROC) curve. After data grouping, eigenvalue screening, and data training, the logistic regression model with the maximum absolute value normalized was constructed. For the training group, the area under the ROC curve (AUC) was 94.4% (95% confidence interval: 0.941–0.977); the sensitivity and specificity were 89.7% and 86.7%, respectively; and the diagnostic accuracy was 87.6%. For the test group, the AUC was 94.2% (95% confidence interval: 0.881–0.999); the sensitivity and specificity were 89.4% and 86.8%, respectively; and the diagnostic accuracy was 87.6%. The CT radiomic model of the entire thyroid gland is highly efficient in differentiating between benign and malignant thyroid nodules.

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.

Radiomics Analysis of [18F]FDG PET/CT Thyroid Incidentalomas: How Can It Improve Patients’ Clinical Management? A Systematic Review from the Literature

Background: We performed a systematic review of the literature to provide an overview of the application of PET-based radiomics of [18F]FDG-avid thyroid incidentalomas and to discuss the additional value of PET volumetric parameters and radiomic features over clinical data. Methods: The most relevant databases were explored using an algorithm constructed based on a combination of terms related to our subject and English-language articles published until October 2021 were considered. Among the 247 identified articles, 19 studies were screened for eligibility and 11 met the criteria, with 4 studies including radiomics analyses. Results: We summarized the available literature based on a study of PET volumetric parameters and radiomics. Ten articles provided accurate details about volumetric parameters and their prospective value in tumour assessment. We included the data provided by these articles in a sub-analysis, but could not obtain statistically relevant results. Four publications analyzed the diagnostic potential of radiomics and the possibility of creating precise predictive models, their corresponding quality score being assessed. Conclusions: The use of PET volumetric parameters and radiomics analysis in patients with [18F]FDG-avid thyroid incidentalomas outlines a great prospect in diagnosis and stratification of patients with malignant nodules and may represent a way of limiting the need for unnecessary invasive procedures; however, further studies need to be performed for a standardization of the method.

Quantitative classification and radiomics of [18F]FDG-PET/CT in indeterminate thyroid nodules

PURPOSE

To evaluate whether quantitative [¹⁸F]FDG-PET/CT assessment, including radiomic analysis of [¹⁸F]FDG-positive thyroid nodules, improved the preoperative differentiation of indeterminate thyroid nodules of non-Hürthle cell and Hürthle cell cytology.

METHODS

Prospectively included patients with a Bethesda III or IV thyroid nodule underwent [¹⁸F]FDG-PET/CT imaging. Receiver operating characteristic (ROC) curve analysis was performed for standardised uptake values (SUV) and SUV-ratios, including assessment of SUV cut-offs at which a malignant/borderline neoplasm was reliably ruled out (≥ 95% sensitivity). [¹⁸F]FDG-positive scans were included in radiomic analysis. After segmentation at 50% of SUV_peak, 107 radiomic features were extracted from [¹⁸F]FDG-PET and low-dose CT images. Elastic net regression classifiers were trained in a 20-times repeated random split. Dimensionality reduction was incorporated into the splits. Predictive performance of radiomics was presented as mean area under the ROC curve (AUC) across the test sets.

RESULTS

Of 123 included patients, 84 (68%) index nodules were visually [¹⁸F]FDG-positive. The malignant/borderline rate was 27% (33/123). SUV-metrices showed AUCs ranging from 0.705 (95% CI, 0.601-0.810) to 0.729 (0.633-0.824), 0.708 (0.580-0.835) to 0.757 (0.650-0.864), and 0.533 (0.320-0.747) to 0.700 (0.502-0.898) in all (n = 123), non-Hürthle (n = 94), and Hürthle cell (n = 29) nodules, respectively. At SUV_max, SUV_peak, SUV_max-ratio, and SUV_peak-ratio cut-offs of 2.1 g/mL, 1.6 g/mL, 1.2, and 0.9, respectively, sensitivity of [¹⁸F]FDG-PET/CT was 95.8% (95% CI, 78.9-99.9%) in non-Hürthle cell nodules. In Hürthle cell nodules, cut-offs of 5.2 g/mL, 4.7 g/mL, 3.4, and 2.8, respectively, resulted in 100% sensitivity (95% CI, 66.4-100%). Radiomic analysis of 84 (68%) [¹⁸F]FDG-positive nodules showed a mean test set AUC of 0.445 (95% CI, 0.290-0.600) for the PET model.

CONCLUSION

Quantitative [¹⁸F]FDG-PET/CT assessment ruled out malignancy in indeterminate thyroid nodules. Distinctive, higher SUV cut-offs should be applied in Hürthle cell nodules to optimize rule-out ability. Radiomic analysis did not contribute to the additional differentiation of [¹⁸F]FDG-positive nodules.

TRIAL REGISTRATION NUMBER

This trial is registered with ClinicalTrials.gov: NCT02208544 (5 August 2014), https://clinicaltrials.gov/ct2/show/NCT02208544 .

Machine intelligence in non-invasive endocrine cancer diagnostics

Artificial intelligence (AI) has illuminated a clear path towards an evolving health-care system replete with enhanced precision and computing capabilities. Medical imaging analysis can be strengthened by machine learning as the multidimensional data generated by imaging naturally lends itself to hierarchical classification. In this Review, we describe the role of machine intelligence in image-based endocrine cancer diagnostics. We first provide a brief overview of AI and consider its intuitive incorporation into the clinical workflow. We then discuss how AI can be applied for the characterization of adrenal, pancreatic, pituitary and thyroid masses in order to support clinicians in their diagnostic interpretations. This Review also puts forth a number of key evaluation criteria for machine learning in medicine that physicians can use in their appraisals of these algorithms. We identify mitigation strategies to address ongoing challenges around data availability and model interpretability in the context of endocrine cancer diagnosis. Finally, we delve into frontiers in systems integration for AI, discussing automated pipelines and evolving computing platforms that leverage distributed, decentralized and quantum techniques.

Texture and shape analysis of diffusion-weighted imaging for thyroid nodules classification using machine learning

PURPOSE

To assess whether the integration between (a) functional imaging features that will be extracted from diffusion-weighted imaging (DWI); and (b) shape and texture imaging features as well as volumetric features that will be extracted from T2-weighted magnetic resonance imaging (MRI) can noninvasively improve the diagnostic accuracy of thyroid nodules classification.

PATIENTS AND METHODS

In a retrospective study of 55 patients with pathologically proven thyroid nodules, T2-weighted and diffusion-weighted MRI scans of the thyroid gland were acquired. Spatial maps of the apparent diffusion coefficient (ADC) were reconstructed in all cases. To quantify the nodules' morphology, we used spherical harmonics as a new parametric shape descriptor to describe the complexity of the thyroid nodules in addition to traditional volumetric descriptors (e.g., tumor volume and cuboidal volume). To capture the inhomogeneity of the texture of the thyroid nodules, we used the histogram-based statistics (e.g., kurtosis, entropy, skewness, etc.) of the T2-weighted signal. To achieve the main goal of this paper, a fusion system using an artificial neural network (NN) is proposed to integrate both the functional imaging features (ADC) with the structural morphology and texture features. This framework has been tested on 55 patients (20 patients with malignant nodules and 35 patients with benign nodules), using leave-one-subject-out (LOSO) for training/testing validation tests.

RESULTS

The functionality, morphology, and texture imaging features were estimated for 55 patients. The accuracy of the computer-aided diagnosis (CAD) system steadily improved as we integrate the proposed imaging features. The fusion system combining all biomarkers achieved a sensitivity, specificity, positive predictive value, negative predictive value, F1-score, and accuracy of 92.9 % (confidence interval [CI]: 78.9 % -- 99.5 % ), 95.8 % (CI: 87.4 % -- 99.7 % ), 93 % (CI: 80.7 % -- 99.5 % ), 96 % (CI: 88.8 % -- 99.7 % ), 92.8 % (CI: 83.5 % -- 98.5 % ), and 95.5 % (CI: 88.8 % -- 99.2 % ), respectively, using the LOSO cross-validation approach.

CONCLUSION

The results demonstrated in this paper show the promise that integrating the functional features with morphology as well as texture features by using the current state-of-the-art machine learning approaches will be extremely useful for identifying thyroid nodules as well as diagnosing their malignancy.

The use of radiomic analysis of magnetic resonance imaging in predicting distant metastases of rectal carcinoma following surgical resection: A systematic review and meta-analysis

AIM

Estimating prognosis in rectal carcinoma (RC) is challenging, with distant recurrence (DR) occurring in up to 30% of cases. Radiomics is a novel field using diagnostic imaging to investigate the tumour heterogeneity of cancers and may have the potential to predict DR. The aim of the study was to perform a systematic review of the current literature evaluating the use of radiomics in predicting DR in patients with resected RC.

METHODS

A systematic review was performed as per PRISMA guidelines to identify studies reporting radiomic analysis of magnetic resonance imaging (MRI) to predict DR in patients diagnosed with RC. Sensitivity and specificity of radiomic analyses were included for meta-analysis.

RESULTS

A total of seven studies including 1497 patients (998 males) were included, seven, five and one of whom reported radiomics, respectively. The overall pooled rate of DR from all included studies was 17.1% (256/1497), with 15.6% (236/1497), 1.3% (19/1497) and 0.2% (3/1497) of patients having hepatic, pulmonary and peritoneal metastases. Meta-analysis demonstrated that radiomics correctly predicted DR with pooled sensitivities and specificities of MRI 0.76 (95% CI: 0.73, 0.78) and 0.85 (95% CI: 0.83, 0.88), respectively.

CONCLUSION

This systematic review suggests the benefit of radiomic analysis of preoperative MRI in identifying patients with resected RC at an increased risk of DR. Our findings warrant validation in larger prospective studies as modalities to predict DR is a significant unmet need in RC. Radiomics may allow for tailored therapeutic strategies for high-risk groups.

Differentiation of benign thyroid nodules from malignant thyroid nodules through miR-205-5p and thyroid-stimulating hormone receptor mRNA

PURPOSE

This study was designed to determine the value of miR-205-5p and thyroid-stimulating hormone receptor (TSHR) mRNA in differentiating benign thyroid nodules from malignant thyroid nodules.

METHODS

A total of 67 patients with malignant thyroid nodules admitted to our hospital from October 2016 to March 2018 were enrolled in the malignant group, and 71 patients with benign thyroid nodules were enrolled in the benign group. Another 56 healthy individuals, as determined by physical examination, were enrolled in the control group. Quantitative real-time polymerase chain reaction (qRT-PCR) was adopted to determine the relative expression of miR-205-5p and TSHR mRNA in serum, and receiver operating characteristic (ROC) curves were used to analyze the diagnostic value of miR-205-5p and TSHR mRNA in thyroid nodules and their benignancy or malignancy. The correlation of miR-205-5p and TSHR mRNA with clinical data of the patients with thyroid carcinoma was analyzed, and multivariate logistic regression was used to analyze risk factors for lymph node metastasis of thyroid carcinoma. In addition, Pearson's test was used to analyze the relationship between miR-205-5p and TSHR mRNA in serum of patients with thyroid carcinoma.

RESULTS

The malignant group showed significantly higher expression of miR-205-5p and TSHR mRNA than the benign group and control group (both p <0.001), and the benign group showed significantly higher expression of these than the control group (p <0.01). miR-205-5p expression was related to tumor size, TNM staging, lymph node metastasis, capsular infiltration, and BRAF mutation (p <0.05), and TSHR mRNA expression was related to lymph node metastasis, capsular infiltration, BRAF mutation, and RAS mutation (p <0.05). Gender, number of lesions, capsular infiltration, miR-205-5p, and TSHR mRNA were independent risk factors for lymph node metastasis of patients with thyroid carcinoma. Moreover, there was a positive correlation between miR-205-5p and TSHR mRNA in patients with thyroid carcinoma (r=0.735, p<0.001).

CONCLUSION

MiR-205-5p and TSHR mRNA may be potential diagnosis indexes of thyroid nodules, their benignancy, or malignancy, while the independent risk factors for patients with thyroid carcinoma include a large number of lesions, occurrence of capsular infiltration, and relatively high expression of miR-205-5p and TSHR mRNA.

False Negative Rates in Benign Thyroid Nodule Diagnosis: Machine Learning for Detecting Malignancy

BACKGROUND

Thyroid nodules are common; up to 67% of adults will show nodules on high-quality ultrasound, and 95% of these nodules are benign. FNA cytology is a crucial step in determining the risk of malignancy, and a false negative diagnosis at this stage delays cancer treatment. The purpose of this study is to develop a predictive model using machine learning which can identify false negative FNA results based on less-invasive clinical data.

MATERIALS AND METHODS

We conducted a retrospective medical record review at one academic and one community center. Inclusion criteria were thyroid nodules evaluated by ultrasound and FNA with a Bethesda II (benign) result or malignancy detected on pathology or FNA. Linear, non-linear, and ensemble models were generated with scikit-learn using 10-fold cross validation with repetition and compared with AUROC. The classification task was the prediction of malignancy using information acquired from less-invasive ultrasound and FNA.

RESULTS

A total of 604 subjects met inclusion criteria; 38 were diagnosed with malignancy. Of all algorithms tested, a Random Forest method achieved the best AUROC (0.64) in separating benign and malignant nodules, though the improvement over other tested algorithms was not statistically significant.

CONCLUSIONS

A Random Forest model performed better than random chance using readily available data obtained via standard evaluation of thyroid nodules. The diagnostic probability threshold of this model can be varied to minimize false positives at the cost of increasing the number of false negatives. Future studies will prospectively evaluate the model's performance.

A Comparative Analysis of Six Machine Learning Models Based on Ultrasound to Distinguish the Possibility of Central Cervical Lymph Node Metastasis in Patients With Papillary Thyroid Carcinoma

Current approaches to predict central cervical lymph node metastasis (CLNM) in patients with papillary thyroid carcinoma (PTC) have failed to identify patients who would benefit from preventive treatment. Machine learning has offered the opportunity to improve accuracy by comparing the different algorithms. We assessed which machine learning algorithm can best improve CLNM prediction. This retrospective study used routine ultrasound data of 1,364 PTC patients. Six machine learning algorithms were compared to predict the possibility of CLNM. Predictive accuracy was assessed by sensitivity, specificity, positive predictive value, negative predictive value, and the area under the curve (AUC). The patients were randomly split into the training (70%), validation (15%), and test (15%) data sets. Random forest (RF) led to the best diagnostic model in the test cohort (AUC 0.731 ± 0.036, 95% confidence interval: 0.664–0.791). The diagnostic performance of the RF algorithm was most dependent on the following five top-rank features: extrathyroidal extension (27.597), age (17.275), T stage (15.058), shape (13.474), and multifocality (12.929). In conclusion, this study demonstrated promise for integrating machine learning methods into clinical decision-making processes, though these would need to be tested prospectively.

Value of MRI-based radiomics analysis for differentiation of benign and malignant epithelial neoplasms in the lacrimal gland: a retrospective study

BACKGROUND

It is essential to distinguish malignant from benign epithelial neoplasms in the lacrimal gland for different treatment options and prognosis.

PURPOSE

To retrospectively assess the performance of magnetic resonance imaging (MRI)-based radiomics features in the differentiation of benign and malignant epithelial neoplasms in the lacrimal gland.

MATERIAL AND METHODS

Seventy-six consecutive patients with histopathology-proven epithelial neoplasms of the lacrimal gland were enrolled in the study, including 41 benign and 35 malignant neoplasms. Radiomics features were extracted from T2-weighted and post-contrast T1-weighted imaging. The least absolute shrinkage and selection operator method was used to select imaging features and reduce data dimension to discriminate malignant from benign neoplasms in the lacrimal gland. Diagnostic performance of the radiomics model was assessed by receive operation characteristic (ROC) curve and compared with that of radiologists.

RESULTS

Four quantitative image features including inverse difference moment normalized (IDMN), mean deviation (MD), standard deviation (SD), and long-run emphasis (LRE) were selected to distinguish malignant from benign epithelial neoplasms in the lacrimal gland. Area under the curve (AUC) of these four features were 0.88, 086, 0.88, and 0.86, respectively, with 0.93 for the combination model. The model identified malignant epithelial neoplasms from benign group with 89% sensitivity, 93% specificity, and 89% accuracy. There was a significant difference in the diagnostic performance of radiomics model and the radiologists, with AUC of 0.70 for radiologists. The diagnostic performance of radiomics is superior to that of radiologists.

CONCLUSION

MRI-based radiomics analysis has potential for differentiation of benign and malignant epithelial neoplasms in the lacrimal gland.

Challenges and opportunities for artificial intelligence in oncological imaging

Imaging plays a key role in oncology, including the diagnosis and detection of cancer, determining clinical management, assessing treatment response, and complications of treatment or disease. The current use of clinical oncology is predominantly qualitative in nature with some relatively crude size-based measurements of tumours for assessment of disease progression or treatment response; however, it is increasingly understood that there may be significantly more information about oncological disease that can be obtained from imaging that is not currently utilized. Artificial intelligence (AI) has the potential to harness quantitative techniques to improve oncological imaging. These may include improving the efficiency or accuracy of traditional roles of imaging such as diagnosis or detection. These may also include new roles for imaging such as risk-stratifying patients for different types of therapy or determining biological tumour subtypes. This review article outlines several major areas in oncological imaging where there may be opportunities for AI technology. These include (1) screening and detection of cancer, (2) diagnosis and risk stratification, (3) tumour segmentation, (4) precision oncology, and (5) predicting prognosis and assessing treatment response. This review will also address some of the potential barriers to AI research in oncological imaging.

Machine learning to identify lymph node metastasis from thyroid cancer in patients undergoing contrast-enhanced CT studies

INTRODUCTION

We compared the diagnostic performance of morphological methods such as the major axis, the minor axis, the volume and sphericity and of machine learning with texture analysis in the identification of lymph node metastasis in patients with thyroid cancer who had undergone contrast-enhanced CT studies.

METHODS

We sampled 772 lymph nodes with histology defined tissue types (84 metastatic and 688 benign lymph nodes) that were visualised on CT images of 117 patients. A support vector machine (SVM), free programming software (Python), and the scikit-learn machine learning library were used to discriminate metastatic-from benign lymph nodes. We assessed 96 texture and 4 morphological features (major axis, minor axis, volume, sphericity) that were reported useful for the differentiation between metastatic and benign lymph nodes on CT images. The area under the curve (AUC) obtained by receiver operating characteristic analysis of univariate logistic regression and SVM classifiers were calculated for the training and testing datasets.

RESULTS

The AUC for all classifiers in training and testing datasets was 0.96 and 0.86, at the SVM for machine learning. When we applied conventional methods to the training and testing datasets, the AUCs were 0.63 and 0.48 for the major axis, 0.70 and 0.44 for the minor axis, 0.66 and 0.43 for the volume, and 0.69 and 0.54 for sphericity, respectively. The SVM using texture features yielded significantly higher AUCs than univariate logistic regression models using morphological features (p = 0.001).

CONCLUSION

For the identification of metastatic lymph nodes from thyroid cancer on contrast-enhanced CT images, machine learning combined with texture analysis was superior to conventional diagnostic methods with the morphological parameters.

IMPLICATIONS FOR PRACTICE

Our findings suggest that in patients with thyroid cancer and suspected lymph node metastasis who undergo contrast-enhanced CT studies, machine learning using texture analysis is high diagnostic value for the identification of metastatic lymph nodes.

Removal of manually induced artifacts in ultrasound images of thyroid nodules based on edge-connection and Criminisi image restoration algorithm

BACKGROUND AND OBJECTIVE

There are various artificial markers in ultrasound images of thyroid nodules, which have impact on subsequent processing and computer-aided diagnosis. The purpose of this study was to develop an approach to automatically remove artifacts and restore ultrasound images of thyroid nodules.

METHODS

Fifty ultrasound images with manually induced artifacts were selected from publicly available and self-collected datasets. A combined approach was developed which consisted of two steps, artifacts detection and removal of the detected artifacts. Specifically, a novel edge-connection algorithm was used for artifact detection, detection accuracy and false discovery rate were used to evaluate the performance of artifact detection approaches. Criminisi algorithm was used for image restoration with peak signal-to-noise ratio (PSNR) and mean gradient difference to evaluate its performance. In addition, computation complexity was evaluated by execution time of relevant algorithms.

RESULTS

Results revealed that the proposed joint approach with edge-connection and Criminisi algorithm could achieve automatic artifacts removal. Mean detection accuracy and mean false discovery rate of the proposed edge-connection algorithm for the 50 ultrasound images were 0.86 and 1.50. Mean PSNR of the 50 restored images by Criminisi algorithm was 36.64 dB, and mean gradient difference of the restored images was -0.002 compared with the original images.

CONCLUSIONS

The proposed combined approach had a good detection accuracy for different types of manually induced artifacts, and could significantly improve PSNR of the ultrasound images. The proposed combined approach may have potential use for the repair of ultrasound images with artifacts.

On the identification of thyroid nodules using semi-supervised deep learning

Detecting malign cases from thyroid nodule examinations is crucial in healthcare particularly to improve the early detection of such cases. However, malign thyroid nodules can be extremely rare and is hard to find using the traditional rule based or expert-based methods. For this reason, the solutions backed by Machine Learning (ML) algorithms are key to improve the detection rates of such rare cases. In this paper, we investigate the application of ML in the healthcare domain for the detection of rare thyroid nodules. The utilized dataset is collected from 636 distinct patients in 99 unique days in Turkey. In addition to the texture feature data of the Ultrasound (US), we have also included the scores of different assessment methods created by different health institutions (e.g., Korean, American and European thyroid societies) as additional features. For detection of extremely rare malign cases, we use auto-encoder based neural network model. Through numerical results, it is shown that the auto-encoder based model can result in an average Recall score of 0.98 and a Sensitivity score of 1.00 for detecting malign and non-malign cases from the healthcare dataset outperforming the traditional classification algorithms that are trained after Synthetic Minority Oversampling Technique (SMOTE) oversampling.

Artificial Intelligence for Personalized Medicine in Thyroid Cancer: Current Status and Future Perspectives

Thyroid cancers (TC) have increasingly been detected following advances in diagnostic methods. Risk stratification guided by refined information becomes a crucial step toward the goal of personalized medicine. The diagnosis of TC mainly relies on imaging analysis, but visual examination may not reveal much information and not enable comprehensive analysis. Artificial intelligence (AI) is a technology used to extract and quantify key image information by simulating complex human functions. This latent, precise information contributes to stratify TC on the distinct risk and drives tailored management to transit from the surface (population-based) to a point (individual-based). In this review, we started with several challenges regarding personalized care in TC, for example, inconsistent rating ability of ultrasound physicians, uncertainty in cytopathological diagnosis, difficulty in discriminating follicular neoplasms, and inaccurate prognostication. We then analyzed and summarized the advances of AI to extract and analyze morphological, textural, and molecular features to reveal the ground truth of TC. Consequently, their combination with AI technology will make individual medical strategies possible.

Diagnostic performance evaluation of different TI-RADS using ultrasound computer-aided diagnosis of thyroid nodules: An experience with adjusted settings

BACKGROUND

Thyroid cancer diagnosis has evolved to include computer-aided diagnosis (CAD) approaches to overcome the limitations of human ultrasound feature assessment. This study aimed to evaluate the diagnostic performance of a CAD system in thyroid nodule differentiation using varied settings.

METHODS

Ultrasound images of 205 thyroid nodules from 198 patients were analysed in this retrospective study. AmCAD-UT software was used at default settings and 3 adjusted settings to diagnose the nodules. Six risk-stratification systems in the software were used to classify the thyroid nodules: The American Thyroid Association (ATA), American College of Radiology Thyroid Imaging, Reporting, and Data System (ACR-TIRADS), British Thyroid Association (BTA), European Union (EU-TIRADS), Kwak (2011) and the Korean Society of Thyroid Radiology (KSThR). The diagnostic performance of CAD was determined relative to the histopathology and/or cytology diagnosis of each nodule.

RESULTS

At the default setting, EU-TIRADS yielded the highest sensitivity, 82.6% and lowest specificity, 42.1% while the ATA-TIRADS yielded the highest specificity, 66.4%. Kwak had the highest AUROC (0.74) which was comparable to that of ACR, ATA, and KSThR TIRADS (0.72, 0.73, and 0.70 respectively). At a hyperechoic foci setting of 3.5 with other settings at median values; ATA had the best-balanced sensitivity, specificity and good AUROC (70.4%; 67.3% and 0.71 respectively).

CONCLUSION

The default setting achieved the best diagnostic performance with all TIRADS and was best for maximizing the sensitivity of EU-TIRADS. Adjusting the settings by only reducing the sensitivity to echogenic foci may be most helpful for improving specificity with minimal change in sensitivity.

Machine Learning by Ultrasonography for Genetic Risk Stratification of Thyroid Nodules

Importance

Thyroid nodules are common incidental findings. Ultrasonography and molecular testing can be used to assess risk of malignant neoplasm.

Objective

To examine whether a model developed through automated machine learning can stratify thyroid nodules as high or low genetic risk by ultrasonography imaging alone compared with stratification by molecular testing for high- and low-risk mutations.

Design, Setting, and Participants

This diagnostic study was conducted at a single tertiary care urban academic institution and included patients (n = 121) who underwent ultrasonography and molecular testing for thyroid nodules from January 1, 2017, through August 1, 2018. Nodules were classified as high risk or low risk on the basis of results of an institutional molecular testing panel for thyroid risk genes. All thyroid nodules that underwent genetic sequencing for cytological results with Bethesda System categories III and IV were reviewed. Patients without diagnostic ultrasonographic images within 6 months of fine-needle aspiration or who received definitive treatment at an outside medical center were excluded.

Main Outcomes and Measures

Thyroid nodules were categorized by the model as high risk or low risk using ultrasonographic images. Results were compared using genetic testing.

Results

Among the 134 lesions identified in 121 patients (mean [SD] age, 55.7 [14.2] years; 102 women [84.3%]), 683 diagnostic ultrasonographic images were selected. Of the 683 images, 556 (81.4%) were used for training the model, 74 (10.8%) for validation, and 53 (7.8%) for testing. Most nodules had no mutation (75 [56.0%]), whereas 43 nodules (32.1%) had a high-risk mutation and 16 (11.9%) had an unknown or a low-risk mutation (χ2 = 39.060; P < .001). In total, 228 images (33.4%) were of nodules classified as genetically high risk (n = 43), and 455 (66.6%) were of low-risk nodules (n = 91). The model performed with a sensitivity of 45% (95% CI, 23.1%-68.5%), a specificity of 97% (95% CI, 84.2%-99.9%), a positive predictive value of 90% (95% CI, 55.2%-98.5%), a negative predictive value of 74.4% (95% CI, 66.1%-81.3%), and an overall accuracy of 77.4% (95% CI, 63.8%-97.7%).

Conclusions and Relevance

The study found that the model developed through automated machine learning could produce high specificity for identifying nodules with high-risk mutations on molecular testing. This finding shows promise for the diagnostic applications of machine learning interpretation of sonographic imaging of indeterminate thyroid nodules.

A pilot study of radiomics signature based on biparametric MRI for preoperative prediction of extrathyroidal extension in papillary thyroid carcinoma

OBJECTIVE

To investigate efficiency of radiomics signature to preoperatively predict histological features of aggressive extrathyroidal extension (ETE) in papillary thyroid carcinoma (PTC) with biparametric magnetic resonance imaging findings.

MATERIALS AND METHODS

Sixty PTC patients with preoperative MR including T2WI and T2WI-fat-suppression (T2WI-FS) were retrospectively analyzed. Among them, 35 had ETE and 25 did not. Pre-contrast T2WI and T2WI-FS images depicting the largest section of tumor were selected. Tumor regions were manually segmented using ITK-SNAP software and 107 radiomics features were computed from the segmented regions using the open Pyradiomics package. Then, a random forest model was built to do classification in which the datasets were partitioned randomly 10 times to do training and testing with ratio of 1:1. Furthermore, forward greedy feature selection based on feature importance was adopted to reduce model overfitting. Classification accuracy was estimated on the test set using area under ROC curve (AUC).

RESULTS

The model using T2WI-FS image features yields much higher performance than the model using T2WI features (AUC = 0.906 vs. 0.760 using 107 features). Among the top 10 important features of T2WI and T2WI-FS, there are 5 common features. After feature selection, the models trained using top 2 features of T2WI and the top 6 features of T2WI-FS achieve AUC 0.845 and 0.928, respectively. Combining features computed from T2WI and T2WI-FS, model performance decreases slightly (AUC = 0.882 based on all features and AUC = 0.913 based on top features after feature selection). Adjusting hyper parameters of the random forest model have negligible influence on the model performance with mean AUC = 0.907 for T2WI-FS images.

CONCLUSIONS

Radiomics features based on pre-contrast T2WI and T2WI-FS is helpful to predict aggressive ETE in PTC. Particularly, the model trained using the optimally selected T2WI-FS image features yields the best classification performance. The most important features relate to lesion size and the texture heterogeneity of the tumor region.

Applications of machine learning and deep learning to thyroid imaging: where do we stand?

Ultrasonography (US) is the primary diagnostic tool used to assess the risk of malignancy and to inform decision-making regarding the use of fine-needle aspiration (FNA) and postFNA management in patients with thyroid nodules. However, since US image interpretation is operator-dependent and interobserver variability is moderate to substantial, unnecessary FNA and/or diagnostic surgery are common in practice. Artificial intelligence (AI)-based computeraided diagnosis (CAD) systems have been introduced to help with the accurate and consistent interpretation of US features, ultimately leading to a decrease in unnecessary FNA. This review provides a developmental overview of the AI-based CAD systems currently used for thyroid nodules and describes the future developmental directions of these systems for the personalized and optimized management of thyroid nodules.

Artificial intelligence and hybrid imaging: the best match for personalized medicine in oncology

Artificial intelligence (AI) refers to a field of computer science aimed to perform tasks typically requiring human intelligence. Currently, AI is recognized in the broader technology radar within the five key technologies which emerge for their wide-ranging applications and impact in communities, companies, business, and value chain framework alike. However, AI in medical imaging is at an early phase of development, and there are still hurdles to take related to reliability, user confidence, and adoption. The present narrative review aimed to provide an overview on AI-based approaches (distributed learning, statistical learning, computer-aided diagnosis and detection systems, fully automated image analysis tool, natural language processing) in oncological hybrid medical imaging with respect to clinical tasks (detection, contouring and segmentation, prediction of histology and tumor stage, prediction of mutational status and molecular therapies targets, prediction of treatment response, and outcome). Particularly, AI-based approaches have been briefly described according to their purpose and, finally lung cancer-being one of the most extensively malignancy studied by hybrid medical imaging-has been used as illustrative scenario. Finally, we discussed clinical challenges and open issues including ethics, validation strategies, effective data-sharing methods, regulatory hurdles, educational resources, and strategy to facilitate the interaction among different stakeholders. Some of the major changes in medical imaging will come from the application of AI to workflow and protocols, eventually resulting in improved patient management and quality of life. Overall, several time-consuming tasks could be automatized. Machine learning algorithms and neural networks will permit sophisticated analysis resulting not only in major improvements in disease characterization through imaging, but also in the integration of multiple-omics data (i.e., derived from pathology, genomic, proteomics, and demographics) for multi-dimensional disease featuring. Nevertheless, to accelerate the transition of the theory to practice a sustainable development plan considering the multi-dimensional interactions between professionals, technology, industry, markets, policy, culture, and civil society directed by a mindset which will allow talents to thrive is necessary.

Thyroid Nodule Classification for Physician Decision Support Using Machine Learning-Evaluated Geometric and Morphological Features

The classification of thyroid nodules using ultrasound (US) imaging is done using the Thyroid Imaging Reporting and Data System (TIRADS) guidelines that classify nodules based on visual and textural characteristics. These are composition, shape, size, echogenicity, calcifications, margins, and vascularity. This work aims to reduce subjectivity in the current diagnostic process by using geometric and morphological (G-M) features that represent the visual characteristics of thyroid nodules to provide physicians with decision support. A total of 27 G-M features were extracted from images obtained from an open-access US thyroid nodule image database. 11 significant features in accordance with TIRADS were selected from this global feature set. Each feature was labeled (0 = benign and 1 = malignant) and the performance of the selected features was evaluated using machine learning (ML). G-M features together with ML resulted in the classification of thyroid nodules with a high accuracy, sensitivity and specificity. The results obtained here were compared against state-of the-art methods and perform significantly well in comparison. Furthermore, this method can act as a computer aided diagnostic (CAD) system for physicians by providing them with a validation of the TIRADS visual characteristics used for the classification of thyroid nodules in US images.

Thyroid Incidentalomas: Practice Considerations for Radiologists in the Age of Incidental Findings

Radiologists very frequently encounter incidental findings related to the thyroid gland. Given increases in imaging use over the past several decades, thyroid incidentalomas are increasingly encountered in clinical practice, and it is important for radiologists to be aware of recent developments with respect to workup and diagnosis of incidental thyroid abnormalities. Recent reporting and management guidelines, such as those from the American College of Radiology and American Thyroid Association, are reviewed along with applicable evidence in the literature. Trending topics, such as artificial intelligence approaches to guide thyroid incidentaloma workup, are also discussed.

Study Processes and Applications of Ultrasomics in Precision Medicine

Ultrasomics is the science of transforming digitally encrypted medical ultrasound images that hold information related to tumor pathophysiology into mineable high-dimensional data. Ultrasomics data have the potential to uncover disease characteristics that are not found with the naked eye. The task of ultrasomics is to quantify the state of diseases using distinctive imaging algorithms and thereby provide valuable information for personalized medicine. Ultrasomics is a powerful tool in oncology but can also be applied to other medical problems for which a disease is imaged. To date there is no comprehensive review focusing on ultrasomics. Here, we describe how ultrasomics works and its capability in diagnosing disease in different organs, including breast, liver, and thyroid. Its pitfalls, challenges and opportunities are also discussed.

Artificial intelligence in radiotherapy: a technological review

Radiation therapy (RT) is widely used to treat cancer. Technological advances in RT have occurred in the past 30 years. These advances, such as three-dimensional image guidance, intensity modulation, and robotics, created challenges and opportunities for the next breakthrough, in which artificial intelligence (AI) will possibly play important roles. AI will replace certain repetitive and labor-intensive tasks and improve the accuracy and consistency of others, particularly those with increased complexity because of technological advances. The improvement in efficiency and consistency is important to manage the increasing cancer patient burden to the society. Furthermore, AI may provide new functionalities that facilitate satisfactory RT. The functionalities include superior images for real-time intervention and adaptive and personalized RT. AI may effectively synthesize and analyze big data for such purposes. This review describes the RT workflow and identifies areas, including imaging, treatment planning, quality assurance, and outcome prediction, that benefit from AI. This review primarily focuses on deep-learning techniques, although conventional machine-learning techniques are also mentioned.

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.

Ensemble Deep Learning Model for Multicenter Classification of Thyroid Nodules on Ultrasound Images

Background

Thyroid nodules are extremely common and typically diagnosed with ultrasound whether benign or malignant. Imaging diagnosis assisted by Artificial Intelligence has attracted much attention in recent years. The aim of our study was to build an ensemble deep learning classification model to accurately differentiate benign and malignant thyroid nodules.

Material/Methods

Based on current advanced methods of image segmentation and classification algorithms, we proposed an ensemble deep learning classification model for thyroid nodules (EDLC-TN) after precise localization. We compared diagnostic performance with four other state-of-the-art deep learning algorithms and three ultrasound radiologists according to ACR TI-RADS criteria. Finally, we demonstrated the general applicability of EDLC-TN for diagnosing thyroid cancer using ultrasound images from multi medical centers.

Results

The method proposed in this paper has been trained and tested on a thyroid ultrasound image dataset containing 26 541 images and the accuracy of this method could reach 98.51%. EDLC-TN demonstrated the highest value for area under the curve, sensitivity, specificity, and accuracy among five state-of-the-art algorithms. Combining EDLC-TN with models and radiologists could improve diagnostic accuracy. EDLC-TN achieved excellent diagnostic performance when applied to ultrasound images from another independent hospital.

Conclusions

Based on ensemble deep learning, the proposed approach in this paper is superior to other similar existing methods of thyroid classification, as well as ultrasound radiologists. Moreover, our network represents a generalized platform that potentially can be applied to medical images from multiple medical centers.

False-Positive Malignant Diagnosis of Nodule Mimicking Lesions by Computer-Aided Thyroid Nodule Analysis in Clinical Ultrasonography Practice

This study aims to test computer-aided diagnosis (CAD) for thyroid nodules in clinical ultrasonography (US) practice with a focus towards identifying thyroid entities associated with CAD system misdiagnoses. Two-hundred patients referred to thyroid US were prospectively enrolled. An experienced radiologist evaluated the thyroid nodules and saved axial images for further offline blinded analysis using a commercially available CAD system. To represent clinical practice, not only true nodules, but mimicking lesions were also included. Fine needle aspiration biopsy (FNAB) was performed according to present guidelines. US features and thyroid entities significantly associated with CAD system misdiagnosis were identified along with the diagnostic accuracy of the radiologist and the CAD system. Diagnostic specificity regarding the radiologist was significantly (p < 0.05) higher than when compared with the CAD system (88.1% vs. 40.5%) while no significant difference was found in the sensitivity (88.6% vs. 80%). Focal inhomogeneities and true nodules in thyroiditis, nodules with coarse calcification and inspissated colloid cystic nodules were significantly (p < 0.05) associated with CAD system misdiagnosis as false-positives. The commercially available CAD system is promising when used to exclude thyroid malignancies, however, it currently may not be able to reduce unnecessary FNABs, mainly due to the false-positive diagnoses of nodule mimicking lesions.

Computer-aided diagnosis system of thyroid nodules ultrasonography: Diagnostic performance difference between computer-aided diagnosis and 111 radiologists

To evaluate the diagnostic efficiency of computer-aided diagnosis (CAD) system and 111 radiologists with different experience in identifying benign and malignant thyroid nodules, and to summarize the ultrasound features that may affect the diagnostic of CAD and radiologists.Fifty thyroid nodules and 111 radiologists were enrolled in this study. All the 50 nodules were diagnosed by the 111 radiologists and the CAD system simultaneously. The diagnostic performance of the CAD system, senior and junior radiologists with the maximum accuracy were calculated and compared. Interobserver agreement for different ultrasound characteristics between the CAD and senior radiologist were analyzed.CAD system showed a higher specificity than junior radiologist (87.5% vs 70.4%, P = .03), and a lower sensitivity than the senior radiologist and junior radiologist but the statistics were not significant (76.9% vs 86.9%, P > .5; 76.9% vs 82.6%, P > .5). The CAD system and senior radiologist got larger AUC than junior radiologist but the differences were not statistically significant (0.82 vs 0.76, respectively; P = .5). The interobserver agreement for the US characteristics between the CAD system and senior radiologist were: substantial agreement for hypoechoic and taller than wide (kappa value = 0.66, 0.78), and moderate agreement for irregular margin and micro-calcifications (kappa value = 0.52, 0.42).The CAD system achieved equal diagnostic accuracy to the senior radiologists and higher accuracy than the junior radiologists. The interobserver agreements in the US features between the CAD system and senior radiologist were substantial agreement for hypoechoic and taller than wide; moderate agreement for irregular margin and micro-calcifications. The location of a thyroid nodule and the feature of macrocalcification with wide acoustic shadow may influence the analysis of the CAD system.

Preliminary study on identification of estrogen receptor-positive breast cancer subtypes based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) texture analysis

Background

Currently, breast cancer is divided into Luminal A, Luminal B, HER-2 overexpression (HER-2) and basal cell at genetic level. However, the differential diagnosis of estrogen receptor (ER)-positive breast cancer subtypes is rare. Therefore, we aimed to investigate the feasibility of identifying the ER-positive breast cancer subtypes based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) texture analysis.

Methods

A retrospective analysis was performed for clinical data of 51 patients with ER-positive breast invasive ductal carcinoma confirmed by surgery and pathology from January 20 to October 2018. FireVoxel texture analysis software was used to delineate the tumor boundary layer by layer. The differences in the above characteristics between Luminal A and Luminal B breast cancer were compared, and the diagnostic efficacy of statistically significant texture parameters for ER-positive breast cancer subtypes was analyzed.

Results

There were no significant differences in mean, standard deviation (SD), skewness and tumor size between Luminal A and Luminal B groups (P>0.05). The kurtosis, inhomogeneity and entropy could effectively distinguish between the two groups with statistically significant difference (P=0.001, P=0.000, and P=0.000). The area under the receiver operating characteristic (ROC) curve (AUC) of kurtosis, inhomogeneity and entropy diagnosed with malignant mass were 0.832, 0.859 and 0.891, respectively (P<0.01). In addition, the entropy was the best among the three indicators. When the entropy was ≤4.22, the sensitivity of the diagnosis Luminal B was 90.62% and the specificity was 78.95%.

Conclusions

The texture analysis features based on DCE-MRI can help to identify ER-positive breast cancer subtypes. Entropy can be the best single texture indicator.

Computer-aided diagnostic system for thyroid nodule sonographic evaluation outperforms the specificity of less experienced examiners

PURPOSE

Computer-aided diagnosis (CAD) may improve interobserver agreement in the risk stratification of thyroid nodules. This study aims to evaluate the performance of the Korean Thyroid Imaging Reporting and Data System (K-TIRADS) classification as estimated by an expert radiologist, a senior resident, a medical student, and a CAD system, as well as the interobserver agreement among them.

METHODS

Between July 2016 and 2018, 107 nodules (size 5-40 mm, 27 malignant) were classified according to the K-TIRADS by an expert radiologist and CAD software. A third-year resident and a medical student with basic imaging training, both blinded to previous findings, retrospectively estimated the K-TIRADS classification. The diagnostic performance was calculated, including sensitivity, specificity, positive and negative predictive values, and the area under the receiver operating characteristic curve.

RESULTS

The CAD system and the expert achieved a sensitivity of 70.37% (95% CI 49.82-86.25%) and 81.48% (61.92-93.7%) and a specificity of 87.50% (78.21-93.84%) and 88.75% (79.72-94.72%), respectively. The specificity of the student was significantly lower (76.25% [65.42-85.05%], p = 0.02).

CONCLUSION

In our opinion, the CAD evaluation of thyroid nodules stratification risk has a potential role in a didactic field and does not play a real and effective role in the clinical field, where not only images but also specialistic medical practice is fundamental to achieve a diagnosis based on family history, genetics, lab tests, and so on. The CAD system may be useful for less experienced operators as its specificity was significantly higher.

Application of machine learning to ultrasound images to differentiate follicular neoplasms of the thyroid gland

Purpose

This study was conducted to evaluate the diagnostic performance of machine learning in differentiating follicular adenoma from carcinoma using preoperative ultrasonography (US).

Methods

In this retrospective study, preoperative US images of 348 nodules from 340 patients were collected from two tertiary referral hospitals. Two experienced radiologists independently reviewed each image and categorized the nodules according to the 2015 American Thyroid Association guideline. Categorization of a nodule as highly suspicious was considered a positive diagnosis for malignancy. The nodules were manually segmented, and 96 radiomic features were extracted from each region of interest. Ten significant features were selected and used as final input variables in our in-house developed classifier models based on an artificial neural network (ANN) and support vector machine (SVM). The diagnostic performance of radiologists and both classifier models was calculated and compared.

Results

In total, 252 nodules from 245 patients were confirmed as follicular adenoma and 96 nodules from 95 patients were diagnosed as follicular carcinoma. As measures of diagnostic performance, the average sensitivity, specificity, and accuracy of the two experienced radiologists in discriminating follicular adenoma from carcinoma on preoperative US images were 24.0%, 84.0%, and 64.8%, respectively. The sensitivity, specificity, and accuracy of the ANN and SVM-based models were 32.3%, 90.1%, and 74.1% and 41.7%, 79.4%, and 69.0%, respectively. The kappa value of the two radiologists was 0.076, corresponding to slight agreement.

Conclusion

Machine learning-based classifier models may aid in discriminating follicular adenoma from carcinoma using preoperative US.

A Novel Risk Stratification System for Thyroid Nodules With Indeterminate Cytology-A Pilot Cohort Study

Background: Thyroid ultrasound (US), fine needle aspiration biopsy (FNAB), and molecular testing have been widely used to stratify the risk of malignancy in thyroid nodules. The goal of this study was to investigate a novel diagnostic approach for cytologically indeterminate thyroid nodules (ITN) based upon a combination of US features and genetic alterations. Methods: We performed a pilot cohort study of patients with ITN (Bethesda III/IV), who underwent surgical treatment. Based on standardized sonographic patterns established by the American Thyroid Association (ATA), each ITN received an US score (X_US), ranging between 0 and 0.9 according to its risk of thyroid cancer (TC). DNA and RNA were extracted from pathologic material, available for all patients, and subjected to Oncomine™ Comprehensive Assay v2 (OCAv2) next-generation sequencing. Each genetic alteration was annotated based on its strength of association with TC and its sum served as the genomic classifier score (X_GC). The total risk score (TRS) was the sum of X_US and X_GC. ROC curves were generated to assess the diagnostic accuracy of X_US, X_GC, and TRS. Results: The study cohort consisted of 50 patients (39 females and 11 males), aged 47.5 ± 14.8 years. Three patients were excluded due to molecular testing failure. Among the remaining 47 patients, 28 (59.6%) were diagnosed with TC. BRAFV600E was the most common mutation in papillary TC, PAX8-PPARG fusion was present in NIFTP, pathogenic variants of SLX4, ATM, and NRAS were found in Hürthle cell TC and RET mutations in medullary TC. The diagnostic accuracy of X_GC and TRS was significantly higher compared with X_US (88 vs. 62.5%, p < 0.001; 85.2 vs. 62.5%, p < 0.001, respectively). However, this increased accuracy was due to significantly better sensitivity (80.7 vs. 34.6%, p < 0.001; 84.6 vs. 34.6%, p < 0.001, respectively) without improved specificity (94.7 vs. 90%, p = 0.55; 85.7 vs. 90%, p = 0.63, respectively). Conclusion: Molecular testing might not be necessary in ITN with high-risk US pattern (X_US = 0.9), as specificity of TC diagnosis based on Xus alone is sufficient and not improved with molecular testing. OCAv2 is useful in guiding the management of ITN with low-to-intermediate risk US features (X_US < 0.9), as it increases the accuracy of TC diagnosis.

An investigation of machine learning methods in delta-radiomics feature analysis

Purpose

This study aimed to investigate the effectiveness of using delta-radiomics to predict overall survival (OS) for patients with recurrent malignant gliomas treated by concurrent stereotactic radiosurgery and bevacizumab, and to investigate the effectiveness of machine learning methods for delta-radiomics feature selection and building classification models.

Methods

The pre-treatment, one-week post-treatment, and two-month post-treatment T1 and T2 fluid-attenuated inversion recovery (FLAIR) MRI were acquired. 61 radiomic features (intensity histogram-based, morphological, and texture features) were extracted from the gross tumor volume in each image. Delta-radiomics were calculated between the pre-treatment and post-treatment features. Univariate Cox regression and 3 multivariate machine learning methods (L1-regularized logistic regression [L1-LR], random forest [RF] or neural networks [NN]) were used to select a reduced number of features, and 7 machine learning methods (L1-LR, L2-LR, RF, NN, kernel support vector machine [KSVM], linear support vector machine [LSVM], or naïve bayes [NB]) was used to build classification models for predicting OS. The performances of the total 21 model combinations built based on single-time-point radiomics (pre-treatment, one-week post-treatment, and two-month post-treatment) and delta-radiomics were evaluated by the area under the receiver operating characteristic curve (AUC).

Results

For a small cohort of 12 patients, delta-radiomics resulted in significantly higher AUC than pre-treatment radiomics (p-value<0.01). One-week/two-month delta-features resulted in significantly higher AUC (p-value<0.01) than the one-week/two-month post-treatment features, respectively. 18/21 model combinations were with higher AUC from one-week delta-features than two-month delta-features. With one-week delta-features, RF feature selector + KSVM classifier and RF feature selector + NN classifier showed the highest AUC of 0.889.

Conclusions

The results indicated that delta-features could potentially provide better treatment assessment than single-time-point features. The treatment assessment is substantially affected by the time point for computing the delta-features and the combination of machine learning methods for feature selection and classification.

The Diagnostic Efficiency of Ultrasound Computer-Aided Diagnosis in Differentiating Thyroid Nodules: A Systematic Review and Narrative Synthesis

Computer-aided diagnosis (CAD) techniques have emerged to complement qualitative assessment in the diagnosis of benign and malignant thyroid nodules. The aim of this review was to summarize the current evidence on the diagnostic performance of various ultrasound CAD in characterizing thyroid nodules. PUBMED, EMBASE and Cochrane databases were searched for studies published until August 2019. The Quality Assessment of Studies of Diagnostic Accuracy included in Systematic Review 2 (QUADAS-2) tool was used to assess the methodological quality of the studies. Reported diagnostic performance data were analyzed and discussed. Fourteen studies with 2232 patients and 2675 thyroid nodules met the inclusion criteria. The study quality based on QUADAS-2 assessment was moderate. At best performance, grey scale CAD had a sensitivity of 96.7% while Doppler CAD was 90%. Combined techniques of qualitative grey scale features and Doppler CAD assessment resulted in overall increased sensitivity (92%) and optimal specificity (85.1%). The experience of the CAD user, nodule size and the thyroid malignancy risk stratification system used for interpretation were the main potential factors affecting diagnostic performance outcomes. The diagnostic performance of CAD of thyroid ultrasound is comparable to that of qualitative visual assessment; however, combined techniques have the potential for better optimized diagnostic accuracy.

Quantitative imaging biomarkers in nuclear medicine: from SUV to image mining studies. Highlights from annals of nuclear medicine 2018

OBJECTIVE

Quantification in medical imaging is one of the main goals in research and clinical practice since it allows immediate understanding, objective communication, and comparison. Our aim was to summarize relevant investigations on quantification in nuclear medicine studies published in the volume 32 of Annals of Nuclear Medicine.

METHODS

In this article, we summarized the data of 14 selected papers from international research groups that were published between January and December 2018. This is a descriptive review with an inherently subjective selection of articles.

RESULTS

We discussed the role of parameters ranging from standardized uptake value to ratios, to flow within a region of interest, to volumetric parameters and to texture indices in different clinical scenarios in oncology, cardiology, and neurology.

CONCLUSIONS

In all the medical disciplines in which nuclear medicine examinations play a role, quantification is essential both in research and in clinical practice. Standardization and high-quality protocols are crucial for the success and reliability of imaging biomarkers.