Endocrine Tumor Classification via Machine-Learning-Based Elastography: A Systematic Scoping Review

Abstract

Elastography complements traditional medical imaging modalities by mapping tissue stiffness to identify tumors in the endocrine system, and machine learning models can further improve diagnostic accuracy and reliability. Our objective in this review was to summarize the applications and performance of machine-learning-based elastography on the classification of endocrine tumors. Two authors independently searched electronic databases, including PubMed, Scopus, Web of Science, IEEEXpress, CINAHL, and EMBASE. Eleven (n = 11) articles were eligible for the review, of which eight (n = 8) focused on thyroid tumors and three (n = 3) considered pancreatic tumors. In all thyroid studies, the researchers used shear-wave ultrasound elastography, whereas the pancreas researchers applied strain elastography with endoscopy. Traditional machine learning approaches or the deep feature extractors were used to extract the predetermined features, followed by classifiers. The applied deep learning approaches included the convolutional neural network (CNN) and multilayer perceptron (MLP). Some researchers considered the mixed or sequential training of B-mode and elastographic ultrasound data or fusing data from different image segmentation techniques in machine learning models. All reviewed methods achieved an accuracy of ≥80%, but only three were ≥90% accurate. The most accurate thyroid classification (94.70%) was achieved by applying sequential training CNN; the most accurate pancreas classification (98.26%) was achieved using a CNN–long short-term memory (LSTM) model integrating elastography with B-mode and Doppler images.

Computer-aided screening of aspiration risks in dysphagia with wearable technology: a Systematic Review and meta-analysis on test accuracy

“…Accuracy is the ratio of correct tests to the total number of tests. Sensitivity shows the proportion of positive diagnoses from the index test that are also detected as positive by the reference test, while specificity indicates the proportion of negative diagnoses from Highly imbalanced classes are a prevalent issue in healthcare and medicine (Jothi and Husain, 2015;Mao et al, 2022;Mao et al, 2023) since it is natural to have fewer positive than negative cases (i.e., nonhealthy cases are often underrepresented), which was also reflected in our review (Table 4). It should be noted that some studies defined "adjusted accuracy" by taking a simple average of sensitivity and specificity and claimed that the parameter could resolve the imbalanced class issue, with which we disagreed.…”

“…Highly imbalanced classes are a prevalent issue in healthcare and medicine ( Jothi and Husain, 2015 ; Mao et al, 2022 ; Mao et al, 2023 ) since it is natural to have fewer positive than negative cases (i.e., non-healthy cases are often underrepresented), which was also reflected in our review ( Table 4 ). It should be noted that some studies defined “adjusted accuracy” by taking a simple average of sensitivity and specificity and claimed that the parameter could resolve the imbalanced class issue, with which we disagreed.…”

En–DeNet Based Segmentation and Gradational Modular Network Classification for Liver Cancer Diagnosis

“…For making comparisons between approaches, it is important to evaluate the performance of classification algorithms, choose the best one, learn about the constraints of the system, and notice areas for potential improvement [27].…”

Applications of Deep Learning Models on the Medical Images of Osteonecrosis of the Femoral Head (ONFH): A Comprehensive Review

“…IoU is the ratio of overlapped and union regions of the manual and predicted segmentation (( 21) and Fig. 6), while the Dice coefficient is the ratio of intersected manual (A) and predicted (Â) segmented regional area to the total area (22). Li, et al [39] achieved a voxel-based F1-score of 97.7% and a Hausdorff distance (HD) of 7.52, whereas Wang, et al [46] produced a higher HD performance at 1.491.…”

“…Computer-aided diagnosis has been emerging as an indispensable tool in the field of medical imaging. It has been extensively utilized for disease classification, and pathological region segmentation (especially tumor) across various imaging modalities, including X-ray [13,14], magnetic resonance imaging (MRI) [15][16][17][18], computer tomography (CT) [17,19], and ultrasonography [20][21][22][23]. Notably, the diagnosis and prognosis analysis of ONFH has also been benefited from this advance [24].…”