Johannes Uhlig1, Annemarie Uhlig2, Meike Kunze1, Tim Beissbarth3, Uwe Fischer4, Joachim Lotz1,5, Susanne Wienbeck1. 1. 1 Department of Diagnostic and Interventional Radiology, University Medical Center Goettingen, Robert Koch Strasse 40, Goettingen 37075, Germany. 2. 2 Department of Urology, University Medical Center Goettingen, Goettingen, Germany. 3. 3 Department of Medical Statistics, University Medical Center Goettingen, Goettingen, Germany. 4. 4 Diagnostic Breast Center Goettingen, Goettingen, Germany. 5. 5 German Cardiovascular Research Center (DZHK), Partnersite Goettingen, Goettingen, Germany.
Abstract
OBJECTIVE: The purpose of this study is to evaluate the diagnostic performance of machine learning techniques for malignancy prediction at breast cone-beam CT (CBCT) and to compare them to human readers. SUBJECTS AND METHODS: Five machine learning techniques, including random forests, back propagation neural networks (BPN), extreme learning machines, support vector machines, and K-nearest neighbors, were used to train diagnostic models on a clinical breast CBCT dataset with internal validation by repeated 10-fold cross-validation. Two independent blinded human readers with profound experience in breast imaging and breast CBCT analyzed the same CBCT dataset. Diagnostic performance was compared using AUC, sensitivity, and specificity. RESULTS: The clinical dataset comprised 35 patients (American College of Radiology density type C and D breasts) with 81 suspicious breast lesions examined with contrast-enhanced breast CBCT. Forty-five lesions were histopathologically proven to be malignant. Among the machine learning techniques, BPNs provided the best diagnostic performance, with AUC of 0.91, sensitivity of 0.85, and specificity of 0.82. The diagnostic performance of the human readers was AUC of 0.84, sensitivity of 0.89, and specificity of 0.72 for reader 1 and AUC of 0.72, sensitivity of 0.71, and specificity of 0.67 for reader 2. AUC was significantly higher for BPN when compared with both reader 1 (p = 0.01) and reader 2 (p < 0.001). CONCLUSION: Machine learning techniques provide a high and robust diagnostic performance in the prediction of malignancy in breast lesions identified at CBCT. BPNs showed the best diagnostic performance, surpassing human readers in terms of AUC and specificity.
OBJECTIVE: The purpose of this study is to evaluate the diagnostic performance of machine learning techniques for malignancy prediction at breast cone-beam CT (CBCT) and to compare them to human readers. SUBJECTS AND METHODS: Five machine learning techniques, including random forests, back propagation neural networks (BPN), extreme learning machines, support vector machines, and K-nearest neighbors, were used to train diagnostic models on a clinical breast CBCT dataset with internal validation by repeated 10-fold cross-validation. Two independent blinded human readers with profound experience in breast imaging and breast CBCT analyzed the same CBCT dataset. Diagnostic performance was compared using AUC, sensitivity, and specificity. RESULTS: The clinical dataset comprised 35 patients (American College of Radiology density type C and D breasts) with 81 suspicious breast lesions examined with contrast-enhanced breast CBCT. Forty-five lesions were histopathologically proven to be malignant. Among the machine learning techniques, BPNs provided the best diagnostic performance, with AUC of 0.91, sensitivity of 0.85, and specificity of 0.82. The diagnostic performance of the human readers was AUC of 0.84, sensitivity of 0.89, and specificity of 0.72 for reader 1 and AUC of 0.72, sensitivity of 0.71, and specificity of 0.67 for reader 2. AUC was significantly higher for BPN when compared with both reader 1 (p = 0.01) and reader 2 (p < 0.001). CONCLUSION: Machine learning techniques provide a high and robust diagnostic performance in the prediction of malignancy in breast lesions identified at CBCT. BPNs showed the best diagnostic performance, surpassing human readers in terms of AUC and specificity.
Entities:
Keywords:
breast neoplasm; cone-beam CT; contrast media; decision support techniques; machine learning
Authors: Sangmi Lee; Myeongkyun Kang; Keunho Byeon; Sang Eun Lee; In Ho Lee; Young Ah Kim; Shin-Wook Kang; Jung Tak Park Journal: J Digit Imaging Date: 2022-04-11 Impact factor: 4.903