Abstract:Background: Breast cancer (BC) is the most common malignancy in women and the second cause of cancer death. In recent years, there has been a strong development in artificial intelligence (AI) applications in medical imaging for several tasks. Our aim was to evaluate the potential of transfer learning with convolutional neural networks (CNNs) in discriminating suspicious breast lesions on ultrasound images. Methods: Transfer learning performances of five different CNNs (Inception V3, Xception, Densenet121, VGG 16, and ResNet50) were evaluated on a public and on an institutional dataset (526 and 392 images, respectively), customizing the top layers for the specific task. Institutional images were contoured by an expert radiologist and processed to feed the CNNs for training and testing. Postimaging biopsies were used as a reference standard for classification. The area under the receiver operating curve (AUROC) was used to assess diagnostic performance. Results: Networks performed very well on the public dataset (AUROC 0.938-0.996). The direct generalization to the institutional dataset resulted in lower performances (max AUROC 0.676); however, when tested on BI-RADS 3 and BI-RADS 5 only, results were improved (max AUROC 0.792). Good results were achieved on the institutional dataset (AUROC 0.759-0.818) and, when selecting a threshold of 2% for classification, a sensitivity of 0.983 was obtained for three of five CNNs, with the potential to spare biopsy in 15.3%-18.6% of patients. Conclusion: In conclusion, transfer learning with CNNs may achieve high sensitivity and might be used as a support tool in managing suspicious breast lesions on ultrasound images. Relevance statement: Transfer learning is a powerful technique to exploit the performances of well-trained CNNs for image classification. In a clinical scenario, it might be useful for the management of suspicious breast lesions on breast ultrasound, potentially sparing biopsy in a non-negligible number of patients. Key points: Properly trained CNNs with transfer learning are highly effective in differentiating benign and malignant lesions on breast ultrasound. Setting clinical thresholds increased sensitivity. CNNs might be useful as support tools in managing suspicious lesions on breast ultrasound.

Fully automatic classification of automated breast ultrasound (ABUS) imaging according to BI-RADS using a deep convolutional neural network

Breast Cancer Classification in Automated Breast Ultrasound Using Multiview Convolutional Neural Network with Transfer Learning.

A Novel Computer-Aided-Diagnosis System for Breast Ultrasound Images Based on BI-RADS Categories

Multiview deep learning networks based on automated breast volume scanner images for identifying breast cancer in BI-RADS 4

Diagnostic accuracy of automated ACR BI-RADS breast density classification using deep convolutional neural networks

Application of Convolution Neural Network Algorithm Based on Multicenter ABUS Images in Breast Lesion Detection

Computer-aided Diagnosis System for Breast Ultrasound Images Using Deep Learning

Automated assessment of BI-RADS categories for ultrasound images using multi-scale neural networks with an order-constrained loss function

Application of Deep Learning to Establish a Diagnostic Model of Breast Lesions Using Two-Dimensional Grayscale Ultrasound Imaging.

Study on automatic detection and classification of breast nodule using deep convolutional neural network system

Computer-aided detection of cancer in automated 3-D breast ultrasound

Deep Learning for Describing Breast Ultrasound Images with BI-RADS Terms

Automatic Breast Ultrasound Tumor Segmentation Via One-Stage Hierarchical Target Activation Network

Ultrasound-based Deep Learning in the Establishment of a Breast Lesion Risk Stratification System: a Multicenter Study

Two-stage Augmentation for Detecting Malignancy of BI-RADS 3 Lesions in Early Breast Cancer

The Clinical Application of Artificial Intelligence Assisted Contrast-Enhanced Ultrasound on BI-RADS Category 4 Breast Lesions.

Segmentation-based BI-RADS ensemble classification of breast tumours in ultrasound images

Tumor Classification in Automated Breast Ultrasound (ABUS) Based on a Modified Extracting Feature Network

Two-stage CNNs for Computerized BI-RADS Categorization in Breast Ultrasound Images

Transfer learning classification of suspicious lesions on breast ultrasound: is there room to avoid biopsies of benign lesions?

Detecting and classifying lesions in mammograms with Deep Learning