Impact of Ultrasound Image Reconstruction Method on Breast Lesion Classification with Deep Learning

Abstract

Deep learning algorithms, especially convolutional neural networks, have become a methodology of choice in medical image analysis. However, recent studies in computer vision show that even a small modification of input image intensities may cause a deep learning model to classify the image differently. In medical imaging, the distribution of image intensities is related to applied image reconstruction algorithm. In this paper we investigate the impact of ultrasound image reconstruction method on breast lesion classification with neural transfer learning. Due to high dynamic range raw ultrasonic signals are commonly compressed in order to reconstruct B-mode images. Based on raw data acquired from breast lesions, we reconstruct B-mode images using different compression levels. Next, transfer learning is applied for classification. Differently reconstructed images are employed for training and evaluation. We show that the modification of the reconstruction algorithm leads to decrease of classification performance. As a remedy, we propose a method of data augmentation. We show that the augmentation of the training set with differently reconstructed B-mode images leads to a more robust and efficient classification. Our study suggests that it is important to take into account image reconstruction algorithms implemented in medical scanners during development of computer aided diagnosis systems.

Deep-Learning-Based Computer-Aided Systems for Breast Cancer Imaging: A Critical Review

“…Most of these DL architectures use a large data set; thus, it is required to apply an augmentation technique to avoid overfiting and to have better performance during classification. In this sense, the researchers mentioned in Table 6 [145,168,180,181] and Table 7 [35,49,62,66,152,182] the authors used transfer learning and ensemble methods, such as data augmentation, to improve the performance of the CNN network, reaching an 89.86% accuracy and 0.9578% AUC in DM, and an AUC of 0.68% on US images. Furthermore, Singh et al [165] showed that the results obtained with a GAN for breast tumor segmentation outperformed the UNet model, and the SegNet and ERFNet models yielded the worst segmentation results on US images.…”

Recognition of Benign and Malignant Breast Ultrasound Images Based on Deep Transfer Learning

“…Michał Byra et al [15]studied the influence of ultrasound image reconstruction algorithm on the classification of breast lesions based on transfer learning. In order to minimize the degradation of classification performance caused by the reconstruction algorithm, they proposed a data expansion method and compared the three image reconstruction thresholds, which are 40dB, 50dB, and 60dB.…”

Breast ultrasound lesion classification based on image decomposition and transfer learning

“…The reported metrics were as follows: accuracy was 90%; and sensitivity, specificity, and the area under the curve (AUC) were 86%, 96%, and 0.95, respectively. Byra et al 10 investigated an ultrasound image reconstruction method for use in breast lesion classification. The authors used the InceptionV3 model and the reported accuracy, sensitivity, specificity, and AUC were 78%, 77%, 78%, and 0.857, respectively.…”