Automated lesion detection and segmentation in digital mammography using a u-net deep learning network

Abstract

Computer-aided detection or decision support systems aim to improve breast cancer screening programs by helping radiologists to evaluate digital mammography (DM) exams. Commonly such methods proceed in two steps: selection of candidate regions for malignancy, and later classification as either malignant or not. In this study, we present a candidate detection method based on deep learning to automatically detect and additionally segment soft tissue lesions in DM. A database of DM exams (mostly bilateral and two views) was collected from our institutional archive. In total, 7196 DM exams (28294 DM images) acquired with systems from three different vendors (General Electric, Siemens, Hologic) were collected, of which 2883 contained malignant lesions verified with histopathology. Data was randomly split on an exam level into training (50%), validation (10%) and testing (40%) of deep neural network with u-net architecture. The u-net classifies the image but also provides lesion segmentation. Free receiver operating characteristic (FROC) analysis was used to evaluate the model, on an image and on an exam level. On an image level, a maximum sensitivity of 0.94 at 7.93 false positives (FP) per image was achieved. Similarly, per exam a maximum sensitivity of 0.98 at 7.81 FP per image was achieved. In conclusion, the method could be used as a candidate selection model with high accuracy and with the additional information of lesion segmentation.

Weakly supervised pneumonia localization in chest X‐rays using generative adversarial networks

“…We also performed free‐response ROC (FROC) analysis to evaluate the localization performance at the level of individual abnormalities (as opposed to DR analysis above, which is an image‐level measure of detection success). The FROC methods were based on connected components analysis 67,68 and intensity peaks of the abnormality maps 69 . An abnormality‐level detection TPR was computed for successful detection of individual abnormalities (as opposed to whole image as in DR analysis or pixel‐level TPR as in the ROC analysis) across all images and measured against the average number of false positive detections per image, at each intensity threshold of our abnormality map.…”

Deep learning approach for breast cancer diagnosis

“…The developed network architecture consists of CNN with fully connected network at the last layers. U-net architecture is used in the work presented in [12]. Results indicate a usefulness of U-net in mammogram diagnosis, however, several false positive results are reported.…”

“…Convolutional neural networks in particular leads to a remarkable impact in image analysis and understanding especially in image segmentation, classification and analysis [4]. Several models employ deep learning are already developed for diagnosis and identification of breast cancer through analysis of digital mammography [5][6][7][8][9][10][11][12][13][14][15][16][17].…”

Automated Segmentation of Abnormal Tissues in Medical Images

“…As mentioned previously, second category of breast cancer detection methods bypasses feature extraction from candidate regions of interests. One representative method of this category uses a modified deep network architecture known as u-net [37] for segmentation of masses in digital mammograms [38]. Output of the network is a probability map binarized by thresholding.…”