Detecting Cancer Metastases on Gigapixel Pathology Images

TL;DR

This paper introduces a CNN-based framework for automatic detection and localization of cancer metastases in gigapixel pathology images, significantly improving accuracy over previous automated methods and approaching human expert performance.

Contribution

The authors develop a novel deep learning approach that achieves state-of-the-art metastasis detection accuracy on large pathology images, addressing the challenge of small tumor localization.

Findings

Detects 92.4% of tumors at 8 false positives per image

Achieves over 97% image-level AUC on test sets

Identifies labeling errors in the dataset

Abstract

Each year, the treatment decisions for more than 230,000 breast cancer patients in the U.S. hinge on whether the cancer has metastasized away from the breast. Metastasis detection is currently performed by pathologists reviewing large expanses of biological tissues. This process is labor intensive and error-prone. We present a framework to automatically detect and localize tumors as small as 100 x 100 pixels in gigapixel microscopy images sized 100,000 x 100,000 pixels. Our method leverages a convolutional neural network (CNN) architecture and obtains state-of-the-art results on the Camelyon16 dataset in the challenging lesion-level tumor detection task. At 8 false positives per image, we detect 92.4% of the tumors, relative to 82.7% by the previous best automated approach. For comparison, a human pathologist attempting exhaustive search achieved 73.2% sensitivity. We achieve image-level AUC scores above 97% on both the Camelyon16 test set and an independent set of 110 slides. In addition, we discover that two slides in the Camelyon16 training set were erroneously labeled normal. Our approach could considerably reduce false negative rates in metastasis detection.