Self-Supervised Transfer Learning Based on Domain Adaptation for Benign-Malignant Lung Nodule Classification on Thoracic CT

The spatial heterogeneity is an important indicator of the malignancy of lung nodules in lung cancer diagnosis. Compared with 2D nodule CT images, the 3D volumes with entire nodule objects hold richer discriminative information. However, for deep learning methods driven by massive data, effectively capturing the 3D discriminative features of nodules in limited labeled samples is a challenging task. Different from previous models that proposed transfer learning models in a 2D pattern or learning from scratch 3D models, we develop a self-supervised transfer learning based on domain adaptation (SSTL-DA) 3D CNN framework for benign-malignant lung nodule classification. At first, a data pre-processing strategy termed adaptive slice selection (ASS) is developed to eliminate the redundant noise of the input samples with lung nodules. Then, the self-supervised learning network is constructed to learn robust image representations from CT images. Finally, a transfer learning method based on domain adaptation is designed to obtain discriminant features for classification. The proposed SSTL-DA method has been assessed on the LIDC-IDRI benchmark dataset, and it obtains an accuracy of 91.07% and an AUC of 95.84%. These results demonstrate that the SSTL-DA model achieves quite a competitive classification performance compared with some state-of-the-art approaches

I. Introduction

Lung cancer represents the primary reason of cancer-related death all over the world. The 2020 Global Cancer Statistics show that lung cancer accounts for approximately 11.4% of 19.3 million new cancer cases and 18% of almost 10.0 million cancer-related deaths in 2020 [1]. Early lung cancer screening provides the best opportunity for a cure and improves the survival rate of patients [2]. The National Lung Screening Trial reports that three annual screenings with low-dose computed tomography (CT) result in 20% lower mortality from lung cancer among high-risk persons than screening with the use of chest radiography [3]. A critical task in CT-based lung cancer screening is the benign-malignant discrimination of lung nodule [4], [5]. To make a correct diagnosis on the malignancy of a lung nodule, radiologists are usually required to read the thoracic CT slice by slice, while this process is highly labor intensive and prone to operator bias [6].