Classification of Benign and Malignant Pulmonary Nodules Based on the Multiresolution 3D DPSECN Model and Semisupervised Clustering

Siyuan Tang,Yingchun Bai,Shijun Chen,Rong Ma,Qingqian Li

doi:10.1109/access.2021.3060178

Abstract

Deep learning model training requires a large number of labeled samples, but the acquisition of labeled samples is time-consuming and laborious in the medical field. To solve this problem, a semisupervised clustering algorithm combined with a 3D convolutional neural network model is proposed to improve the classification performance for benign and malignant pulmonary nodules. The research contents are as follows: Firstly, a multiresolution 3D dual path squeeze excitation deep learning network model is constructed. Then, the feature extractor in the network model is used to extract the high-level features of the image, and semisupervised clustering is applied to the extracted image features. The corresponding pseudolabels can be obtained for the unlabeled samples, and the categories of unlabeled samples are determined and utilized. Finally, the oversampling algorithm is used to balance the data categories of different types of samples, and the benign and malignant pulmonary nodules are classified by a classifier constructed by a 3D dual path squeeze excitation network. The experimental results show that the proposed semisupervised clustering algorithm can label the categories of unlabeled samples. The proposed network model can learn more characteristics related to pulmonary nodules and can effectively improve the classification performance of pulmonary nodules. The proposed network model was tested using Lung Image Database Consortium (LIDC-IDRI) dataset, and an accuracy of 94.4% and an AUC of 0.931 were obtained. Compared with some existing classification models, the proposed method can achieve a better classification effect of pulmonary nodules.

Highlights

Lung cancer is the leading cause of death among all types of cancer and kills millions of people every year
After the pixel spacing in the X, Y and Z dimensions of the collected CT image data was interpolated to approximately equal, pulmonary nodules were extracted, and the resulting pulmonary nodules were guaranteed to be approximately cubic in structure
According to the 3D characteristics of pulmonary nodules, a multiresolution method was proposed, and the method was integrated into the 3D network model to construct a 3D dual path squeeze excitation converged network model based on multiresolution

Summary

Introduction

Lung cancer is the leading cause of death among all types of cancer and kills millions of people every year. Early intervention, early diagnosis and treatment are important measures to reduce the mortality rate of lung cancer. The earliest manifestation of lung cancer is pulmonary nodules [1]. If pulmonary nodules can be quickly determined to be benign or malignant (or pathological), doctors can take more timely appropriate treatment measures to delay the invasion and spread of malignant lesions. Patients’ pulmonary nodule data detected by CT images are rapidly accumulating, and most of these data need to be analyzed manually. If we rely entirely on manual reading, we will be at a disadvantage in terms of time consumption, the identification rate of pulmonary nodules, false positive rates, labor costs, the avoidance of subjectivity and repetition, etc. Based on the above analysis, this paper uses the advantages of deep learning in big data processing. A large number of high-resolution lung CT images were used to construct a VOLUME XX, 2017

Methods

Results

Conclusion