Application of Deep Learning in Automated Analysis of Molecular Images in Cancer: A Survey.

Yong Xue,Shihui Chen,Hanwei Chen,Bingsheng Huang,Jing Qin,Yong Liu

doi:10.1155/2017/9512370

Abstract

Molecular imaging enables the visualization and quantitative analysis of the alterations of biological procedures at molecular and/or cellular level, which is of great significance for early detection of cancer. In recent years, deep leaning has been widely used in medical imaging analysis, as it overcomes the limitations of visual assessment and traditional machine learning techniques by extracting hierarchical features with powerful representation capability. Research on cancer molecular images using deep learning techniques is also increasing dynamically. Hence, in this paper, we review the applications of deep learning in molecular imaging in terms of tumor lesion segmentation, tumor classification, and survival prediction. We also outline some future directions in which researchers may develop more powerful deep learning models for better performance in the applications in cancer molecular imaging.

Highlights

With increasing incidence and mortality, cancer has always been a leading cause of death for many years
Deep learning-based automated analysis tools can greatly alleviate the heavy workload of radiologists and physicians caused by the popularity of molecular imaging in early diagnosis of cancer as well as enhance the diagnostic accuracy, especially when there exist subtle pathological changes that cannot be detected by visual assessment
The stacked nonnegativityconstrained autoencoders (SNCAE)-based method proposed by Reda et al [41] has achieved excellent classification performance on the diffusion-weighted magnetic resonance images (DW-MRI) data from 53 subjects, but this method still needs several preprocessing steps leveraging hand-crafted features, which may greatly affect the computational load of the classification

Summary

Introduction

With increasing incidence and mortality, cancer has always been a leading cause of death for many years. Deep learning is promising in a wide variety of applications including cancer detection and prediction based on molecular imaging, such as in brain tumor segmentation [8], tumor classification, and survival prediction. Deep learning-based automated analysis tools can greatly alleviate the heavy workload of radiologists and physicians caused by the popularity of molecular imaging in early diagnosis of cancer as well as enhance the diagnostic accuracy, especially when there exist subtle pathological changes that cannot be detected by visual assessment. This review contains 25 papers and is organized according to the application of deep learning in cancer molecular imaging, including tumor lesion segmentation, cancer classification, and prediction of patient survival. Biomedical engineering researchers may benefit from this survey by acquiring the state of the art in this field or inspiration for better models/methods in future research

Deep Learning in Tumor Lesion Segmentation

Results

Deep Learning in Cancer Classification

Deep Learning in Survival Prediction

Trends and Challenges

Conclusion