Radiomics in Early Lung Cancer Diagnosis: From Diagnosis to Clinical Decision Support and Education

Abstract

Lung cancer is the most frequent cause of cancer-related death around the world. With the recent introduction of low-dose lung computed tomography for lung cancer screening, there has been an increasing number of smoking- and non-smoking-related lung cancer cases worldwide that are manifesting with subsolid nodules, especially in Asian populations. However, the pros and cons of lung cancer screening also follow the implementation of lung cancer screening programs. Here, we review the literature related to radiomics for early lung cancer diagnosis. There are four main radiomics applications: the classification of lung nodules as being malignant/benign; determining the degree of invasiveness of the lung adenocarcinoma; histopathologic subtyping; and prognostication in lung cancer prediction models. In conclusion, radiomics offers great potential to improve diagnosis and personalized risk stratification in early lung cancer diagnosis through patient-doctor cooperation and shared decision making.

Keywords: ground-glass nodules; lung cancer screening; overdiagnosis; radiomics; subsolid nodules.

Figure 1

The workflow of radiomics analysis in early lung cancer diagnosis. Because lung nodules in early-stage lung cancer usually manifest with ground-glass or part-solid nodules, automatic nodular contour segmentation is usually not accurate. The manual approach to ROI analysis for early lung cancer diagnosis is highly demanding in terms of time and radiomics expertise.

Figure 2

Flowchart describing the workflow process of radiomic texture analysis and modeling development for early lung cancer diagnosis with the application of the radiomics quality score (RQS), which was used to assesses the characteristics and the quality of the radiomics studies and report guidelines. Detailed RQS scores with 16 domains were recorded (domain 1: image protocol quality +1~2; domain 2: multiple segmentation +1; domain 3: phantom study +1; domain 4: imaging at multiple time points +1; domain 5: feature reduction or adjustment for multiple testing −3~+3; domain 6: multivariable analysis +1; domain 7: biological correlates +1; domain 8: cut-off analysis +1; domain 9: discrimination statistics +1~2; domain 10: calibration statistics +1~2; domain 11: prospective study +7; domain 12: validation −5~+5; domain 13: comparison to ‘gold standard’ +2; domain 14: potential clinical applications +2; domain 15: cost-effectiveness analysis +1; domain 16: open science and data +1~4.).

Figure 3

Four key elements required for the successful implementation of personalized medicine in early lung cancer diagnosis, including patient education, professional physician education, radiomic-based diagnostics, and SDM (shared decision making).