Imaging Biomarkers in Lung Cancer = Imaging Biomarkers in Lung Cancer

다국어 초록 (Multilingual Abstract)

Lung cancer screening with low-dose CT (LDCT) has shown to reduce lung cancer mortality in several randomized clinical trials. Most nodule management systems are based on the nodule type (solid, part-solid, nonsolid) and nodule size. Traditionally the nodule size was measured manually and expressed as one-dimensional diameter or two-dimensional average diameter. Recently, many studies have shown the value of nodule volumetry obtained with software in reducing interobserver variability and predicting nodule malignancy. Calculating the volume doubling time can be used in this process. In addition to the assessment of lung nodule, LDCT provides valuable information on smoking-related diseases such as emphysema and coronary artery disease. Management based on this information can contribute to reducing all-cause mortality. With the advance of machine learning techniques, quantitative analysis of LDCT in lung cancer screening will be more effectively explored.
Radiomics is defined as high-throughput extraction, analysis, and interpretation of quantitative features from medical images. Radiomics have been applied various aspect of lung cancer imaging: differentiation benign from malignant nodules, histologic correlation in terms of tumor grade, prediction of mutation status, prognosis, clinical outcomes, and treatment response. Many studies have suggested that more heterogeneous tumors are more aggressive and associated with poorer prognosis/survival and tumors that become more homogeneous during treatment seem to be responding to therapy. Although radiomics showed many potentials, there are still challenges and limitation. Because of the nonstandardized protocols and relative lack of multicenter studies, the results are still hard to be generalized. Standardized data collection, establishment of evaluation criteria, and reporting guidelines are required.