Q Sun1, Y Huang1, J Wang1, S Zhao1, L Zhang1, W Tang1, N Wu2. 1. Department of Diagnostic Radiology, National Cancer Center, Cancer Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, 100021, China. 2. Department of Diagnostic Radiology, National Cancer Center, Cancer Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, 100021, China; PET-CT Center, National Cancer Center, Cancer Clinical Research Center for Cancer, Cancer Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, 100021, China. Electronic address: cjr.wuning@vip.163.com.
Abstract
AIM: To analyse subsolid nodules (SSNs) detected during low-dose (LD) computed tomography (CT) screening and investigated whether CT texture analysis parameters can predict the malignancy and growth trends of GGNs. MATERIALS AND METHODS: In this retrospective study, 89 SSNs were detected in 86 LDCT screening participants, including 42 pure ground-glass nodules (GGNs) and 47 part-solid GGNs. In these participants, 28 SSNs were diagnosed as lung cancer at histopathology, and 61 SSNs from participants who underwent at least two LDCT imaging studies. All nodules were divided into three groups: cancer group, growth group, and non-growth group. The nodule size, volume, attenuation, volume doubling time (VDT), and texture parameters (mean value, uniformity, entropy and energy) were assessed, respectively. RESULTS: The entropy of the cancer group was significantly higher than that of the growth and non-growth groups (pure GGNs: p=0.009, 0.001; part-solid GGNs: p=0.012, 0.004). The energy of the cancer group was significantly lower than that of the other groups (pure GGNs: p=0.043, 0.021; part-solid GGNs: p=0.001, 0.002). A good positive correlation was found between uniformity and VDT (p=0.022). CONCLUSION: Different CT texture parameters show good predictive value for SSNs detected at LDCT screening: the entropy and energy differences between malignant pulmonary nodules and others could be a helpful quantitative index to predict the malignancy of SSNs. Uniformity could be used to predict the growth probability of pure GGNs at baseline to pay more attention to these nodules. Moreover, the follow-up and treatment plan could be more targeted.
AIM: To analyse subsolid nodules (SSNs) detected during low-dose (LD) computed tomography (CT) screening and investigated whether CT texture analysis parameters can predict the malignancy and growth trends of GGNs. MATERIALS AND METHODS: In this retrospective study, 89 SSNs were detected in 86 LDCT screening participants, including 42 pure ground-glass nodules (GGNs) and 47 part-solid GGNs. In these participants, 28 SSNs were diagnosed as lung cancer at histopathology, and 61 SSNs from participants who underwent at least two LDCT imaging studies. All nodules were divided into three groups: cancer group, growth group, and non-growth group. The nodule size, volume, attenuation, volume doubling time (VDT), and texture parameters (mean value, uniformity, entropy and energy) were assessed, respectively. RESULTS: The entropy of the cancer group was significantly higher than that of the growth and non-growth groups (pure GGNs: p=0.009, 0.001; part-solid GGNs: p=0.012, 0.004). The energy of the cancer group was significantly lower than that of the other groups (pure GGNs: p=0.043, 0.021; part-solid GGNs: p=0.001, 0.002). A good positive correlation was found between uniformity and VDT (p=0.022). CONCLUSION: Different CT texture parameters show good predictive value for SSNs detected at LDCT screening: the entropy and energy differences between malignant pulmonary nodules and others could be a helpful quantitative index to predict the malignancy of SSNs. Uniformity could be used to predict the growth probability of pure GGNs at baseline to pay more attention to these nodules. Moreover, the follow-up and treatment plan could be more targeted.