1.Department of Computed Tomography and Magnetic Resonance, the Fourth Hospital of Hebei Medical University, Shijiazhuang 050011, China; 2.Department of Radiology, Hebei General Hospital, Shijiazhuang 050051, China; 3.Department of Urology, the Second Hospital of Hebei Medical University, Shijiazhuang 050000, China; 4.Department of Pathology, Hebei General Hospital, Shijiazhuang 050051, China
GUO Xiaowan,JIA Xudong,ZHANG Danqing, et al. Consistency analysis of the three-dimensional volume and mass measurement in pulmonary subsolid nodules[J]. JOURNAL OF WEZHOU MEDICAL UNIVERSITY, 2023, 53(12): 974-979.
Abstract：Objective: To investigate the reproducibility of volume and mass measurements assessed by commercially available pulmonary nodule segmentation software and to compare intraobserver and interobserver differences. Methods: From June 1, 2016 to December 31, 2019, image data of 163 SSNs found in 117 patients who underwent chest CT examination in Hebei Provincial People’s Hospital were retrospectively collected. The two-dimensional (2D) volume, three-dimensional (3D) volume and mass of the enrolled SSNs were measured with volumetric software by two radiologists. Bland-Altman method was used to compare intraobserver and interobserver variability, and Wilcoxon method was used to compare the variance between the two. Results: A total of 163 SSNs were included. The intra-observer correlation coefficient of 3D mass and volume measurement variability was 0.999 and the interobserver correlation coefficient of 3D mass and volume measurement variability was 0.999. As for two-dimensional volume measurement, intraobserver and interobserver variability ranged from -25.1% to 24.6% and from -32.0% to 30.3% respectively, in contrast to the range from -16.7% to 13.7% and from -15.0% to 13.5%, respectively, for the three-dimensional volume measurement. As for the three-dimensional mass measurement, the intrao-bserver and interobserver variability ranged from -7.8% to 8.2% and from -11.0% to 9.7%, respectively. No significant difference was found for intra-observer and interobserver variability between 3D volume and mass measurements (P>0.05). Conclusion: The commercially available semi-automatic pulmonary nodule segmentation software has good repeatability in the quality measurement of SSNs, which can be used as a quantitative assessment method for the follow-up growth of SSNs.
