Fractal analysis for classification of ground-glass opacity on high-resolution CT: an in vitro study.

Purpose: Fractal analysis based on the fractional Brownian motion model was applied to ground-glass opacity on high-resolution CT (HRCT) to investigate its usefulness in distinguishing ground-glass opacity caused by nonfibrotic disease processes and that caused by fibrotic disease processes, confirming pathology. Method: Twenty-one postmortem lungs inflated and fixed by Heitzman's method were evaluated. By correlating HRCT and pathology, the lungs were classified into nonfibrotic disease processes and fibrotic disease processes. Fractal analysis based on the fractional Brownian motion model provides the parameter H, which is a statistical measure related to the psychophysical perception of roughness. For regions of interest positioned over ground-glass opacities on HRCT, conventional statistics (mean value and SD) and the estimated H values were calculated using a workstation. Results: Pathologically, 10 lung specimens were categorized as nonfibrotic disease processes and 11 as fibrotic disease processes. Whereas the conventional statistics had considerable overlap in two disease processes, the overlapping was drastically reduced in the H values. The H values of fibrotic disease processes (mean ± SD, 0.423 ± 0.064) were significantly greater than those of nonfibrotic disease processes (0.297 ± 0.036) (p < 0.001). Conclusion: Fractal analysis based on the fractional Brownian motion model may provide a new promising scheme for assessing ground-glass opacity on HRCT caused by either nonfibrotic or fibrotic disease processes.