Two-dimensional multi-criterion segmentation of pulmonary nodules on helical CT images

A multi-criterion algorithm for automatic delineation of small pulmonary no dules on helical CT images has been developed. In a slice-by-slice manner, the algorithm uses density, gradient strength, and a shape constraint of th e nodule to automatically control segmentation process. The multiple criter ia applied to separation of the nodule from its surrounding structures in l ung are based on the fact that typical small pulmonary nodules on CT images have high densities, show a distinct difference in density at the boundary , and tend to be compact in shape. Prior to the segmentation, a region-of-i nterest containing the nodule is manually selected on the CT images. Then t he segmentation process begins with a high density threshold that is decrea sed stepwise, resulting in expansion of the area of nodule candidates. This progressive region growing approach is terminated when subsequent threshol ds provide either a diminished gradient strength of the nodule contour or s ignificant changes of nodule shape :from the compact form. The shape criter ion added to the algorithm can effectively prevent the high density surroun ding structures (e.g., blood vessels) from being falsely segmented as nodul e, which occurs frequently when only the gradient strength criterion is app lied. This has been demonstrated by examples given in the Results section. The algorithm's accuracy has been compared with that of radiologist's manua l segmentation, and no statistically significant difference has been found between the nodule areas delineated by radiologist and those obtained by th e multi-criterion algorithm. The improved nodule boundary allows for more a ccurate assessment of nodule size and hence nodule growth over a short time period, and for better characterization of nodule edges. This information is useful in determining malignancy status of a nodule at an early stage an d thus provides significant guidance for further clinical management. (C) 1 999 American Association of Physicists in Medicine.