Computer‐aided diagnosis of pulmonary nodules on CT scans: Segmentation and classification using 3D active contours

Abstract

We are developing a computer-aided diagnosis (CAD) system to classify malignant and benign lung nodules found on CT scans. A fully automated system was designed to segment the nodule from its surrounding structured background in a local volume of interest (VOI) and to extract image features for classification. Image segmentation was performed with a three-dimensional (3D) active contour (AC) method. A data set of 96 lung nodules (44 malignant, 52 benign) from 58 patients was used in this study. The 3D AC model is based on two-dimensional AC with the addition of three new energy components to take advantage of 3D information: (1) 3D gradient, which guides the active contour to seek the object surface, (2) 3D curvature, which imposes a smoothness constraint in the z direction, and (3) mask energy, which penalizes contours that grow beyond the pleura or thoracic wall. The search for the best energy weights in the 3D AC model was guided by a simplex optimization method. Morphological and gray-level features were extracted from the segmented nodule. The rubber band straightening transform (RBST) was applied to the shell of voxels surrounding the nodule. Texture features based on run-length statistics were extracted from the RBST image. A linear discriminant analysis classifier with stepwise feature selection was designed using a second simplex optimization to select the most effective features. Leave-one-case-out resampling was used to train and test the CAD system. The system achieved a test area under the receiver operating characteristic curve (A z ) of 0.83±0.04. Our preliminary results indicate that use of the 3D AC model and the 3D texture features surrounding the nodule is a promising approach to the segmentation and classification of lung nodules with CAD. The segmentation performance of the 3D AC model trained with our data set was evaluated with 23 nodules available in the Lung Image Database Consortium (LIDC). The lung nodule volumes segmented by the 3D AC model for best classification were generally larger than those outlined by the LIDC radiologists using visual judgment of nodule boundaries.

Automatic 3D pulmonary nodule detection in CT images: A survey

“…After this analysis, only 38 articles [24,4,18,22,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58] were identified as 3D automated algorithms to segment lung nodules in CT images, the target of this work.…”

“…Some of the selected works are not fully automatic, as require user intervention, or nor indicate the values of sensitivity and/or about false positives: [33], [34], [36], [40], [41], [42], [43], [44], [48], [49], [50], [53], [54], [55], [56] and [58]. Although these works are not the focus of this review, they are briefly introduced here.…”

“…Way et al [43] proposed a 3D active contours model based on twodimensional active contours (AC) with the addition of three new energy components to take advantage of the 3D information: gradient, curvature and energy mask. The search for the best energy weights in the 3D AC model is guided by the simplex optimization method.…”

Computational methods for the analysis of functional 4D-CT chest images.

“…These approaches are adopted to evolve 3D surface boundary directly. In Way et al [199], an approach to derive volumetric segmentation by 2D ACs applied to successive axial slices with 3D gradient, 3D curvature and mask energy terms in order to facilitate continuity along slice depths.…”

“…Technical approaches previously reported for volumetric lung nodule segmentation can be roughly classified into the following eleven categories: 1) thresholding [167,168,[183][184][185][186][187], 2) mathematical morphology [95,98,[188][189][190][191], 3) region growing [174,189,190,[192][193][194], 4) deformable model [195][196][197][198][199][200][201], 5) dynamic programming [202][203][204], 6) spherical/ellipsoidal model fitting [205][206][207][208][209] [220,221], and 11) watersheds [222]. Next, an overview of the technical approaches for lung nodule segmentation is given.…”