Tumor Burden Analysis on Computed Tomography by Automated Liver and Tumor Segmentation

Abstract

The paper presents the automated computation of hepatic tumor burden from abdominal CT images of diseased populations with images with inconsistent enhancement. The automated segmentation of livers is addressed first. A novel three-dimensional (3D) affine invariant shape parameterization is employed to compare local shape across organs. By generating a regular sampling of the organ's surface, this parameterization can be effectively used to compare features of a set of closed 3D surfaces point-to-point, while avoiding common problems with the parameterization of concave surfaces. From an initial segmentation of the livers, the areas of atypical local shape are determined using training sets. A geodesic active contour corrects locally the segmentations of the livers in abnormal images. Graph cuts segment the hepatic tumors using shape and enhancement constraints. Liver segmentation errors are reduced significantly and all tumors are detected. Finally, support vector machines and feature selection are employed to reduce the number of false tumor detections. The tumor detection true position fraction of 100% is achieved at 2.3 false positives/case and the tumor burden is estimated with 0.9% error. Results from the test data demonstrate the method's robustness to analyze livers from difficult clinical cases to allow the temporal monitoring of patients with hepatic cancer.

Automatic segmentation of subcutaneous mouse tumors by multiparametric MR analysis based on endogenous contrast

“…While the threshold-based method is likely to require re-definition of the set thresholds for each tumor type, automatic selection of clusters could possibly also be performed on the basis of features more specific to tumors in general, for example shape, size [18], and tissue homogeneity [9]. Recently, Linguraru et al [34] showed that liver tumors can be accurately detected on contrastenhanced computed tomography images by using a set of features that describe the intensity, shape, size, and homogeneity of identified objects of interest in the liver. Apart from their potential utility for cluster selection, such tumor features could also be used to select the tumor volume from the results from morphological component analysis.…”

Computer Aided Diagnosis for Liver Cancer Using Statistical Model

“…If the features extracted are carefully chosen it is expected that the features set will extract the relevant information from the input data in order to perform the desired task using this reduced representation instead of the full size input [5], [12].…”

Automated segmentation methods for liver analysis in oncology applications

“… Casciaro et al [82] published a method that could achieve TPR=92.3% (no information about FPC).  The evaluation of the method presented by Linguraru et al [83] demonstrated TPR=100% with FPC=2.3.  The paper of Wu et al [84] demonstrated TPR=90% with FPC=2.6.…”

“…In addition to the main cancer types the detection and segmentation of brain pathologies [66] in MR images [67], lymph nodes in CT images [68], and liver tumours in CT or MR images were also focused on in many publications. The three main motivations for liver lesion detection are lesion classification [51, 69-73, 88, 90], lesion segmentation and quantification [75][76][77][78][79][80][81][82][83][84][85]89], and follow-up [86]. The following paragraphs summarize the recent methods and results related to liver lesion detection.…”

“…The method presented by Casciaro et al [82] is based on graph-cut and gradient flow active contour, which has a relatively high computation cost (10+ seconds per slice). Linguraru et al [83] published a technique that uses graph cuts to segment the hepatic tumours using shape and enhancement constraints. This method takes a couple of minutes to run.…”