Diagnosis of Liver Neoplasms by Computational and Statistical Image Analysis

Abstract

Background: Distinguishing well-differentiated hepatocellular carcinoma (WD-HCC), hepatocellular adenoma (HA) and non-neoplastic liver tissue (NNLT) solely on morphology is often challenging. The purpose of this study was to evaluate the use of computational image analysis to distinguish WD-HCC, HA and NNLT.

Methods: Seventy-seven cases comprising of WD-HCC (n = 26), HA (n = 23) and NNLT (n = 28) were retrieved and reviewed. A total of 485 hematoxylin and eosin (H&E) photomicrographs (× 400, 0.09 µm²) of WD-HCC (n = 183), HA (n = 173), NNLT (n = 129) and nine whole-slide scans (three of each diagnosis) were obtained, color deconvoluted and digitally transformed. Quantitative data including nuclear density, nuclear sphericity, nuclear perimeter, and nuclear eccentricity from each image were acquired. The data were analyzed by one-way analysis of variance (ANOVA) with Tukey post hoc test, followed by unsupervised and supervised (Chi-square automatic interaction detection (CHAID)) cluster analysis.

Results: Unsupervised cluster analysis identified three well defined clusters of WD-HCC, HA and NNLT. Employing the four most discriminating nuclear features, supervised analysis was performed on a training set of 383 images, and validated on the remaining 102 test images. The analysis identified WD-HCC (sensitivity 100%, specificity 98%), HA (sensitivity 71%, specificity 85%) and NNLT (sensitivity 70%, specificity 86%). An analysis of whole-slide images identified WD-HCC with sensitivity and specificity of 100%.

Conclusions: We have successfully demonstrated that computational image analysis of nuclear features can differentiate WD-HCC from non-malignant liver with high accuracy, and can be used to assist in the histopathological diagnosis of hepatocellular carcinoma.

Keywords: Computational analysis; Hepatocellular adenoma; Hepatocellular carcinoma; Image analysis; Liver.

Figure 1

Illustration of work flow and methods for research.

Figure 2

Representative whole slide scan of resected liver tissue showing region of interest selected for analysis demarcated in yellow. Note that it excludes portal tracts.

Figure 3

Representative photomicrographs of non-neoplastic liver tissue (NNLT), hepatocellular adenoma (HA), inflammatory type hepatocellular adenoma (HA-I), and well-differentiated hepatocellular carcinoma (WD-HCC) in sequential order: hematoxylin and eosin (H&E), and digitally transformed to hematoxylin blue, luminance, and morphological opening (× 400).

Figure 4

Distribution of nuclear density (a), median nuclear perimeter (b), median nuclear sphericity (c), and standard deviation (SD) nuclear eccentricity (d) of non-neoplastic liver tissue (NNLT) (n = 129), well-differentiated hepatocellular carcinoma (WD-HCC) (n = 183), and hepatocellular adenoma (HA) (n = 173) (P < 0.001; one-way analysis of variance (ANOVA)). Bars indicate 5th and 95th percentiles; boxes represent the 25th and 75th percentiles; lines inside the boxes are medians. Asterisks (*) indicate significant (P < 0.05) differences by Tukey post hoc test.

Figure 5

Unsupervised clustering. (a) Dendrogram developed with hierarchical clustering analysis. Dissimilarity is decided with euclidean distance. (b) Distances between the final cluster centers developed with k-means clustering analysis. (c) Composition of each cluster developed via k-means clustering analysis.

Figure 6

Decision tree from Chi-square automatic interaction detection (CHAID) analysis. Columns left to right: root nodes; decision nodes; terminal nodes. Shades of red from dark to light indicate well-differentiated hepatocellular carcinoma (WD-HCC), hepatocellular adenoma (HA), and non-neoplastic liver tissue (NNLT), respectively. Purity of each diagnosis is given in terminal nodes. D: nuclear density; E: standard deviation (SD) nuclear eccentricity; P: median nuclear perimeter; S: median nuclear sphericity. ∊: a set (number 1, number 2); ): number included in analysis; [: number excluded from analysis.