Three-dimensional quantitative analysis of cell nuclei for grading renal cell carcinoma

In this paper, we have proposed a method for renal cell carcinoma (RCC) grading, using a three-dimensional (3D) quantitative analysis of cell nuclei based on digital image cytometry. We acquired volumetric RCC data for each grade using confocal laser scanning microscopy (CLSM) and developed a method for grading RCC using 3D visualization and quantitative analysis of cell nuclei. First, we used a method of segmenting cell nuclei based on Pun's method. Second, to determine quantitative features, we used a 3D labeling method based on slice information. After applying the labeling algorithm, we determined the measurements of cell nuclei using 3D quantitative analysis. To evaluate which of the quantitative features provided by 3D analysis could contribute to diagnostic information and could increase accuracy in nuclear grading, we analyzed statistical differences in 3D features among the grades. We compared features measured in two dimensions (diameter, area, perimeter, and circularity) with features measured in three dimensions (volume, surface area, and spherical shape factor) between identical cell nuclei by using regression analysis. For 3D visualization, we used a contour-based method for surface rendering. We found a statistically significant correlation between the nuclear grade and the 3D morphological features. Comparing our results to an ideal RCC grading system, we found that our nuclear grading system based on the 3D features of a cell nucleus provides distinct dividing points between grades and also provides data that can be easily interpreted for diagnoses. 3D visualization of cell nuclei offers a realistic display and additional valuable medical information that can lead to an objective diagnosis. This method could overcome the limitations inherent in 2D analysis and could improve the accuracy and reproducibility of quantification of cell nuclei. Our study showed that a nuclear grading system based on the 3D features of a cell nucleus might be an ideal grading system.

[1]  John R. Hand,et al.  Carcinoma of the Kidney: The Degree of Malignancy in Relation to Factors Bearing on Prognosis1 , 1932 .

[2]  N. Goldstein,et al.  The current state of renal cell carcinoma grading , 1997, Cancer.

[3]  C Decaestecker,et al.  Classification strategies for the grading of renal cell carcinomas, based on nuclear morphometry and densitometry , 1997, The Journal of pathology.

[4]  Joan S. Weszka,et al.  A survey of threshold selection techniques , 1978 .

[5]  Ewert Bengtsson,et al.  Image Analysis Based Grading of Bladder Carcinoma. Comparison of Object, Texture and Graph Based Methods and Their Reproducibility , 1997, Analytical cellular pathology : the journal of the European Society for Analytical Cellular Pathology.

[6]  Thierry Pun,et al.  Entropic thresholding, a new approach , 1981 .

[7]  Ewert Bengtsson,et al.  A Feature Set for Cytometry on Digitized Microscopic Images , 2003, Analytical cellular pathology : the journal of the European Society for Analytical Cellular Pathology.

[8]  J. Portillo,et al.  Nuclear morphometry in prognosis of renal adenocarcinoma. , 1992, Urology.

[9]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[10]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[11]  S. Fuhrman,et al.  Prognostic significance of morphologic parameters in renal cell carcinoma , 1982, The American journal of surgical pathology.

[12]  Jeha Ryu,et al.  Contour-based algorithms for generating 3D CAD models from medical images , 2003 .

[13]  S. Aktaş,et al.  Volume-weighted mean nuclear volume in renal cell carcinoma. , 1998, Urology.

[14]  Joakim Lindblad,et al.  Algorithms for Cytoplasm Segmentation of Fluorescence Labelled Cells , 2002, Analytical cellular pathology : the journal of the European Society for Analytical Cellular Pathology.

[15]  An Integrated System for Feature Evaluation of 3D Images of a Tissue Specimen , 2002 .

[16]  K Fujikawa,et al.  Role of volume weighted mean nuclear volume for predicting disease outcome in patients with renal cell carcinoma. , 1997, The Journal of urology.

[17]  Chun-Jen Chen,et al.  A linear-time component-labeling algorithm using contour tracing technique , 2004, Comput. Vis. Image Underst..

[18]  C Decaestecker,et al.  Improving the prognostic value of histopathological grading and clinical staging in renal cell carcinomas by means of computer‐assisted microscopy , 1999, The Journal of pathology.

[19]  Luc Vincent,et al.  Watersheds in Digital Spaces: An Efficient Algorithm Based on Immersion Simulations , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[20]  P. Hermanek,et al.  Histological grading of renal cell carcinoma. , 1976, European urology.

[21]  R. Colvin,et al.  Diagnosis and management of renal cell carcinoma A clinical and pathologic study of 309 cases , 1971, Cancer.

[22]  Thierry Pun,et al.  A new method for grey-level picture thresholding using the entropy of the histogram , 1980 .

[23]  G Brugal,et al.  Improving accuracy in the grading of renal cell carcinoma by combining the quantitative description of chromatin pattern with the quantitative determination of cell kinetic parameters. , 2000, Cytometry.

[24]  A W Partin,et al.  Nuclear morphometry adds significant prognostic information to stage and grade for renal cell carcinoma. , 1999, Urology.

[25]  P.K Sahoo,et al.  A survey of thresholding techniques , 1988, Comput. Vis. Graph. Image Process..