RESUMO
This study deals with the application of methods of second-order stereology to investigate the spatial distribution of nuclei in normal prostate, prostatic intraepithelial neoplasia and adenocarcinoma. We aimed to identify differences related to the progression of premalignant lesions (PIN) to carcinoma, as well as the spatial changes in relation to tumour grade. Estimation of second-order stereology parameters, such as g(r), (pair correlation function), statistic M, and Clark-Evans aggregation index (CEAI) were employed to investigate the distribution of nuclei. Linear discriminant analysis (LDA) with M and CEAI as model variables was implemented to classify the cancer cases into two groups according to Gleason score. We found that the point processes of the nuclei in prostatic cancer and normal tissues differed by first-order as well as by second-order properties. In the PIN the mean g-values were intermediate between normal and cancer. The LDA indicates that M and CEAI were able to classify into the correct group of Gleason score more than 90% of the cases analysed. Cancer cases showing a higher degree of disorder in the spatial distribution of nuclei were significantly classified into the group of higher Gleason score. The nuclei in both normal and pathological prostate were not Poisson distributed. Additionally, we found that the progression from normal tissue to carcinoma was accompanied by a progressive increase in spatial disorder which is intermediate in pre-malignant lesions (PIN). The parameters employed were able to classify the cancer cases according to the Gleason score
No disponible
