Assessment of colorectal cancer prognosis through nuclear morphometry

Abstract

Due to the fact that different biological parameters play a key role in colonic malignant behavior, with nuclear morphometry being a prognostic marker in many malignancies, then predictive approaches in colorectal cancer (CRC) carried out on histologically well-defined groups may prevent interpretative errors. Subsequently, in the present study, CRC patients were screened according to the morphometric features of tumor cell nuclei, using an achúrate histotechnical approach, to analyze their clinical evolution according to Dukes? stratification. Materials and methods. A total of 66 cases were grouped according to Dukes? classification (5 y of follow-up). The perimeter, nuclear area, and shape factor of 50 interphase carcinoma nuclei were recorded through microphotographs obtained from each subject. Nuclei boundaries were drawn by an electronic pencil and examined by a computerized system. Data were submitted to a variance analysis, and a multiregression model compared results. Results. The sample was made up of 44 males (66.67%) and 22 females (33.33%) aged 59.7 +- 6 y old. Forty-nine patients (74.24%) were classified as stage B, and 17 (25.76%) as stage C. Nuclear homogeneity was confirmed by analysis of variance. The nuclear parameters were (mean +- SD): area (3.17 +- 1.74), perimeter (6.72 +- 1.83), and shape factor (0.82 +- 0.03). A multiple logistic regression model showed that stage C subjects had a higher risk of developing a worse clinical evolution than those at stage B (P < 0.02), independent of sex and age. Conclusions. Dukes' classification remains the best predictor of evolution. Although nucleomorphometric suitability is still controversial, this technique has become an additional tool to establish CRC prognosis.