用逐步回归法对乳腺单纯癌癌细胞形态结构进行定量分析

A Quantitative Analysis of Morphology of the Simple Cancer Cells in Breast Carcinoma with Mathematics Statistical Stepwise Method

用体视学方法对女性晚期乳腺单纯癌癌细胞、细胞核、核仁的１７ 个形态参数进行测试， 并用逐步回归法分析， 得最优方程 Ｙ＝ ０６０９ ＋ ００２８ Ｎ１４ － ０４７５ Ｎ１ － ０５８７ Ｎ１７ ， 结果筛选出区分正常及癌细胞的最佳参数： 核质比 Ｖｎｐ 、核平均截面积 Ａｎ 、核仁平均周长 Ｂｕ ． 提示了３ 个形态参数变化程度越大， 可能癌症病人的预后越不好．

Stereological method was applied to the analysis of 17 morphological paramenters of the cells, nuclei and nucleoli in scirrhous breast carcinoma. The best equation set up with mathematics statistical stepwise method was as follows: Y =0 609+0 028 N 14 -0 475 N 1-0 587 N 17 . Due to above equation, three morphometric parameters which could be used to distinguish any normal adjacent glandular cells from carcinomatous cells were selected, namely, the ratio of nucleus to cytoplasm V np, average section of nucleus A n and average perimeter of nucleolus B u. It is suggested that the greater the three morphological parameters change, the worse the prognosis of the cancer patients will be.