Statistical Analysis of Histopathological Parameters for Assessing Prognosis of Oral Squamous Cell Carcinoma Cases

对10例口腔鳞状细胞癌(SCC)福尔马林固定石蜡包埋材料进行了DNA倍体分析，并用5微米切片显微分光光度法和分离的肿瘤细胞荧光分光光度法对活检材料进行了选择。两种方法显示相同DNA直方图的病例很少。这是由于显微分光光度法测单细胞DNA含量不够准确所致。25例舌鳞癌患者接受了放射治疗，并有足够的临床和随访资料。从石蜡包埋材料中分离的肿瘤细胞进行DADI染色，以50~100个肿瘤细胞的核DNA含量为DNA直方图。根据Auer等人的标准，将DNA多倍性分为4种类型。单二倍体(2C)峰的I型和二倍体(2C)和/或四倍体(4C)峰的II型被归类为低倍体组。类型III，DNA含量在2C~4C之间的细胞数量较多，代表增殖细胞；类型IV，DNA分布广泛，范围从2C到8C以上，被归类为高倍体组。两组生存曲线(Kaplan-Meier法)显示高倍体患者预后较差，但分析病例数不足，经Logrank检验或Cox-Mantel检验，两组生存曲线差异无统计学意义。高倍体组死亡发生率(9/14)高于低倍体组(4/11)。

DNA ploidy analysis of formalin-fixed, paraffin embedded materials of 10 cases of oral squamous cell carcinoma(SCC) was performed by two different methods, microspectrophotmetry with 5 micron sections and cytofluorometry with isolated tumor cells to choose the method for the biopsy materials. There were few cases which showed same DNA histogram pattern in two methods. This was due to the insufficient accuracy in measurement of single cell DNA content with the microspectrophotometric method.Twenty-five cases with the tongue SCC,which had recieved radiation therapy and had sufficient clinical and follow-up data, were available for this study. Isolated tumor cells from the paraffinembedded materials were stained with DADI and DNA histogram was made of nuclear DNA content of 50-100 tumor cells. According to the criteria of Auer et al., DNA polidy patterns were divided into 4 types. Type I with a single diploid (2C)peak and type II with a diploid(2C) and/or a tetraploid(4C)peak were summarized as low ploidy group. Type III with a sizable amount of cells with DNA content between 2C and 4C representing proliferating cells and type IV showing a broad DNA distribution ranging from 2C to beyond 8C were summarized as high ploidy group. Survival curves (Kaplan-Meier method) of two groups demonstrated the more unfavorable prognosis in high ploidy cases, however, the number of analysed case was not sufficient so that the difference between two survival curves was not statistically significant with logrank test or Cox-Mantel test. Death was more frequently encountered in the high ploidy cases(9/14) than in the low ploidy cases(4/11).