Fisher E R, Redmond C, Fisher B
Institute of Pathology, Shadyside Hospital, Pittsburgh, PA 19104.
J Natl Cancer Inst Monogr. 1992(11):151-8.
Twenty-two pathologic (including estrogen and progesterone receptor status) and four clinical features of 950 node-negative stage I invasive breast cancers from 950 women enrolled in the National Surgical Adjuvant Breast and Bowel Project protocol B-06 were analyzed for their possible prognostic significance. Univariate analyses revealed 10 characteristics that were significant at the 1% level. Only three of these--notably nuclear grade, histologic tumor type, and race--were found to be significant when entered into a Cox regression model. Patients whose tumors exhibited a good nuclear grade fared significantly better than those whose tumors were scored as poor. Similarly, a significantly better prognosis was noted when the histologic type of cancer was found to be "favorable" (mucinous, tubular, or papillary) than when it was "intermediate" (NOS, "Not Otherwise Specified," combination; typical medullary; and lobular invasive) or "unfavorable" (NOS pure and atypical medullary). Blacks exhibited a worse prognosis than whites. Survival was 94% at 8 years when the nuclear grade was good and the tumor type favorable, but only 54% when the nuclear grade was poor and tumor type unfavorable. Patients with one favorable and one unfavorable feature exhibited an intermediate survival. A brief overview as well as our own preliminary experience indicates that the combined use of these two prognostic pathologic parameters may be as good as and in some instances a better predictor of survival in node-negative patients than information derived from more "objective" methodologies such as flow cytometry, receptor analyses and tumor labeling indices or the demonstration of oncogene overexpression. Assessment of the pathologic parameters is simple, universally available, and quick and requires only modest training to be reproducible.
对参加国家外科辅助乳腺和肠道项目协议B - 06的950名女性的950例淋巴结阴性I期浸润性乳腺癌的22项病理特征（包括雌激素和孕激素受体状态）和4项临床特征进行了分析，以探讨其可能的预后意义。单因素分析显示有10项特征在1%水平上具有显著性。当将这些特征纳入Cox回归模型时，仅发现其中3项特征具有显著性，即核分级、组织学肿瘤类型和种族。肿瘤核分级良好的患者预后明显好于核分级差的患者。同样，当癌症的组织学类型为“有利”（黏液性、管状或乳头状）时，其预后明显好于“中等”（未另行规定、组合；典型髓样；小叶浸润）或“不利”（未另行规定的纯型和非典型髓样）类型。黑人的预后比白人差。当核分级良好且肿瘤类型有利时，8年生存率为94%，而当核分级差且肿瘤类型不利时，生存率仅为54%。具有一项有利特征和一项不利特征的患者生存率处于中间水平。简要概述以及我们自己的初步经验表明，对于淋巴结阴性患者，联合使用这两个预后病理参数在预测生存率方面可能与从更“客观”的方法（如流式细胞术、受体分析、肿瘤标记指数或癌基因过表达的检测）获得的信息一样好，在某些情况下甚至更好。评估病理参数简单、普遍适用且快速，只需要适度的培训就能实现可重复性。
