Lack of correlation between flow cytometric and immunohistologic proliferation measurements of tumors.

Abstract

We compared different means of assaying tumor proliferative activity by either flow cytometric or immunohistologic methods in formalin-fixed, paraffin-embedded blocks. A series of 84 Dukes' stage B colorectal carcinomas were examined to identify high-risk patients who may potentially benefit from adjuvant therapy. Flow cytometric analysis was performed by a modified Hedley method with a combined S+G2/M phase proliferative fraction calculated by means of a rectangular model after debris subtraction. Immunohistologic tumor proliferative activity was analyzed by means of serial step sections from the same blocks used for flow cytometric examination with antibodies to proliferating cell nuclear antigen (PCNA) and Ki-67 (MIB-1). Mean with standard deviation and (range) tumor proliferative activity measurements were: flow cytometric analysis: proliferative fraction = 14.8% +/- 5.3 (5-27%), PCNA = 43.2% +/- 21.2 (4-90%), and MIB-1 = 16.2% +/- 10.8 (2-47%). No correlation was found between flow cytometric proliferative fraction and immunohistologic tumor proliferation measurement or between PCNA and MIB-1 staining indices. Lack of correlation between flow cytometric and immunohistologic findings may be related to the use of archival formalin-fixed paraffin-embedded tissue for flow cytometric evaluation, with resultant increased debris and decreased accuracy of cell cycle calculations. Discordance between PCNA and MIB-1 may reflect inherent problems with anti-PCNA antibody staining of formalin-fixed tissues whereby anti-PCNA clone PC-10 detects non-replicon associated PCNA in formalin-fixed tissues. Prospective studies using fresh tissue with two-color multiparameter flow cytometric analysis and histogram-dependent background fitting may help to clarify the relationship between findings of tumor proliferation as analyzed by flow cytometric and by immunohistologic techniques.