Proliferation markers Ki-67 and p105 in soft-tissue lesions. Correlation with DNA flow cytometric characteristics

Frozen tissue immunoreactivity with Ki-67, a monoclonal antibody that recognizes a nuclear antigen in nonresting or proliferating cells, was compared to DNA flow cytometry results (from fresh tissue) in a diverse group of 60 soft-tissue lesions. Both DNA index and Ki-67 score were independently reported to be associated with grade and prognosis in sarcomas, but no direct comparison of these two variables was made. It was attempted to measure proliferative activity in fixed paraffin-embedded tissues immunohistochemically in a subset of lesions using an antibody to another nuclear proliferation antigen, p105. Lesions were given a grade according to lesion category (reactive, 1; benign, 2; low-grade malignant, 3; and high-grade malignant, 4). Ki-67 reactivity correlated relatively well with this grading system (r = 0.59); benign lesions usually exhibited a low Ki-67 score and malignant lesions usually but not always exhibited a high score. For example, some malignant fibrous histiocytomas contained only rare positive cells. Some disparity between Ki-67 score and grade and within histologic types indicates some independence from these features, a fact that may be important when correlation with prognosis is performed. However Ki-67 did not correlate well with flow data such as percentage S phase (r = 0.30), percentage S + G2M phases (r = 0.37), or DNA index (r = 0.39). This probably is due to the fact that Ki-67 also marks cells in the G1 phase, whereas these are excluded in flow data analyses. Anti-p105 highlighted almost all nuclei in all cases tested, including fibromatosis, and did not correlate with Ki-67 score, histologic grade or DNA flow cytometric data. Results with p105 could not be favorably affected by titration experiments. It is reasonable to conclude that the Ki-67 score is a variable related to but independent of histologic grade, histologic type, and DNA flow values. Whether it is prognostically important in human sarcomas, as has been suggested, awaits further clinicopathologic study.