Prognostic implications of ploidy and proliferative activity in human solid tumors.

Ploidy and cell cycle compartment distribution were measured by DNA flow cytometry in 261 patients with a variety of different tumors. Eighty-one percent of all tumors were aneuploid, and 72% were hyperdiploid. Ploidy levels spanned a wide range from hypodiploid (maximum 30% less than diploid controls) to hyperoctaploid (440% in excess of diploid controls) with a mean and median values coinciding at a near-triploid DNA content. The proportion of cells with G1 DNA content decreased with increasing hyperdiploid abnormality. While unrelated to biopsy site and to a number of host factors such as age, sex and race, both ploidy and cytokinetic parameters were markedly affected by histopathologic diagnosis. Patients with metastatic lung, breast and GI cancer had higher ploidy levels than individuals with the corresponding primary tumors. Ploidy (except for one patient) remained constant, and G1/100 proportions showed only minor variation by disease site and over a median observation time of 6 months. Prognostic factor analysis was performed in the subgroup of patients studied within 6 months from diagnosis. The adverse impact of low tumor G1/100 proportion on survival was lost as the proportional hazard analysis was extended to include diagnostic subgroups. Accounting for histopathologic diagnosis, stage of disease, ploidy, and the proportion of tumor G1/100 cells, the following sequence of adverse prognostic factors in order of their relative ranks was established: (1) absence of breast cancer (p=0.0001), (2) hypertriploid DNA index (p=0.049), and (3) presence of metastatic disease (p=0.079). Our study demonstrates that DNA content-derived information on instrinsic tumor cell features pertaining to cytogenetics and cytokinetics may provide an objective means of biologically relevant cancer classification.