Background: The prognostic value of quantitative histopathology was investigated in a retrospective study of 71 patients with ductal carcinoma of the breast. All patients were treated according to a standardized protocol. The median follow-up was 6 years.
Methods: Measurements were performed in microscopic fields that were sampled systematically from the whole tumor area of a routine histologic section. The unbiased stereologic method of point-sampled intercepts was used to estimate the mean nuclear volume, [vv(nuc)]. Using a test system with points and counting frames, estimates were obtained of the mean nuclear profile area, [aH(nuc)], the nuclear volume fraction, [Vv(nuc/tis)], the nuclear profile density (ND), the mitotic profile frequency (MF), and the mitotic profile density (MD). Traditional clinicopathologic parameters and biochemical estrogen receptor status were recorded.
Results: Single-factor survival analyses were significant regarding regional lymph node status, tumor dimension, clinical stage, age, aH(nuc), and vv(nuc) (P < or = 0.03). A tendency for prognostic value of MF was found (P = 0.10), whereas Vv(nuc/tis), ND, MD, histologic grade, and estrogen receptor status were insignificant. In a multivariate Cox analysis of patients with positive lymph nodes including the variables of tumor dimension, age, ND, vv(nuc), and MF, only vv(nuc) (P = 0.01) or MF (P = 0.004) were parameters of independent prognostic value.
Conclusion: The present feasibility study suggests that stereologic estimates of the mean nuclear volume and morphometric estimates of the mitotic profile frequency are of independent prognostic value for patients with ductal breast cancer with positive axillary lymph nodes. The prognostic information resulting from the two variables are correlated closely and cannot be separated in this study. Consequently, larger studies are needed. In addition, the independent value of quantitative histopathology in patients with lymph node negative breast cancer should be assessed.