Disease-free survival of node-positive breast cancer patients. Improved prognostication by cytometrical parameters.

Feulgen stained cytologic samples from 225 node-positive breast cancers were investigated by means of an image analysis system. From each tumor sample, 100 cells were scanned and several DNA, morphometrical and textural parameters were evaluated. The meaning of the cytometric parameters for prediction of distant metastases within five years was investigated by the stepwise Cox regression analysis. Most of the investigated DNA- and morphometrical parameters, as well as one textural feature, showed a significant univariate correlation with the clinical course. In the multivariate approach, the lymph node status (pN) was the strongest prognostic factor, followed by the histogram type, the tumor size (pT) and a textural parameter (heterochromatin area). By the linear combination of these selected variables a multivariate prognostic factor was calculated for each individual patient. Using this factor, the patients could be splitted into four groups according to their risk for distant metastases. For this, the continuous range of the multivariate factor was subdivided so that about 35% of the patients were in the middle groups and about 15% of the patients in each of the border groups with highest and lowest factors, respectively. Thus a low risk group (lowest factors) of node-positive patients could be identified with a 5-year distant recurrence-rate of only 6.5%, as well as a group of patients with a considerably worse prognosis (highest factors) and a distant recurrence-rate of 67%. Therefore, DNA, morphometrical and textural parameters can provide powerful prognostic information in node-positive breast carcinomas. Using the multivariate combination of clinical and relevant cytometrical parameters may allow a more appropriate selection of patients for adjuvant therapy.