Grading and prognosis of invasive ductal mammary carcinoma by nuclear image analysis in tissue sections

In primary tumours of 84 patients with invasive ductal carcinomas the prognostic significance of the nuclear image was assessed. Feulgen stained tissue sections were investigated by image cytometry. At least 150 tumour cell nuclei as well as 50 lymphocytes or trout erythrocytes for diploid reference were measured in each specimen. 35 morphological features, including those of the chromatin pattern, were derived from the digitized nuclear images. By means of multivariate discriminant analysis, significant differences could be found between tumours of 41 node-negative and 43 node-positive patients. The differences were sufficiently large to distinguish subgroups with low and high risk of lymph node involvement. Irrespective of lymph node status, 66 patients with an adequate 5 year follow-up were divided into two groups according to their survival time. The classification results revealed a clear relationship between characteristics of the nuclear image and prognosis. By image cytometry it is thus possible to separate both node-positive and node-negative patients into subgroups with increasing risk of recurrence. The results indicate that nuclear image analysis provides prognostic information in addition to established prognostic factors, such as histological grade and tumour stage.