Staging procedures used to detect metastatic breast cancer at the time of diagnosis are bone scan (BS), chest X-ray (CXR), liver ultrasonography (LUS) and laboratory parameters (LP). These procedures are expensive and not all patients need them. We aimed to identify groups of patients with different risks for metastatic disease.

We reviewed data from 1,218 consecutive cases of breast cancer. Pathological and biological parameters and instrumental procedures performed at the time of diagnosis and during 6 months of follow-up were recorded. True positive and negative, false positive and negative cases were evaluated. All cases were grouped on the basis of tumour size, nodal involvement, biological characteristics, menopausal status and age.

We observed 46 (3.8%) true positive cases with metastatic disease at the time of diagnosis. Documentation relating to BS, CXR and LUS was available for 1,193, 1,206 and 1,206 patients, respectively, with 37 (3.1%), 8 (0.7%) and 10 (0.8%) true positive tests. Logistic regression analysis showed significant odds ratio estimates for pT status and nodal status, thus highlighting the role of these morphological data. These findings suggest that breast cancer patients can be divided into two subgroups: first group pT1-3N0-1. with < or = 3 involved nodes, and second group pT1-3N1 with > or = 4 involved nodes, pT4 and pN2 (metastases detection rate 1.46 and 10.68%, respectively). In the former group the appropriate procedures of staging would only be laboratory parameters, whereas in the latter group BS, CXR, LUS, LP and tumour markers CEA and CA 15.3 would.be necessary.

The standard staging procedures to detect metastatic disease at breast cancer diagnosis require modification. On the basis of the literature data and our findings, the full staging procedure is appropriate in the second group of patients.