PubMed1996
OBJECTIVE: To determine the accuracy and cost-effectiveness of skeletal scintigraphy in women with early, node-negative (T1-2N0M0) breast cancer. DESIGN: Retrospective, where scintigraphic prediction of metastases was compared with the criterion standard of radiological confirmation during a follow-up of 5 - 10 years. SETTING: Tertiary referral breast clinic at Groote Schuur Hospital. PATIENTS: Six hundred and seventy-three women with clinical T1-2N0M0 breast cancer who had skeletal scintigraphy between 1974 and 1987, and who had been followed up for more than 5 years. INTERVENTIONS: Initial skeletal scintigraphy, annual follow-up with radiological examination of symptomatic areas. MAIN OUTCOME MEASURES: Correlation of the sites indicated by scintigraphy with the initial presence or later development of metastases at 1 - 10 years, and the cost. RESULTS: Five hundred and sixty-one (83.4%) scans were normal, 35 (5.2%) indicated benign processes, and 77 (11.4%) were suggestive or diagnostic of metastatic disease, with radiological confirmation in 3 (initial detection rate 3/673, 0.44%; accuracy rate 3/77, 3.9%). Of the remaining 74 abnormal scans without radiological confirmation of metastases, 62 has a focus at a single site, and 45 were of low intensity and equivocal, with no apparent explanation. The cumulative sensitivity for predicting site of metastases at 1 year was 33% (3/9) and the positive predictive value 4.0% (3/75). At 10 years the sensitivity was 5.0% (3/60) and the positive predictive value 5.0% (3/65). The total cost of screening was calculated to be R323 460.00, suggesting that the cost for each patient in whom metastases were detected was R64 629.00. CONCLUSION: While scintigraphy may be of value in symptomatic or more advanced disease, screening of node-negative women had a minimal detection rate, was expensive and cannot be supported.