99Tcm-MIBI scintimammography in 300 consecutive patients: factors that may affect accuracy.

We evaluated the diagnostic yield of 99Tcm-MIBI scintimammography in a relatively large series of consecutive patients referred for breast surgery on the basis of physical examination or mammogram. 99Tcm-MIBI uptake was correlated to tumour size, receptor status, neovascularity, proliferating activity, P-170 glycoprotein expression and the patient's gonadal state. Three hundred consecutive patients referred to our institution, with either a positive mammogram or a palpable mass, were entered into the study. All patients underwent 99Tcm-MIBI scintimammography. Pathological status was obtained after surgery in all patients. Breast cancer was diagnosed in 218 (73%) patients. The MIBI scan was positive in 89% (194/218) cancer patients and in 17% (14/82) of patients with benign masses (false-positives); the scan was negative in 24 (11%) cancer patients (false-negatives). The sensitivity of MIBI scintigraphy was higher for tumours > 1 cm (95 vs 48% in lesions < or = 1 cm) and in pre-menopausal women (95 vs 85%). Conversely, the specificity was better for lesions < 1 cm (100%) and in post-menopausal women (89%). The positive predictive value of MIBI scan was good both in small (< 1 cm) and large tumours (100% and 93%, respectively) and slightly modified by gonadal state (89% and 96% in pre- and post-menopausal state). The negative predictive value was unsatisfactory, especially in small tumours and in older patients. The diagnostic performance increased stratifying data for tumour size, indicating that lesion size is a major determinant in the diagnostic accuracy of MIBI scintimammography. We conclude that 99Tcm-MIBI scintimammography is useful in the diagnostic evaluation of young patients, because it can select patients for further invasive diagnostic procedures. In older patients, a positive 99Tcm-MIBI scan is highly suggestive of malignancy and might be an indication for surgery. In the case of a negative scan, biopsy is advisable given the poor negative predictive value. Small tumour size and a well-differentiated histotype characterize false-negative cases.