Tc-99m MIBI scintimammography (SMM) is known to be a useful diagnostic tool for primary breast cancer. We conducted this study to compare the quantitative and visual analysis for detection of primary breast cancer and also investigated the incremental role of quantitative index of SMM. METHODS 520 highly suspected breast cancer patients (malignant 370: palpable 232, nonpalpable 138; benign 150: palpable 67, nonpalpable 83) were included in this study. The SMM was performed 10 min after the injection of 750 MBq of Tc-99m MIBI. For visual analysis, three-scoring based method was used. The lesions to non-lesion ratios (L/Ns) were calculated. Receiver operating characteristic curve (ROC) analyses were performed to determine the optimal visual grade, to calculate cut-off value of L/N and to investigate whether the L/N could provide incremental value additive to visual analysis. The patients were divided into four groups according to the tumor size (group A: size </= 1cm; group B: 1cm < size </= 3 cm; group C: 3 cm < size </= 5 cm; group D: size > 5 cm). Also, the patients were divided into two groups according to age (under and over 45 years) to investigate incremental value of quantitative analysis. RESULTS When visual grade 3 was used as the cut-off grade, the sensitivity and specificity were 82.7 and 78%, respectively. The L/N of malignant breast disease was significantly higher than that of benign (2.00 +/- 1.88 versus 0.60 +/- 0.7, P < 0.01). When L/N of 1.27 was used as the cut-off value, the sensitivity and specificity of SMM were 77.6 and 83.3%, respectively. When the L/N was added to visual grade, the area under curve (AUC) of visual + quantitative (V + Q) analyses was higher than that of visual (V) and quantitative (Q) analyses (AUC 0.874 versus 0.803, P < 0.01). In group A and B, the AUCs of V + Q (0.861, 0.895) were higher than those of V (0.808, 0.781) and Q (0.808, 0.813). In group C, the AUC of V + Q (0.847) was higher than that of Q (0.803, P = 0.041). However, the AUC of V (0.915) was higher than that of V + Q (P = 0.009). In group D, there were no statistical differences between V + Q (0.685) and V (0.570, P = 0.058) and Q (0.620, P = 0.145). The V + Q revealed incremental value in the detection of primary breast cancer in both age groups. Also, in older age patients group (over 45 years), the specificities of Q and V + Q were higher than younger group (under 45 years). For axillary lymph node involvement, the sensitivity, specificity, and accuracy of SMM were 66.9, 70.1, and 68%, respectively. CONCLUSIONS From this study, the optimal visual grade for diagnosis of breast cancer was grade 3 and cut-off value of L/N was 1.27. Also, we found that the L/N provides incremental value additive to visual analysis. Especially, when the tumor is small (size </=3 cm), the L/N should be obtained for the diagnosis of breast cancer. The V + Q revealed incremental value in the detection of primary breast cancer in both age groups. Also, in older age patients group, the specificities of Q and V + Q were higher than younger group.