PURPOSE: To reveal the rate of complete therapeutic effect of radiofrequency ablation (RFA) and its correlation with tumor size by the histopathological examination of surgically resected early breast cancers. METHODS: For 28 patients who received RFA and subsequent surgical therapies for early breast cancer treatment, the effect of RFA was evaluated by both histopathological examination and nicotinamide adenine dinucleotide (NADH)-diaphorase staining of resected tumor specimens according to the criteria described by Seki et al. (this issue). The correlation of 100% RFA effect with tumor parameters including tumor size and the presence of extensive intraductal component (EIC) was examined. RESULTS: The mean size and invasive size of the primary tumors were 2.21 cm (ranging from 0.6 to 5.0 cm) and 1.44 cm (ranging from 0 to 5.0 cm), respectively. By examining hematoxylin-eosin (HE) sections, the effectiveness of RFA was found to be 100% in 16 tumors (57%). However, the effectiveness of RFA was found to be 100% in 22 cases (79%) examined by NADH-diaphorase staining of frozen sections containing part of tumorous and nontumorous tissues. The accuracy of diagnosis of complete RFA effect using NADH-diaphorase staining with reference to HE was 79% (22 of 28) with 100% (16 of 16) sensitivity and 50% (6 of 12) specificity. The rate of 100% RFA effect by HE examination was higher in EIC(-) tumors (13 of 17, 76%) than in EIC(+) tumors (1 of 9, 11%) (P = 0.0022), and was higher in tumors of ≤ 1.5 cm (10 of 11, 91%) than in tumors of >1.5 cm (6 of 17, 35%; P = 0.0034). All five tumors of ≤ 1.0 cm showed 100% RFA effect, but 3 (27%) of 11 tumors of >1.0 and ≤ 2.0 cm and 9 (75%) of 12 tumors of > 2.0 cm showed suboptimal RFA effect by HE. CONCLUSIONS: Tumor size of ≤ 1.5 cm, strictly ≤ 1.0 cm, could be an indication for RFA if a complete histological therapeutic effect is mandatory.