数字乳腺断层融合X线成像对乳腺病变的诊断价值

Value of digital breast tomosynthesis （DBT） in the diagnosis of breast lesions

目的评价数字乳腺断层融合X线成像（DBT）对乳腺病变的诊断价值。方法选择经病理证实的253个乳腺病灶，术前分别采用全视野数字化乳腺摄影（FFDM）、DBT、DBT联合FFDM进行独立诊断，以病理诊断作为金标准．评价DBT对乳腺病变的诊断价值．比较DBT和FFDM在不同年龄、不同乳腺腺体密度和不同激素水平亚组中对乳腺病变的诊断效能，比较DBT和FFDM在不同病理类型乳腺癌和不同大小乳腺癌中的敏感度。比较DBT和FFDM显示乳腺癌病灶的最长径与病理测量值的相关性。结果DBT、FFDM和两者联合诊断乳腺病变的受试者工作特征（ROC）曲线下面积分别为0．890、0．833和0．890。DBT、DBT联合FFDM诊断乳腺病变的ROC曲线下面积均大于FFDM（均P〈0．05）。在乳腺密度〉50％、年龄≤50岁和未绝经组中，DBT、DBT联合FFDM诊断乳腺病变的ROC曲线下面积大于FFDM（P〈0．05）；在乳腺密度450％、年龄〉50岁和已绝经组中，DBT、DBT联合FFDM诊断乳腺病变的差异无统计学意义（P〉0．05）。DBT和FFDM在原位癌中的敏感度均为90．9％（10／11），DBT和FFDM在非原位癌中的敏感度分别为92．3％（120／130）和83．8％（109／130）。在病灶≤10mm、10mm〈病灶≤20mm、20mm〈病灶≤30mm、30mm〈病灶440mm组中，DBT和FFDM诊断乳腺癌的敏感度分别为51．7％（4／7）和51．7％（4／7）、93.8％（61／65）和78．5％（51／65）、96．7％（30／31）和93．5％（29／31）、100％（11／11）和100％（11／11）。DBT和FFDM显示乳腺癌病灶最长径与病理测量值的相关系数分别为0．905和0．849（P〈0．001）。结论与FFDM相比，在乳腺密度〉50％、年龄450岁、未绝经、非原位癌和10mm〈癌灶长径≤20mm的患者中，DBT诊断乳腺癌更具优势，DBT显示乳腺癌病灶最长径更准确。

Objective To assess the value of digital breast tomosynthesis （DBT） in the diagnosis of breast lesions. Methods Two hundred and fifty-three breast lesions in 250 patients were selected in this study. All lesions were confirmed pathologically. Preoperative diagnosis was performed independently with full-field digital mammography （FFDM） , DBT and DBT plus FFDM, respectively. The diagnostic value of DBT for breast lesions was evaluated based on the pathological diagnosis as the gold standard. The diagnostic performance of DBT and FFDM for breast lesions was compared between the groups with different ages, mammary gland densities and hormone levels. The sensitivity of DBT and FFDM was compared between the groups with different pathological types and different sizes of breast cancer. The correlation between the longest diameter of breast cancers and pathological measurements shown on DBT and FFDM was analyzed. Results The areas under ROC curves were 0.890, 0.833 and 0.890 for DBT, FFDM and DBT plus FFDM, respectively. The areas under ROC curves for DBT or DBT plus FFDM were significantly greater than that for FFDM （P〈0.05）. In the group with breast density〉50%, group with age ≤ 50 and non-menopause group, all the areas under ROC curves for DBT or DBT plus FFDM were all significantly larger than that for FFDM （P〈0.05）. No significant differences were observed in the group with breast density ≤ 50%, group with age 〉 50 and menopause group （P〉0.05）. The sensitivity for both DBT and FFDM in the diagnosis of carcinomain situ was 90.9% （ 10/11 ）. The sensitivity for DBT and FFDM in the diagnosis of non-carcinoma in situ was 92.3% （120/130） and 83.8% （109/130）, respectively. The sensitivity in the groups with the longest diameter of foci 〉0 mm but ≤10 mm, 〉10 mm but ≤20 mm, 〉20 mm but ≤30 mm, and 〉30 mm but ≤ 40 mm were 51.7% （4/7） , 93.8% （61/65）, 96.7% （30/31） and 100% （ 11/11 ） , respectively, for DBT, and were 51.7% （4/7）, 78.5% （51/65）, 93.5% （29/31）, and 100% （11/11） , respectively, for FFDM. The correlation coefficients between the longest diameter of breast cancers and pathological measurements shown on DBT and FFDM were 0.905 and 0. 849, respectively （P〈 0.001 ）. Conclusions Compared with FFDM, DBT shows a higher diagnostic efficiency in patients with breast density 〉50%, age ≤50 years and non-menopause, non-carcinoma in situ, and the longest diameter of lesions 〉 10 mm but ≤ 20 mm. The longest diameter of breast lesions is more accurately shown on DBT.