Abstract
To systematically determine the diagnostic accuracy of digital breast tomosynthesis (DBT) in combination with digital mammography (DM) compared with DM alone for breast cancer. We searched multiple English and Chinese databases from January 2010 to October 2023. Eligible diagnostic studies comparing DBT/DM with DM alone for breast cancer in women were included. QUADAS-2 tool was used to assess the risk of bias. The summary sensitivity, specificity, and diagnostic odds ratio (DOR) were estimated using a bivariate model. 27 studies with 45,694 individuals and 59,493 breast lesions were included. The pooled sensitivity and specificity of DBT/DM were 87% (95% confidence interval [CI], 84-90%) and 84% (95% CI, 78-88%), respectively. These values were higher than those of DM alone, with pooled sensitivity of 76% (95% CI, 72-79%) and specificity of 80% (95% CI, 73-86%). The pooled DORs for DBT/DM and DM alone were 34.74 (95% CI, 22.18-54.39) and 12.96 (95% CI, 8.80-19.08), respectively. Additionally, adding DBT to DM improved diagnostic performance more significantly in Asians (DORs, 59.47 vs. 16.10) than in Westerners (DORs, 19.48 vs. 10.30). Meta-regression analysis revealed that country, reference standard, and published language were significant factors (P < 0.05) affecting heterogeneity for DBT/DM. This meta-analysis demonstrated that DBT/DM has higher diagnostic accuracy than DM alone for breast cancer. Diagnosing breast cancer with DBT/DM significantly increased sensitivity and slightly increased specificity compared with DM alone. Furthermore, adding DBT to DM increased diagnostic performance more significantly in Asians than in Westerners.