Summary Background Mammography is the only proven method for breast cancer screening that reduces mortality, although it is inaccurate in young women or women with dense breasts. We investigated the efficacy of adjunctive ultrasonography. Methods Between July, 2007, and March, 2011, we enrolled asymptomatic women aged 40–49 years at 42 study sites in 23 prefectures into the Japan Strategic Anti-cancer Randomized Trial (J-START). Eligible women had no history of any cancer in the previous 5 years and were expected to live for more than 5 years. Randomisation was done centrally by the Japan Clinical Research Support Unit. Participants were randomly assigned in 1:1 ratio to undergo mammography and ultrasonography (intervention group) or mammography alone (control group) twice in 2 years. The primary outcome was sensitivity, specificity, cancer detection rate, and stage distribution at the first round of screening. Analysis was by intention to treat. This study is registered, number UMIN000000757. Findings Of 72 998 women enrolled, 36 859 were assigned to the intervention group and 36 139 to the control group. Sensitivity was significantly higher in the intervention group than in the control group (91·1%, 95% CI 87·2–95·0 vs 77·0%, 70·3–83·7; p=0·0004), whereas specificity was significantly lower (87·7%, 87·3–88·0 vs 91·4%, 91·1–91·7; p<0·0001). More cancers were detected in the intervention group than in the control group (184 0·50% vs 117 0·32%, p=0·0003) and were more frequently stage 0 and I (144 71·3% vs 79 52·0%, p=0·0194). 18 (0·05%) interval cancers were detected in the intervention group compared with 35 (0·10%) in the control group (p=0·034). Interpretation Adjunctive ultrasonography increases sensitivity and detection rate of early cancers. Funding Ministry of Health, Labour and Welfare of Japan.