To explore the diagnostic value of full-field digital mammography (FFDM) combined with digital breast tomosynthesis (DBT) for nonpalpable breast lesions.

This prospective study included 126 patients with a total of 137 breast lesions which was nonpalpable in clinic but confirmed by imaging examinations in Zhenjiang Women and Children’s Hospital from June 2017 to December 2019. All patients underwent FFDM and FFDM+ DBT before operation. After enrollment, the guide wire positioning excision biopsy, puncture biopsy or vacuum-assisted resection was performed, and the specimens were sent for pathological examination. With pathological results as the gold standard, the receiver operating characteristic (ROC) curve was drawn and the area under the curve (AUC) was calculated. The McNemar test, Fisher’s exact test or χ² test was used to compare diagnostic power of the two methods. The Bowker test was used to analyze the differences in the edge characteristics between the two methods.

The detection rate of 137 untouchable breast lesions by FFDM and FFDM+ DBT was 86.9% (119/137) and 94.2% (129/137), indicating a significant difference (χ²=4.249, P=0.039) The detection rates of lesions in dense breasts was 91.0% (81/89) by FFDM+ DBT, significantly higher than79.8% (71/89) by FFDM (χ²=4.504, P=0.034); all lesions in less dense breasts were detected in both examinations. Taking pathological result as the gold standard, the diagnostic accuracy of FFDM+ DBT for all lesions was significantly higher than that of FFDM [92.0% (126/137) vs. 78.8% (108/137), χ²=9.485, P=0.002]. For the diagnosis of all nonpalpable lesions, the AUC of FFDM+ DBT was 0.976±0.010 (95%CI: 0.934-0.994), and the AUC of FFDM was 0.821±0.037 (95%CI: 0.747-0.882), indicating a significant difference (Z=4.131, P<0.001). For diagnosis of lesions in dense breasts, the sensitivity, accuracy and negative predictive value of FFDM+ DBT were significantly higher than those of FFDM (P=0.008, 0.012, 0.037); for diagnosis of lesions in less dense breasts, the two methods showed no significant difference (all P>0.050). For 112 mass-type lesions, the specificity, accuracy and positive predictive value of FFDM+ DBT were significantly higher than those of FFDM (P=0.011; χ²= 9.333, 5.829; P=0.002, 0.016); among them, for 62 benign lesions, the detection rate of lesions with clear edges was 54.8% (34/62) by FFDM and 80.6% (50/62) by FFDM+ DBT, indicating a significant difference (Bowker T=18.000, P<0.001); for 50 malignant lesions, the detection rate of lesions with clear edges was 50% (25/50)by FFDM and 72% (36/50) by FFDM+ DBT, indicating a significant difference (Bowker T=12.000, P= 0.017). For 45 suspicious calcification lesions, the specificity and accuracy of FFDM+ DBT were significantly higher than those of FFDM [83.9%(26/31) vs 61.3%(19/31), P=0.046; 88.9%(40/45) vs 71.1%(32/45), χ²=4.444, P=0.035]. For structurally tangled lesions, the detection rate was 3/9 by FFDM and 9/9 by FFDM+ DBT, indicating a significant difference (P=0.003).

The combination of DBT and FFDM is superior to FFDM in the diagnosis of nonpalpable breast lesions, especially in the dense breasts.