Diagnostic Value of Combined Automated Breast Ultrasound System and Conventional Two-Dimensional Ultrasound for Intraductal Breast Lesions

Technical Summary: Diagnostic Value of Combined Automated Breast Ultrasound System and Conventional Two-Dimensional Ultrasound for Intraductal Breast Lesions

Problem Statement
Intraductal breast lesions, ranging from benign hyperplasia to ductal carcinoma in situ (DCIS) and microinvasive cancer, often present with non-specific clinical manifestations such as nipple discharge or palpable nodules. Accurate preoperative differentiation between benign and malignant lesions is critical for individualized treatment planning. While conventional two-dimensional (2D) ultrasound is the primary imaging modality due to its non-invasive nature and ability to display lesion morphology and blood flow, it is limited by its planar imaging capability. It often fails to fully visualize the three-dimensional architecture of the ductal system and the overall extent of lesion infiltration, leading to diagnostic challenges for tiny or atypical lesions. Conversely, Automated Breast Ultrasound System (ABUS) offers 3D reconstruction and coronal imaging, which can better delineate ductal structures and specific malignant signs, yet few studies have systematically combined 2D and ABUS indicators to build a comprehensive diagnostic model for these specific lesions.

Methodology
This retrospective study enrolled 200 patients with pathologically confirmed intraductal breast lesions (132 benign, 68 malignant) treated at the First Medical Center of Chinese PLA General Hospital between April 2025 and April 2026. All subjects underwent both conventional 2D ultrasound and ABUS scanning.

• Imaging Protocols: 2D ultrasound was performed using a Mindray Resona 7 system, evaluating lesion size, margin, morphology, echo, calcification, duct dilatation, and blood flow (Adler grading). ABUS scanning utilized an IBUS BE3 system to generate coronal, sagittal, and transverse reconstructions, focusing on features such as the convergence sign, nipple-oriented extension, spiculation/angulation, and peripheral hyperechoic rims.
• Analysis: Two experienced sonographers independently reviewed images under blind conditions. Statistical analysis included univariate and multivariate logistic regression to identify independent risk factors for malignancy. A combined diagnostic model was established, and its performance was evaluated using Receiver Operating Characteristic (ROC) curves, comparing the Area Under the Curve (AUC), sensitivity, and specificity against standalone 2D and ABUS examinations.

Key Contributions
The study identifies and validates seven independent ultrasonic risk factors for malignancy in intraductal lesions:

1. Irregular lesion morphology
2. Presence of microcalcification
3. Adler grade II–III blood flow (hypervascularity)
4. Duct dilatation ≥ 5 mm
5. Convergence sign (radial convergence of parenchyma toward the lesion on coronal view)
6. Lesion extension toward the nipple
7. Spiculation or angulation at the lesion margin

The primary contribution is the establishment of a combined diagnostic model integrating these 2D and ABUS features, demonstrating that the synergy of planar detail and 3D volumetric reconstruction significantly outperforms single-modality approaches.

Results

• Feature Distribution: Malignant lesions showed significantly higher incidences of irregular shape, microcalcification, abundant blood flow, severe duct dilatation (≥5 mm), convergence sign, nipple-oriented extension, and spiculation/angulation compared to benign lesions (P < 0.05).
• Diagnostic Performance:
  • Combined Model: Achieved an AUC of 0.90 (95% CI: 0.85–0.94), with a sensitivity of 90.7% and specificity of 89.2%.
  • Standalone 2D Ultrasound: AUC of 0.81 (95% CI: 0.74–0.87), sensitivity 81.5%, specificity 84.2%.
  • Standalone ABUS: AUC of 0.86 (95% CI: 0.80–0.91), sensitivity 87.0%, specificity 88.3%.
  • The combined model demonstrated statistically superior diagnostic efficiency compared to either modality used alone (P < 0.05).

Significance and Claims
The paper claims that the combination of ABUS and conventional 2D ultrasound holds prominent clinical value for differentiating benign and malignant intraductal breast lesions. By compensating for the respective limitations of each modality—2D's lack of 3D context and ABUS's potential lack of internal echo detail—the combined approach improves preoperative diagnostic accuracy. The study posits that utilizing the seven identified risk factors allows sonographers to establish a standardized imaging-pathology cognition, thereby facilitating better clinical decision-making and treatment strategies.

The authors acknowledge limitations, noting the single-center, retrospective nature of the study with a sample size of 200, which may introduce selection bias. They also note the absence of head-to-head comparisons with mammography or MRI and the potential for subjective bias in image interpretation despite blind evaluation. Consequently, the authors suggest that future multi-center prospective research with external validation cohorts is necessary to verify the diagnostic model, and they propose that radiomics and artificial intelligence could be applied in future work to develop automatic intelligent diagnostic systems.