Practical Challenges of DBT-Guided VABB: Harms and Benefits, from Literature to Clinical Experience

Comparison of two visual localization biopsy methods for suspicious breast lesions under mammography: a multicenter cohort study

Background

Compared with traditional mammography, digital breast tomosynthesis (DBT) increases the detection of breast lesions by reducing the overlapping masking effect of glands through multi-angle exposures of a rotating X-ray tube. Breast biopsy under the guidance of mammography can effectively diagnose breast lesions. The main methods of mammography-guided biopsy include wire positioning, core needle biopsy (CNB), and vacuum-assisted breast biopsy (VABB). Currently, it is mainly guided by stereotactic or DBT-guided localization. In this study, we conducted a retrospective study to compare the clinical efficiency and performance of DBT-guided and prone stereotactic (PS)-guided breast lesion biopsies.

Methods

We performed a retrospective analysis of 406 patients who underwent mammogram-guided biopsies from three hospitals between August 2020 and August 2024. The cohort comprised 234 cases of DBT-guided biopsies and 172 cases of PS-guided biopsies. The DBT-guided biopsy methods included wire positioning, CNB, and VABB. The PS-guided biopsy methods included wire positioning and CNB. Statistical analyses utilized the Chi-squared or Mann-Whitney U tests to compare the effectiveness of biopsy methods acquired under DBT guidance and PS guidance.

Results

In wire positioning and CNB, both the DBT-guided group and PS-guided group had higher biopsy success rates (100% vs. 92.6%, 100% vs. 96.2%), with no statistically significant difference (P=0.154 and P=0.127, respectively). Compared to the PS-guided group, the DBT-guided group had shorter total intervention (P<0.001) and lesion targeting time (P<0.001), less time to obtain the first effective positioning images (P<0.001), and fewer exposure times (P<0.001). The incidence of complications was lower in both DBT-guided and PS-guided groups, with no statistically significant difference (P=0.851 and 0.861, respectively). In addition, we successfully performed 69 cases of DBT-guided VABB. Compared with wire positioning and CNB under DBT guidance, this method also had shorter total intervention (19.49±4.75 minutes) and lesion targeting time (5.00, 7.00 minutes), and reduced exposure time (4.00, 5.00 exposures).

Conclusions

DBT-guided biopsies outperformed PS-guided biopsies in clinical efficiency, achieving quicker procedures and requiring fewer exposures while enabling the biopsy of a broader range of non-calcified breast lesions.

Deep-AutoMO: Deep automated multiobjective neural network for trustworthy lesion malignancy diagnosis in the early stage via digital breast tomosynthesis

Breast cancer is the most prevalent cancer in women, and early diagnosis of malignant lesions is crucial for developing treatment plans. Digital breast tomosynthesis (DBT) has emerged as a valuable tool for early breast cancer detection, as it can identify more lesions and improve the early detection rate. Deep learning has shown great potential in medical image-based cancer diagnosis, including DBT. However, deploying these models in clinical practice may be challenging due to concerns about reliability and robustness. In this study, we developed a novel deep automated multiobjective neural network (Deep-AutoMO) to build a trustworthy model and achieve balance, safety and robustness in a unified way. During the training stage, we introduced a multiobjective immune neural architecture search (MINAS) that simultaneously considers sensitivity and specificity as objective functions, aiming to strike a balance between the two. Each neural network in Deep-AutoMO comprises a combination of a ResNet block, a DenseNet block and a pooling layer. We employ Bayesian optimization to optimize the hyperparameters in the MINAS, enhancing the efficiency of the model training process. In the testing stage, evidential reasoning based on entropy (ERE) approach is proposed to build a safe and robust model. The experimental study on DBT images demonstrated that Deep-AutoMO achieves promising performance with a well-balanced trade-off between sensitivity and specificity, outperforming currently available methods. Moreover, the model's safety is ensured through uncertainty estimation, and its robustness is improved, making it a trustworthy tool for breast cancer diagnosis in clinical settings. We have shared the code on GitHub for other researchers to use. The code can be found at https://github.com/ChaoyangZhang-XJTU/Deep-AutoMO.

Breast Cancer: A Multi-Disciplinary Approach from Imaging to Therapy

Breast cancer (BC) is the most prevalent form of cancer among women worldwide, accounting for over 2 million diagnoses annually [...].