Can An Intelligent Vacuum-Assisted Biopsy Algorithm Reliably Identify Patients With Breast Cancer Who Have a Pathologic Complete Response to Neoadjuvant Therapy?

Get Permission

In a study reported in the Journal of Clinical Oncology, Pfob et al developed a machine learning algorithm–based (intelligent) vacuum-assisted biopsy model that could identify patients with pathologic complete response (ypT0 and ypN0) to neoadjuvant therapy for breast cancer who may be able to avoid breast and axilla surgery.

Study Details

In the study, the investigators trained, tested, and validated a machine learning algorithm using patient, imaging, tumor, and vacuum-assisted biopsy variables to detect residual cancer after neoadjuvant systemic treatment (ypT1 or in situ or ypN1) before surgery. The intelligent vacuum-assisted biopsy model discovery set consisted of 318 women with cT1–3, cN0 or 1, HER2-positive, triple-negative, or high-proliferative luminal B–like breast cancer who underwent vacuum-assisted biopsy before surgery. The model was further assessed in 45 patients belonging to an external validation set.

Intelligent vacuum-assisted biopsy performance was compared with histopathologic evaluation of surgical specimens. False-negative rate (proportion of patients with missed residual disease) and specificity (identification of pathologic complete response [ypT0 and ypN0] among patients with pathologic complete response) of the intelligent vacuum-assisted biopsy were the main outcome measures.

Key Findings

In the development set, the intelligent vacuum-assisted biopsy had a false-negative rate of 5.2% (95% confidence interval [CI] = 2.4%–9.6%), specificity of 37.5% (95% CI = 29.6%–45.9%) and receiver operating characteristic area under the curve (AUC) of 0.92 (95% CI = 0.90–0.94) for detecting residual cancer (ypT1 or in situ or ypN1) after neoadjuvant treatment.

In the validation set, intelligent vacuum-assisted biopsy had a false-negative rate of 0.0% (95% CI = 0.0%–13.7%), specificity of 40.0% (95% CI = 19.1%–63.9%), and AUC of 0.91 (95% CI = 0.82–0.97) for detecting residual cancer. Overall, the predictive probability of the model was found to be well calibrated with observed probability (z score = −0.746, P = .228).

The false-negative rate of the intelligent vacuum-assisted biopsy was lower than rates observed with imaging after neoadjuvant treatment (24.4% in discovery set, 24.0% in validation set), vacuum-assisted biopsy alone (32.8% and 28.0%), and with imaging after neoadjuvant treatment plus vacuum-assisted biopsy (16.7% and 12.0%).

The investigators concluded, “An intelligent vacuum-assisted biopsy algorithm can reliably exclude residual cancer after neoadjuvant systemic treatment. The omission of breast and axillary surgery for these exceptional responders may be evaluated in future trials.”

Joerg Heil, MD, of the Department of Obstetrics & Gynecology, Heidelberg University, is the corresponding author for the Journal of Clinical Oncology article.

Disclosure: The study was supported by the German Research Foundation. For full disclosures of the study authors, visit

The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.