Search

Recently, convolutional neural network (CNN) models have been proposed to automate the assessment of breast density, breast cancer detection or risk stratification using single image modality. However, analysis of breast density using multiple mammographic types using clinical data has not been reported in the literature. In this study, we investigate pre-trained EfficientNetB0 deep learning (DL) models for automated assessment of breast density using multiple mammographic types with and without clinical information to improve reliability and versatility of reporting. 120,000 for-processing and for-presentation full-field digital mammograms (FFDM), digital breast tomosynthesis (DBT), and synthesized 2D images from 5032 women were retrospectively analyzed. Each participant underwent up to 3 screening examinations and completed a questionnaire at each screening encounter. Pre-trained EfficientNetB0 DL models with or without clinical history were optimized. The DL models were evaluated using BI-RADS (fatty, scattered fibroglandular densities, heterogeneously dense, or extremely dense) versus binary (non-dense or dense) density classification. Pre-trained EfficientNetB0 model performances were compared using inter-observer and commercial software (Volpara) variabilities. Results show that the average Fleiss’ Kappa score between-observers ranged from 0.31–0.50 and 0.55–0.69 for the BI-RADS and binary classifications, respectively, showing higher uncertainty among experts. Volpara-observer agreement was 0.33 and 0.54 for BI-RADS and binary classifications, respectively, showing fair to moderate agreement. However, our proposed pre-trained EfficientNetB0 DL models-observer agreement was 0.61–0.66 and 0.70–0.75 for BI-RADS and binary classifications, respectively, showing moderate to substantial agreement. Overall results show that the best breast density estimation was achieved using for-presentation FFDM and DBT images without added clinical information. Pre-trained EfficientNetB0 model can automatically assess breast density from any images modality type, with the best results obtained from for-presentation FFDM and DBT, which are the most common image archived in clinical practice.

Background Targeted axillary dissection (TAD) involves locating and removing both clipped nodes and sentinel nodes for assessment of the axillary response to neoadjuvant chemotherapy (NAC) by clinically node-positive breast cancer patients. Initial reports described radioactive seeds used for localization, which makes the technique difficult to implement in some settings. This trial was performed to determine whether magnetic seeds can be used to locate clipped axillary lymph nodes for removal. Methods This prospective registry trial enrolled patients who had biopsy-proven node-positive disease with a clip placed in the node and treatment with NAC. A magnetic seed was placed under ultrasound guidance in the clipped node after NAC. All the patients underwent TAD. Results Magnetic seeds were placed in 50 patients by 17 breast radiologists. All the patients had successful seed placement at the first attempt (mean time for localization was 6.1 min; range 1-30 min). The final position of the magnetic seed was within the node (n = 44, 88%), in the cortex (n = 3, 6%), less than 3 mm from the node (n = 2, 4%), or by the clip when the node could not be adequately visualized (n = 1, 2%). The magnetic seed was retrieved at surgery from all the patients. In 49 (98%) of the 50 cases, the clip and magnetic seed were retrieved from the same node. Surgeons rated the transcutaneous and intraoperative localization as easy for 43 (86%) of the 50 cases. No device-related adverse events occurred. Conclusions Localization and selective removal of clipped nodes can be accomplished safely and effectively using magnetic seeds.

Neoadjuvant systemic therapy (NST) for triple-negative breast cancer and HER2-positive breast cancer yields a pathological complete response in approximately 60% of patients. A pathological complete response to NST predicts an excellent prognosis and can be accurately determined by percutaneous image-guided vacuum-assisted core biopsy (VACB). We evaluated radiotherapy alone, without breast surgery, in patients with early-stage triple-negative breast cancer or HER2-positive breast cancer treated with NST who had an image-guided VACB-determined pathological complete response.This multicentre, single-arm, phase 2 trial was done in seven centres in the USA. Women aged 40 years or older who were not pregnant with unicentric cT1-2N0-1M0 triple-negative breast cancer or HER2-positive breast cancer and a residual breast lesion less than 2 cm on imaging after clinically standard NST were eligible for inclusion. Patients had one biopsy (minimum of 12 cores) obtained by 9G image-guided VACB of the tumour bed. If no invasive or in-situ disease was identified, breast surgery was omitted, and patients underwent standard whole-breast radiotherapy (40 Gy in 15 fractions or 50 Gy in 25 fractions) plus a boost (14 Gy in seven fractions). The primary outcome was the biopsy-confirmed ipsilateral breast tumour recurrence rate determined using the Kaplan-Meier method assessed in the per-protocol population. Safety was assessed in all patients who received VACB. This study has completed accrual and is registered with ClinicalTrials.gov, NCT02945579.Between March 6, 2017, and Nov 9, 2021, 58 patients consented to participate; however, four (7%) did not meet final inclusion criteria and four (7%) withdrew consent. 50 patients were enrolled and underwent VACB following NST. The median age of the enrolled patients was 62 years (IQR 55-77); 21 (42%) patients had triple-negative breast cancer and 29 (58%) had HER2-positive breast cancer. VACB identified a pathological complete response in 31 patients (62% [95% CI 47·2-75·4). At a median follow-up of 26·4 months (IQR 15·2-39·6), no ipsilateral breast tumour recurrences occurred in these 31 patients. No serious biopsy-related adverse events or treatment-related deaths occurred.Eliminating breast surgery in highly selected patients with an image-guided VACB-determined pathological complete response following NST is feasible with promising early results; however, additional prospective clinical trials evaluating this approach are needed.US National Cancer Institute (National Institutes of Health).

BACKGROUND Heterogeneity exists in the response of triple-negative breast cancer (TNBC) to standard anthracycline (AC)/taxane-based neoadjuvant systemic therapy (NAST), with 40% to 50% of patients having a pathologic complete response (pCR) to therapy. Early assessment of the imaging response during NAST may identify a subset of TNBCs that are likely to have a pCR upon completion of treatment. The authors aimed to evaluate the performance of early ultrasound (US) after 2 cycles of neoadjuvant NAST in identifying excellent responders to NAST among patients with TNBC. METHODS Two hundred fifteen patients with TNBC were enrolled in the ongoing ARTEMIS (A Robust TNBC Evaluation Framework to Improve Survival) clinical trial. The patients were divided into a discovery cohort (n = 107) and a validation cohort (n = 108). A receiver operating characteristic analysis with 95% confidence intervals (CIs) and a multivariate logistic regression analysis were performed to model the probability of a pCR on the basis of the tumor volume reduction (TVR) percentage by US from the baseline to after 2 cycles of AC. RESULTS Overall, 39.3% of the patients (42 of 107) achieved a pCR. A positive predictive value (PPV) analysis identified a cutoff point of 80% TVR after 2 cycles; the pCR rate was 77% (17 of 22) in patients with a TVR ≥ 80%, and the area under the curve (AUC) was 0.84 (95% CI, 0.77-0.92; P < .0001). In the validation cohort, the pCR rate was 44%. The PPV for pCR with a TVR ≥ 80% after 2 cycles was 76% (95% CI, 55%-91%), and the AUC was 0.79 (95% CI, 0.70-0.87; P < .0001). CONCLUSIONS The TVR percentage by US evaluation after 2 cycles of NAST may be a cost-effective early imaging biomarker for a pCR to AC/taxane-based NAST.