What question did this study set out to answer?

To develop and evaluate multiregional 2D and 2.5D deep learning models based on dynamic contrast-enhanced MRI for noninvasive prediction of axillary lymph node pathological complete response following neoadjuvant chemotherapy.

September 3, 2026BMC Medical ImagingOpen Access

Multichannel deep learning using multiregional DCE-MRI to predict axillary pathological complete response after neoadjuvant chemotherapy in breast cancer

Q: What does this research mean for the field?

A 2.5D deep learning model combining multiregional DCE-MRI features with clinicopathological factors accurately predicts axillary pathological complete response after neoadjuvant chemotherapy in breast cancer, potentially reducing unnecessary axillary lymph node dissections. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

View Full Paper

Ask AI

Bookmark

Authors

JLJun Liao MXMengting XuHenan Provincial People's Hospital RLRuimin LiFudan University Shanghai Cancer Center

Discussion

Member takes

Implication

Retrospective study reveals deep learning on DCE-MRI predicts axillary complete response in breast cancer, indicating potential to avoid unnecessary lymph node surgery.

Key Points

To develop and evaluate multiregional 2D and 2.5D deep learning models based on dynamic contrast-enhanced MRI for noninvasive prediction of axillary lymph node pathological complete response following neoadjuvant chemotherapy.
Retrospectively analyzed 305 patients with invasive breast cancer and ipsilateral axillary lymph node metastasis, randomized into a training cohort (n = 214) and a validation cohort (n = 91).
Extracted DCE-MRI features from primary tumors and axillary nodes, selecting optimal features via Mann–Whitney U test, Spearman correlation, mRMR, and LASSO.
Trained and evaluated 2D and 2.5D deep learning architectures alone and in combination with independent clinicopathological predictors (ER status, HER2 status, tumor MP grade) using ROC and decision curve analyses.
Pathological complete response of axillary lymph nodes occurred in 46.6% (142/305) of patients.
The 2.5D tumor–axillary lymph node model outperformed its 2D counterpart in the validation set (AUC 0.834 vs. 0.815), while the combined 2.5D tumor–node–clinicopathologic model achieved the highest discrimination (AUC 0.861).
Model-guided simulation reduced the estimated rate of unnecessary axillary lymph node dissections from 40.7% to 8.8% and increased the estimated overall benefit rate from 59.3% to 74.7%.

Cite This Study

Liao et al. (2026) studied this question.

synapsesocial.com/papers/6a99352e636c6408cfa7d229 https://doi.org/10.1186/s12880-026-02736-y

View Full Paper

Ask AI

Bookmark

Implication

Retrospective study reveals deep learning on DCE-MRI predicts axillary complete response in breast cancer, indicating potential to avoid unnecessary lymph node surgery.

Multichannel deep learning using multiregional DCE-MRI to predict axillary pathological complete response after neoadjuvant chemotherapy in breast cancer

Discussion

Member takes

Implication

Key Points

Cite This Study

Multichannel deep learning using multiregional DCE-MRI to predict axillary pathological complete response after neoadjuvant chemotherapy in breast cancer

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider