Truace
Subscribe

Multimodal Gene Expression Deep Learning for Predicting Sentinel Lymph Node Macro-metastasis in Early Breast Cancer: Development and Validation in the SCAN-B Cohort

Background This study evaluates DL using GEX and preoperatively available clinical data (PreopClinic) to predict SLNM, and explores their potential for guiding axillary surgery and prognostic assessment. Methods We retrospectively included 6,836 clinically node-negative T1-T2 patients with invasive breast cancer who underwent primary surgery from the SCAN-B. Three DL models-a multilayer perceptron, a pathway-informed sparse neural network, and a transformer-were developed using the development set (n=4,625) and…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1718· cite & verify
Multimodal Gene Expression Deep Learning for Predicting Sentinel Lymph Node Macro-metastasis in Early Breast Cancer: Development and Validation in the SCAN-B Cohort
"Capillaritis-Acute hemorrhage - Milk allergy Case 107" by Pulmonary Pathology is licensed under CC BY-SA 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/2.0/.

In brief

Researchers used data from 6,836 clinically node-negative T1-T2 invasive breast cancer patients in the SCAN-B cohort to build deep learning models that predict whether sentinel lymph nodes contain macro-metastasis. They trained a multilayer perceptron, a pathway-informed sparse neural network and a transformer on 4,625 patients, then compared them with XGBoost in an independent test set of 2,211.

The work matters because better preoperative estimates of nodal spread could eventually help guide axillary surgery and prognosis. The reported accuracy is still modest, and the study is retrospective, so it does not show that the tool changes surgical decisions or patient outcomes. Biopsy-based use remains a future prospect.

Main points

  1. The study retrospectively included 6,836 clinically node-negative T1-T2 invasive breast cancer patients from SCAN-B.
  2. The transformer outperformed the other methods for gene expression modeling and minimized prior gene selection.
  3. In the independent test set, adding gene expression raised ROC AUC from 0.596 to 0.693.
  4. Agreement between paired biopsy and surgical specimens supports the feasibility of future biopsy-based preoperative use.

The gain

Adding tumor gene expression to preoperative clinical data in a deep learning model improved prediction of sentinel lymph node macro-metastasis in early breast cancer over clinical data alone.

The rundown

The data were split into a development set of 4,625 patients and an independent test set of 2,211 patients. Three deep learning architectures were compared against the XGBoost baseline.

The authors say the transformer was robust against noise and captured informative transcriptomic features without extensive prior gene selection, which may simplify how such models are built.

The authors explore potential uses in guiding axillary surgery and prognostic assessment. Paired biopsy and surgical specimens showed agreement, suggesting gene expression from a preoperative biopsy might one day be used.

Sources

  1. Peer-reviewedClinical Cancer Research2026-10-08

The debate