Predicting Axillary Lymph Node Metastasis in Early Breast Cancer Using Deep Learning on Primary Tumor Biopsy Slides

Abstract

Objectives: To develop and validate a deep learning (DL)-based primary tumor biopsy signature for predicting axillary lymph node (ALN) metastasis preoperatively in early breast cancer (EBC) patients with clinically negative ALN.

Methods: A total of 1,058 EBC patients with pathologically confirmed ALN status were enrolled from May 2010 to August 2020. A DL core-needle biopsy (DL-CNB) model was built on the attention-based multiple instance-learning (AMIL) framework to predict ALN status utilizing the DL features, which were extracted from the cancer areas of digitized whole-slide images (WSIs) of breast CNB specimens annotated by two pathologists. Accuracy, sensitivity, specificity, receiver operating characteristic (ROC) curves, and areas under the ROC curve (AUCs) were analyzed to evaluate our model.

Results: The best-performing DL-CNB model with VGG16_BN as the feature extractor achieved an AUC of 0.816 (95% confidence interval (CI): 0.758, 0.865) in predicting positive ALN metastasis in the independent test cohort. Furthermore, our model incorporating the clinical data, which was called DL-CNB+C, yielded the best accuracy of 0.831 (95%CI: 0.775, 0.878), especially for patients younger than 50 years (AUC: 0.918, 95%CI: 0.825, 0.971). The interpretation of DL-CNB model showed that the top signatures most predictive of ALN metastasis were characterized by the nucleus features including density (p = 0.015), circumference (p = 0.009), circularity (p = 0.010), and orientation (p = 0.012).

Conclusion: Our study provides a novel DL-based biomarker on primary tumor CNB slides to predict the metastatic status of ALN preoperatively for patients with EBC.

Keywords: axillary lymph node metastasis; breast cancer; core-needle biopsy; deep learning; whole-slide images.

Figure 1

Patient recruitment workflow.

Figure 2

The overall pipeline of the deep learning core-needle biopsy incorporating the clinical data (DL-CNB+C) model to predict axillary lymph node (ALN) status between N0 and N(+). (A) Multiple training bags were built based on clinical data and the cropped patches from the selected tumor regions of each core-needle biopsy (CNB) whole-slide image (WSI). (B) DL-CNB+C model training process included two phases of feature extraction and multiple-instance learning (MIL), and finally the weighted features fused with clinical features were used to predict classification probabilities and calculate the cross-entropy loss. (C) The predicted probabilities of each bag from a raw CNB WSI were merged to guide the final ALN status classification between N0 and N(+).

Figure 3

Overview on interpretability methods of deep learning core-needle biopsy (DL-CNB) model based on nucleus morphometric features. (A) The selected tumor regions of each whole-slide image (WSI) was cropped into patches. (B) For each patch, we processed nucleus segmentation (a weakly supervised segmentation framework was applied to obtain the nucleus), defined multiple nucleus morphometric features (such as major axis, minor axis, area, orientation, circumference, density, circularity, and rectangularity, which are denoted as f ₁, f ₂, f ₃, …, f _n), and extracted n feature parameters correspondingly. (C) All n kinds of feature parameters from a WSI were quantized into n distribution histograms and saved to n feature matrices (m ₁, m ₂, m ₃, …, m _n). (D) The matrices from a WSI were considered as instances of a bag and served as the input of DL-CNB model; the re-trained DL-CNB model could generate scores of features (instances) in the bag, which represented the weight of each feature in pathological diagnosis.

Figure 4

Comparison of receiver operating characteristic (ROC) curves between different models for predicting disease-free axilla (N0) and heavy metastatic burden of axillary disease (N(+)). Numbers in parentheses are areas under the receiver operating characteristic curve (AUCs).

Figure 5

The confusion matrix of predicting axillary lymph node (ALN) status between disease-free axilla (N0), low metastatic burden of axillary disease (N+(1 − 2)), and heavy metastatic burden of axillary disease (N+(≥3)).

Figure 6

The interpretability of the deep learning core-needle biopsy (DL-CNB) model of two patients. (A, B) The heat maps and nuclear segmentation from core-needle biopsy (CNB) whole-slide images (WSIs) of the N0 and the N(+) separately, and the red regions show greater contribution to the final classification. (C) The statistical analysis of three nuclear characteristics most relevant to diagnosis of all patients.