Development and validation of a point-based scoring system for predicting axillary lymph node metastasis and disease outcome in breast cancer using clinicopathological and multiparametric MRI features

Xiaofeng Chen, Zhiqi Yang, Ruibin Huang, Yue Li, Yuting Liao, Guijin Li, Mengzhu Wang, Xiangguang Chen, Zhuozhi Dai, Weixiong FanCancer 2023 · DOI

TL;DR

Researchers have developed a non-invasive, point-based scoring system (PSS) that predicts whether breast cancer has spread to the axillary lymph nodes (ALN) using a combination of clinical data and quantitative MRI features. Beyond just predicting metastasis, a high score on this system (>19.5) serves as a significant red flag for poorer long-term recurrence-free and overall survival.

What Problem Does This Paper Address?

In breast cancer treatment, knowing if the cancer has spread to the armpit (axillary) lymph nodes is critical for staging and deciding on surgery or chemotherapy. Currently, the gold standard for this assessment is invasive: surgeons perform a Sentinel Lymph Node Biopsy (SLNB) or Axillary Lymph Node Dissection (ALND). While effective, these procedures carry risks like lymphedema, nerve damage, and infection.

Non-invasive alternatives like ultrasound or PET/CT often suffer from high false-negative rates. While complex mathematical models (nomograms) exist to predict node status, they are often too cumbersome for daily clinical use. This paper seeks to bridge that gap by creating a simple, reliable "point system" that clinicians can use to assess risk and prognosis without immediate surgery.

Methods

The researchers conducted a retrospective study of 219 breast cancer patients. They combined two types of data:

  1. Clinicopathological features: Standard metrics like AJCC stage, T stage (tumor size), and receptor status (HER-2, ER).
  2. Multiparametric MRI (mpMRI): They used Dynamic Contrast-Enhanced (DCE) MRI and Diffusion-Weighted Imaging (DWI) to extract quantitative parameters. These included $K_{ep}$ (reflux rate constant), $V_e$ (extracellular volume fraction), and TTP (time to peak enhancement).

Using multivariate logistic regression, they identified the most significant risk factors. To make the model "clinician-friendly," they converted the statistical coefficients into a point-based system (rounding to the nearest 0.5). Patients were then stratified into low-risk and high-risk groups based on their total score.

Datasets

  • Source: 219 female patients from Meizhou People’s Hospital (January 2016 to May 2019).
  • Cohorts: The data was split into a Primary Cohort ($n=153$) to build the model and a Validation Cohort ($n=66$) to test it.
  • Ground Truth: Histopathological results from surgical specimens were used to confirm the actual presence of metastasis.

Key Insights and Ideas

  • The Power of "Leakiness": The study found that quantitative MRI parameters—specifically $K_{ep}$ and TTP—were significant predictors. These values reflect the tumor's microvascular permeability (how "leaky" the blood vessels are), which is a known biological driver of metastasis.
  • Clinical Weighting: Interestingly, clinical factors like AJCC stage and Estrogen Receptor (ER) status were assigned much higher point values (up to 17.5 points) than individual MRI parameters (0.5 to 1.0 points). This suggests that while imaging adds precision, the fundamental pathology remains the strongest predictor.
  • Prognostic "Double-Duty": The PSS isn't just a diagnostic tool for the present; it's a crystal ball for the future. A score above 19.5 was a robust predictor of shorter recurrence-free survival (RFS) and overall survival (OS), effectively identifying patients who may need more aggressive systemic therapy.

Key Results

  • Accuracy: The PSS achieved an Area Under the Curve (AUC) of 0.799 in the primary cohort and 0.713 in the validation cohort, indicating good discriminative power.
  • Risk Stratification: The risk of metastasis scaled dramatically with the score. For example, a score of 15 carried only a 4.7% risk, while any score of 31 or higher represented a 100% risk in this dataset.
  • Survival Outcomes: In the primary cohort, the high-risk group (>19.5 points) had significantly worse outcomes:
  • RFS: $P < 0.0001$
    • OS: $P = 0.042$
  • These survival trends were successfully confirmed in the validation cohort ($P = 0.032$ for RFS and $P = 0.029$ for OS).

Limitations and Future Work

  • Single-Center Bias: The study was conducted at a single institution. The authors acknowledge that multi-center studies are required to ensure the scoring system works across different populations and MRI equipment.
  • Retrospective Design: As a retrospective study, there is an inherent risk of selection bias.
  • Follow-up Duration: Some patients were followed for less than five years, which may limit the strength of the long-term survival data.
  • MRI-Pathology Mapping: A common challenge in this field is the difficulty of perfectly matching a specific lymph node seen on an MRI scan with the exact node removed during surgery. Future work using targeted markers could improve this alignment.

Summary updated

Abstract

PSS could predict the ALN metastasis risk of BC. A PSS greater than 19.5 was demonstrated to be a predictor of short RFS and OS.
BibTeX
@inproceedings{chen2023development,
  author = {Xiaofeng Chen and Zhiqi Yang and Ruibin Huang and Yue Li and Yuting Liao and Guijin Li and Mengzhu Wang and Xiangguang Chen and Zhuozhi Dai and Weixiong Fan},
  title = {{Development and validation of a point-based scoring system for predicting axillary lymph node metastasis and disease outcome in breast cancer using clinicopathological and multiparametric MRI features}},
  booktitle = {Cancer},
  year = {2023},
  doi = {10.1186/s40644-023-00564-9},
}
Figure