Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Guiding Treatment Strategies: The Role of Adjuvant Anti-Her2 Neu Therapy and Skin/Nipple Involvement in Local Recurrence-Free Survival in Breast Cancer Patients

Joe Omatoi, Abdul M Mohammed, Dennis Trujillo

TL;DR

This paper addresses the limited generalizability of randomized trials by applying causal inference to observational breast cancer data. Using DirectLiNGAM/LiNGAM on the Duke MRI Breast Cancer dataset (n=922) with a rigorous preprocessing and feature-selection pipeline, the study identifies causal factors for local recurrence-free survival. It finds that Adjuvant Anti-Her2 Neu Therapy increases days to last recurrence-free assessment by approximately 169 days, while Skin/Nipple involvement decreases it by about 351 days, with robustness supported by backdoor and refutation analyses. The work demonstrates how causal techniques can extract actionable insights from observational data to inform personalized treatment strategies and highlights avenues for applying these methods to additional datasets.

Abstract

This study explores how causal inference models, specifically the Linear Non-Gaussian Acyclic Model (LiNGAM), can extract causal relationships between demographic factors, treatments, conditions, and outcomes from observational patient data, enabling insights beyond correlation. Unlike traditional randomized controlled trials (RCTs), which establish causal relationships within narrowly defined populations, our method leverages broader observational data, improving generalizability. Using over 40 features in the Duke MRI Breast Cancer dataset, we found that Adjuvant Anti-Her2 Neu Therapy increased local recurrence-free survival by 169 days, while Skin/Nipple involvement reduced it by 351 days. These findings highlight the therapy's importance for Her2-positive patients and the need for targeted interventions for high-risk cases, informing personalized treatment strategies.

Guiding Treatment Strategies: The Role of Adjuvant Anti-Her2 Neu Therapy and Skin/Nipple Involvement in Local Recurrence-Free Survival in Breast Cancer Patients

TL;DR

This paper addresses the limited generalizability of randomized trials by applying causal inference to observational breast cancer data. Using DirectLiNGAM/LiNGAM on the Duke MRI Breast Cancer dataset (n=922) with a rigorous preprocessing and feature-selection pipeline, the study identifies causal factors for local recurrence-free survival. It finds that Adjuvant Anti-Her2 Neu Therapy increases days to last recurrence-free assessment by approximately 169 days, while Skin/Nipple involvement decreases it by about 351 days, with robustness supported by backdoor and refutation analyses. The work demonstrates how causal techniques can extract actionable insights from observational data to inform personalized treatment strategies and highlights avenues for applying these methods to additional datasets.

Abstract

This study explores how causal inference models, specifically the Linear Non-Gaussian Acyclic Model (LiNGAM), can extract causal relationships between demographic factors, treatments, conditions, and outcomes from observational patient data, enabling insights beyond correlation. Unlike traditional randomized controlled trials (RCTs), which establish causal relationships within narrowly defined populations, our method leverages broader observational data, improving generalizability. Using over 40 features in the Duke MRI Breast Cancer dataset, we found that Adjuvant Anti-Her2 Neu Therapy increased local recurrence-free survival by 169 days, while Skin/Nipple involvement reduced it by 351 days. These findings highlight the therapy's importance for Her2-positive patients and the need for targeted interventions for high-risk cases, informing personalized treatment strategies.
Paper Structure (9 sections, 4 equations, 6 figures, 2 tables)

This paper contains 9 sections, 4 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Dataset and Preprocessing
  4. Feature Selection
  5. Causal Inference
  6. Evaluation
  7. Results
  8. Discussion
  9. Conclusion

Figures (6)

  • Figure 1: Feature importance plot showing the absolute coefficients of various clinical and treatment variables in predicting breast cancer outcomes. The variables are ranked by their coefficient magnitude, with 'Adjuvant Radiation Therapy_No' having the strongest predictive power and 'Multicentric/Multifocal_Yes' having the least impact.
  • Figure 2: Illustration of three possible causal structures under acyclicity and absence of hidden common causes assumptions
  • Figure 3: Correlation matrix showing the relationships between different clinical variables in a medical study. The values range from -1 to 1, with blue colors indicating positive correlations and darker colors indicating negative or weaker correlations. The diagonal shows perfect correlation (1.0) as variables correlate perfectly with themselves.
  • Figure 4: P-value matrix displaying the statistical significance of relationships between medical variables. The values range from 0 to 1, with lighter colors indicating more statistically significant relationships (lower p-values). Values closer to 0.05 or less typically indicate statistical significance.
  • Figure 5: Backdoor criterion example
  • ...and 1 more figures