1 of 5

Survival Impact of Hormone Receptor Status in Breast Cancer Patients

Julia Durocher | julia.durocher@student.chamainde.edu

Chaminade University of Honolulu

​

​

Introduction:�Breast cancer is one of the most common cancers worldwide, and its growth can be influenced by hormones like estrogen and progesterone. Some breast tumors contain special hormone receptors — estrogen receptors (ER) and progesterone receptors (PR) — which allow them to respond to hormone-based therapy (Chen et al., 2022). This study explores whether patients who are positive for both ER and PR tend to live longer. Additionally, we apply machine learning techniques to build a classification model that predicts long-term survival based on clinical features.

Hypothesis:�Patients with breast cancer that tests positive for both ER (estrogen receptor) and PR (progesterone receptor) tend to survive longer than those whose cancer tests negative for either receptor.

Discussion/ Conclusion:

The results show that patients who are positive for both ER and PR tend to live longer than those who are not. This trend is consistent across all the visualizations, and the difference is statistically significant based on the Mann-Whitney U test. The Kaplan-Meier survival curves especially highlight this, clearly showing that ER+/PR+ patients have a higher probability of survival over time.

Acknowledgements�Project Lead: Connor Flynn, Dr. Rylan Chong, Dr. Chrystie Naeole, Kahoalii Keahi, Dr. Kelly Gaither�Mentors: Lela Gi, Clara Slate-Liu

Methods

  • Used a public breast cancer dataset
  • Created visualizations: barplots and Kaplan-Meier survival curves
  • Compared survival between ER+/PR+ and non-ER+/PR+ patients
  • Performed Mann-Whitney U test to confirm statistical significance
  • Used a Random Forest Classifier to build a machine learning model predicting 5-year survival
  • Evaluated model performance using accuracy, precision, and feature importance

Next Steps:

  • Use AutoML to build predictive models based on full patient profiles
  • Explore how treatment types influence survival outcomes
  • Analyze survival differences across demographic factors like race and age
  • Visualize correlations between hormone status and other clinical variables (e.g., tumor size)

References:

Namdari, R. (2022). Breast Cancer [Data set]. Kaggle. https://www.kaggle.com/datasets/reihanenamdari/breast-cancer�Chen, M., Wu, J., Liu, D., Chen, W., Lin, C., Andriani, L., Ding, S., Huang, O., He, J., Chen, X., Chen, W., Li, Y., Shen, K., & Zhu, L. (2022). Combined estrogen receptor and progesterone receptor level can predict survival outcome in human epidermal growth factor receptor 2‑positive early breast cancer. Clinical Breast Cancer, 22(2), e147–e156. https://doi.org/10.1016/j.clbc.2021.05.012

Calculated P-value:�Mann-Whitney U test, p = 0.0001 (statistically significant)

Figure 1.

Figure 2.

Figure 3.

2 of 5

Figure 1.

3 of 5

Figure 2.

4 of 5

Figure 3.

5 of 5

Survival Impact of Hormone Receptor Status in Breast Cancer Patients

Julia Durocher | julia.durocher@student.chamainde.edu

Chaminade University of Honolulu

​

​

Introduction:�Breast cancer is one of the most common cancers worldwide, and its growth can be influenced by hormones like estrogen and progesterone. Some breast tumors contain special hormone receptors — estrogen receptors (ER) and progesterone receptors (PR) — which allow them to respond to hormone-based therapy (Chen et al., 2022). This study explores whether patients who are positive for both ER and PR tend to live longer. Additionally, we apply machine learning techniques to build a classification model that predicts long-term survival based on clinical features.

Hypothesis:�Patients with breast cancer that tests positive for both ER (estrogen receptor) and PR (progesterone receptor) tend to survive longer than those whose cancer tests negative for either receptor.

Discussion/ Conclusion:

The results show that patients who are positive for both ER and PR tend to live longer than those who are not. This trend is consistent across all the visualizations, and the difference is statistically significant based on the Mann-Whitney U test. The Kaplan-Meier survival curves especially highlight this, clearly showing that ER+/PR+ patients have a higher probability of survival over time.

Acknowledgements�Project Lead: Connor Flynn, Dr. Rylan Chong, Dr. Chrystie Naeole, Kahoalii Keahi, Dr. Kelly Gaither�Mentors: Lela Gi, Clara Slate-Liu

Methods

  • Used a public breast cancer dataset
  • Created visualizations: barplots and Kaplan-Meier survival curves
  • Compared survival between ER+/PR+ and non-ER+/PR+ patients
  • Performed Mann-Whitney U test to confirm statistical significance
  • Used a Random Forest Classifier to build a machine learning model predicting 5-year survival
  • Evaluated model performance using accuracy, precision, and feature importance

Next Steps:

  • Use AutoML to build predictive models based on full patient profiles
  • Explore how treatment types influence survival outcomes
  • Analyze survival differences across demographic factors like race and age
  • Visualize correlations between hormone status and other clinical variables (e.g., tumor size)

References:

Namdari, R. (2022). Breast Cancer [Data set]. Kaggle. https://www.kaggle.com/datasets/reihanenamdari/breast-cancer�Chen, M., Wu, J., Liu, D., Chen, W., Lin, C., Andriani, L., Ding, S., Huang, O., He, J., Chen, X., Chen, W., Li, Y., Shen, K., & Zhu, L. (2022). Combined estrogen receptor and progesterone receptor level can predict survival outcome in human epidermal growth factor receptor 2‑positive early breast cancer. Clinical Breast Cancer, 22(2), e147–e156. https://doi.org/10.1016/j.clbc.2021.05.012

Calculated P-value:�Mann-Whitney U test, p = 0.0001 (statistically significant)

Figure 1.

Figure 2.

Figure 3.