Evaluating Related Factors to the Number of Involved Lymph Nodes in Patients With Breast Cancer Using Zero-Inflated Negative Binomial Regression Model Publisher



Younespour S¹ ; Maraghi E² ; Malehi AS² ; Raissizadeh M² ; Seghatoleslami M³ ; Hosseinzadeh M⁴ Show Affiliations
Source: Journal of Biostatistics and Epidemiology Published:2020

Abstract

Background and aims: In Iran, breast cancer accounts for 24.4% of all cancers and contributes to 14.2% of cancer-associated mortality in women. A major challenge facing the health system is to examine the health status of patients with breast cancer, which often involves the axillary lymph nodes. The number of involved nodes should be clinically predicted to ascertain postoperative radiotherapy and chemotherapy. The present study employed regression models to investigate the determinants of the number of lymph nodes involved in patients with breast cancer. Methods: This retrospective study recruited patients diagnosed with breast cancer during 2005-2015 referring to Shafa Hospital affiliated to Ahvaz Jundishapur University of Medical Sciences, Ahvaz, Iran. The outcome variable was the number of involved lymph nodes. Regression models for count outcomes, were utilized for investigating the related factors to the number of involved lymph nodes in patients with breast cancer. Results: A sample of 165 patients was eligible for the present study. The Akaike information criterion (AIC) of the zero-inflated negative binomial (ZINB) model was the lowest. The logistic part showed that absence of metastasis significantly increased the chance of node-negative breast cancer (P=0.027). The negative binomial part revealed an increase of 86% in the risk of a greater number of involved nodes in stage III breast cancer compared to stages I and II, suggesting that the patients were at a high risk (P=0.006). Conclusion: Metastasis status and tumor grade significantly relate to the number of lymph nodes involved in breast cancer. Determining the factors associated with nodal involvement is crucial for the early diagnosis of breast cancer by clinicians. © 2020, Tehran University of Medical Sciences. All rights reserved.