Multicentric Validation of a Complete Blood Count-Based Model for Breast Cancer Risk Stratification in Women Age...

Area of Science:

• Oncology
• Hematology
• Biostatistics

Background:

• Recent breast cancer (BC) screening guidelines expanded eligibility to women aged 40-49.
• Risk stratification in this younger demographic remains a significant challenge.
• Complete blood count (CBC) parameters offer a potential low-cost risk assessment tool.

Purpose of the Study:

• To evaluate a CBC-based BC risk model's performance in women aged 40-49.
• To assess the model's efficacy across diverse Brazilian clinical settings.
• To determine the impact of CBC timing relative to diagnostic procedures on model performance.

Main Methods:

• Retrospective analysis of CBC data from 239,089 women aged 40-49.
• Utilized a ridge regression model incorporating age, neutrophil-to-lymphocyte ratio (NLR), and red blood cell (RBC) count.
• Assessed model performance using area under the receiver operating characteristic curve (AUC), sensitivity, and specificity with cross-validation.

Main Results:

• The CBC model demonstrated stable performance for CBCs collected within 3 months of diagnosis (AUC ≈ 0.61-0.63).
• Model performance slightly declined for CBCs collected 91-180 days prior to diagnosis (AUC ≈ 0.58).
• At 70% specificity, the model identified 47% of cancers within the top 30% risk group, with an elevated relative risk of 2.07.

Conclusions:

• A CBC-based model offers a cost-effective method for BC risk stratification in women aged 40-49.
• Optimal performance is achieved when CBCs are collected within three months of evaluation.
• This opportunistic screening strategy is particularly valuable in resource-limited settings with restricted mammography access.