Support Vector Machine-Based Formula for Detecting Suspected α Thalassemia Carriers: A Path toward Universal...

Area of Science:

• Hematology
• Medical Diagnostics
• Genetics

Background:

• Alpha thalassemia carriers (α-thal) present similar blood counts to beta thalassemia carriers, lacking elevated Hemoglobin A2 (HbA2).
• Accurate identification of α-thal carriers is crucial for genetic counseling and preventing hemoglobinopathies in offspring.
• Existing diagnostic methods may require further refinement for efficient carrier screening.

Purpose of the Study:

• To evaluate the efficacy of mathematical formulas in detecting suspected cases of alpha thalassemia.
• To assess the sensitivity and negative predictive value (NPV) of specific mathematical models for α-thal screening.
• To determine if computational approaches can improve the identification of at-risk couples for hemoglobinopathies.

Main Methods:

• Analysis of Red Blood Cell (RBC) indices from 1334 blood counts suspected of α-thal.
• Application of mathematical formulas, including Shine and Lal and Support Vector Machine (SVM), to RBC data.
• Calculation of sensitivity and negative predictive value (NPV) for the evaluated formulas.

Main Results:

• The Shine and Lal and SVM formulas demonstrated high sensitivity (85.54% and 99.33%) and NPV (98.93% and 99.93%) in detecting suspected α-thal.
• Molecular defects confirmed α-thal in 291 individuals, while 81 had normal α genes; 962 samples lacked molecular analysis.
• These mathematical models effectively flagged potential α-thal carriers within the analyzed cohort.

Conclusions:

• Mathematical formulas, particularly Shine and Lal and SVM, are effective tools for detecting suspected α-thal carriers.
• Integrating these reliable formulas into automated analyzers can enhance screening programs for hemoglobinopathies.
• Improved carrier detection can increase physician awareness and prompt further genetic testing for at-risk individuals.