Project

A Machine Learning Approach for Early Prediction of Blood Culture Positivity in Neutropenia Patients Using Medical History and Hematological Parameters

Cancer patients who are treated with multiple chemotherapy sessions exhibit a decline in peripheral neutrophils, which leads to immunodeficiency. Internal infections that affect these patients and are left untreated result in neutropenicsepsis. Neutropenicsepsis is a life-threatening condition and an immense cause of cancer chemotherapy-related mortality, which is incited by the breach in the first line of defence against microbes constituted by a decrease in the absolute neutrophil count. Cancer patients who are affected by neutropenic sepsis are generally treated with Broad Spectrum Antibiotics (BSA) as a first-line medication that destroys common bacteria and microbes. Broad-spectrum antibiotics introduce Antimicrobial resistance(AMR),which will subsequently reduce the effect of even specific antibiotic drugs. The physicians resort to using BSA due to the fact that a blood culture result would typically take 2–5 days. The study postulated that machine learning using hematological parameters would enable early prediction of the presence of bacterial growth in blood cultures and its characterization before culture results.

Proposed Future Work Details

1. Predict class organisms in positive blood culture – whether it is gram-negative, Gram – positive or budding yeast.
2. Scoring system to determine positive culture using blood lab values and biomarkers.