Optimization of Rituximab Therapy in Adult Patients With PLA2R1-Associated Membranous Nephropathy With Artificial...

Area of Science:

• Nephrology
• Pharmacology
• Machine Learning

Background:

• Rituximab is a first-line treatment for membranous nephropathy.
• Urinary drug loss in nephrotic syndrome can limit rituximab exposure.
• Rituximab underdosing (serum level <2 μg/ml at month-3) is a risk factor for treatment failure.

Purpose of the Study:

• To develop a machine learning algorithm to predict the risk of rituximab underdosing.
• To investigate the relationship between predicted underdosing risk and cumulative rituximab dose for remission.

Main Methods:

• A machine learning algorithm was developed using patient characteristics at rituximab infusion.
• Rituximab concentrations were measured at month-3 in 92 adult patients with primary membranous nephropathy.
• A forward-backward selection procedure identified optimal predictive variables, with performance evaluated via 10-fold cross-validation.

Main Results:

• The best predictive model included age, gender, BSA, anti-PLA2R1 antibody titer, serum albumin, and serum creatinine.
• The algorithm achieved high accuracy (79.4% training, 79.2% testing), sensitivity, and specificity.
• Patients with high predicted underdosing risk required longer time to remission and higher cumulative rituximab doses.

Conclusions:

• The developed algorithm accurately predicts rituximab underdosing risk in membranous nephropathy.
• Early identification of high-risk patients allows for potential regimen intensification.
• Optimizing rituximab dosing can increase the likelihood of achieving remission.