A new clinical tool predicts complete remission in patients with lupus nephritis, according to a recent report.
Using artificial intelligence and artificial neural networks, investigators created a machine learning model that gauged lupus nephritis prognosis based on key laboratory and histopathologic characteristics. Patient variables include erythrocyte sedimentation rate, C-reactive protein, serum albumin value, triglycerides levels, complement C3 and C4 levels, presence of antinuclear antibodies, urine protein to creatinine ratio (UPCR), and data from histopathological examinations of biopsy samples.
Investigators randomly assigned 58 patients with lupus nephritis and baseline UPCR greater than 1.0 mg/mg to a training or testing set. Patients were treated with 6 intravenous pulses of cyclophosphamide (500 mg each) followed by oral mycophenolate mofetil.
The predictive models assessed with 91.67% accuracy the probability of achieving complete remission, defined as UPCR less than 0.5 mg/mg and stable renal function at 6 months of follow-up, Andrzej Konieczny, MD, PhD, of Wroclaw Medical University in Poland, and colleagues reported in BMC Nephrology. The area under the receiver operative curve for the model’s discriminatory ability was 0.9375.
“We emphasize the possibility of using this solution in a pilot program after conducting further observations on a larger research group.”
Stojanowski J, Konieczny A, Rydzyńska K, et al. Artificial neural network – an effective tool for predicting the lupus nephritis outcome. BMC Nephrology 23(1):381. doi:10.1186/s12882-022-02978-2
This article originally appeared on Renal and Urology News