The initial treatment of lupus nephritis is usually based on a renal biopsy. Subsequent disease flares, however, are often treated without the benefit of kidney pathology because repeat biopsies are infrequent. A noninvasive, real-time method to assess renal pathology would be useful to adjust treatment and improve outcome. To develop such a method we collected urine samples at or close to the time of 64 biopsies from 61 patients with lupus nephritis to identify potential biomarkers of tubulointerstitial inflammation and correlated these to biopsy parameters scored by a renal pathologist using a semiquantitative scale. Linear discriminant analysis was used to weight variables and derive composite biomarkers that identified the level of tubulointerstitial inflammation based on urine concentrations of monocyte chemotactic protein-1, hepcidin (a marker of active lupus), and liver fatty acid-binding protein. The discriminant function that described the most accurate composite biomarkers included urine monocyte chemotactic protein-1 and serum creatinine as the independent variables. This composite had sensitivity, specificity, positive predictive value, and negative predictive value of 100, 81, 67, and 100%, respectively. Only 14% of the biopsies were misclassified. Thus, specific renal pathologic lesions can be modeled by composite biomarkers to noninvasively follow and adjust the treatment of lupus nephritis reflecting renal injury.