Abstract
INTRODUCTION: We evaluated a panel of novel urinary and serum biomarkers (BMs) for early and sensitive detection of cisplatin drug-induced kidney injury (DIKI) in patients with cancer, comparing their diagnostic accuracy with standard BMs (SBMs). METHODS: In this prospective exploratory observational study, 105 patients treated with cisplatin ("treated" with > 65 mg/m(2)/cycle), 20 non-cisplatin treated cancer controls ("nontreated"), and 34 "healthy" controls were enrolled. The treated group's serum and urine samples were collected predose, after 12 hours, and on days 1, 2, 4, 7, 14, and 21. SBMs and novel BMs (NBMs; 8 urinary, 1 serum) were measured, comparing accuracy, percent changes from baseline (PCFBs), and median time to peak values between treated patients and nontreated cancer controls. Blinded adjudication of the treated group's BM profiles occurred at 2 stages for DIKI diagnosis. RESULTS: All urinary NBMs had significant PCFBs in the treated group compared with the nontreated cancer control group; most accurately detected cisplatin exposure (area under the receiver operating characteristics [ROC] curve [AUROC] > 0.8). NBMs peaked earlier. In stage 1 adjudication (SBMs) of the treated group, PCFB of urinary NBMs showed no difference between DIKI (n = 24) and no-DIKI (n = 71) groups except for neutrophil gelatinase-associated lipocalin (NGAL) and cystatin C (CYSC). In treated participants, all BMs showed greater PCFBs than control groups, regardless of stage 1 DIKI adjudication. Stage 2 (SBMs and NBMs), DIKI incidence (n = 63) increased by 41%, with most BMs having an AUROC > 0.80 compared with the nontreated cancer control group. CONCLUSION: NBMs accurately and timely detected cisplatin exposure and identified "sub-clinical" DIKI undetected by standard acute kidney injury (AKI) criteria, highlighting the limitations of current functional BMs in estimating the true DIKI incidence.