Detection of chronic allograft nephropathy by quantitative Doppler imaging

Background. Chronic allograft nephropathy (CAN) is the major cause of graft loss, and early detection is desirable to avoid irreversible graft damage. We have evaluated a new technique of color Doppler quantification using Cineloop (Philips Medical Systems, Bothell, WA) imaging for noninvasive diagnosis of CAN. Methods. Provisional normal ranges were defined by pilot study (n=13) and prospectively tested in stable recipients in whom CAN was independently quantified by contemporaneous histology (n=67), using the Banff schema. Results. The maximal fractional area (MFA, systolic color pixels/total area) was 28.7 +/- 9.7% in normal subjects and reduced to 18.8 +/- 8.0% in grade 1 and 12.5 +/- 6.4% in grade 2 CAN (both P < 0.001). The minimum color fractional area was reduced from 10.3 +/- 5.3% in normal subjects to 3.1 +/- 2.6% in grade 2 CAN (P < 0.001), but was less useful. Distance from peripheral color pixels to capsule increased in CAN grade 2 versus 0 (6.0 +/- 1.6 vs. 3.9 +/- 1.0 mm, respectively; P < 0.001). Calcineurin inhibitor nephrotoxicity reduced MFA (18.0 +/- 9.3 vs. 26.9 +/- 10.7%; P < 0.001) and other dynamic measurements. Parenchymal damage exerted minimal effect on resistance index, mean variance, and peak Doppler velocity. MFA (cutoff<17.3%) can diagnose CAN (sensitivity 69%, specificity 88%, positive predictive value 86%) and severe CAN (sensitivity 87%, specificity 71%, negative predictive value 95%). Distance to capsule >5 mm was less sensitive (49%) but more specific (91% alone, and 97% combined with MFA). Conclusions. In conclusion, quantitative Doppler ultrasound can reliably detect CAN and, although imperfect at correctly grading, allows recognition of significant tubulointerstitial damage for initiation of a confirmatory needle core biopsy.