Novel reliability criteria for controlled attenuation parameter assessments for non‐invasive evaluation of hepatic steatosis

Abstract

BACKGROUND: There is conflicting evidence regarding reliability criteria for the controlled attenuation parameter (CAP; a marker for hepatic steatosis [HS]). Thus, we assessed the diagnostic performance of CAP according to different reliability criteria based on real-world data from an academic centre. METHODS: Patients undergoing measurement of CAP and liver biopsy (±6 months) at the Medical University of Vienna were included. HS was assessed according to SAF score. RESULTS: = 29, 9.1%). Histological steatosis and fibrosis stages were: S0: 93 (29.2%), S1: 100 (31.3%), S2: 67 (21.0%), and S3: 59 (18.5%); F0/F1: 150 (47.0%), F2: 47 (14.7%), and F3/F4: 122 (48.3%). In the overall cohort, the area under the receiver operating characteristic curve (AUC) of CAP was 0.843 (95% confidence interval [CI]: 0.798-0.887) for diagnosing HS ≥ S1), 0.789 (95%CI: 0.740-0.839) for ≥S2, and 0.767 (95%CI: 0.712-0.823) for S3. CAP corrections as suggested by Karlas et al. did not improve the diagnostic performance. Importantly, the AUC of CAP for HS ≥ S1 was numerically highest in patients with CAP-IQR/median<0.10 or <0.20 (obtained in 37.9% and 74.9%), in whom CAP also had better diagnostic performance, as compared with patients not meeting these criteria. Moreover, it was substantially higher in 288 (90.3%) patients with CAP-IQR/median<0.3: 0.856 (95%CI: 0.809-0.903) vs. patients not meeting this criterion (0.530 [95%CI: 0.309-0.751]). In contrast, the previously suggested reliability criterion of CAP-IQR<40 dB/m was not associated with an improved diagnostic performance for HS≥S1 (0.866 [95%CI: 0.812-0.920] vs. 0.799 [95%CI: 0.717-0.881]) and was only obtained in 199 (62.4%) patients. CONCLUSION: CAP-IQR/median<0.1, <0.2, and <0.3 identify reliable measurements for diagnosing any hepatic steatosis (≥S1). Importantly, CAP-IQR/median<0.3 has a considerably higher applicability in clinical practice, as compared with the previously suggested CAP-IQR<40 dB/m criterion.