We read with great interest the article by Shen et al. The authors describe the diagnostic utility of CK18 (M30, M65 and M65ED) for the diagnosis of NAFLD and NASH.
CK 18 was first described in this context by Wieckowska et al. in 2007. In their seminal paper, CK18 fragments were shown to be markedly increased in patients with NASH as compared with simple steatosis and control. In patients with NAFLD, CK18 achieved an AUC of 0.93 for the diagnosis of NASH. In a further, larger scale (n = 139) validation study, Feldstein et al. found that CK18 fragments achieved an AUC of 0.83 for diagnosing NASH. Both studies described the progressive relationship between CK18 and fibrosis grade. Since then, CK 18 has been validated in a number of other studies, with AUC values ranging from 0.83 to 0.88.[4-6]
In this current study, Shen et al. report impressive AUC values for the diagnosis of NAFLD from a control population for all three biomarkers (0.92–0.94); however, the AUC for NASH diagnosis drops to a disappointing 0.66–0.71. These are values lower than reported in other validation studies. A potential explanation is the unclear description of NASH. NASH was diagnosed in the presence of steatosis, lobular inflammation and hepatocyte ballooning, but it was not clear if this constituted an NAS score ≥5, or if it included the NAS score 3–4 cases.
This paper introduces an additional use for CK18 fragments; they measured the three biomarkers in paired patient samples 36 months apart, a novel observation. Of the 51 patients analysed, 10 progressed from non-NASH to NASH. Within this cohort, the three biomarkers had AUC of 0.74–0.82 for the prediction of progression. This is novel for CK18-based markers and, despite the small numbers included in this sub-analysis, is worthy of note and future study.