On behalf of the TARGET-NASH investigators, we thank Dr Calzadilla Bertot for his interest and comments1 on our original manuscript describing liver biopsy reporting, interpretation, and concordance with a pragmatic clinical definition, in our real-world cohort of patients from the TARGET-NASH study.2

The important take-away message from our study is that, while liver biopsy and histological interpretation are the referent standard for the diagnosis and assessment of disease severity in non-alcoholic fatty liver disease (NAFLD), the practical application of this tool often falls short in routine clinical practice. Liver biopsy will always have an important role in making a diagnosis of liver disease with an atypical presentation. However, most clinicians should be comfortable diagnosing NAFLD based on history, steatosis on imaging and the laboratory exclusion of other common liver diseases. The difficulty has been accurately identifying the population of patients with the progressive phenotype of non-alcoholic steatohepatitis (NASH).

We agree that liver biopsy is currently a necessary tool for inclusion of NASH patients in clinical trials and for establishing end points for drug efficacy. Liver biopsy specimens in these clinical trials are interpreted by experienced hepatopathologists from major academic centres. The challenge lies in how the refined histological assessments made in clinical trials are potentially applied to millions of patients with NASH in clinical practice. In our study, we demonstrated that essential descriptors of NASH, like steatosis, ballooning and inflammation were not mentioned in between 21% and 46% of pathology reports. Furthermore, when reported, concordance with a centralised expert pathologist ranged from slight to fair for the inflammatory features of NASH. Fortunately, fibrosis assessment, especially advanced fibrosis or cirrhosis, was much better. If biopsy is the metric for making treatment decisions once new pharmacotherapy is available, efficacy seen in clinical trials may not translate to clinical effectiveness in the real world due to potential misclassification of patients.

The clinical definition of NASH used in the TARGET-NASH cohort is based on clinical information available in routine clinical practice.3 When compared to patients with either definite or probable (satisfying all criteria except ballooning) NASH on a centrally interpreted biopsy, the concordance was 91% and sensitivity 95%. As the decision to biopsy a patient in routine clinical practice can vary by provider and there may be bias in which patients are selected for biopsy,4 a pragmatic clinical definition for NASH that is applicable to all, even if slightly imperfect, is essential for patient care.

We agree with Dr Calzadilla Bertot that continued efforts are required to improve the non-invasive testing (NIT), diagnosis and assessment of NASH. The most useful NITs will combine accuracy with accessibility; such advances will facilitate treatment decisions and enhance patient care.

The authors' declarations of personal and financial interests are unchanged from those in the original article2.