In their interesting review of noninvasive assessment of liver fibrosis,1 Martínez et al. note that “the introduction of noninvasive methods in clinical practice is making such slow headway in the field of hepatology”. However, their statement that “an exception to this rule is France, where three well-validated methods (FibroTest, Fibrometer, and FibroScan) have been approved by the public health system and are routinely used in clinical practice” has to be discussed.
Certainly, there is no doubt about wide use of these methods in clinical practice, despite the restrictions stated by the French National Authority for Health (Haute Autorité de Santé [HAS]).2 In February 2011, the Social Health Insurance finalized the decisional process of listing those techniques for eligibility for reimbursement, subject to very strict rules according to this advice. Transient elastography, for example, will be reimbursed only when performed by specialists, for the follow-up of untreated patients with chronic hepatitis C and no evidence of cirrhosis, once a year only (with the exception of patients presenting a risk of rapid evolution to cirrhosis).
One should also be careful when qualifying noninvasive methods of liver fibrosis assessment as “well-validated” in “large cohorts” of patients. First, numerous accuracy studies of these methods were carried out and published by their manufacturers or inventors. The largest independent one, to date, is the Fibrostic study comparing the diagnostic accuracy of biomarkers to that of transient elastography (FibroScan) in one and the same population of 1307 patients with chronic hepatitis,3, 4 but it was unfortunately published too recently for Martínez et al. to take it into account in their review. Second, performance of noninvasive methods is certainly good for diagnosing cirrhosis but poor for significant fibrosis; the Fibrostic study did not even reach the minimum values of 85% sensitivity and specificity deemed high enough by Martínez et al. Moreover, the Fibrostic study results showed the ability of noninvasive tests to confirm or rule out significant fibrosis was satisfactory in limited ranges of high or low values only.3 Third, accuracy is only a step toward a possible usefulness of a test.5 As rightly stressed by Martínez et al. at the end of their article, the improvement of patient outcomes is more relevant. However, the prediction of clinical endpoints by noninvasive methods was only recently investigated by very few studies, and to our knowledge, no study assessed their ability to predict response to therapy.
We all wish that noninvasive methods could allow us to avoid liver biopsy while ensuring that patients will be managed as well or better than with old techniques. Therefore, studies are needed in order to specify their contribution to the choice of patient management for improving clinical endpoints or treatment response.