Controlled attenuation parameter (CAP): a new device for fast evaluation of liver fat?

Abstract

The diagnosis of liver steatosis has several implications in chronic liver diseases. Indeed, liver steatosis is associated with liver fibrosis progression and a decreased rate of sustained viral response in chronic hepatitis C 1, 2. Donor liver macrovesicular steatosis is independently associated with graft failure at 1 year after liver transplantation 3. After major hepatic resection, liver steatosis induces a two-fold increased risk of post-operative complications, and excessive steatosis a three-fold increased risk of death 4. Finally, liver steatosis is the primary lesion of non-alcoholic fatty liver disease (NAFLD) which, as a consequence of the worldwide burden of visceral obesity, is now the main cause of chronic liver disease in western countries 5. At present, the histological examination of a liver biopsy remains the reference for evaluating liver steatosis. However, this procedure is invasive, impaired by sampling bias, has imperfect reproducibility and allows only for a semiquantitative grading of steatosis 6. The non-invasive diagnosis of liver steatosis, especially imaging techniques and blood tests, has been extensively developed in the last decade, but is not yet recommended for use in clinical practice. Ultrasonography is considered the imaging technique of choice for steatosis screening 7, but its sensitivity in detecting fatty liver is only 60–94% and operator dependent 8. Computed tomography induces radiation exposure, lacks sensitivity for small amounts of fat and is impaired by intermachine variability 8. Proton magnetic resonance spectroscopy and magnetic resonance imaging offer high accuracy for quantification of liver fat, but are limited by low availability, high cost and lack of standardization 9. Finally, blood tests have been implemented for binary diagnosis or quantification of steatosis, but their diagnostic accuracy remains to be validated 10. Moreover, pre-analytical and analytical conditions, such as fasting samples, need to be observed for blood tests. Similarly, blood tests including serum glucose and/or lipid(s) are sensitive to medical interventions, such as the use of antidiabetic or lipid-lowering drugs. Controlled attenuation parameter (CAP) is a new technology that measures ultrasonic attenuation in the liver using signals acquired by a transient elastography probe (FibroScan, Echosens, Paris, France). CAP was developed on the postulate that fat affects ultrasound propagation 11. CAP is measured at the same time, on the same volume and on the same signal as a valid liver stiffness measurement by Fibroscan. Results are expressed as decibels per meter (dB/m) for a given ultrasound frequency and range from 100 to 400 dB/m. Two studies evaluating CAP for the non-invasive assessment of liver steatosis in patients with chronic liver diseases are presented in this issue of Liver International 12, 13. In both works, pathological examination of liver biopsy was used as the reference for steatosis evaluation. As they included quite different populations, it is difficult to directly compare the results of the de Ledinghen and Myers studies. Nevertheless, they both showed that CAP has good overall accuracy for the diagnosis of >10% steatosis with AUROC at 0.84 in the de Ledinghen study and 0.81 in the Myers study. Moreover, despite different calculations, diagnostic cut-offs for the detection of steatosis were also very close respectively: 266 dB/m and 283 dB/m. De Ledinghen et al. also showed that CAP could exclude (when <238 dB/m) or affirm (when ≥263 dB/m) the presence of steatosis with ≥80% negative or positive predictive values. These two thresholds were very close and the rate of patients remaining in the grey zone between them was only 11%. The de Ledinghen and Myers studies compared CAP to 3 blood steatosis tests: Steatotest, Fatty Liver Index and Hepatic Steatosis Index. CAP provided higher AUROCs than the blood tests for the diagnosis of steatosis (S) grades S ≥ 1 (>10% hepatocytes containing lipid vesicles), S ≥ 2 (>33%) or S ≥ 3 (>66%), but the difference was clearly significant only in the de Ledinghen study. CAP has several advantages compared to blood tests: it provides immediate results and is probably less sensitive to medicinal intervention such as antidiabetic or lipid-lowering treatments. However, continuing this comparison, it also has several limitations. Firstly, measurement failure with the Fibroscan M probe increases in parallel with BMI 14, whereas it is theoretically 0% with blood tests. Consequently, CAP measurement failure may be frequent in NAFLD patients, a population of interest for steatosis evaluation. This limitation might be circumvented by the development of CAP measurement with the new XL probe recently developed for liver stiffness measurement in obese patients 15, 16. Secondly, the reproducibility of blood fibrosis tests has been shown to be excellent 17, but that of liver stiffness measurement by Fibroscan is only fair in low stiffness values and high BMI 18. Thus, the reproducibility of blood steatosis tests and that of CAP measurement have to be similarly evaluated. Finally, CAP will likely be less widely available than blood tests, which can be performed by all practitioners. For all these reasons, further studies directly comparing CAP to blood tests for the detection of steatosis, on an intention-to-diagnose basis according to STARD statements, are warranted in large samples of patients. Accurate steatosis quantification is important for the evaluation of medical interventions in clinical practice and treatment responses in therapeutic trials. Although CAP was well correlated with pathological grades of steatosis in both the de Ledinghen and Myers studies, the results overlapped between grades and, consequently, CAP showed poor accuracy for the differentiation of adjacent grades of steatosis. In addition, calculated cut-offs in the Myers study were close for the diagnosis of S ≥ 1, S ≥ 2 or S ≥ 3 grades. Finally, by defining four classes of CAP results according to the three calculated cut-offs in the de Ledignhen study, only 25/112 patients were correctly classified for steatosis grading. We have previously shown that such combinations of diagnostic cut-offs based on a binary diagnosis are hand-crafted statistics and result in decreased diagnostic accuracy 19. At first sight, all these results may suggest that CAP is not accurate for the quantification of liver steatosis. In fact, classical optical examination of a liver biopsy by a pathologist is not the best reference for liver steatosis measurement: the determination of the percentage of hepatocytes containing lipid vesicles is highly subjective, and steatosis grading corresponds only to a semiquantitative scale. Thus, the de Ledinghen and Myers studies were not designed to assess the ability of CAP to quantify steatosis, although they certainly did show that the method is accurate for detecting it. For quantification, better references are required, such as morphometry, which determines the area of steatosis on liver biopsy, or imaging techniques like proton magnetic resonance spectroscopy or resonance magnetic imaging. Because of its invasiveness, liver biopsy is usually reserved for a subset of highly selected patients. Consequently, populations included in studies based on biopsy as reference are usually not representative of patients seen in clinical practice, making the extrapolation of study results to clinical practice difficult. For this reason, future studies aiming to evaluate CAP for the quantification of liver steatosis should use accurate and non-invasive imaging techniques as reference rather than morphometry on liver biopsy. At present, there is no specific test for NAFLD diagnosis, this latter being based on the presence of liver steatosis when other steatosis causes and other chronic liver diseases have been excluded. For this purpose, liver steatosis is usually evaluated by ultrasonography in clinical practice. Interestingly, in the de Ledinghen study, CAP demonstrated better accuracy, and even more so, better sensitivity than ultrasonography for the detection of >10% steatosis. As CAP is non-invasive, safe and acceptable, this suggests that CAP might replace ultrasonography as the reference examination for NAFLD screening. Further studies including large cohorts of patients with NAFLD are now required to validate this preliminary result. CAP should also be evaluated in the other main causes of chronic liver disease, such as viral hepatitis or alcohol abuse. In this setting, Sasso et al. have recently shown that CAP was accurate for steatosis diagnosis in 615 patients with chronic hepatitis C 20. Finally, as it has been done for liver stiffness, the signification of CAP measurement (liver fat content, macro versus microvacular steatosis, etc.) and its potential confounding factors have to be determined for an accurate interpretation of its results. In conclusion, the accurate assessment of liver steatosis is crucial in clinical practice for the management of patients with chronic liver disease, and in clinical research for epidemiological and therapeutic studies. The ideal method for liver steatosis evaluation must be widely available, non-invasive, safe, sensitive, accurate for quantification, reproducible and inexpensive. Currently, non-invasive and accurate methods are available, but they don't fulfil all these characteristics. In this context, the de Ledinghen and Myers studies demonstrate that CAP holds promise. Nevertheless, before being considered as the Holy Grail of steatosis evaluation in chronic liver diseases, CAP needs further evaluation in studies using adequate references and including large samples of patients in a single cause. It also has to be directly compared to the most accurate steatosis quantification methods. Conflict of interest: P. Calès is a consultant for BioLiveScale (FibroMeter).