Article InfoChang T, Wu Y, Tung S, et al. Alpha-fetoprotein measure benefits hepatocellular carcinoma surveillance in patients with cirrhosis. Am J Gastroenterol 2015;110:836–844.
Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death worldwide and one of the leading causes of death in patients with cirrhosis (Gastroenterology 2012;142:1264–1273 e1). Several societies recommend HCC surveillance for patients with cirrhosis given its association with higher rates of early tumor detection, curative treatment received, and overall survival (Hepatology 2010;53:1–35). Ultrasonography and alpha-fetoprotein (AFP) are the most common tests for HCC surveillance in clinical practice. However, the accuracy of AFP has been critically challenged, and there is growing debate about its continued use in HCC surveillance programs.
The study by Chang et al addressed this important issue through a retrospective cohort study of patients with cirrhosis who were undergoing HCC surveillance in Taiwan between 2002 and 2010 (Am J Gastroenterol 2015;110:836–844). Among 5664 potentially eligible patients, 1597 were retained for analyses after excluding those with history of HCC, prevalent HCC within 18 months of enrollment, concurrent extrahepatic neoplasms, and/or follow-up duration of <18 months. Patients routinely received HCC surveillance with ultrasonography and AFP every 3-6 months, with diagnostic workup typically initiated when mass lesions were seen on ultrasonography or AFP levels were increased.
Most patients (n = 1398; 88%) had hepatitis B virus (HBV) and/or hepatitis C virus (HCV) cirrhosis, and most (n = 1233, 77%) had Child-Pugh class A cirrhosis. Over a median of 4.8 years (range, 1.4–14), 363 (22.7%) developed HCC, of whom 213 (59%) were early or very early stage HCC as defined by Barcelona Clinic Liver Cancer criteria. HCC was the factor most strongly associated with an increased AFP (relative risk [RR], 13.2; 95% CI, 9.6–18.2); however, false-positive AFP levels were significantly more likely in HCV- (RR, 6.9; 95% CI, 1.6–29.5) or HBV-infected (RR, 5.6; 95% CI, 1.3–24.7) patients than those with nonviral liver disease.
Most patients had HCC (334/363; 92%) detected by ultrasonography, regardless of AFP level. However, 29 patients (8%) presented with increasing AFP alone (ie, no mass on ultrasonography). The sensitivity and specificity of ultrasonography alone were 92.0% (95% CI, 89.2–94.8) and 74.2% (95% CI, 71.8–76.7), respectively. Combining ultrasonography and AFP, at a cutoff of 20 ng/mL, significantly increased sensitivity (P < .001) compared with ultrasonography alone. In subgroup analysis, the significant increase in sensitivity was observed in patients with viral cirrhosis (P < .001) but not those with nonviral cirrhosis (P > .10). The combination of ultrasonography and AFP had a sensitivity of 99.2% (95% CI, 98.2–100) with minimal decrease in specificity (68.3%, 95% CI, 65.7–70.9). Adding an additional criterion of AFP to increase ≥2 times from nadir maintained sensitivity at 99.2% (95% CI, 98.2–100) and increasing specificity to 71.6% (95% CI, 69.0–74.0).
CommentThe effectiveness of HCC surveillance in clinical practice is contingent on the availability of surveillance tools with high sensitivity for early HCC detection. The most recent guidelines from the American Association for the Study of Liver Diseases and European Association for the Study of the Liver recommend using ultrasonography alone to achieve this goal given concerns about the suboptimal sensitivity and specificity of AFP (Hepatology 2010;53:1–35). Although ultrasonography is highly efficacious in well-controlled settings, this study by Chang et al contributes to the growing evidence base that its effectiveness in clinical practice may be more heterogeneous (Am J Gastroenterol 2015;110:836–844).
In addition to the operator-dependent nature of ultrasonography, its performance may be influenced by patient characteristics, such as liver nodularity. A prior study using a multicenter Italian database found that ultrasonography failure rate for detecting HCC varied between centers (range, 10%–64%), with higher failure rates associated with male gender, Child B or C cirrhosis, high AFP levels, and surveillance intervals of >6 months (Clin Gastroenterol Hepatol 2014;12:1927–1933.e2). Inadequate sensitivity of surveillance tools is the most common reason for late-stage tumor detection among patients followed in tertiary care centers (Am J Gastroenterol 2013;108:425–432). With less than one-half of all HCC being detected at an early stage, it is clear that better surveillance tests are needed. Although several novel biomarkers are being investigated, most are in phase II-III biomarker development and years away from being available widely for use in clinical practice (Best Pract Res Clin Gastroenterol 2014;28:843–853). AFP is currently the only biomarker to have undergone all 5 phases of biomarker development. Similarly, ultrasonography is the only imaging modality to be evaluated adequately as a surveillance imaging modality. Although some advocate use of CT or MRI for HCC surveillance, their widespread use is limited by costs and harms, such as radiation exposure. Therefore, current efforts should be focused on optimizing the delivery and effectiveness of surveillance using ultrasonography and AFP.
The study by Chang et al suggests that using ultrasonography and AFP in combination significantly increases surveillance sensitivity for HCC compared with ultrasonography alone (92% vs 99%; P < .001; Am J Gastroenterol 2015;110:836–844). However, these data describe the performance of ultrasonography and AFP for HCC detection at any stage, whereas the goal of HCC surveillance is to detect tumors at an earlier stage, when curative options are available (Hepatology 2010;53:1–35). A prior effectiveness trial suggested adding AFP to ultrasonography can also increase significantly surveillance sensitivity for early HCC (32% vs 63%; P < .001), but these data still require validation in other studies (Cancer Epidemiol Biomarkers Prev 2012;21:793–799). Given the correlation between AFP and HCC tumor stage/size, it is possible that AFP could detect advanced HCC missed on ultrasonography without improvement in early tumor detection. Only 59% of HCC in the study by Chang et al were detected at an early stage, and it is unclear if the 29 cases detected by AFP alone were early or advanced tumors (Am J Gastroenterol 2015;110:836–844). Future studies evaluating ultrasonography and AFP for HCC surveillance should report test performance characteristics for early HCC detection.
Although adding AFP ≥20 ng/mL to ultrasonography significantly increased sensitivity compared with ultrasonography alone, the best performance characteristics were observed when using longitudinal AFP measurements. Requiring an increase in AFP level of ≥2 from its nadir in the prior year maintained high sensitivity of surveillance while increasing specificity. This finding confirms prior studies that suggested longitudinal AFP measurements may have greater diagnostic accuracy than 1-time measurements for HCC detection, particularly in patients with HCV cirrhosis, where AFP elevations are often due to disease activity rather than HCC (Clin Gastroenterol Hepatol 2013;11:437–440). Alternatively, others have suggested the accuracy of AFP for HCC surveillance can be improved by using different cutoffs based on liver etiology or adjusting AFP level for patient age, alanine aminotransferase level, and platelet count (Clin Gastroenterol Hepatol 2014;12:870–877; Gastroenterology 2014;146:1249–1255.e1).
In subgroup analysis, AFP increased sensitivity for HCC detection in patients with HCV or HBV cirrhosis, but not those with nonviral liver disease. The authors concluded AFP may not be beneficial in patients with alcohol or nonalcoholic steatohepatitis (NASH); however, lack of significance in this subgroup may simply be related to small sample size (n = 199). In fact, patients with NASH who are obese may be particularly prone to ultrasonography failure and stand to benefit from the addition of serum biomarkers. Further studies are needed to evaluate the sensitivity of ultrasonography and potential benefit of AFP in patients with NASH because the burden of NASH is growing worldwide.
Increasing surveillance sensitivity must be weighed carefully with trade offs in specificity, because high rates of false-positive tests can result in unnecessary diagnostic evaluation and resultant harms. In this study, adding AFP to ultrasonography resulted in a minimal decrease in specificity, which could be mitigated further by incorporating longitudinal AFP data. In fact, most patients who underwent CT or MRI for false-positive surveillance tests did so because of ultrasonography results (91%), with only 33 patients (9%) undergoing diagnostic evaluation for false-positive AFP levels. These data highlight the possibility that the harms of ultrasonography may be greater in clinical practice than prior efficacy studies given “unnecessary” diagnostic evaluation for indeterminate ultrasonography results. Similarly, the harms of AFP may be lower in clinical practice, because providers may not pursue diagnostic evaluation in all cases after accounting for factors, such as AFP trend, alanine aminotransferase levels, and liver disease etiology.
Overall, the study by Chang et al tackles an important issue in HCC surveillance. Ultrasonography was able to detect 92% of HCC, which was significantly improved by combining with AFP with minimal loss in specificity. However, nearly 40% of tumors were diagnosed beyond an early stage, highlighting the need for better surveillance tools in the future. In the interim, using ultrasonography and AFP in combination may be the best surveillance strategy to optimize early HCC detection.
| 欢迎光临 肝胆相照论坛 (http://hbvhbv.info/forum/) | Powered by Discuz! X1.5 |