Diagnostik des hepatozellulären Karzinoms - Eine Metaanalyse

Mit über 600 000 Todesfällen pro Jahr ist der Leberkrebs weltweit unter den 10 häufigsten Krebstodesursachen zu finden [59]. Der mit bis zu 85% häufigste histologische Subtyp ist das hepatozelluläre Karzinom, das somit bereits im Jahr 1990 für über 5% aller Krebsfälle weltweit verantwortlich war [12...

Full description

Saved in:
Bibliographic Details
Main Author: Pado, Janina
Contributors: Sitter, Helmut (Priv.-Doz. Dr.) (Thesis advisor)
Format: Doctoral Thesis
Published: Philipps-Universität Marburg 2015
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents: Liver cancer is one of the ten most common causes of cancer death worldwide with an estimated 600.000 deaths each year [59]. Hepatocellular carcinoma as a histological subtype represents 85% of the liver cancers and thereby accounted for up to 5% of the worldwide cancer burden in the year 1990 [12]. Because of its 5 year survival rate of 6.5%, HCC is associated with a very poor prognosis [25]. In the last few years the incidence was rising constantly, especially in countries with a low prevalence of HCC like the USA or Western Europe. However, in parts of Asia or Africa where the prevalence of HCC is high, there was a decrease in the number of new cases of HCC [27]. It occurs almost twice as often in men than in women [40]. The most important risk factor for the development of HCC is the cirrhosis of the liver. A particularly high incidence of HCC occurs in patients with cirrhosis who also suffer from hepatitis C infection [28]. An early diagnosis is the decisive factor for the patients given the few therapeutic options. The aim of this dissertation was to outline the currently available data for the diagnosis of HCC, to evaluate the existing studies concerning their methodology and quality and to determine their level of evidence in order to quantitatively summarize the results in a meta-analysis. Therefore a systematic literature research was conducted on PubMed, the thereby obtained studies were compared to predetermined inclusion and exclusion criteria and were excluded if necessary. The eligible studies were assigned to an evidence level by two independent reviewers. Any disagreement resulted in a consensus through discussion. Subsequently, a meta-analysis was conducted which included an overall analysis as well as several subgroup analyses. The diagnostic accuracies of each study were assessed by using the bivariate model to account for the two-dimensional effect of sensitivity and specificity. In total, 46 studies were included, of these 8 studies used CT as a diagnostic test, 10 used ultrasound, 11 used MRI and 4 studies made use of AFP. 13 studies used more than one diagnostic test. None of the studies were assigned to a higher evidence level than 2B. In many cases the reasons for downgrading in quality were the lack of information about the independence of the reference standard as well as about the handling of inconclusive test results. The bivariate analysis of ultrasound (7927 patients) resulted in a sensitivity of 66.2% (CI 54.0% - 76.5%) with a related specificity of 95.7% (CI 91.1% - 98.0%). Hence, the cost-efficient and fast method of US provides a satisfying diagnostic accuracy and is therefore convenient as a screening method in everyday clinical practice. For the analysis of computed tomography (1578 patients) the analysis showed a sensitivity of 72.2% (CI 63.5% - 79.5%) and a specificity of 90.3% (CI 84.8% - 93.9%). The subgroup analysis did not display any advantages for studies with a better level of evidence or a higher impact factor, but it also showed good results for each subgroup. The best results in the bivariate analysis were found for the magnetic resonance imaging (882 patients): sensitivity 82% (CI 71.9% - 89%), specificity 90.3% (CI 83.5% - 94.4%). With a specificity value similar to the one of the CT examination, MRI shows a considerably higher sensitivity value. The subgroup analysis revealed equally good results in each subgroup, however there were no divergence from the overall result. The bivariate analysis of the AFP diagnosis (2154 patients) with a cut-off value of >200 ng/ml resulted in a sensitivity of 39.6% (CI 23.9% - 57.8%) and a related specificity of 98.9% (CI 97.9% - 99.4%). As a very unspecific diagnostic marker the relatively poor benefit of AFP for the clinical practice to diagnose HCC was confirmed. For the diagnosis of HCC, good diagnostic accuracies are found particularly for CT and MRI examinations. Their advantages are limited by these procedures being costly and time-consuming, furthermore there is a significant radiation exposure to be considered for the CT examination. For the screening of patients at risk of developing HCC, the ultrasound imaging therefore still plays an important role. Its significance can be enhanced by combining it with the measurement of the biomarker AFP. For the exact comparison of the different diagnostic imaging techniques, well conducted diagnostic studies of high quality will be needed in the future.