We are requesting the reference set, which includes 50 HCC cases and 50 cirrhotic
controls. In our preliminary study, AFP had a AUROC of 0.66 while the AUROC for the 5
glycoproteins was 0.81. The sensitivity and specificity for the 5 glycoproteins was 79% and
72% at the point that maximizes sensitivity+specificity in the ROC curve, and it was 79%
and 35%, respectively, for AFP at the same point in the ROC curve. The reference set will
allow us to determine the best performance of the 5 glycoproteins by themselves or whether
their combination has a better sensitivity and/or specificity and AUROC. While a direct
comparison with AFP will be made, the reference set will not allow a robust comparison due
to the low sample size. If the glycoproteins are complementary or have better performance
than AFP, then the next step would be to test them in the entire phase 2 hepatocellular
In this study, we used the χ2 and t test to compare the distribution differences of demographic and clinical variables among the two groups of cases and controls for discrete and continuous values, respectively. Spearman correlation test was applied to determine correlations of the tested values from different markers as well as the processing platforms. To justify performances for HCC early detection, the curves of receiver operating characteristic (ROC) were first plotted for all these markers. The area under the ROC curve (AUC) and the corresponding 95% confidence intervals (CI) were used to judge their overall performances. In addition, the logistic regression models with the forward selection were used to know if and how these biomarkers could be combined together to predictive positive responses. For the purpose to estimate their best performances, the potential cutoff values were proposed as all the empirical factor levels and their combination linear scores. The maximum sum value of sensitivity and specificity crossing different cutoff values was then employed to set up the optimal cutoff value. With the optimal cutoff value, the sensitivity, specificity and their 95% CIs were further calculated to contrast the resulting performances for these tested biomarkers
There are currently no biomarkers annotated for this protocol.
No datasets are currently associated with this protocol.
Thank you to everyone who helped make the 27th EDRN Steering Committee Meeting a success. We look forward to seeing everyone at the 9th EDRN Scientific Workshop from September 8-11, 2014 in Washington D.C.