The combination of AFP and Des-gamma carboxy-prothrombin yielded only a slight increase of sensitivity to 70%.[22] In clinical practice, AFP is considered relatively insensitive in the detection of HCC recurrence, with sensitivities ranging from

41% to 65%.[10, 23-26] This low sensitivity can be explained in part by the predominance of non-AFP-producing HCC. In our study, AFP only has 12% sensitivity for recurrent NVP-AUY922 purchase non-AFP-producing tumors. Despite its low sensitivity, AFP is still being utilized for HCC surveillance and recurrence detection. It currently serves as a complementary test to imaging studies because it is simple, inexpensive and widely available. However, the results of our study suggest that measuring serum AFP in non-AFP-producing HCC may not be helpful and should be omitted

because of its poor sensitivity. In contrast, AFP is still highly sensitive and specific in the detection of recurrent AFP-producing HCC and therefore has great clinical potential. These findings support previous studies, which have shown that AFP is useful in detecting early HCC recurrence in patients who have high pretreatment AFP values (AFP-producing HCC).[3, 4] Liver inflammation, as indicated by elevated serum ALT, is a condition which can cause non-specific AFP elevation thereby contributing to false positive HCC recurrence. Interestingly, the majority of these false positive AFP elevations in our study were within 100 ng/mL (IQR = 30–102 ng/mL) while the AFP level from true HCC recurrence were usually more than 100 ng/mL Lapatinib cost (IQR = 171–2261). Therefore, raising the AFP cutoff to 100 ng/mL for both the pretreatment AFP and HCC recurrence criteria in cases with abnormal ALT levels has dramatically increased sensitivity, specificity, and accuracy to 100%, 84.6% and 89.2%, respectively (Table 6). The improved performance of AFP by the modified criteria may 上海皓元医药股份有限公司 be explained by elimination of false negative HCC cases which were erroneously categorized as AFP-producing HCC,

and the elimination of some of the false positive cases which had abnormal ALT levels at the time of HCC recurrence. However, the trade-off for this adjustment is the number of eligible HCC was decreased to 37 cases (25%). Normal laboratory AFP cutoff in the normal population is less than 10 ng/mL; this cutoff can be applied only to patients with normal liver status.[27] Many studies have suggested an optimal AFP cutoff value of 20 ng/mL helps to reduce false positive rates in patients with cirrhosis and underlying liver disease.[5, 10, 28] However, our study shows that in patients with active liver inflammation (abnormal ALT) at the time of serum AFP measurement, the AFP cutoff should be further increased to distinguish false positives from active liver inflammation. Our results shows that the optimal AFP cutoff value of 100 ng/mL resulted in high sensitivity, specificity and accuracy.