NEW YORK (360Dx) – Researchers at South Korea's Seoul National University College of Medicine have developed a mass spectrometry-based assay for measuring the liver cancer biomarker lens culinaris agglutinin-reactive fraction of-fetoprotein (AFP-L3).
Described in a paper published this month in Clinical Chemistry, the assay offers improved sensitivity compared to existing tests for the marker, which could make it more useful for diagnosing and monitoring liver cancer, said Youngsoo Kim, professor of biomedical engineering at SNUCM and senior author on the study.
As the authors note, AFP is a marker for hepatocellular carcinoma, but it has limited clinical utility due to its poor sensitivity. Researchers have looked into whether measuring specific glycosylated forms of AFP might improve the marker's performance and have found that AFP-L3 is more effective for early detection of HCC than is total AFP.
The standard method for measuring AFP-L3 is liquid-phase binding assay (LiBA), with a US Food & Drug Administration-cleared LiBA test available from Wako Diagnostics. LiBA testing for AFP-L3 is hampered by poor sensitivity, however, which makes false negative results an issue.
To address this challenge, Kim and his colleagues turned to multiple-reaction monitoring mass spectrometry with the aim of developing an MRM-MS test that could improve upon the sensitivity of the LiBA method.
The test measures AFP-L3 in patient serum, using a monoclonal antibody to AFP to enrich the target protein followed by use of LCA lectin to separate out the AFP-L3 fraction of the sample. The isolated molecules were then deglycosylated and digested with trypsin. The sample was then separated via liquid chromatography using an Agilent 1260 Capillary LC system and the levels of two peptides, one a surrogate for total AFP and the other a surrogate for AFP-L3, were quantified using MRM on an Agilent 6490 triple quadrupole instrument.
Comparing the sensitivity of the MRM assay to the LiBA test, the researchers found that the former had a lower limit of quantification for AFP-L3 of 0.051 ng/mL versus 0.3 ng/mL for LiBA.
They also compared the performance of the tests in patient samples, looking at their abilities to distinguish patients with HCC from patients with other liver diseases (chronic hepatitis and liver cirrhosis). Looking at samples from 100 patients with hepatitis, 100 patients with cirrhosis, and 200 with HCC, the researchers found that the MRM assay distinguished between the HCC and non-HCC patients with an area under the curve of .85, sensitivity of 81 percent and specificity of 89.5 percent. The LiBA test had an AUC of .77, sensitivity of 61.5 percent, and specificity of 90 percent. The MRM test also detected AFP-L3 in 39 HCC samples where levels were below the limit of detection of the LiBA test.
Kim noted that several other groups have previously developed mass spec assays to measure AFP-L3 but said that those assays were still mainly at the research stage.
"Our mass spec assays for AFP-L3 is almost ready to be applied to the clinical laboratory," he said.
Kim cited several aspects of the assay that contribute to its suitability for the clinic. One key component is the use of a full-length stable isotope-labeled protein as an internal standard, as opposed to SIL peptide standards. Generally, thinking among experts in mass spec-based protein quantitation is that full-length peptides provide better controls analytical variation as they are able to control for variation within sample prep processes like protein digestion.
Another important element is the use of an "addition only" workflow, which can be easily adapted to automated liquid handling systems, which are key to enabling high-throughput, highly reproducible mass spec assays.
In measurements taken in four QC samples over the course of six days, the total coefficients of variation for the assay ranged between 4.5 percent and 12.3 percent. As performed in the Clinical Chemistry paper, the assay required around 12 hours for sample prep and around 12 minutes per sample for the LC-MS run, indicating that it could be run at the sort of throughput necessary for a clinical test.
Kim said the he believed that in addition to being more sensitive, the MRM-MS test would likely be less expensive than the LiBA test, though he said that it was still to early in the assay development process to do a thorough cost comparison.
He noted that he has no plans to commercialize the assay himself but said that he would be interested in licensing the test to an outside company.