SAN ANTONIO – The development of a serum glycoprotein microarray has led to the discovery of four proteins for which antibodies significantly predict nodal metastases in patients with melanoma.
If validated, the data could form the basis of a blood test to select patients for sentinel lymph node biopsy, Dr. Michael Sabel said during a symposium sponsored by the Society of Surgical Oncology.
In an internal validation cohort of 79 patients with newly diagnosed melanoma, 22 patients (28%) had serum antibodies to acid ceramidase (ASAH1), cathepsin D (CTSD), or lactate dehydrogenase B (LDHB) that were significantly associated with being node negative (P value less than .0001; hazard ratio, 0.173).
"This isn't a small percentage of patients, so if validated, this test could significantly reduce the number of patients to whom we offer sentinel node biopsy," he said in an interview.
Antibodies to a fourth protein, GRP94 (also known as heat shock protein 90), were detected in nine patients (11%), and were significantly associated with being node positive (P less than .0001; HR, 3.04).
"Theoretically, this could be used in patients who are on the borderline for the procedure, where the presence of antibodies to GRP94 could prompt you to use sentinel node biopsy," Dr. Sabel said.
Surprisingly, the presence of antibodies to GRP94 did not correlate with the volume of disease burden within the sentinel node. The antibodies were equally present in patients who had microscopic disease and in those with more significant tumor burden within their sentinel node.
The bigger picture, however, is the potential for glycoprotein microarray analysis to identify predictive and prognostic biomarkers in melanoma, suggested Dr. Sabel, a surgical oncologist with the University of Michigan Health Systems in Ann Arbor.
Studies are ongoing to find other antibodies and their relevant glycoproteins that may differentiate benign from malignant skin lesions, identify patients at high risk of distant metastases, and signal response to adjuvant therapies.
"Glycoproteins have apparently important roles in melanoma progression, and also may serve as targets for therapy," he said.
The use of serum autoantibody profiling would be particularly attractive in clinical practice because it does not require primary tumor tissue. Unlike most other solid tumors, roughly 90% of a primary melanoma tumor is removed during the initial biopsy, leaving little tissue for subsequent testing once the diagnosis has been made, Dr. Sabel said.
Sentinel node biopsy with selective lymph node dissection was endorsed in the latest American Joint Commission on Cancer staging guidelines for melanoma, but it is positive in only 20% of patients with melanoma, he observed.
The researchers studied glycoproteins rather than a broad array of proteins because glycoproteins have a strong interaction with the immune system. This is particularly advantageous in melanoma because the interaction between the immune system and melanoma is well documented. In addition, glycoproteins undergo distinct changes during cancer progression, including precancerous states and posttranslational modifications that are associated with disease progression, Dr. Sabel explained.
In order to perform serum autoantibody profiling, the researchers used dual-lectin affinity and reverse-phase chromatography to separate out glycoproteins from a whole cell lysate created from a metastatic melanoma cell line. The glycoproteins were spotted on a microarray and tested using serum from 43 melanoma patients, with anti–human IgG used to detect response, Dr. Sabel said.
Wilcoxon rank-sum testing and outlier analysis identified nine fractions that significantly separated 27 node-negative patients from 16 node-positive patients. Mass spectrometry was then used to identify the relevant proteins of interest. Prevalidation testing with recombinant proteins boiled down the nine fractions to the four proteins used in the validation set of 79 patients.
Interestingly, although LDHB, CTSD, and GRP94 have strong associations with melanoma, the ASAH1 protein has been associated with breast, thyroid, and prostate cancers, but not melanoma, Dr. Sabel said.
In multivariate analysis of the validation cohort, there was no significant correlation among the four proteins or between the proteins and known melanoma prognostic factors. Only a mild, negative correlation was observed between ASAHI and increasing age, he said.
All four antibodies remained significant after adjustment for age, sex, Breslow thickness, ulceration, and mitotic rate, with a receiver operating characteristic area under the curve of 0.8690.
Dr. Sabel disclosed research funding from Merck Oncology. His coauthors report no conflicts.