SAN FRANCISCO—A new gene expression profile (GEP) test more accurately predicts the risk of metastasis and survival in patients with cutaneous melanoma compared with the standard predictive method currently used to guide patient management, researchers reported here at the American Academy of Dermatology Annual Meeting. With more testing, they and other experts said, it may allow oncologists to know which patients need early treatment to prevent disease progression.
“There is great variability in prognosis in stage 2 melanoma patients, and these patients benefit from increased surveillance or better treatment of early metastatic disease. The value of the GEP test to an oncologist is to know which patients to worry about, and to treat these patients more carefully,” the lead author, Laura Ferris, MD, PhD, Director of Clinical Trials in the Department of Dermatology and Associate Professor of Dermatology at the University of Pittsburgh, said in an interview.
“The ability to accurately predict the risk of metastasis and death among patients with cutaneous melanoma is of the utmost importance in guiding clinical decision making. Together with the previously published data, our study shows that the GEP test can help to identify high-risk patients.”
She presented the results of a study that showed that the noninvasive, 31-gene GEP test, DecisionDx-Melanoma, performed significantly better than the American Joint Committee on Cancer (AJCC) Individualized Melanoma Patient Outcome Prediction Tool in predicting five-year distant metastasis-free and overall survival.
The GEP test stratifies patients into low (class 1) or high (class 2) risk, and provides a linear probability score from 0 to 1 associated with the similarity of the biological signature to an established and clinically useful training set.
She and her co-researchers examined the records from 355 node-negative melanoma patients who met AJCC criteria. As a means of comparing the GEP and AJCC tests, five-year survival rate cutoff scores of 79 percent and 68 percent were used. The cutoffs reflect five-year survival rates for Stages IIA or IIB melanoma patients, which can result in significantly different treatment strategies, according to National Comprehensive Cancer Network guidelines.
“The GEP test predicted melanoma risk more accurately than, and independent of, the AJCC online prediction tool,” Ferris said, noting that 15 patients who were predicted to be low risk in AJCC were class 2, and died of melanoma. Only two patients predicted to be class 1 in the GEP test were high risk in AJCC.
“Most patients who are high risk in the AJCC criteria are likely to be class 2 with the GEP test, and those who are low risk in the AJCC criteria are likely to be class 1 in the GEP test,” she said. “The GEP test is more likely to predict metastatic melanoma within five years.”
Patients had slightly more than twice the risk of distant metastases-free survival if they had a high-risk GEP score compared with a high-risk AJCC score. For overall survival, patients were 1.5 times more likely to have a poor outcome if they had a high-risk GEP than a high-risk AJCC score.
‘Needs More Testing’
Asked for his opinion, though, Klaus J. Busam, MD, Director of both the Dermatopathology Service and the Dermatopathology Fellowship Program of Memorial Sloan Kettering Cancer Center, cautioned: “The GEP has not yet been adequately tested for outcomes prediction. I welcome efforts to improve prognostication, and it is a very good start, but we have only a few hundred cases of the GEP test. The AJCC has 60,000 cases. We need good prospective studies of the GEP test.”
The study's first author, Pedram Gerami, MD, Associate Professor in Dermatology, Pathology, and Pediatrics-Dermatology at Northwestern University Feinberg School of Medicine, echoed his coauthor: “The GEP test is an independent predictor of metastasis for stage 1 or 2 melanoma and is a stronger predictor of metastatic recurrence than AJCC stage.”
Second Validation Study
In a separate session at the meeting, Gerami presented the results of a second validation study of the GEP test to predict the metastatic risk associated with cutaneous melanoma. (A first validation study had found that the GEP test was an objective tool that accurately predicted metastatic risk in sentinel lymph node [SLN] biopsy-eligible patients in a multicenter cohort of 217 melanoma patients, and found a five-year disease-free survival rate of 97 percent for GEP class 1 patients versus 31 percent for GEP class 2 patients.)
The second validation study found that for every category, there was a bigger separation of class 1 versus class 2 compared with SLN biopsy positive or negative patients, he said. The negative predictive value was higher with the GEP test (79%) than with SLN biopsy (56%).
“If I tell a melanoma patient that he is GEP class 1, that's more reassuring than telling him he has a negative SLN biopsy. For GEP class 2 patients it does not matter if the SLN is positive or negative. The overall survival rate is about the same,” Gerami said.
About one-third of the patients were GEP class 1 and had a negative SLN biopsy; they had an overall five-year disease-free survival rate of 83 percent. In a small cohort of patients who had a GEP class 1 and positive SLN biopsy, the overall five-year disease-free survival rate was 53 percent.
Gerami admits more studies are needed to verify the results of the GEP test, which has been available for more than a year.
“The main indication for the GEP test is to identify node-negative patients who may have other tumor characteristics that are worrisome for high-risk disease. A huge proportion of patients who die from melanoma are SLN biopsy negative. We have a lot of preventive therapies available. We can inform patients who are class 1 on the GEP test that it is likely they do not have metastatic disease, and for GEP class 2 patients, we can increase their imagining schedules. Why not consider adjuvant therapy for stage 2 patients even if we do not find melanoma?”
‘Use in Combination with SLNB’
Based on data from this study and the prior validation study, Gerami said he and his colleagues propose that the GEP test be used in combination with SLN biopsy to help clinicians further stratify patients as higher versus lower risk. The combined use of these two prognostic tests may identify more than 80 percent of patients who are at risk for metastasis compared with SLN status alone, providing a significant opportunity for improved patient treatment, he said.
Still, Busam expressed concerns: “The biggest problem I have with the rollout is the proposed value of the GEP test—that is, to pick up patients who will benefit from SNL biopsy. SLN biopsy is a surgical prognostic test. If we want a good prognostic test, it should help us avoid SLN biopsy. If the company trusts class 1 results so much, then the class 1 patient does not need SLN biopsy. If it is better than SLN biopsy, why use the test to recruit patients to have surgery to give a test result that is inferior to the one you propose? I'm having trouble designating clinical value to the test.
“I'm all for using any means to improve tumor prognostication to avoid the need for surgery, but it is not therapeutic and does not help survival,” he continued.
“The GEP test has prognostic value, mostly if a patient participates in a clinical trial. For personal life planning and treatment purposes, I wish it was robust enough to avoid SLN biopsy. If we want to upstage patients and have more scans—say, five times a year instead of once a year—what are the scans really doing? They are not saving lives. I caution against using the test if a patient is class 2. The patient may become depressed and have unnecessary tests for years to come that may not be beneficial.
“We already have ways to improve prognostication. We can add in clinical information, get tumor-infiltrating lymphocyte levels, and mutational status of NRAS and BRAF, which are targets that can be treated. The GEP test should also be measured against less expensive parameters that are known to make prognostication of a primary tumor more accurate. Don't forget the power of visual recognition. The microscope is still a great tool.”
Also asked for his opinion, John Votto, MD, Professor of Surgery at Oregon Health & Science University, generally agreed: “We need a test about how and when a patient will fail. Can we specify a gene profile for a pattern of failure? We need to use genetics to determine when a patient should have SLN biopsy or go straight to systemic therapy, or have both. These data are early and exciting, but are only on a few hundred patients. We need confirmation in a prospective trial.”
The next step, Ferris said, is to use the GEP test in the treatment of patients with high-risk, stage 2 melanoma. “Most patients in our cohort had stage 2 melanoma and were GEP class 2. Some patients with stage 2 melanoma do well, showing improved survival with targeted therapies.”
There are more false positives with the GEP test than with SLN biopsy. “We speculate that when we find early melanoma, some lesions are thinner, less aggressive tumors and others are thicker, more aggressive ones. There is a window of cure. I predict the GEP test has the potential to catch melanomas in the window of cure before they metastasize,” she said.
Outside of the research setting, Ferris said she has used the GEP test on 10 melanoma patients who have thinner, 0.5 to one millimeter lesions. The patients she has selected for the GEP test “have an increased mitotic rate and evidence of ulceration, and are questionable candidates for SLN biopsy. Others are patients who want SLN biopsy, but do not meet the criteria. The GEP test can assure these patients they do not need an additional procedure. If patients have very deep, four to five millimeter lesions, the GEP test is unlikely to be necessary since we know that these are high-risk patients.
“There is no good consensus on how to deal with intermediate-risk melanoma,” she continued. “We need a consensus statement on how the GEP test fits in. I believe it is predictive. Most SLN biopsies are negative. If a patient is class 1, the chance of a negative SLN biopsy is low. Then we would only do SLN biopsies on class 2 patients. This is reasonable, but requires the entire melanoma community to come together. This would save a lot of procedures and money.”
Cost is also a factor, she noted: A sentinel lymph node biopsy costs $12,000 to $15,000 versus $3,000 to $7,000 for the GEP test. Some insurance companies pay a reimbursement rate of $1,500 for the GEP test, and if insurance does not cover the test, the developer, Castle Biosciences, told OT that it will not go to the patient for payment.