Gene Methylation Patterns Predict Early-Stage NSCLC Recurrence

By Michael Smith
BALTIMORE, March 12 -- A group of four genes can be used to predict the risk of recurrence of stage I non-small-cell lung cancer, researchers here said.The finding could change how physicians stage such tumors, according to Malcolm Brock, M.D., of the Johns Hopkins Kimmel Comprehensive Cancer Center, and colleagues.Treatment for patients with early-stage non-small-cell lung cancer is surgery with curative intent, Dr. Brock and colleagues noted, but up to 40% of patients with discrete lesions and histologically negative lymph nodes (T1-2N0) die of recurrent disease.The current research is based on a retrospective, single-institution study and needs prospective confirmation before it moves to the clinic, Dr. Brock and colleagues said in the March 13 issue of the New England Journal of Medicine.
Nevertheless, the method -- which relies on detecting methylation of the promoter regions of the four genes -- opens the door to what the researchers called "molecular staging" of tumors.
It seems likely that the surgery misses micrometastases beyond the margins of the resection, the researchers said, and knowing whether they exist might alter the way patients are treated.
The technique can be thought of as "DNA forensics" for cancer, Dr. Brock said.
"Much as forensic sciences have been revolutionized by DNA, it is possible that we now in medicine are in a phase of showing that DNA can contain information that the microscope is not picking up," Dr. Brock said.
The researchers tested the association between methylation of the promoter regions of seven genes -- essentially a sign they're turned off -- and the risk of recurrence within 40 months of surgery.
Dr. Brock and colleagues began with all 715 patients who had lobectomy or greater resections for T1-2N0 cancer at the Johns Hopkins Hospital between Jan. 1, 1986, and July 31, 2002.
Of those, 71 had recurrence within 40 months and for the first phase of the study, the researchers compared tissue samples from 51 of them to samples from 116 matched controls whose cancer did not recur.
Analysis of methylation patterns narrowed the seven original genes down to four that had the largest differences compared to controls -- the cyclin-dependent kinase inhibitor 2A gene p16, the H-cadherin gene CDH13, the Ras association domain family 1 gene RASSF1A, and the adenomatous polyposis coli gene APC.
Methylation of p16 alone in the primary tumor, regional nodes, or mediastinal lymph nodes, the researchers found, was significantly associated with an increased risk of recurrence, at P=0.001, P=0.008, and P=0.007, respectively.
Methylation of CDH13 in mediastinal lymph nodes but not elsewhere was also associated with a significant elevation in the risk (at P=0.02), while the other two genes were associated with non-significant elevations in risk, when they were looked at separately.
Treated as pairs, however, all of the genes had some predictive power.
For example, methylation of p16 and either CDH13, RASSF1A, or APC in both tumor and mediastinal lymph node samples were associated with early recurrence -- on average at nine months after surgery.
In the absence of methylation of those paired genes, the median time to recurrence was 25 months after surgery -- a difference that was significant at P=0.04.
The most compelling pair was p16 and CDH13, the researchers said.
When both were methylated in the primary tumor alone, the odds ratio for recurrence within 40 months was 8.00 (95% CI 2.50 to 25.51), compared with controls. The difference was significant at P<0.001.
When methylation of both genes was seen in the mediastinal lymph nodes as well as the primary tumor, the odds ratio for recurrence soared to 15.50 (95% CI 1.61 to 185.02). The increased risk was significant at P=0.03.
If the results are confirmed, Dr. Brock said, doctors might want to reclassify cancers with such methylation patterns as advanced, rather than early stage.
Co-author James Herman, M.D., also of Hopkins, said the finding may lead doctors to consider treating high-risk patients more aggressively with chemotherapy after surgery.
Another possibility, Dr. Herman said, is that stripping off the methyl groups hold therapeutic promise. "These marks of aggressive disease also are themselves targets for therapy," he said.
The study was supported by the National Cancer Institute, the Commonwealth Foundation for Cancer Research, the Hodson Trust, and OncoMethylome Sciences. Dr. Brock and Dr. Herman reported financial links with OncoMethylome Sciences, with holds the license for a test of methylation.
Primary source: New England Journal of MedicineSource reference:Brock MV, et al "DNA methylation markers and early recurrence in stage I lung cancer" N Engl J Med 2008; 358: 1118-28.