NEW YORK (GenomeWeb) – Through a genomic characterization of brain metastases and their paired primary tumors, a Harvard Medical School-led team of researchers has found that, while the pairs are genetically related, they also differ in potentially clinically actionable ways.
The researchers performed whole-exome sequencing on 86 matched primary tumor biopsies, brain metastases, and normal tissue, as they reported this weekend in Cancer Discovery. Through phylogenetic analysis, they found that the primary tumors and brain metastases shared a common ancestor, but then evolved independently. But, for 53 percent of patients, their brain metastases had genetic changes not present in the primary tumor that could be targeted by drugs.
"We found genetic alterations in brain metastases that could affect treatment decisions in more than half of the patients in our study," first author Priscilla Brastianos, a neuro-oncologist at Massachusetts General Hospital, said in a statement. "We could not detect these genetic alterations in the biopsy of the primary tumor. This means that when we rely on analysis of a primary tumor we may miss mutations in the brain metastases that we could potentially target and treat effectively with drugs."
About 200,000 cases of brain cancer metastases are reported each year in the US, with 8 percent to 10 percent of cancer patients, especially melanoma and lung or breast carcinoma patients, developing such metastases. More than half of these patients die within a few months of receiving brain metastasis diagnosis.
Brastianos and her colleagues performed exome sequencing on 86 matched primary tumor biopsies, brain metastases, and normal tissue sample trios. These samples were from patients with lung, breast, and renal cell carcinoma primary tumors. For eight cases, they also sequenced samples from regionally and anatomically separated brain metastasis sites, regional lymph nodes, and distal extracranial metastases.
Based on the SNVs and CNVs the researchers uncovered, they constructed phylogenetic trees to examine the evolutionary relationship between the various clones from each patient.
Each brain metastasis, the researchers reported, likely developed from a single clone genetically related to the primary tumor, though both then continued to evolve independently.
For instance, in one patient with a primary renal cell carcinoma, the researchers uncovered a homozygous PTEN nonsense mutation in the brain metastasis that was lacking in the primary tumor. A PTEN mutation, the researchers noted, could indicate sensitivity to PI3K/AKT/mTOR inhibitors.
Similarly, the researchers reported that the breast cancer and brain metastasis samples from another patient shared an ERBB2 amplification as well as a homozygous deletion of TP53. However, the primary tumor harbored a MYC amplification the brain metastasis lacked, and the brain metastasis had a homozygous missense BRCA2 mutation of unknown significance that the breast tumor didn't have.
Brastianos and her colleagues also found potentially clinically actionable mutations in 53 percent of cases. By drawing on the TARGET database of gene variants with clinical implications, the researchers found that 330 of the total 95,431 gene alterations in their dataset could be clinically actionable. Some 46 of their 86 cases had such actionable mutations in their brain metastases, but not in their primary tumors.
Common alterations indicated sensitivity to cyclin-dependent kinase inhibitors, HER2/EGFR inhibitors, and mutations affecting the PI3k/AKT/mTOR pathway, they noted.
Some of the patients in the cohort had multiple brain metastases, and these additional brain metastases — even if regionally and anatomically separated — were genetically homogenous, the researchers reported, and shared nearly all of the variants detected.
For example, Brastianos and her colleagues sampled four regions of a brain metastasis from a patient with HER2-amplified breast cancer and found that each of those regions shared a PIK3CA mutation and ERBB2 amplification with the primary tumor. The metastatic regions also, though, had CCNE1 and EGFR amplifications not present at the primary site. This patient, the researchers added, received a PI3K inhibitor and didn't show evidence of disease progression for eight months.
Similarly, some patients also had distal extracranial metastases or regional lymph node metastases, and these, the researchers noted, were typically divergent from the brain metastases.
This, the researchers said, indicates that biopsies of primary tumors, and even of lymph node or other distal extracranial metastases, don't capture the full suite of clinically actionable mutations that may be present in associated brain metastases.
"When brain metastasis tissue is available as part of clinical care, we are suggesting sequencing and analysis of that sample. It may offer more therapeutic opportunities for the patient," Brastianos noted. "Genetic characterization of even a single brain metastasis may be superior to that of the primary tumor or a lymph node biopsy for selection of a targeted treatment."