Houston, TX—Despite much talk about biomarkers and a field that is exploding, only a few biomarkers can be reliably and routinely used to improve patient care at this time, according to Peter G. Ellis, MD, Deputy Director of Clinical Services, Associate Chief Medical Officer, University of Pittsburgh Medical Center Cancer Centers.
A biomarker is any measure (ie, gene mutation, staining pattern, gene expression microarray) that can be associated with a clinically distinct prognosis, diagnosis, or response to a specific therapy.
“Biomarkers hold promise, and we are clearly making progress, but I say, ‘Not so fast!’ Paradoxically, they present challenges that we must deal with going forward, such as obtaining clinical trial data to prove they actually improve patient care,” said Dr Ellis, speaking at the Second Annual Conference of the Association for Value-Based Cancer Care.
Several biomarkers represent the enhanced understanding of cancer pathogenesis and have led to some remarkable therapeutic advances. These include imatinib in chronic myeloid leukemia, all-trans retinoic acid in acute promyelocytic leukemia, ALK in lung cancer, epidermal growth factor receptor in lung cancer, and BRAF in melanoma, Dr Ellis said.
“But we need to be cautious about which biomarkers are reproducibly predictive of either prognosis or response before we consider them mainstream or primetime,” he suggested. “They may do more harm than good.”
A Google search of “biomarker company” by Dr Ellis brought up almost 1.8 million results. These revealed multiple “omics” (ie, genomics, proteomics) companies with products that are deserving of further investigation, but the products lack sufficient data to be considered standard of care. Further research and expert consensus will be required for the products to be found useful, he said.
Examples of promising markers that require further testing are ERCC1 (which indicates resistance to platinum agents in lung cancer), thymidylate synthase (TS; high TS is associated with lack of response to 5-fluorouracil, capecitabine, and pemetrexed), and SPARC (secreted protein, acidic and rich in cysteine; high SPARC is associated with response to nab-paclitaxel), according to Dr Ellis.
The use of the ERCC1 marker, to spare patients with lung cancer the possibility of futile platinum treatment, is heavily promoted, but because these agents are the mainstay of treatment for lung cancer, Dr Ellis said he would not take a chance by eliminating them in his patients with an ERCC1 mutation before their use is proved.
In fact, a recent investigation points to a potential flaw in the whole biomarker landscape (Gerlinger M, et al. N Engl J Med. 2012;366:883-892). Multiregion sequencing of metastatic renal-cell carcinoma revealed high intratumor heterogeneity; some regions of a single tumor harbored genes suggestive of a good prognosis, whereas other regions expressed genes that suggested the opposite.
“How do we know [that] what we biopsy and the markers we obtain are actually representative of the tumor?” Dr Ellis asked.
Before clinicians begin using a host of biomarkers now being heavily marketed, they should be proved valid in clinical trials. Once that occurs, they should be incorporated into decision-support instruments, Dr Ellis concluded.
