There is seemingly no shortage of markers for cancer cells in vitro. Every week brings a new antigen, genotype, morphology, or molecule for identifying some particular cancer cell line or behavior. The tools available for identifying additional markers are also improving.
High-throughput assays such as next-generation sequencing, cDNA and protein microarrays, and ELISA/purification-assisted mass spectrometry provide a better look into the molecular world of cancer biology. At the same time, computing resources permit analysis across more data sets (different cell lines) and data types (proteins, mRNA, epigenetic modifications, and genomes), generating even more hits.
Cancer Then and Now
While the equipment and procedures for identifying these hits was improving, so too did the understanding of cancer behavior. The immense complexity and diversity of cancer was highlighted as more genotypes and phenotypes were found in patient cells.
Tumors, once thought to be monocultures of malignant cells, actually turned out to be micro-ecosystems consisting of several populations of cells including cancer stem cells. The concept of the tumor microenvironment came to the forefront, and models of metastasis changed as nonmetastatic populations of circulating tumor cells were found.
This new understanding of cancer is immensely helpful in further research, but it does not bode well for individual use of biomarkers. The complicated behavior of so-called cancerous cells has forced questions of how to precisely define cancerous for the purposes of screening, diagnosis, and treatment.
In an editorial in the May 5, 2010, edition of the Journal of the National Cancer Institute, Drs. Laura Esserman and Ian Thompson discuss overdiagnosis of cancer: They suggest that, considering approximately 75% of biopsies produce negative results, more focus should be given to distinguishing aggressive cancers from indolent ones rather than widespread nonspecific screening.
One much-maligned biomarker serves as an excellent example: prostate-specific antigen (PSA). Both the American Cancer Society and the NCI admit that the harm of overtreatment and cost of doing PSA screening outweigh the benefits of the test.
A Plethora of Proteins
It is the complex interplay of normal and cancerous cells that defines future research into biomarkers for cancer progression. The search for a single biomarker to measure or even detect the presence of cancer is over.
Clinical approaches haven’t yet caught up, though, diagnosis and measurement of progression use a multimodal approach, these modes are generally limited to imaging, biopsy, and a few biomarkers. More robust measurements, especially noninvasive types integrating data from several biomarkers, are trickling onto the market.
In late 2009, the FDA approved OVA1 (Vermillion and Quest Diagnostics), a 5-protein serum assay for ovarian tumor malignancy. Genetic tests for risk factors have become more commonplace, and fluorescence in situ hybridization (FISH) tests exist for tumors to determine mutations and thus treatment susceptibility.
The ultimate biomarker-based assay for cancer diagnosis and progression would not only measure circulating tumor cells (as existing product CellSearch, from Veridex and Quest Diagnostics, does) but also check the composition of those tumor cells; certain mutations turn circulating tumor cells into cells with complete capabilities for founding new tumors elsewhere.
Cancer is also an opportunistic disease. Indicator compounds such as abnormal protein levels and metabolites are important but so are indicators of normal function. By connecting large pools of data (as microarray and other multiplex assays become cheaper and easier), a fuller, more individualized picture of cancer pathology can be generated, and treatments can be appropriately tuned.
A treatment that shrinks a tumor but fails to kill its stem cell population will lead to eventual recurrence but may still be useful in making those cancer stem cells more vulnerable to other agents. If your test shows that those stem cells are still alive (or worse, circulating), then different treatments can be applied.
In essence, the rise of highly-multiplexed assays makes single biomarker tests outdated, and makes less-specific/less-sensitive biomarkers more useful as part of a portfolio to account for the myriad ways in which cancer can manifest itself in the body. A 2008 study in Clinical Cancer Research described how measuring six proteins could provide an ovarian cancer assay with 95% sensitivity and 99% specificity, versus 72% sensitivity and 95% specificity for one protein alone.
By improving sensitivity, the fraction of false negatives decreases; by increasing specificity, the fraction of unnecessary biopsies falls as well. It is time for us to use all of our molecular tools side-by-side in diagnosing and measuring the progression of cancer.