Multiple biomarkers in molecular oncology

Several years ago the Director of the National Cancer Institute, Andrew von Eschenbach, put forth the goal of eliminating the suffering and death that results from cancer by as early as 2015. This goal did not explicitly state that a cure for cancer would be found, only that suffering and death resulting from this condition would be eliminated. That goal may best be accomplished by intervening at a number of phases of the disease. The most obvious phase, which probably has the greatest impact on a cancer patient’s quality of life, is early detection. Early detection is a key determinant in the survival rate of cancer patients, affords the greatest number of treatment options for the oncologist and, in general, the best outcomes for patients. Treatment represents another aspect of intervention that can impact a patient’s survival. The most appropriate treatment given at the proper time is also a major factor on the patient’s survival and quality of life. Unfortunately, there is no universal treatment for cancer and, too often, the treatment assignment for an individual patient is categorically determined.

Why has progress for improved patient outcomes significantly lagged in oncology versus heart disease and HIV/AIDS? Since 1950, the death rates due to coronary heart disease for patients aged 40–60 years have improved dramatically. In contrast, there has been minimal improvement in patient outcomes for cancer. Significant progress in oncology has revolved around supportive care measures; specifically, regimens focused on improved pain control, antiemetics and hematologic growth factors. These advances have transformed the practice of oncology from the hospital to an outpatient/community setting. Over the past 20 years, biomedical science has transformed HIV/AIDS from a killer with defined certainty, to a chronic disease that can be effectively managed. An argument can be made that improved outcomes in coronary heart disease and HIV/AIDS are due to effective biomarkers that enable adequate tailoring of care on a patient-by-patient basis.

Biomarker use in clinical cardiology is now incorporated into guidelines. These mandate the use of a variety of biomarkers and biomarker panels for both risk assessment and management. For example, cholesterol panels are used for both risk assessment and coronary disease management. Risk assessment affords pre-emption, allowing the patient and physician to discuss strategies, such as lifestyle changes and begin medication, if indicated. Troponin is a very sensitive and specific marker of cardiac necrosis and is central to the management of myocardial infarctions. Other routine measures used in a biomarker context include blood pressure and inflammatory markers (i.e., C-reactive protein). These quantitative measures enable cardiologists to effectively personalize care to patients.

Biomarker use in HIV/AIDS is also central to clinical practice. Patients with HIV/AIDS now have their immune systems reconstituted routinely. HIV and infectious disease specialists customize complex antiviral regimens to patients with the aid of biomarkers, such as viral load, CD4 counts and, in some cases, viral genotypes. Viral loads and CD4 counts are routinely measured and tracked over time and are vital to guide sophisticated medical therapies. Thus, the incorporation of biomarkers into HIV/AIDS clinical practice enables a disease state molecular feedback, and helps avoid the threats of opportunistic disease, life-threatening sickness and death.

Compared with cardiology and HIV/AIDS, biomarker use in clinical oncology is quite limited. This is not due to a lack of time or effort, but rather a more complex disease process. Interestingly, a few of the true successes in oncology involve the routine clinical use of biomarkers. For example, germ cell tumors have an approximate 95% 5-year survival, and relatively high success rate for cure, especially when detected at an early stage of disease. Diagnosis, treatment response and relapse monitoring all involve biomarker (e.g., α-fetoprotein, β-human chorionic gonadotropin) assessment. Today, breast cancer survival can be extended. Breast cancer treatment choices are dictated by biomarker analysis, specifically, the estrogen receptor and ERBB2 (also known as HER2). These biomarkers are integral in therapy selection and also for enhancing quality of life by avoiding toxic treatments of unlikely benefit. Most recently, cKIT receptor mutational analysis from gastrointestinal stromal tumors (GIST) is now used for predicting response to tyrosine kinase inhibitors (TKIs) and likelihood of resistance. Just a few years ago GIST was a certain death sentence. Today, the use of upfront diagnostic testing coupled to TKI therapy has transformed GIST to a disease of prolonged survival with a minimal degree of toxicity for a myriad of patients, many of whom possess grossly metastatic disease.

The new wealth of molecular methods, scientific tools and resulting data streams has transformed biomedicine into an information science. This phenomenon is certainly true for oncology. Cancer is a complex multifactorial disease. When a cell escapes its normal process of replication and division, there are a plethora of changes at the gene, transcript, protein and metabolite level that accompany this evasion. While science used to be limited in the numbers of molecular changes that could be studied, present technologies have enabled entire genomes, transcriptomes, proteomes and now metabolomes to be analyzed within single experiments. These new techniques also bring much hope and promise to advance all of clinical oncology, from clinical practices based on categorical treatment assignments to more rational therapies based on an individual’s tumor characteristics. The rapidly evolving fields of mass spectroscopy (MS) and clinical proteomics have demonstrated significant promise for exploiting the complex information content contained in tumors to provide necessary scientific insights for clinical advances.

A major thrust area in proteomics is to find novel biomarkers that can be used to diagnose cancer at its earliest stages, as well as determine which course of treatment would most effectively work for an individual patient. Increasing the rate at which early-stage cancers are diagnosed would minimally require the discovery of a biomarker that could be measured from a routinely drawn biospecimen acquired during a routine physical. Unfortunately, past experience suggests that a single biomarker will not suffice. Testing for prostate-specific antigen, for instance, will miss a percentage of men with prostate cancer (false negative), and incorrectly identify some who prove not to have cancer (false positive). This lack of sensitivity and specificity is not surprising given the degree of heterogeneity present in both solid tumors and the human population at large. Thus, a prevailing hypothesis is that a panel of biomarkers would cumulatively possess a higher specificity and sensitivity than any single biomarker.

Efforts to identify panels of biomarkers for diseases such as cancer have been and continue to be pursued. In one of the earliest and most famous studies, Emanuel Petricoin and Lance Liotta compared serum obtained from women with neoplastic and non-neoplastic disease within the ovary. They found that a simple pattern of peaks within a mass spectrum could correctly classify cancer-affected from noncancer, with a sensitivity and specificity of 100 and 95%, respectively [1]. Subsequent to this study, a number of others were able to demonstrate similar diagnostic capabilities with biospecimen taken from patients with other types of cancer. The Petricoin and Liotta study was subsequently cited and criticized for reproducibility issues; however, it was quite disruptive through the scientific community and served to establish the entire concept of MS-based clinical proteomics for biomarker discovery [2]. Key concepts utilizing mass spectral patterns to identify regions of interest, combat dynamic range problems and hunt for biomarkers continue to evolve [3].

New biomarker discovery technologies are sorely needed to advance patient outcomes in cancer treatment. The Strategic Consensus Conference on Biomarker Research in Breast Cancer was a recent groundbreaking meeting of prominent breast cancer scientists and clinicians to discuss strategies to improve biomarker research. A priority-based timeframe for the implementation of both focal and broad recommendations over the next three years was established [4]. At the recent American Association for Cancer research (AACR) special conference on Advances in Proteomics in Cancer Research (February 27 – March 2, 2007), biomarker discovery methods were extensively discussed. In general, there was scientific consensus that MS serum biomarker discovery is best approached employing a method of tissue or proximal fluid analysis, and then using that information to map into the more complex serum proteomic space.

Clinical proteomics is a young science and MS methods continue to rapidly evolve. Modern molecular analyses of tumor tissue result in data streams of enormous size and complexity. The heterogeneity, intricate nature and aberrant pathways of tumors are only beginning to be understood. These challenges coupled to a heterogeneous human population make robust biomarker discovery a daunting task, especially assays based on solitary proteins. Therefore, assay technologies composed of multiple biomarkers are a logical choice for advancing molecular oncology. The discovery phase of the diagnostic assay and biomarker pipeline is well suited for MS, given its sensitivity and flexibility.

As demonstrated in cardiology and HIV/AIDS, the discovery and development of reliable biomarkers enables rapid clinical progress. The continued transformation of medical oncology from categorical treatment designations to rational treatment assignments as a function of individual tumor or serum analyses will very likely help alleviate the suffering and death from cancer. Additionally, it may enable new forms of molecular-based cancer risk assessments, allowing for preemptive strategies and prophylaxis.