Progress in Lung Cancer Diagnostic Biomarker Development

By Peter Mazzone, MD, MPH

Biomarkers are objectively measured indicators of the state of an individuals’ health. They range from commonly measured vital signs to complex molecular signatures. A new biomarker can improve on currently used tests by being more accurate, less invasive, less expensive and/or novel in its intent. To have a clinical impact, the result of the test must affect a decision to the benefit of the patient.

Benefits of biomarkers

Lung cancer molecular biomarkers have the potential to benefit patients. Biomarkers that identify subjects at risk of developing lung cancer will lead to advances in screening and chemoprevention. Biomarkers that improve our management of imaging findings will help to expedite diagnosis, and lower the risk and cost of evaluation. Biomarkers that optimize tumor characterization and monitor treatment response will help to personalize therapy decisions.

Phases of biomarker development

Biomarker development occurs in phases:

• In the discovery phase, new analysis platforms are developed, or relationships between molecular signatures and clinical outcomes are identified.
• During the analytical validation phase, the consistency with which the testing platform measures the specific biomarker is assessed.
• In the clinical validation phase, the accuracy of the test result in the relevant clinical population is evaluated.
• Finally, an assessment of clinical utility is performed to determine if the test improves the benefit to risk balance of an associated clinical decision (biopsy, treatment).

In each of these phases, optimization of test performance can occur in the pre-analytical, analytical, and post-analytical aspects of testing. Cost effectiveness, another important metric, can only be assessed in a speculative manner prior to proven clinical utility.

Advances in lung cancer biomarker development

Until recently, the majority of advances in lung cancer biomarker development occurred in tumor characterization. With the acceptance of lung cancer screening, more attention is focused on biomarkers of risk prediction and early diagnosis. A handful of these biomarkers have completed analytical and clinical validation and are commercially available for use. To date, none of the available diagnostic biomarkers has completed an assessment of their clinical utility.

Table 1: Potential outcome from the binary application (positive or negative) of theoretical lung nodule biomarkers. Three examples of test sensitivity combined with specificity are applied to three probabilities of malignancy. For the low probability example, it is assumed that surveillance imaging alone would occur if the biomarker were not available and thus only a positive result would have an impact on decisions. For the moderate probability example, it is assumed that a positive test will lead to more aggressive management and a negative result will lead to less aggressive management. For the high probability examples, it is assumed that aggressive management will be pursued if the biomarker were not available and, thus, only a negative result would have an impact on decisions. Each row represents a theoretical cohort of 1,000 patients.

The clinical utility of a diagnostic biomarker depends on the accuracy of the test and the manner in which the test results are used by the interpreting clinician. Even biomarkers that are sensitive and specific enough to be considered accurate by most clinicians have the potential to impact clinical care in a negative manner.

For example, if a biomarker is used to characterize a lung nodule that has a very low risk of malignancy, a positive result is more likely to be false positive. If all positive results lead to aggressive testing, then more invasive testing will be performed on benign nodules with few cancers diagnosed earlier.

In addition to the potential physical complications of this approach, other significant social and behavioral consequences may develop as a result of testing. Table 1 illustrates the potential consequences of interpreting a biomarker as a binary positive or negative outcome for various test accuracies and malignancy probabilities.

Tests on the market

Table 2: Overview of diagnostic lung cancer biomarkers. TAA = tumor associated antibodies, MRM = multiple reaction monitoring mass spectrometry, ELISA = enzyme linked immunosorbent assay, RNA = ribonucleic acid, * = clinical audit of test accuracy as used in practice, other clinical validation studies also performed, ** = test alone (not including bronchoscopy results) in intermediate probability of malignancy patients (41% with malignancy), *** = wide range of sensitivity based on histology and stage.

There are at least three blood tests commercially available at the time of this writing, and one test of airway epithelial cells collected during bronchoscopy. The blood tests include a panel of antibodies to tumor associated antigens, a panel of proteins, and a panel combining proteins and an antibody. The test of airway epithelial cells assesses a panel of genomic changes. Each of these tests has completed analytical and clinical validation showing moderate accuracy, variably selecting thresholds that optimize sensitivity or specificity. Thoughtful application of the tests can be considered at this time, while awaiting more definitive assessments of their clinical utility (Table 2).

Partnering on development

We are fortunate to be collaborating with industry biomarker developers in projects that either aim to optimize the performance of their biomarker or to develop evidence of clinical utility. In addition, we continue to work on developing volatile organic compound biomarkers of the breath and urine, and small molecule metabolite biomarkers in the blood. We look forward to the opportunity to improve our management of patients through the use of accurate, fully developed biomarkers.

Dr. Mazzone, Director of the Lung Cancer Program for the Respiratory Institute, Director of the Lung Cancer Screening Program, and Director of the Pulmonary Rehabilitation Program, can be reached at 216.445.4812 or mazzonp@ccf.org.