Originally Published IVD Technology January/February 2006
DIAGNOSTIC OUTSOURCING
Using specialty labs to develop IVDs
Mario R. Ehlers
Specialty labs offer IVD manufacturers expertise in standardization and generation of clinical utility data.
 |
| IVD assays suitable for use in routine clinical diagnostics must be standardized against accepted reference methods, when available, and should be supported by clinical utility data. |
Developing an IVD assay with broad utility as a routine clinical diagnostic requires numerous resources and skill sets. Since IVD manufacturers may lack direct access to all of the resources and expertise required, such as knowledge of regulatory requirements for FDA submissions and conducting clinical trials, they may opt to outsource such development services to specialized third-party firms. This article highlights two components of IVD development that can be outsourced to specialty labs: reference and standardization services, and generation of clinical utility data.
One key concern in IVD assay development is standardizing assays to reference methods or predicate devices. Various clinical and regulatory organizations, such as FDA, the National Institutes of Health (Bethesda, MD), the Centers for Disease Control and Prevention (CDC; Atlanta), and the College of American Pathologists (Northfield, IL), have stipulated that the accuracy and precision of newly developed assays must correlate with established reference methods or predicate devices when they are available. Since such specialized reference and standardization services may not be available in-house, IVD manufacturers could contract such services out to specialty reference laboratories.
A second area of concern is the generation of clinical utility data, which is essential for FDA clearance and the successful launch and adoption of new tests, postapproval. While FDA may require only basic clinical data such as patient reference ranges, accuracy, precision, and correlation with predicate devices, the standard is higher for commercial launches. Broad clinical adoption of IVD products demands additional data and compelling scientific evidence that are obtained from appropriate clinical trials.
For example, for a new IVD assay, inclusion in a pharmaceutical clinical trial can provide baseline demographic data and information on its value in assessing disease progression and monitoring treatment responses. Such clinical utility data are not usually available to IVD manufacturers and can be critical to driving the adoption of a new assay by the clinical community. Equally important to facilitating adoption and driving sales of new assays is obtaining reimbursement codes, which is also expedited by demonstrating clinical need and healthcare benefit. Specialty labs can provide invaluable services by brokering the inclusion of IVD assays in clinical trials sponsored by pharmaceutical and biotech companies.
As defined in this article, specialty labs are clinical labs with expertise in one or more diagnostic or therapeutic areas. Specialty labs may be clinical reference labs associated with academic centers, or commercial labs dedicated to supporting clinical drug and diagnostic product development. By working with multiple IVD manufacturers and maintaining established reference methods, specialty labs can provide unique services to manufacturers. Moreover, because of their specific areas of expertise, IVD and drug manufacturers use specialty labs to support their product development, which leads to useful synergies—especially for generating clinical utility data.
Reference and Standardization Services
The role of specialty labs in providing reference and standardization services is illustrated in the area of cholesterol testing. Measuring serum lipids, especially total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides, is an established element in cardiovascular risk assessment.1 Lipid panels have become routine components in standard clinical chemistry panels. Recently, there has been growing interest in including additional lipoprotein markers, such as apolipoproteins and lipoprotein subclasses, in routine panels. Such developments have prompted IVD manufacturers to develop automated assays for the clinical diagnostics market.
The association between high blood cholesterol and coronary heart disease is one of the best-established medical relationships.2 Given the importance of cholesterol in assessing heart disease risk, getting reliable cholesterol measurements has become a national public health priority and led to the formation of the Laboratory Standardization Panel. This panel recommends that all cholesterol measurements be standardized to a National Reference System for Cholesterol (NRS/CHOL). However, since an estimated 157,000 sites perform clinical laboratory testing in the United States, direct standardization is impractical. A program has been set up to assist manufacturers of cholesterol diagnostics in calibrating their systems.
In 1989, the CDC-affiliated Cholesterol Reference Method Laboratory Network (CRMLN) was created to oversee this program.2 CRMLN labs assist in developing new cholesterol assays by serving as a reference target during product development. CRMLN labs also provide formal certification for total cholesterol, LDL-C, and HDL-C assays using freshly collected serum samples by following the CRMLN certification protocols when the products are ready for commercialization.2 In addition, some CRMLN laboratories provide proficiency testing programs and serum samples with assigned target values as determined by the reference methods to assist with internal method standardization at IVD companies.
 |
Figure 1. A comparison of four direct high-density lipoprotein (HDL) cholesterol assay methods with the designated comparison reference method (DCM) for unusual patient samples indicates variable but significant bias, especially at the lower concentrations. |
The demand for such services is strong because of continued interest in developing direct LDL-C and HDL-C assays, and the challenge of maintaining the standardization of established methods. With physicians worldwide using direct homogeneous assays extensively to assess coronary heart disease risk and make treatment decisions, standardization and correlation with reference methods are important considerations for IVD manufacturers (see Figure 1).
Standardizing cholesterol assays is complicated by the sensitivity of most assay systems to sample matrices. CDC created CRMLN to improve cholesterol measurements by assisting IVD manufacturers in validating the traceability of their reference methods to the NRS/CHOL. CRMLN has served as a model for other clinical analytes in which the traceability of methods is needed. For example, in 1996, a laboratory network modeled after CRMLN was established to standardize the measurement of glycohemoglobin for diabetes. CDC and CRMLN are also working closely with other organizations such as the Joint Committee on the Traceability of Laboratory Medicine and the International Federation of Clinical Chemistry (Milan, Italy) to incorporate CRMLN’s cholesterol certification program in a larger all-laboratory survey program.
The ultimate goal in standardizing cholesterol test results is to achieve accurate and comparable values among patient specimens regardless of manufacturer, assay technology, or instrument platform. As discussed, cholesterol levels are key determinants of coronary heart disease risk and the basis of potentially lifelong treatment decisions. Therefore, physicians need assurances regarding the long-term accuracy and precision of cholesterol assays. Establishing correlations to reference methods through comparisons with fresh patient specimens is a critical component in standardization. CDC reference methods that are maintained by CRMLN labs establish the link and traceability to enable accurate measurements of patient specimens.
With the existence of thousands of clinical laboratories, performing such direct comparisons in routine lab settings is impractical. However, a limited number of IVD manufacturers provide the instruments, reagents, and supplies used in such labs. CRMLN’s program focuses on assisting manufacturers in calibrating their systems properly and having direct traceability to the NRS/CHOL.
For more than 20 years, CRMLN’s system has served as the standard for cholesterol testing. Medical decision points are based on data that were obtained from clinical studies performed in research laboratories, and are standardized and traceable to the reference methods at CDC and CRMLN labs. By focusing on the diagnostic instrument systems, CDC can directly affect the quality of cholesterol tests performed in clinical laboratories. CRMLN also provides a means for clinical laboratories to establish a direct link to the NRS/CHOL by performing split-sample comparisons with the reference methods. Satisfactory completion of the requirements leads to a certificate of traceability.2
Clinical Utility Data
Data proving an IVD’s clinical utility are critical for both regulatory approval and successful market adoption. An example that illustrates this principle is biochemical markers of bone turnover. Bone markers are being recognized as an important adjunct to clinical and radiological assessments in the diagnosis, risk stratification, and treatment monitoring of osteoporosis.3,4 While it has taken more than 10 years to reach this point, routine acceptance of bone markers by physicians will require additional clinical utility data and the endorsement of opinion leaders. Clinical data indicating the usefulness of bone markers as surrogate endpoints during the development of osteoporosis drugs are particularly valuable for such purposes. As with many diseases, physicians are more likely to order tests that provide useful treatment-monitoring information.
Biochemical markers of bone turnover may eventually be used as companion diagnostics, tests that increase the clinical utility of accompanying therapeutic drugs while reducing the risks associated with taking the drugs. When such tests are sold with drugs, they are referred to as combination products or theranostics.
Data from recent osteoporosis clinical trials suggest the value of bone markers in fracture risk prediction and treatment monitoring.3,4 The prerequisites for biomarkers as suitable companion diagnostics include adequate marker response following therapy or disease progression, availability of a robust assay platform, and appropriate reference ranges for guiding patient management.
Pacific Biometrics Inc. (Seattle) assisted Roche Diagnostics Corp. (Indianapolis) in investigating serum beta-crosslaps (CTX), a bone resorption marker, and total serum N-terminal propeptide of type I procollagen (P1NP), a bone formation marker. Roche looked into these markers as possible theranostics for osteoporosis using its Elecsys 2010 automated immunoanalyzer.5
The study monitored the long-term stability of the analytes and the long-term performance of the Elecsys assays. The study also compared individual patient responses following antiresorptive treatment by including the assays in an appropriate pharmaceutical clinical trial. In addition, the study measured CTX and P1NP in nonosteoporotic African Americans and Caucasian Americans to explore appropriate reference ranges for patient management in different racial, gender, and age groups.
These and other studies have established that serum CTX and P1NP undergo dynamic changes in response to antiresorptive therapies, and are more sensitive and rapid in evaluating treatment response than bone scans.5 Moreover, after three years of sample storage at –70°C, the correlation between the original and measured values was verified, indicating excellent long-term precision.
 |
| Figure 2. Serum CTX and P1NP patterns of responders and nonresponders from a six-month trial of antiresorptive treatment in 1 03 osteoporotic subjects. |
These studies concluded that serum CTX and P1NP measured as a panel on the Elecsys are precise, stable, easy to use, respond well to antiresorptive therapies, and are therefore excellent candidates as theranostics or companion diagnostics for osteoporosis treatments (see Figure 2). Significant overlaps in reference values for CTX and P1NP between African Americans and Caucasians, and between men and women in all age groups, were also found. The data derived for CTX and P1NP established not only utility for monitoring treatment response but also reference ranges in different populations. The data enabled Roche to insert specific claims in its product inserts for review by regulatory agencies, and broaden the usefulness of the assays in routine clinical diagnostics.
Another example that demonstrates the value of collecting clinical utility data to support the development of an IVD assay concerns the bone marker serum osteocalcin (OC), a biomarker for bone formation and turnover. At one time, accurate measurement of serum OC was virtually impossible due to the analyte’s instability. The introduction of sandwich methods that determine the intact amino acid molecules, and the major N-terminal fragment and MID-fragment of the protein appeared to address the instability issue in OC determination.
Nevertheless, some specialists in metabolic bone disease still doubted the stability of this bone marker in serum and plasma. Studies have demonstrated that N-MID osteocalcin measured by Roche’s Elecsys analyzer is stable over a wide concentration range using samples from normal subjects and patients who are likely to undergo OC testing.
In one study, samples with OC concentrations of 10, 38, and 63 ng/ml in serum and heparin- and EDTA-treated plasma were stored at –70°C and remained stable for at least 24 months. The measured concentrations were also stable after two freeze-thaw cycles.
Another study measured matched sets of serum, heparin plasma, and EDTA plasma samples from 44 subjects, and analyzed them within one month of collection and again after four years of storage at –70°C using a second aliquot. The correlation of serum, EDTA plasma, and heparin plasma, over a concentration range of 17 to 1133 ng/ml, gave a correlation coefficient of 0.99 with slopes ranging from 0.92 to 1.01. Similar results were obtained from a study on OC levels in women with osteoporosis.
Such data underscore the technical improvements in OC assays with respect to stability and robustness, and provide support for the clinical utility of OC in diagnosing osteoporosis. This example shows the progress that can be made to expand the use of a marker into the general clinical community. An IVD manufacturer improves and automates an assay, and then obtains the relevant clinical utility data by working with a specialty lab that has access to diverse specimens from clinical studies.
While data supporting clinical utility are valuable to IVD manufacturers seeking FDA approval, such data are often difficult to generate in-house. By contracting with specialty labs that are involved with pharmaceutical trials, IVD companies can obtain the critical data needed for FDA clearance and successful marketing of their novel assays postapproval.
 |
| Mario R. Ehlers, MD, PhD, is the chief medical officer at Pacific Biometrics Inc. (Seattle). He can be reached at marioe@pacbio.com. |
Specimens collected during pharmaceutical clinical trials are well characterized with respect to patient demographics, inclusion and exclusion criteria, time-course of placebo-controlled treatment response, and correlation with established surrogate markers and endpoints. Moderately sized phase-2 trials cost several million dollars, while large phase-3 trials can run into the tens and sometimes hundreds of millions of dollars. Since such costs far exceed the budgets allocated for the clinical product development phase of IVD assays, gaining access to such specimens by contracting with specialty labs can represent significant cost savings. Reductions in development times may also occur if IVD manufacturers can complete the required database of clinical information from a single study rather than assembling the data piecemeal.
Conclusion
Although IVD manufacturers can sponsor limited clinical trials to obtain data in normal and affected populations, including responses to approved drugs, they generally do not have the resources to sponsor large clinical trials that provide the data required to demonstrate clinical utility to FDA or to practicing physicians. IVD manufacturers also cannot sponsor trials involving investigational drugs, which provide the most compelling data for IVD assay marketing because use of the assay can be included in the supporting information on the drug label. Specialty labs can play a role in brokering relationships between diagnostics and pharmaceutical companies that can give the former access to data from which they are ordinarily excluded. This allows IVD manufacturers to generate data on theranostic use, which supports the publication of studies and facilitates marketing activities to drive rapid adoption. Moreover, the role of specialty labs in IVD product development is likely to grow as biopharmaceutical companies place increasing emphasis on biomarker discovery and the development of companion diagnostics for novel therapeutics.
References
1. P Kwiterovich, “The Metabolic Pathways of High-Density Lipoprotein, Low-Density Lipoprotein, and Triglycerides: A Current Review,” American Journal of Cardiology 86, no. 12A (2000): 5L–10L.
2. G Myers et al., “A Reference Method Laboratory Network for Cholesterol: A Model for Standardization and Improvement of Clinical Laboratory Measurements,” Clinical Chemistry 46, no. 11 (2000): 1762–1772.
3. R Eastell et al., “Relationship of Early Changes in Bone Resorption to the Reduction in Fracture Risk with Risedronate,” Journal of Bone and Mineral Research 18, no. 6 (2003): 1051–1056.
4. P Ravn et al., “Biochemical Markers for Prediction of 4-Year Response in Bone Mass During Bisphosphonate Treatment for Prevention of Postmenopausal Osteoporosis,” Bone 33, no. 1(2003): 50–158.
5. E Leary et al., “Serum CTX and P1NP by Roche Elecsys as Theranostics in the Treatment of Osteoporosis,” Journal of Bone Mineral Research 20, supp. 1 (2005): S227.
Copyright ©2006 IVD Technology
