Essential Characteristics of Imaging Biomarkers

Earlier we highlighted the use of imaging in biological research. In such studies, when a correlation is demonstrated between an image-based measurement and a particular physiological characteristic or disease process, it is tempting to identify that measurement as an imaging biomarker. However, this is not always useful. Here we identify several additional characteristics of imaging biomarkers that will generally become essential as an image-based functional measure progresses through the process of validation and clinical qualification (Figure 2 ).

Scientific Characteristics For an image-based measurement to be an effective tool in drug development, the "model" linking the measurement to a

Assay validation

Clinical qualification

Image-based Measure

Biological Function or

Disease Process

Clinical Endpoint

Figure 2 Two-step process for imaging biomarker validation/qualification. Assay validation establishes the link between the image-based measure and a specific biological function or disease process by way of a biological, biophysical, or molecular "model." This includes characterizing the measurement's intra- and interobserver and test-retest variability (measurement reliability), as well as correlating the measurement to an established (nonimaging or invasive) standard. The goal of the second step, clinical qualification , is to establish empirically, by clinical trials, the relationship between the image -based functional measure and a relevant clinical endpoint or outcome. (See insert for color reproduction of the figure. )

Image-based Measure

Biological Function or

Disease Process

Clinical Endpoint

Figure 2 Two-step process for imaging biomarker validation/qualification. Assay validation establishes the link between the image-based measure and a specific biological function or disease process by way of a biological, biophysical, or molecular "model." This includes characterizing the measurement's intra- and interobserver and test-retest variability (measurement reliability), as well as correlating the measurement to an established (nonimaging or invasive) standard. The goal of the second step, clinical qualification , is to establish empirically, by clinical trials, the relationship between the image -based functional measure and a relevant clinical endpoint or outcome. (See insert for color reproduction of the figure. )

biological or disease process must be characterized, which is termed assay validation. This process includes:

• Characterizing the validity of the model by correlation with an accepted standard

• Characterizing the reliability of the measurement:

• Intraobserver variability

• Interobserver variability

• Test-retest variability

• Defining the conditions under which the image-based measurement will result in the expected model validity and measurement reliability:

• Biological conditions

• State and stage of disease

• Class of therapy

• Optimizing the image acquisition and analysis systems for the task of parameter estimation as opposed to classification

Basic research studies of image- based biological measures may provide only partial validation by attempting to establish the validity of the model and occasionally reporting some measure of reliability in a limited sample size, often in preclinical studies. Additional development may be necessary to fully characterize and optimize the measure in both animal models and in humans. Like the underlying image-based methodology, the imaging biomarker is necessarily quantitative, which requires parameter estimation from the images, a process that should aim to be as efficient and objective as possible.

Clinical Characteristics In addition to the scientific concerns of assay validation, several clinical issues must be addressed. The relationship of an image-based measure of biological function to a specified clinical endpoint must be established empirically through clinical trials, which may be termed clinical qualification (Figure 2). Clinical qualification seeks to:

• Define a relevant clinical endpoint or outcome

• Establish empirically the relationship between the image-based measure and the clinical endpoint

• Characterize the propagation of measurement uncertainty to uncertainty in clinical outcome

• Establish the safety of the imaging procedure

These aims are often hindered by the need for large amounts of data to support clinical qualification. The imaged-based measure may be more sensitive to biological changes than the corresponding clinical endpoint, reflecting fluctuations that are not clinically relevant; or the clinical endpoint might be somewhat subjective, subject to large variability, or require lengthy trials.

Additional Criteria Apart from the scientific and clinical issues of bio-marker qualification, further logistical and economic concerns must be addressed as an imaging biomarker progresses through clinical qualification.

• The necessary acquisition hardware, methods, and image analysis tools must be readily available.

• The image acquisition and analysis methods must be standardized across hardware platforms and imaging centers.

• The imaging biomarker must be cost-effective and efficient compared to other available imaging and nonimaging biomarkers with similar purpose.

Development Stages It is clear that most imaging biomarkers today fall short of achieving all of these characteristics. To describe the development stage of imaging biomarkers, the FDA has identified four distinct classes of imaging biomarkers which are focused on drug development applications: pre -biomarkers, biomarkers, surrogate biomarkers. and clinical diagnostic surrogate biomarkers [38]. According to this classification,pre-biomarkers are those at an incomplete stage of assay validation; biomarkers are those that have completed assay validation along with an appropriate safety profile; surrogate biomarkers are those for which sufficient clinical evidence exists to qualify the biomarker for use in drug development; and clinical diagnostic surrogate biomarkers are those that have been validated and qualified to such an extent as to be "integrated into an approved/licensed therapeutic regimen" and are able to "assure/improve safety and/or efficacy" [38].

Imaging biomarkers are useful tools for a variety of tasks in addition to drug development. Therefore, another perspective, which perhaps includes a wider spectrum of methods, is to stage imaging biomarkers according to their intended use [39]. There are at least four distinct categories of use for imaging biomarkers in the wider sense:

1. Elucidate biological function and disease processes in vivo (basic biological research).

2. Evaluate novel therapies in vivo for industry decision making (drug development—industry decisions).

3. Evaluate novel therapeutics in vivo for government licensure (drug development—government registration).

4. Enhance routine diagnosis of disease with better disease management, prognosis, and therapeutic decision making (clinical practice).

The four stages correspond roughly to the FDA imaging biomarker classes described above in terms of the level of validation and qualification within each stage (Table 2). They also parallel the conceptual biomarker hierarchy

TABLE 2 Classification of Imaging Biomarkers According to Their Intended Usea

and the Corresponding Stage of Validation and Qualification

Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook


Post a comment