Photography: Dynamic Graphics Illustration: Paul A. Belci
|
Diabetes is one of the largest therapeutic areas in terms of global pharmaceutical sales. The International Diabetes Foundation
estimates in 2007 nearly U.S. $44 billion will be spent on diabetes pharmaceutical sales. Pharmaceutical companies are actively researching new therapies in the area of diabetes and are seeking strong partners to
run efficient and effective studies. Currently, more than 1000 clinical trials are being conducted in the area of diabetes.
There are many important aspects to consider in conducting a diabetic trial. However, relating to labs, the most important
is imaging markers.
Centralized imaging in trials
Utilizing imaging endpoints in diabetic clinical trials can be extremely effective. An example of this is a trial published
in the November 13, 2006, edition of JAMA, "Effect of Pioglitazone Compared With Glimepiride on Carotid Intima-Media Thickness in Type 2 Diabetes." This study successfully
utilized Carotid Intima-Media Thickness (CIMT) to demonstrate the effectiveness of pioglitazone with subjects who have type
2 diabetes.
There are multiple imaging endpoints that can be considered in diabetic clinical trials. Imaging modalities that would be
utilized include ultrasound, CT, dual-energy, x-ray absorptiometry (DXA), MRI, retinal angiography, angiography, and quantitative
coronary angiography (QCA). Carotid artery intima-media thickness (IMT) is a marker of coronary atherosclerosis and independently predicts cardiovascular
events. Image acquisition for IMT studies is more challenging than for sonographic exams performed for clinical routine. It
is helpful when an imaging partner has arranged a specific training program for sonographers so they are able to conduct various
types of acquisitions and get reproducible results.
In addition, specialized acquisition workstations that include a "mask" function allow the sonographer to find the same acquisition
zone used during a previous exam. This technique has been widely used in long-term clinical trials and has allowed accurate
retrieval of the artery segment for the measurement of IMT in carotid arteries.
Carotid IMT values vary tremendously during the cardiac cycle, sometimes by more than 50%, with a stable value for only a
few milliseconds during the cycle. Stability occurs during the end diastole, which is why measurements must be done during
this interval. Before reading, it is important to process video clips (exams) to extract the best end-diastolic image. Centralized
readings assist in avoiding bias and inconsistency in the image preparation.
Computer-assisted techniques for the semi-automatic measurement of IMT avoid manual measurement. This type of software should
be extensively validated against phantoms, cadaveric arteries, and echo-tracking. For improved consistency and efficacy, central
reading should be divided into two steps: prereading performed by a clinical research technician and reading validation performed
by a trained reader.
Precise and reproducible measurements
In trials on diabetic or obese patients, there is a need for reliable assessment of fat accumulation, fat loss, and lipodystrophy.
These assessments are usually made by clinical examination and anthropometric evaluations, such as limb circumference. These
methods, however, are imprecise and operator dependent. The use of noninvasive imaging techniques such as dual-energy x-ray
absorptiometry (DXA) or high resolution CT (HRCT) allow a precise and reproducible quantitation of body fat and lipodystrophy,
especially when coupled with computer-assisted measurement methods.
Figure 1. Example of abdominal scan: automatic calculation by the computer of subcutaneous fat (yellow) and peri-visceral
fat (red).
|
Each CT slice is evaluated to determine the cross-sectional area of body tissues. The most effective method is to utilize
a semi-automated approach that employs image segmentation algorithms to define areas of adipose tissue (fat) and nonadipose
tissue (muscles, bone, organs). The results should then be reviewed by a radiologist to ensure correct classification of tissues.
The radiologist would adjust the computer algorithm as needed, and adipose tissue is further segmented into subcutaneous and
visceral compartments (see Figure 1).
