|
|
|
|
 |
|
Systems for the measurement of the endothelium-dependent flow mediated dilation (FMD) are assuming considerable importance in cardiovascular medicine. The examination, which is based on ultrasound imaging, consists in the measurement of the brachial artery diameter at rest and after reactive hyperemia induced by ischemia of the forearm. An impaired FMD response has been related to cardiovascular risk factors such as smoking, hypercholesterolemia, hypertension, diabetes and hyperhomocysteinemia. |
|
 |
||
|
The procedure must be initialized by providing two approximated starting borders, an operation that also defines the region of interest where the diameter will be computed. This operation is carried out in a GUI window where the ultrasound images are displayed in real-time. The elaboration starts immediately afterwards. The value of the diameter is shown by a numerical display and is plotted on two graphs: the first graph shows the instantaneous measure over a time scale of 5 seconds; the second graph shows the mean value of the diameter computed over 2 seconds and displayed over a time scale of 9 minutes. Once the examination is completed, the system automatically provides: |
||
| Baseline: mean value of the diameter in the first minute. Maximum: maximum value of the diameter. Post baseline: mean value of the diameter in the last minute. FMD: Flow Mediated Dilation (with respect to baseline value). FMDp: Flow Mediated Dilation with respect to baseline post value. |
 |
|
These data, as well as the diameter curve, can be uploaded on a remote workstation were a report can be created. |
||
 |
||||
The method is applied to B-mode images of a longitudinal section of the vessel and is based on the localization of gray level discontinuities. A new edge detector (mass center of the gray level variability) was used. This mathematical operator was used in a wall tracking algorithm which was especially designed for the brachial artery. Since the shift of the artery wall is only a few pixels, a temporal filter was added to constrain the movement of the contours. Moreover, the algorithm was modified with some ad hoc adjustments so as to exploit some peculiarities of the images under investigation. |
||||
| UPDATED 24/01/2007 | ||||
 |
|
The method was tested on a total of 270 FMD examinations, taken from three clinical studies. The agreement with manual gold-standard was tested on a set of 120 images which were extracted from twenty FMD sequences. Regression analysis and Bland Altman plots were used to compare the results. The inter/intra-observer variability was tested on four FMD. The dataset was analyzed by three observers, each of whom repeated the measurement four times in independent runs. The robustness was quantified by classifying the whole set of 270 sequences. |
|
 |
||||
|
The availability of results in real-time is of particular importance in FMD studies. In these examinations, the quality of the image is a critical component that can compromise the success of the measurement. Indeed, a proper image must be maintained for several minutes to best quantify the transitory response induced by the endothelium. For this reason, adjustments of the position of the probe may be required during the examination, especially to compensate small movements of the patient. The sonographer is remarkably helped in this task by an immediate feedback from the measurement system. Another advantage of on-line analysis is the absence of the drawbacks associated with video storing. Recording the video by a VCR means a reduction of the image quality, while an acquisition on a PC requires a great amount of memory. Finally, a real-time examination is obviously less time consuming and this is also an important aspect in clinical applications. |
||||
 |
||||
|
FMD Studio has been adopted by the SIIA (Italian Society of Hypertension) and SIC (Italian Society of Cardiology) for the multi-center research project entitled “The clinical significance of the evaluation of the endothelial function”. |
||||