Computer-aided ophthalmic artery waveform analysis in healthy individuals and glaucoma patients

Abstract

Purpose : Arterial waveform parameters (WPs) are commonly used to monitor and diagnose systemic diseases. Color Doppler Imaging (CDI) is a consolidated technique to measure blood velocity profile in some of the major ocular vessels. This study proposes a computer-aided manipulation process of ophthalmic artery (OA) CDI images to classify and quantify WPs that might be significant in the assessment of glaucoma and ocular vascular diseases.

Methods : The Siemens Antares Stellar Plus™, probe VFX 9-4 MHz vascular linear array, was used to obtain 50 CDI images acquired by 4 different operators on 9 healthy individuals at the IRCCS Policlinico San Matteo of Pavia. The Philips HDI 5000 SonoCT Ultrasound System with the microvascular small parts clinical option (Philips Medical Systems, Bothell, Washington, USA), 7.5 MHz linear probe, was used to obtain CDI images of 22 glaucoma patients within the Indianapolis Glaucoma Progression Study. An ad-hoc semi-automated image processing code was implemented to detect the digitalized OA velocity waveform and to extract the following WPs: peak systolic velocity (PSV), end diastolic velocity (EDV), resistive index (RI), area under the wave (A), period of a cardiac cycle (T), difference between the PSV time and the dicrotic notch time (Dt) and the area ratio (f), Fig 1.

Results : When compared to healthy individuals, glaucoma patients show:
1) significantly lower values of PSV (25.02±11.29 vs 39.50±11.16 cm/s, p=2e-6), EDV (4.56±2.82 vs 6.02±2.33 cm/s, p=0.014), A (11.54±5.29 vs 14.53±4.80 cm, p=0.012) and Dt (0.22±0.04 vs 0.25±0.01 s, p=4e-7);
2) significantly higher values of f (0.77±0.08 vs 0.56±0.06, p=3e-19);
3) no statistical difference in RI (p=0.053) and T (p=0.107).
All comparisons were made with a two-sample t-test with a 5% significance level. When comparing multiple CDI images for the same healthy individual, T, Dt and f resulted to be more consistent than PSV, EDV, RI and A (average coefficient of variation <8% vs >20%), Fig 2.

Conclusions : The proposed computer-aided manipulation of OA-CDI images allowed to identify novel reliable WPs that vary significantly among healthy individuals and glaucoma patients (Dt and f). In future studies, this technique will be used to further assess the clinical relevance of these findings in the assessment of glaucoma and ocular vascular diseases.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

 

Fig 1. Computer-aided image manipulation process

 

Fig 2. OA-CDI digitalized profiles