DYNAMICS OF EXTERNAL OCULAR BLOOD FLOW STUDIED BY SCANNING ANGIOGRAPHIC MICROSCOPY L. DAVID ORMEROD, ENRIQUE FARIZA and ROBERT H. WEBB Boston, Massachusetts SUMMARY circumepiscleral plexus anterior to the rectus muscle The scanning angiographic microscope (SAM) provides insertions. A few large branches of the anterior a solution to the considerable technical difficulties ciliary arteries pass through scleral channels into the associated with conventional episcleral Buorescein eye, where they arborise almost immediately to angiography. Standardised anterior segment fluores­ supply the iris, intramuscular circle, ciliary muscle, cein videoangiograms were performed using the SAM and anterior choriocapillaris.5,6 in each episcleral quadrant of the right eye in 6 normal The intraocular component of the anterior seg­ subjects; frame-by-fram� analysis proved important. ment blood supply6-8 consists of the medial and Centripetal Bow was seen in all 37 scleral perforating lateral long posterior ciliary arteries, which are the arteries investigated. Other features were the marked main source for the major arterial circle of the iris; individual variability, much larger vertical anterior they also supply the ciliary processes, and give ciliary arteries, the high frequency of arteriovenous branches to the ciliary muscle. A primate study has anastomoses, the complex Bow patterns, the absence of suggested that an extensive collateral system may a 'watershed' zone between anterior ciliary and occur in the intramuscular circle between the posterior episcleral circulations, a characteristic and anterior ciliary and long posterior ciliary circula­ discontinuous distribution of 'leaky' episcleral veins, tions,S although this has been disputed in two human and the primacy of venous drainage into the plexus of 6 muscular veins. Reports of retrograde blood Bow in the studies. ,7 anterior ciliary arteries in most Buorescein angiographic It is generally accepted that the anterior ciliary studies are probably incorrect, the result of unappre­ arteries provide a major part of the vascular supply ciated methodological problems. The SAM is an to the anterior uvea via their scleral perforating important advance on previous anterior segment branches.1-3 This view is supported by the develop­ fluorescein angiography techniques. ment of anterior segment ischaemia following tenotomy of the vertical rectus muscles, particularly The blood supply to the anterior segmentl-3 is when combined with additional tenotomies,4,9-11 and derived from two circulations, one external to the by the results of studies involving corrosion vascular eye and one with an intraocular route. The external castings,5-8 the injection of labelled micro­ circulation comprises the anterior ciliary arteries, spheres,12,l3 and extraocular muscle tenotomies in which are continuations of the muscular arteries of primates.14--16 the rectus muscles. The anterior ciliary arteries However, several investigations of the circulatory emerge from the rectus muscles posterior to the dynamics using segment fluorescein angiography transition kom muscle to tendon,4 and run forward have concluded that blood flow in the anterior along the tendon before dividing within the episcleral ciliary arteries is principally centrifugal or retro­ tissues into branches5 that supply an extensive grade in direction,17-25 and that the anterior ocular surface is predominantly supplied from within the From: Schepens Eye Research Institute, Department of globe. These angiographic observations appear to be Ophthalmology, Harvard Medical School, Boston, Massachu­ at odds with most of the non-angiographic clinical setts, USA. and experimental data. Conventional angiographic Correspondence to: L. David Ormerod, FRCOphth, MRCP, Department of Ophthalmology, University of South Florida, studies use off-axis light sources (to reduce high­ Tampa, FL 33612-4799, USA. lights) that produce markedly uneven illumination Eye (1995) 9, 605-614 © 1995 Royal College of Ophthalmologists 606 L. D. ORMEROD ET AL. across the curved episcleral surface. The resulting died the transit of fluorescein in normal human variable fluorescence excitation may be exacerbated subjects using this new technology. by differences in scleral reflectance at increasingly acute angles of incidence of excitation light. Furthermore, most studies have the scleral perforat­ MATERIALS AND METHODS ing vessels in the central, highly illuminated parts of We performed anterior segment fluoresceinvideoan­ the field and the anterior ciliary arteries in the less giograms in each of the episcleral quadrants of the illuminated periphery. The interpretation of photo­ right eye in 6 normal human volunteers (4 men, 2 graphic records is restricted further by their women; age range 21-42 years). The individuals had dependence upon slow flash-recycle times and the no ocular or systemic disease and were not taking highly non-linear characteristics of photographic film medication. The results of ocular examination were normal in each subject. Colour photographs of all at low light intensities. The low light levels that are a four quadrants aided interpretation of the videoan­ hallmark of a fluorescent light source also lead to a giography. The protocol followed the tenets of the shallow depth of field when' using conventional Helsinki Declaration and was appro�ed by the photographic or video close-up cameras, with the institutional human experimentation committee. All periphery of the field commonly out of focus. Finally, subjects gave their informed consent. light reflexes from the tear film and 'pseudofluores­ The methodology has been described previously?6 cence' commonly interfere with angiography. Briefly, 2 ml of 25% sodium fluorescein was injected These problems are controlled by the configura­ rapidly into an antecubital vein via a three-way tap tion of the scanning angiographic microscope during a sustained period of inspiration. The cannula (SAM),26 a recent development of the scanning was flushed immediately with a 5-ml bolus of laser ophthalmoscope?7,28 The field is co-axially balanced salt solution coincidental with a rapid scanned by a 488-nm laser (Siemens, Munich, expiration?O Video recording with an audio hookup Germany) raster with a long depth of focus?6 Tight was begun before fluorescein injectionand continued filtering provides a dark (scleral) background with­ for 2-3 minutes. out light reflexes. The returned light is collected by a The four quadrants were studied at four separate 1-mm avalanche photodiode detector and integral sessions. Standard visual fixation points were amplifier (RCA, Vaudreinline, Quebec, Canada), a provided by red, light-emitting diodes mounted at highly sensitive detection system with high signal-to­ visual angles of 42° on either side of the midline from noise ratio. Real-time TV operation and SNHS the neutral eye position, and at 30° elevation and 45° video recording are other important advantages. depression in the midline. Insertion of a lid speculum The scanning functions of the SAM are tightly after topical proparacaine 0.5% anaesthesia was integrated electronically with the TV monitor image necessary for the superior quadrant angiograms. signal, resulting in point-to-point reciprocity?9 Each Movement of the headrest during the fluorescein pixel of the monitor corresponds to a particular point transit permitted extension of the fields into the in the illumination raster. A modest limitation of fornices. image resolution compared with conventional photo­ The scanning angiographic microscope was con­ graphy results from restrictions inherent in a 500-line figured to give a 10 X 10 mm field and X25 TV raster. No useful information appears to be lost magnification at the TV monitor. Scleral irradiance of the scanned 488-nm blue laser was standardised at in SAM anterior segment videoangiography, and 100-110 IJ-W/cm2. A diagram of the SAM is shown in there is greatly improved image uniformity across the Fig. 1. Because the tight filtration in the collection field.26 The recorded SNHS video image is a little . optics provided a black background, a second degraded compared with the 'live' image on the 'focusing and alignment' channel using a blue 6 monitor.2 The video frame rate of 30 Hz provides 30 reflector dichroic beam splitter and separate moni­ frames (60 fields) per second, rather than the one tor was provided to image the episclera in unfiltered field everyfew seconds in conventional photographic blue light. The instrumentation of the SAM has been angiography. The analysis of SAM video images is a discussed in detail elsewhere?6 The fluorescein dynamic and not a static process, but, for purposes of angiograms were recorded on S/VHS videotape publication, images are obtained by direct photo­ using a Panasonic (Secaucus, NJ, USA) SNHS graphy of the monitor screen, collecting approxi­ AG-7300 video recording unit. mately seven video frames over an 0.25-second Analysis of the fluoresceintransits was undertaken exposure?6 These photographed images are signifi­ according to a detailed protocol. The angiographic cantly inferior to the video image. sequence was timed from the completion of In view of the conflicting opinions regarding the fluorescein injection. Interpretation of the video vascular dynamics of the anterior ciliary and images was enhanced by the frame-by-frame episcleral vasculature, we have systematically stu- advance and reverse playback features of the DYNAMICS OF EXTERNAL OCULAR BLOOD FLOW 607 MO ITOR I MO ITOR2 RESULTS