Fig. 4.1
Top, Color illustration of retina maturation at 33–36 weeks post-menstrual age (PMA) with sample spectral-domain optical coherence tomography (SDOCT) cross-sectional B-scan of a former 1640 g, 30 week PMA gestational age infant imaged at 33 weeks PMA with no retinopathy of prematurity (ROP). Note the shallow foveal pit, persistence of inner retinal layers, and lack of inner segment band at the center of the fovea. Bottom, Color illustration of retina maturation at 40–42 weeks PMA with sample SDOCT B-scan of a former 880 g, 28 week PMA gestational age infant with stage 2, pre-plus ROP imaged at 41 weeks PMA. The foveal pit has greater depth and there is inner retinal layers displacement from the foveal center when compared to top B-scan
Recent advances in enhanced depth imaging OCT have allowed for better visualization of the choroid-scleral junction and an appreciation for the development of the macula’s vascular supply [32]. Moreno et al. measured significantly thinner choroid in young preterm infants (less than 37 weeks PMA) versus term-aged preterm infants (37–42 weeks PMA), term infants, and adult subjects. In addition, they found term-aged preterm infants had thinner choroid than term infants [26]. Choroid thickness then continues to increase from early childhood through adolescence [33].
Changes in neurosensory tissue of the retina over time can also be appreciated with SDOCT. Samarawickrama et al. measured larger cup-to-disc ratios on OCT in children with a history of low birth weight, short birth length, and a small head circumference [34]. However, analysis of color fundus photography of full term infants did not reveal any differences in optic disc measurements by sex or birth weight [35]. Additionally, RNFL is thinner in school-aged children who were born preterm which correlates with foveal thickness and visual acuity [36]. Optical coherence tomography has also identified differences in optic nerve head parameters by race [37]. Recently, Tong et al. found a larger cup and cup-to-disc ratio in preterm infants imaged with HHOCT compared to fullterm infants as well as a correlation between preterm optic cup-to-disc ratio and cognitive development at 18–24 months PMA [38]. Recently, these differences in RNFL thickness between term and preterm infants have been measured while still in the nursery, with thinner RNFL relating to both brain structure and neurodevelopment doi: 10.1016/j.ajo.2015.01.017 and 10.1016/j.ajo.2015.09.015.
Optical Coherence Tomography in ROP
Because there is such a strong correlation between structure and functional outcomes in patients with ROP, OCT promises to revolutionize the treatment of ROP by providing valuable subclinical information about eye morphology [39, 40]. Joshi et al. used time domain OCT to confirm absence of macular involvement of stage 4A retinal detachments before repair by pars plana vitrectomy and first suggested that subclinical anatomic abnormalities that can be observed by OCT may explain differences in outcomes for stage 4A vitrectomies [15]. Chavala et al. further demonstrated the utility of SDOCT as an adjunct to traditional examination techniques. They identified retinoschisis and preretinal structures believed to represent preretinal fibrovascular proliferation that were not identified by conventional examination in three patients with advanced ROP (Fig. 4.2, top left) [41]. Muni et al. similarly report a case series where ROP progressed after laser photocoagulation presumably due to tractional retinoschisis observed by OCT that remained undetected by indirect ophthalmoscopy [42].
Fig. 4.2
Top left, Three-dimensional representation of clinically undetected preretinal structures identified using spectral-domain optical coherence tomography (SDOCT) in a former 650 g, 23 week post-menstrual age (PMA) gestational age infant imaged at 37 weeks PMA after receiving laser treatment for aggressive posterior retinopathy of prematurity (ROP). Top right, SDOCT B-scan reveals cross section of a former 940 g, 27 week PMA gestational age imaged at 46 weeks PMA demonstrating severe cystoid macular edema. Note the elongated hypo-reflective spaces with intermittent hyper-reflective stranding within the inner nuclear layer. Bottom left, Three-dimensional representation of retinal vessels created from SDOCT scan of a former 700 g, 25 week PMA gestational age infant imaged at 42 week PMA with plus disease. Tortuosity can be appreciated in all dimensions, as opposed to traditional en face examination. Bottom right, retinal surface map of a former 478 g, 23 week PMA gestational age infant imaged at 34 weeks PMA with plus disease represents vessel elevation, as a consequence of vessel dilation, from the inner surface of the retina. Image was produced by segmentation of SDOCT scan and corresponds with adjacent en face vascular pattern
A more thorough study by Lee et al. evaluated 76 eyes of 38 infants by both HHOCT and traditional indirect ophthalmoscopy during routine ROP exams. Masked graders then qualitatively graded the OCT scans for the presence of preretinal and retina findings. They found that 39% of examinations revealed intraretinal cystoid structures and 32% had an ERM on OCT while these features were never identified by conventional exam. However, indirect ophthalmoscopy was able to identify ROP stage and status of plus disease while OCT did not reveal this clinical information [43].
Vinekar et al. similarly compared the SDOCT imaging of patients with ROP to traditional indirect ophthalmoscopy. They identified and described intraretinal cystoid structures in 29% of subjects with stage 2 ROP, none of whom had cystoid macular edema (CME) identified on clinical exam. They further characterized these intraretinal cystoid structures as either elongated with hyper-reflective septae between them and causing foveal deformation and bulging (Fig. 4.2, top right) or fewer intraretinal cystoid structures which did not cause foveal deformation. Additionally, they note that no eyes with stage 1 ROP or without ROP had these findings [16].
Maldonado et al. further characterized this CME in preterm infants. They found CME in 50% of the 42 preterm infants included in their study which most commonly appeared as multiple elongated cystoid structures within the inner nuclear layer (INL). The presence of CME was not associated with ROP outcomes; however, central foveal thickness, the thickness of the INL, and the foveal-to-parafoveal ratio all served as markers of CME severity. While these measures were greater in eyes that required laser photocoagulation or developed plus disease or stage 3 ROP, there was wide overlap of measurements between ROP severities. However, there were no associations between CME and any systemic parameters [44].
The etiology of subclinical CME in subjects with ROP remains unexplained. Some hypothesize that a higher concentration of vascular endothelial growth factor (VEGF) increases vascular permeability and the subsequent formation of cystoid structures. Maldonado et al. posit the CME “may also reflect the contribution of increased intracapillary hydrostatic pressure as a result [of] vascular congestion from plus disease” [44]. In another report, one infant developed CME after bevacizumab treatment (used off-label as anti-VEGF therapy for ROP) while three infants did not experience CME resolution after bevacizumab therapy. Another possibility is these cystoid structures form secondary to mechanical traction exerted on the retina [16]. Although there are no recognized markers of systemic health associated with CME, a case report of an infant with hemochromatosis and severe, bilateral CME that resolved with liver transplant suggests there may be other unidentified systemic health factors associated with CME [45]. It appears that CME does resolve on its own; however, the long term clinical implications remain unknown and require further study [16, 44, 46] doi: 10.1016/j.ophtha.2014.09.022 and 10.1097/IAE.0000000000000579.
OCT has increased our understanding of abnormal development of multiple tissue layers of the infant retina that can be affected by premature birth or by ROP. The most prominent retinal finding on OCT in children with a history of preterm birth and ROP is the notable persistence of inner retinal layers. Recchia et al, in older children and adults, reported an abnormal foveal contour and the persistence of inner retinal layers in those with a history of ROP [47]. Optical coherence tomography allows for an appreciation for the shallower and wider foveal pit in this population [48]. Additionally, there is persistence of inner retinal layers regardless of maximum ROP stage observed [46]. Retinopathy of prematurity appears to affect the development of neurosensory tissue as well. There is thinner RNFL in school-aged children born prematurely with stage 3 and 4 ROP [49]. Tong et al. measured a shallower cup depth in stage 3 ROP or ROP requiring treatment using OCT [38]. Last, while Wu et al. noted a thinner choroid in patients who had undergone treatment for ROP as well as an inverse relationship between choroid thickness and best corrected visual acuity, Park et al. only noted a trend in thinner temporal choroid for children with worsening ROP stage [50, 51].
Evaluation of vasculature is critical to the diagnosis and management of ROP. To date, OCT has only served as a useful adjunct to the gold standard of indirect ophthalmoscopy when characterizing the vasculature and evaluating treatment options [52]. One limitation of this technology has been the inability to reliably characterize ROP stage and determine the status of plus disease. Maldonado et al. recently created a Vascular Abnormality Score on OCT (VASO) in order to quantify vascular and perivascular OCT findings associated with ROP. In contrast to en face color fundus photography or indirect ophthalmoscopy, OCT appreciates three-dimensional vessel distortion, including in the anterior–posterior direction (Fig. 4.2, bottom left). This VASO tool is useful to qualitatively evaluate OCT scans for vessel elevation and severity of elevation, scalloping of the inner plexiform layer and outer plexiform layer, hypo-reflective vessels, and perivascular spaces. Patients with plus disease had a significantly higher VASO due to a greater incidence of vessel elevation, hypo-reflective vessels, and scalloped retinal layers. In addition, retinal surface maps of subjects with plus disease demonstrated greater surface elevation (Fig. 4.2, bottom right). While the VASO system and retinal surface maps require further validation before being incorporated into clinical practice, these tools suggest OCT can provide new information regarding the vascular abnormalities of ROP [53].
Optical coherence tomography has rapidly evolved over the past 20 years and its clinical applications will likely expand in the future. While the initial cost of equipment may limit some health care systems from implementing this technology, the availability of OCT continues to spread. Intraoperative OCT can assist in the management of the complications of ROP, such as retinal detachment and ERMs [41, 54]. OCT was first introduced in the operating room with the same portable, hand-held system brought to the nursery to image infants and children undergoing exams under anesthesia [55]. It has now been integrated into the surgical microscope to allow live cross-sectional imaging while operating [56–58]. Color Doppler OCT, first described in 1997, images fluid flow which allows for in vivo imaging of blood flow and subsequent three-dimension reconstruction of blood vessels [59–61]. While Doppler OCT has not yet been investigated in the context of ROP, one can easily appreciate the utility of cross-sectional visualization of blood flow in this population. In addition, advances to the actual OCT apparatus such as the development of a hand-held swept source OCT system will allow for faster data acquisition with greater precision [62]. With so many advances on the horizon, OCT promises to revolutionize the diagnosis and management of ROP.
Wide Field Imaging and Fluorescein Angiography
What Is Wide Field Imaging?
Evaluation of the peripheral retina is important when diagnosing and monitoring treatment for patients with many pathologies, including ROP. When discussing the field of view obtained by an imaging device, one refers to the external angle of light as it enters the eye, which is determined by the focal length, dimensions of the imaged object, and lens power. Fundus imaging has traditionally been limited to approximately 30° field of view [63]. Fields of view have expanded as individuals have altered the optic systems of devices. For example, Pomerantzeff could image nearly 150° of the retina with his contact lens-based system that relied on both trans-pupillary and trans-scleral illumination [64].
The most commonly accepted wide-field imaging tool currently in practice for the ROP population is the RetCam 120 (Clarity Medical Systems, Inc., Pleasanton, CA) which obtains up to 130o digital fundus photographs with the assistance of different contact lenses. First commercially available in 1997, RetCam was quickly adopted by the pediatric community as it allows portable, supine imaging of the peripheral retina. Because the image quality depends on having a clear lens, RetCam typically produces much poorer images in the adult population [63]. The newer nonmydriatic Optomap Panoramic 200 scanning laser ophthalmoscope (Optos P200, Optos, Dumferline, Scotland, United Kingdom) allows for 180o–200o fundus imaging and is less susceptible than fundus photo to media opacities such as cataract [65]. The Optos’s ellipsoid mirror creates two focal points which allows for greater fields of view [63].
Several advantages and disadvantages exist when comparing the feasibility of the RetCam and Optos P200 MA to image infants. The Optos P200 MA allows for greater field of view while the RetCam will have more difficulty imaging multiple areas simultaneously. As RetCam relies on incandescent light while Optos uses dual laser illumination with a more uniform focus, RetCam may have a more difficult time imaging infants with media opacities or small pupils. Image acquisition time, very important when imaging noncooperative infants with eye movement, is longer for the RetCam system as the user must refocus and reapply coupling media. However, the RetCam system is portable which allows easy imaging of supine infants in unique settings, such as the nursery; the Optos system is traditionally limited to outpatient imaging and requires infants to be held in an upright position. The noncontact image acquisition of Optos may reduce the drawback of changes in vessel appearance from ocular compression but also may introduce artifacts if there is matter between the device and eye [66]. Because there are two laser wavelengths used with Optos, this imaging system does not accurately portray colors.
Wide Field Imaging and ROP Telemedicine
The growing field of telemedicine has benefited from these digital wide-field imaging tools. These devices obtain fundus photographs and videos of patients so they may be analyzed off-site. While these technologies can be performed by a technician or nurse with a more flexible schedule than a pediatric ophthalmologist, the initial cost of the equipment can be expensive and they offer less complete visualization of the fundus and subsequently documentation of ROP compared to the gold standard of indirect ophthalmoscopy [67]. By comparing off-site digital photograph reading to traditional indirect ophthalmoscopy and determining the sensitivity and specificity in detecting ROP, we can decide if digital wide-field imaging can serve as a screening tool for ROP.
Multiple studies have demonstrated variable sensitivity and specificity of ROP diagnoses by RetCam versus indirect ophthalmoscopy [68–71]. Discrepancies in study protocols and aims when comparing published results, such as sensitivity and specificity of ROP diagnoses by RetCam, has led to a standardized, prospective, multicenter trial entitled “Telemedicine Approaches to Evaluating Acute-phase ROP (e-ROP, www.clinicaltrials.gov identifier NCT01264276).” The trial will “evaluate the validity, reliability, feasibility, safety, and relative cost-effectiveness of a retinopathy of prematurity (ROP) telemedicine evaluation system to detect eyes of at-risk babies who meet referral warranted ROP (RW-ROP) criteria and therefore need a diagnostic evaluation by an ophthalmologist experienced in ROP.” With a planned enrollment of 2500 preterm infants independently examined via indirect ophthalmoscopy and by RetCam digital images at a remote location, this study should determine whether digital wide-field imaging can effectively serve as a surrogate for traditional on-site ophthalmologic examination.
Wide-Field Fluorescein Angiography and ROP
The incorporation of FA into wide-field imaging promises to change how we monitor treatment of ROP. In 2006, Ng et al. first presented the feasibility of imaging the peripheral vasculature by applying the fluorescein angiographic RetCam digital system [72]. Lepore et al. have since provided a more thorough atlas of the vascular abnormalities associated with severe ROP. They obtained images before and after laser photocoagulation and believe this helped determine the efficacy of the ablation. Wide variability in the time lapse from intravenous injection to the arteriolar phase and then venular phase suggests great variability in blood flow with these infants, including in the choroid. The authors note irregular branching patterns at the vascular–avascular junction and even more variable vascular findings, including dilatation of capillaries, capillary tuft formation, and rosary-bead-like hyperfluorescent lesions in the vascular retina and posterior pole [73]. Purcaro et al. provide an additional case series demonstrating the efficacy of RetCam FA to characterize vessel branching at the vascular–avascular junction and suggest dye leakage may be a sign of progression to severe ROP [74]. Additionally, RetCam FA can identify vascular changes before they are detected with traditional indirect ophthalmoscopy [75].
Several case series suggest wide-field FA is a useful perioperative tool when monitoring the treatment for different sequelae of ROP. For example, Yokoi et al. obtained RetCam 120 FA images before and after laser and surgical treatment for six eyes with aggressive posterior ROP. Pre-operatively, they could appreciate capillary nonperfusion, shunting in vascularized retina, and a circumferential demarcation line while after treatment they could appreciate a capillary-free zone, poorly developed disorganized vessels, and an inhomogeneous capillary bed [76]. The same research group reports the utility of RetCam120 FA when securing sclera buckles due to stage 4A ROP. All five eyes that underwent scleral buckling demonstrated pre-operative fluorescein leakage from fibrovascular tissue. Between 7 and 20 days after successful surgery, four eyes had decreased leakage while one eye had no leakage [77].
The introduction of bevacizumab as an off-label treatment for ROP offers another opportunity to advance patient care with the assistance of wide-field FA. Wide-field FA can successfully confirm the absence of pathologic neovascularization and document vascular density and evolution of anomalous vascular patterns after bevacizumab treatment [78]. However, additional studies suggest that wide-field FA can help monitor for disease progression that may occur after treatment because of the potential for transient blockage of VEGF with bevacizumab versus the long-acting down-regulation associated with laser photocoagulation (Fig. 4.3). Hoang et al. provide a case report of circumferential anastomosis at the bevacizumab-induced regression site with future anterior involvement observed by RetCam FA. They report an infant with zone II, stage 3 plus disease responded well to intravitreous bevacizumab. However, 2 months after treatment, a second stage 3 complex formed anterior to the original lesion with RetCam FA demonstrating “anterior extraretinal fibrovascular proliferation with leakage and a more posterior circumferential vascular ring with associated telangiectasis but without leakage” [79]. Hu et al. have additionally documented this late reactivation of disease after transient regression following bevacizumab as observed with RetCam FA [80]. Visualization of peripheral retina hemodynamics with the assistance of wide-field FA clearly serves as a clinically relevant adjunct to traditional indirect ophthalmoscopy.
