In this study, UWF imaging provided adequate visualization of the posterior segment in Kpro implanted patients. This imaging modality allowed for noninvasive longitudinal monitoring of retinal and optic nerve clinical pathology in this patient population. Serial visualization and documentation of glaucomatous optic discs and peripheral retinal diseases were successfully achieved by incorporating UWF imaging. Poor vision, ocular comorbidities, and the development of postoperative complications may impact image acquisition and quality in this sample of patients. In addition, since different technicians may have acquired images, the quality of images could also have been influenced by the technician’s skill. Images may therefore be better or worse depending on the patient, the person taking the images, and when the image was acquired. Despite these differences many clinically useful images were captured.

Visualization of posterior segment structures in Kpro patients is often difficult with indirect ophthalmoscopy and with standard fundus photography due to patient-specific ocular comorbidities, post-surgical complications, and the small aperture of the Kpro implant [6, 7, 14,15,16]. The results of this study support existing data suggesting UWF imaging provides meaningful clinical utility in the diagnosis and follow-up of posterior segment disease in Kpro patients. In a case series of 10 Kpro-implanted patients, Kornberg et al. found that UWF imaging detected 100% of pathology identified on clinical examination [17]. The authors also reported improved detection of posterior segment disease compared with clinical examination in patients with RPM, a complication seen in up to 65% of Kpro implanted patients [9, 15, 17]. In our study, one image from an RPM patient was rated as “Poor” and deemed ungradable by both observers. The remaining four images from patients with a RPM provided at least some degree of clinically meaningful information. Imaging obstacles, such as a RPM or the presence of silicone oil, appeared to impact the quality of some images. In addition, capturing the anterior retina in all four quadrants simultaneously was difficult to achieve in this patient population. These findings are consistent with those reported in past studies [17,18,19]. Employing eye steering techniques, changing patient positioning, and acquiring multiple images in a single imaging session may improve the success rate of high quality images in this patient population [20,21,22,23].

Velez-Montoya et al. investigated the interobserver agreement of UWF imaging in a sample of 13 type I Kpro patients [18]. In their prospective study, patients underwent indirect ophthalmoscopy by two experienced retina specialists who then completed a 30-question questionnaire assessing posterior segment clinical pathology and anatomy. Afterwards, a minimum of three UWF images were taken for each patient. Forty-eight hours after the initial examination, the same retina specialists repeated the questionnaire using the UWF images. The authors reported moderate to high interobserver reliability when evaluating posterior segment anatomy with UWF imaging, which was consistent with the findings of our study [18]. Furthermore, we found moderate to high agreement when evaluating posterior segment anatomy between two retina fellow observers, providing additional support of UWF imaging use in this patient population.

Past studies investigating the use of UWF imaging in Kpro implanted patients included only the type I Kpro [17, 18]. In this report we included the use of UWF imaging in a patient implanted with a type II Kpro. In this patient, some anatomic structures, such as the optic nerve, were visible with meaningful clarity with UWF imaging. This was significant as this patient had a history of glaucoma that could no longer be assessed upon clinical examination as there was no view of the posterior segment.

In difficult to image patient populations, UWF imaging was found to be a superior modality for viewing the posterior segment compared to clinical examination and standard fundus photography [24]. Previous studies have evaluated the sensitivity and specificity to detect retinal lesions using UWF imaging in various patient populations [22, 23, 25, 26]. We did not evaluate the sensitivity and specificity of UWF imaging to dilated fundus examination or standard fundus photography in this study. While dilated fundus examination is considered the gold standard for detection of posterior segment disease, in our cohort of Kpro patients, there was often no view of the posterior segment on clinical examination and standard fundus photography was not routinely performed due to low clinical utility [22]. It would be difficult to determine the sensitivity and specificity of UWF imaging to detect posterior segment pathology in the Kpro patient population due to the lack of a reliable and valid standard of care [18].

A recent report of 169 type I Kpro implanted eyes found posterior segment complications using a combination of clinical examination, B-scan ultrasonography, and optical coherence tomography in approximately 40% of eyes, with the highest incidence complications being ERM (16.6%), CME (12.4%), vitreitis (11.2%), retinal detachment (9.5%), and endophthalmitis (4.1%) [16]. Chew et al. reported preexisting glaucoma in 73% of Kpro patients with increased intraocular pressure (38%) and glaucoma progression (14%) being two common postoperative complications after Kpro implantation [8]. Neither study included the use of standard fundus photography or UWF imaging as a means to assist in the diagnosis or monitoring of posterior segment disease. Here, 73.7% of eyes had glaucoma and 47.4% of eyes had a retinal condition at the time of UWF imaging. Visualization of the optic nerve, macula, and anterior retina with discernable details was possible using UWF imaging alone with demonstrated agreement between observers. Longitudinal imaging facilitated management of complex ocular conditions, including retinal detachment and secondary glaucoma following the reimplantation of a type I Kpro in the setting of fungal endophthalmitis and failed penetrating keratoplasties.

A limitation of this study was the lack of a formal observer training period and no expert consensus scale for image quality grading in this patient population. This may explain the limitation in interobserver agreement findings related to image quality. While our results only found fair interobserver agreement regarding image quality, full agreement between the two observers was noted in most of the images. Other limitations of this study included those inherent to a retrospective, non-controlled, non-randomized study design, the small sample size, and no defined imaging acquisition guidelines or protocols. The consecutive enrollment strategy employed in this study may have been a source of selection. In our study, Boston type I and type II Kpro patients were enrolled; therefore, these results may not extend to patients with other types of Kpro, such as in those with an osteo-odonto-keratoprosthesis. Owing to the retrospective study design and the lack of defined imaging protocols for this specific patient population, the results may not reflect the highest possible quality images. Thus, the true clinical utility of UWF imaging in Kpro patients may be understated based on the results of this study. Limitations associated with the Optos scanning laser ophthalmoscope imaging system include the need for a trained technician, imaging artifacts associated with the keratoprosthesis, and patient-specific fixation difficulties resulting in the need to acquire multiple images during a single imaging session.

Despite these limitations, the results of this study are in concordance with past studies. While the sample size was relatively small, the number of enrolled eyes was greater than those enrolled in previous studies. This study also adds evidence that UWF imaging can be used as a rapid and reliable method to serially image the posterior segment and may provide clinically meaningful findings in both type I and type II Kpro patients. UWF imaging serves as a valuable tool for the physician to monitor and manage these challenging patients.