Prospective Evaluation of iCam-OCT

This is a prospective, non-significant-risk clinical study sponsored by Siloam Vision, Inc. to evaluate whether ophthalmologists can assess retinopathy of prematurity (ROP) using images from the iCam-OCT device as effectively as with the current standard, binocular indirect ophthalmoscopy (BIO). Results will support an FDA 510(k) submission. A separate follow-up study (undilated OCT imaging, with dilation only before the BIO exam) is planned once this study completes.

Device. The iCam-OCT is a handheld, swept-source OCT (SS-OCT) system operating at 1040nm and 400kHz, with 7-micron depth resolution and a 12mm scan depth. It captures an ultra-widefield (197°) field of view - the widest reported for OCT - in a contact-based, non-mydriatic mode, with scans under one second and a full bilateral exam averaging 1-2 minutes. It's designed for supine imaging of infants/small children who can't use conventional tabletop OCT. It's a Class 1 laser product conforming to IEC 60601-1, IEC 60601-1-2, IEC 60825-1, and ANSI Z80.36-2021. Over a dozen prior publications using earlier iCam-OCT prototypes (500+ patients, 1700+ imaging sessions) reported no adverse events and found the images clinically useful for ROP staging, zone assessment, and detection of related pathologies.

Sites and population. Two sites: Oregon Health \& Science University (Doernbecher NICU) and Children's Hospital Colorado Anschutz. Both sites are new to the device and unaffiliated with the sponsor. Enrollment target is a minimum of 40 premature infants meeting AAP ROP-screening criteria (birth weight <1500g or gestational age ≤30 weeks; or birth weight 1500-2000g or gestational age >30 weeks if deemed at-risk). At least 30 infants must have ROP (stage 1 or worse in any zone) and at least 10 without ROP, with at least 20 subjects per site. Main exclusions are inability to obtain parental consent or clinical instability precluding research imaging.

Study design. After informed consent, each infant undergoes both a BIO exam (by an independent, ROP-experienced ophthalmologist unaffiliated with the sponsor) and iCam-OCT imaging, in randomized order, following pupil dilation per routine care. Imaging is performed by a trained/certified operator (ophthalmologist, nurse, technician, or coordinator) using a standardized protocol with up to six suggested scan views (posterior, temporal, nasal, temporal/nasal ora serrata, posterior HD). Anonymized OCT images are then graded independently by three masked, ROP-experienced ophthalmologists for image quality (0-3 scale) and clinical utility (usefulness for confirming normal structures or detecting pathology), and separately for ROP zone/stage/plus disease, without knowledge of the BIO findings.

Endpoints and performance goals.

Primary: percent of OCT image sets graded acceptable (score ≥2) for both quality and clinical utility - goal ≥85%.

Secondary: sensitivity/specificity of OCT-based ROP detection vs. BIO (goals >90% sensitivity, >70% specificity); documentation/visualization of full retinal vascularization to the ora serrata (same goals); and documentation of the vascular/avascular border in eyes with incomplete vascularization (same goals).

Safety: tracking of systemic AEs (nurse-initiated pauses), ocular AEs (corneal abrasion, conjunctival irritation, infection, need for topical medication), and any other AEs.

Discordant cases - infants classified "no ROP" by BIO but flagged as stage 1+ by OCT (potential false positives relative to the BIO gold standard) - get extra scrutiny: a coordinator reviews subsequent hospital records for later-diagnosed ROP, and all three graders reconvene to reach consensus on whether the OCT findings represent true early-stage ROP that BIO missed, versus immature retina with a defined vascular border. Clinical management throughout the study is driven by BIO only, not by OCT findings.

Statistics. The eye is the unit of analysis; each eye is graded independently by all three masked graders, and sensitivity/specificity are computed per grader and averaged. Sample size (n=30 for the ROP-positive group) was powered via a non-inferiority test assuming 85% BIO sensitivity (from a cited 2018 Biten et al. study), a 5% non-inferiority margin, and one-sided alpha of 0.05, yielding ~80% power. Supportive analyses include per-grader McNemar's tests for OCT/BIO discordance and non-inferiority testing of OCT sensitivity against the fixed 85% literature benchmark (not a within-study BIO comparison).

Safety monitoring. Risks are characterized as low (corneal abrasion, subconjunctival hemorrhage, infection - similar to standard contact ROP exams) plus general risks of apnea/bradycardia inherent to examining fragile premature infants. Dr. Benjamin Young serves as the independent safety monitor overseeing adverse event review; AEs/SAEs must be reported to the IRB within 5 days.