Preterm Infant Stress During Noncontact Handheld Optical Coherence Tomography and Contact Fundus Photography

Abstract

Importance:

Limiting the stress of retinopathy of prematurity (ROP) assessments may be beneficial for preterm infants.

Objective:

To compare stress in preterm infants from non-contact optical coherence tomography (OCT) versus contact fundus photography during ROP examinations

Design:

Prospective observational study

Setting:

Intensive care nursery at a single tertiary care center

Participants:

Infants enrolled in STudy of Eye imaging in Preterm infants (BabySTEPS2, NCT04995341) from June 2023--January 2025

Exposure:

During clinically indicated ROP examinations, research imaging with an investigational handheld non-contact OCT system (Duke University) and a contact fundus camera (RetCam3^®, Natus, USA) were performed in randomized order in the study eye by expert imagers. Stress data were captured by bedside research nurses.

Main Outcomes and Measures:

Modified CRIES neonatal pain score (cry score 0–4, facial expression score 0–2), heart rate, and oxygen saturation and adverse events (AEs) were collected ≤5 minutes before, during, and for 1–5 minutes after imaging with each device.

Results:

Fifty-nine preterm infants with mean gestational age of 26.1 weeks underwent 397 OCT and fundus photography imaging sessions. Regardless of device order, for OCT compared with fundus photography there was less change in 1) cry score from baseline to during imaging (0.12 vs 1.70; mean difference −1.59 [95% confidence interval (CI), −1.75 to −1.42]) and after (−0.07 vs 0.31; −0.38 [95% CI, −0.47 to −0.28]) (P<.001 for both), 2) facial expression score from baseline to during (16.8% vs 75.8%) and after (2.6% vs 22.5%) imaging (P<.001 for both) and 3) heart rate (bpm) from baseline to during (−5.2 vs 10.9; −16.1 [95% CI, −19.2 to −13.0]) and after (−7.2 vs 18.8; −25.9 [95% CI, −29.3 to −22.6]) imaging (P<.001 for both). Thirty-two AEs (tachycardia, bradycardia, hypoxia, or emesis) occurred in 28/399 (7.0%) imaging sessions; 26/32 (81.3%) were during or after fundus photography. No ocular, unexpected, or severe AEs occurred.

Conclusions and Relevance:

While the role of handheld OCT in ROP screening remains under investigation, this study provides evidence to support that non-contact handheld OCT is less stressful to preterm infants undergoing ROP examination and results in fewer AEs compared with contact fundus photography.

Introduction:

Retinopathy of prematurity (ROP) is the leading cause of preventable childhood blindness worldwide.¹ With improved survival of increasingly premature infants, the prevalence of ROP continues to rise each year.² Binocular indirect ophthalmoscopy (BIO) has been the standard of care for ROP examinations, and over 70,000 infants in the United States (US) receive one or more ROP examinations with BIO each year.³ The growing number of infants requiring screening and the national shortage of pediatric ophthalmologists has led to a lack of qualified ophthalmologists to perform in-person ROP examinations, especially in remote nurseries.^4,5 Access to ROP screening has been expanded through telemedicine with fundus photography.⁶ However, contact fundus photography can be stressful on infants, as it requires topical anesthetic, speculum placement, eye manipulation, and bright light exposure.^4,7–11 Non-contact fundus photography (e.g. Pictor, Volk, USA) is safer in infants compared with BIO but still emits a bright white light.¹² Neonatal stressors are known to lead to poorer neurodevelopmental outcomes.¹³

Handheld swept source OCT has been used and shown to be safe in neonates and pediatric patients across multiple studies.^14–16 Handheld OCT imaging in awake infants at the bedside is thought to be less stressful than conventional contact fundus photography or BIO, because the near-infrared illumination of the OCT is invisible to the infant and thus does not elicit an aversion or lid closure response.^17,18 We have previously shown that handheld non-contact OCT causes less behavioral and physiological stress in preterm infants when compared with BIO.^18,19 A recent study also showed that laser speckle contrast imaging, a non-contact imaging device that can visualize and quantify retinal blood flow and does not emit bright light, was less stressful to preterm infants compared with BIO.²⁰

Handheld OCT has emerged as a promising imaging tool in infants with ROP, providing visualization of retinal microanatomy and cross-sectional vitreoretinal features not apparent on BIO.^21–24 Various OCT biomarkers have been associated with prematurity and ROP including cystoid macular edema, delayed foveal development, and thin choroid.^18,21,25 Handheld OCT can be used to detect the presence of plus or pre-plus disease and even predict the clinical stage of ROP or presence of referral-warranted (RW)-ROP^21,26–32 There has not been a published study to date comparing the behavioral and physiological responses of preterm infants to handheld OCT and fundus photography during ROP examinations. We hypothesize that non-contact handheld OCT is less stressful to preterm infants compared with contact fundus photography. The head-to-head comparison of stress caused by these two imaging modalities is one step towards determining if using handheld OCT can be feasible for diagnosing and monitoring ROP.

Methods:

This prospective observational cohort study is part of the Retinal Microanatomy in ROP to Improve Care (BabySTEPS2), which was approved by the Duke University Health System Institutional Review Board, registered with clinicaltrials.gov (NCT04995341, Cohort 3), and adheres to the Health Insurance Portability and Accountability Act and the Declaration of Helsinki. Infants were eligible to participate in BabySTEPS2 if eligible for ROP screening, birthweight ≤1000 grams and/or gestational age (GA) <29 weeks, and written consent by a parent or legal guardian; approval by the bedside nurse (based on health status) was required for research participation on the day of imaging. Subjects did not receive any incentives or stipends for their participation.

The head-to-head study period started in June 2023 and went through the end of the enrollment of the BabySTEPS2 Cohort 3 in January 2025. Because of the delay in start of the head-to-head study period, 69 infants were enrolled rather than the 102 intended. Of those enrolled, 9 were withdrawn from the study (5 transferred to another nursery and 4 died) and one infant aged out of this study by the time of startup. The remaining 59 infants had 399 imaging sessions. Infants were imaged on the same day as clinically indicated ROP examinations (either for screening or follow up after treatment), after both eyes were pharmacologically dilated and examined with BIO as part of clinical care. Under the head-to-head study design, rather than having both eyes imaged with OCT as in BabySTEPS, one eye of each infant was imaged at the bedside by a certified imager with the non-contact investigational ultracompact UC5 OCT system (Duke University, Durham, North Carolina) and with the commercial RetCam 3 (Natus Sensory, Schaumburg, IL). The eye to be imaged was randomly selected at the first visit and that same eye was imaged throughout the study. The order of imaging device was randomized for each visit. During OCT imaging, scans centered on the macula, optic nerve and papillomacular bundle were captured along with the temporal periphery and nasal/inferior/superior periphery if possible. For fundus photography, an image centered on the optic nerve was captured along with views of the temporal, nasal, superior, and inferior periphery. Imaging typically involved swaddling the infant and a change in infant positioning for access. The following were used as needed (at the discretion of the physician, imager and/or bedside nurse): anesthetic eye drops (proparacaine 0.5%), an eyelid speculum, eye lubrication with preservative free artificial tears or gel, oral sucrose, or pacifier. (Both imaging modalities were typically performed within 15 minutes.

Infant stress was assessed by bedside research nurses based on three items: 1) a modified version of the validated CRIES (C-Crying; R-Requires increased oxygen administration; I-Increased vital signs; E-Expression; S-Sleeplessness) neonatal pain assessment tool,³³ 2) prespecified vital sign abnormalities, and 3) adverse events (AEs). The modified CRIES scale (detailed in Table 2) consisted of: crying (0–4 scale), facial expression (0–2 scale), heart rate (HR), oxygen saturation, respiratory support, and fraction of inspired oxygen (FiO2). Prespecified vital sign abnormalities included bradycardia, tachycardia, and oxygen desaturation based on alarm limits set per neonatal intensive care unit (NICU) protocol for each infant for that day. Finally, the incidence of AEs including prolonged bradycardia or tachycardia (HR < 80 or HR > 200 for longer than 30 seconds), sustained hypoxia (oxygen sat under 80% for longer than 30 seconds), and emesis were recorded. Any unanticipated AEs were also documented. Data were collected at the bedside by NICU research personnel and recorded: 1) up to 5 minutes before, 2) during, and 3) for 1–5 minutes after retinal imaging with each device. The bedside nurse was contacted after a minimum of 30 minutes and up to 2 hours to capture additional AEs.

Table 2:

Behavioral and Physiological Stress Outcomes Associated with Non-contact Handheld Optical Coherence Tomography (OCT) and Contact Fundus Photography

	Before			During			After
	OCT	RetCam	P **	OCT	RetCam	P **	OCT	RetCam	P **
Cry Score
0=No crying	363 (91.2%)	379 (95.2%)		326 (82.1%)	46 (11.6%)		381 (96.0%)	294 (73.9%)
1=Moans or cries minimally	28 (7.0%)	17 (4.3%)		54 (13.6%)	105 (26.4%)		14 (3.5%)	69 (17.3%)
2=Appropriate crying (not irritable)	4 (1.0%)	2 (0.5%)		14 (3.5%)	158 (39.7%)		2 (0.5%)	30 (7.5%)
3=High pitched	3 (0.8%)	0 (0.0%)		2 (0.5%)	78 (19.6%)		0 (0.0%)	5 (1.3%)
4= Inconsolable	0 (0.0%)	0 (0.0%)		1 (0.3%)	11 (2.8%)		0 (0.0%)	0 (0.0%)
Mean	0.11	0.05		0.23	1.76		0.05	0.36
Mean difference (95% CI)**	0.06 (0.01, 0.11)		.01	−1.53 (−1.69, −1.36)		<.001	−0.32 (−0.40, −0.23)		<.001
Facial Expression
0=No grimace	357 (89.7%)	369 (92.7%)		294 (74.1%)	63 (15.8%)		376 (94.7%)	291 (73.1%)
1=Grimace alone	34 (8.5%)	23 (5.8%)		61 (15.4%)	173 (43.5%)		16 (4.0%)	90 (22.6%)
2=Grimace and grunt	3 (0.8%)	4 (1.0%)		22 (5.5%)	104 (26.1%)		3 (0.8%)	11 (2.8%)
9=Unable to assess	4 (1.0%)	2 (0.5%)		20 (5.0%)	58 (14.6%)		2 (0.5%)	6 (1.5%)
Mean	0.10	0.08		0.28	1.12		0.06	0.29
Mean difference (95% CI)**	0.02 (−0.02, 0.07)		.29	−0.84 (−0.94, −0.74)		<.001	−0.23 (−0.29, −0.17)		<.001
Heart Rate, beats/min
Mean	174.40	166.30		169.20	177.19		167.27	185.06
Mean difference (95% CI)**	8.13 (6.20, 10.06)		<.001	−7.98 (−10.54, −5.42)		<.001	−17.79 (−20.13, −15.44)		<.001
O2 Saturation
Mean	96.64	97.15		96.66	96.14		96.84	95.51
Mean difference (95% CI)**	−0.52 (−0.88, −0.15)		.01	0.51 (−0.05, 1.08)		.08	1.33 (0.85, 1.81)		<.001
FiO2
Mean	25.96	25.89		26.00	25.95		26.04	26.12
Mean difference (95% CI)**	0.05 (−0.12, 0.22)		.56	0.04 (−0.15, 0.23)		.69	−0.09 (−0.39, 0.21)		.54
Number (%) of Bradycardia events *	6 (1.5%)	2 (0.5%)	.16	16 (4.0%)	22 (5.5%)	.47	2 (0.5%)	6 (1.5%)	.16
Number (%) of Tachycardia events *	26 (6.5%)	9 (2.3%)	.001	32 (8.1%)	146 (36.7%)	<.001	15 (3.8%)	120 (30.2%)	<.001
Number (%) of Desaturation events *	49 (12.3%)	39 (9.8%)	.18	86 (21.7%)	99 (24.9%)	.25	29 (7.3%)	61 (15.3%)	<.001

^*Bradycardia, tachycardia, and oxygen desaturation events were based on preset alarm levels for each infant

^**Difference was calculated as mean in OCT group minus mean in RetCam group. 95% CI= 95% confidence interval. Generalized estimating equations (GEE) were used to account for repeated visits and measurements.

Data were analyzed with the two-sample t-tests for normal data, Wilcoxon rank sum tests for skewed data, and Chi-square tests for categorical measures. 95% confidence intervals (95% CI) for the difference in stress measures between OCT and Retcam groups were calculated from generalized regression models for data with repeated measures, with generalized estimating equations used to account for correlation from repeated measurements. All analyses were performed in SAS (v9.4, SAS Institute, Cary, North Carolina) and two-sided P<0.05 was considered statistically significant. All P values were two-sided without adjustment for multiple analyses. STROBE reporting guidelines were followed for preparation of the manuscript.

To determine whether prior studies had evaluated preterm infant stress from OCT versus contact fundus photography during ROP examinations, a systematic literature search was conducted in PubMed/MEDLINE for all articles through January 9^th, 2025. The search string (for keywords in the title or abstract) was as follows: (“optical coherence tomography” OR “indirect ophthalmoscopy” OR “retinal imaging” OR “fundus photography” “imaging”) AND (“retinopathy of prematurity”) AND (“stress” OR “safety” OR “pain”). A total of 51 articles were identified and reviewed. None of the articles addressed the study question.

Results:

The 59 included infants had a mean (SD) gestational age of 26.1 (2.0) weeks and birthweight of 790 (252) grams (Table 1). Mean PMA of the infant at time of imaging was 38.5 (4.9) weeks (range: 31 to 55). Of 399 imaging sessions, 190 were with OCT first and 209 with fundus imaging first. Both OCT and fundus photography were successfully captured in 397 sessions. In the two unsuccessful sessions, either OCT (n=1) or fundus photography (n=1) was not performed per bedside nurse recommendation. The mean (SD) duration of imaging was 4.2 (2.2) minutes for OCT and 2.0 (1.2) minutes for fundus photography (P<.001) (Table 1 and eFigure 1). There was no correlation between the amount of time spent on imaging on either device or any stress measures (eTable 1). Infants were on respiratory support in 278 of 399 sessions (69.7%); this included 217 sessions with oxygen by nasal cannula, 47 with continuous positive airway pressure (CPAP), and 14 with mechanical ventilation. Contact fundus photography required instilling anesthetic eye drop and inserting a lid speculum and lubricating eye gel in all imaging sessions; whereas an anesthetic eye drop and lid speculum were used in 14 (3.5%) of OCT imaging sessions.

Table 1:

Descriptive Analysis of Infant Cohort Characteristics and Imaging Sessions. There were a total of 59 infants and 399 imaging sessions.

Gender
Female	24 (40.7%)
Male	35 (59.3%)
Birth weight (g)
Mean (SD)	790.1 (252.4)
Range	440–1545
Gestational Age (weeks)
Mean (SD)	26.06 (1.97)
Range	22 weeks - 30 weeks 2 days
Required ROP Treatment
No	46 (78.0%)
Yes	13 (22.0%)
PMA at imaging (weeks)
Mean (SD)	38.46 (4.89)
Range	31–55
Imaging Order
OCT first	190 (47.6%)
Retcam first	209 (52.4%)
Duration of Imaging (mins)
OCT: mean (SD)	4.22 (2.23)
Retcam: mean (SD)	1.97 (1.17)
Total time: mean (SD)	6.19 (2.78)

Stress Factors

Stress outcome measures captured before, during, and after imaging with each device are reported for each time interval in Table 2 along with number (%) of prespecified vital sign abnormalities. Across all imaging sessions, the mean cry score was lower for OCT compared with fundus photography during (0.23 vs 1.76; −1.53 [95% CI, −1.69 to −1.36]) and after imaging (0.05 vs 0.36; −0.32 [95% CI, −0.40 to −0.23]) (P<.001 for both). Mean facial expression score (grimace) was lower for OCT both during (0.28 vs 1.12; −0.84 [95% CI, −0.94 to −0.74]) and after imaging (0.06 vs 0.29; −0.23 [95% CI, −0.29 to −0.17]) (P<.001 for both). In the few instances where facial expression could not be assessed (Table 2), it was due to limited view of the face. The cry score before imaging was higher for OCT compared with fundus photography. This appeared to be due to persistent crying from infants who were imaged with fundus photography followed by OCT. In these infants undergoing OCT second, the mean cry score before OCT was 0.19 compared with 0.04 in OCT first infants (0.15 [95% CI, 0.08 to 0.21], P<.001). The mean cry score before Retcam was 0.079 in infants who were imaged with Retcam second compared with 0.029 in those imaged with Retcam first (0.05 [95% CI, −0.10 to 0.001], P=.06) (eTable 2).

Mean heart rate was lower for OCT compared with fundus photography during imaging (169.2 beats per min [bpm] vs 177.2 bpm, −7.98 [95% CI −10.54 to −5.42]) and after imaging (167.3 bpm vs 185.1 bpm, −17.8 [95% CI −20.13 to −15.44]) (P<.001 for both, Table 2). Based on preset alarm levels for each infant, a smaller proportion of infants were tachycardic with OCT compared with fundus photography both during (8.1% vs 36.7%) and after (3.8% vs 30.2%) imaging (P<.001 for both). The mean heart rate and incidence of tachycardia before OCT was higher than before fundus photography. Infants who underwent OCT second (after fundus photography) had higher mean heart rate compared with those who had OCT first (180 vs 168, 12.12 [95% CI 9.01 to 15.23]) (P<.001). Infants that had OCT following fundus photography also had a higher incidence of tachycardia (10.1%) before OCT compared with those that had OCT first (2.6%) (P=.003). There was no difference in mean heart rate and incidence of tachycardia for infants who underwent fundus photography second compared with fundus photography first (P=.35 and P=.63 respectively) (eTable 2).

There was no difference in oxygen saturation during imaging with either device, but mean oxygen saturation was higher after imaging with OCT (96.8 [4.6]%) compared with fundus photography (96.8% vs 95.5%, 1.33 [95% CI 0.85 to 1.81]) (P<.001). Based on preset alarm levels, fewer infants desaturated after OCT compared with after fundus photography (7.3% vs 15.3%; P<.001). The mean O2 saturation before imaging was slightly lower for OCT compared with fundus photography (96.6 vs 97.2; −0.52 [95% CI, −0.88 to −0.15], P=.01). There was no difference in O2 saturation before OCT based on whether OCT was done first or second (P=.69), and O2 saturation before fundus photography was also similar whether fundus photography was done first or second (P=.29) (eTable 2). Finally, there was no difference between FiO2s or between the incidence of bradycardia during or after imaging by the two devices.

Change in stress factors from baseline

The change in stress factors relative to baseline (before imaging) are summarized in Table 3. For all stress factors except FiO2, change from before to during imaging and from before to after imaging was less for OCT compared with fundus photography (Table 3). During imaging, the change in mean (SD) cry score from baseline was 0.12 (0.60) for OCT compared with 1.70 (1.00) for fundus photography (P<.001) and the percent of sessions with worsening of facial expression score was 16.8% for OCT compared with 75.8% for fundus photography (P<.001). The mean change in heart rate during imaging was −5.2 (16.1) bpm for OCT and 10.9 (23.1) bpm for fundus photography (P<.001). Oxygen saturation change was minimal during OCT (0.02 [4.16]) and decreased during fundus photography (−1.0 [4.6]) (P=.004). The changes from baseline to after imaging were similar to those found during imaging in cry score, facial expression score, and heart rate and oxygen saturation (Figure 1 and Table 3). There was no difference in mean change in FiO2 from baseline between OCT and fundus photography both during and after imaging. In 14 (3.5%) of the 399 imaging sessions, there was a change in FiO2: one infant went from CPAP to no CPAP and 13 had an increase in FiO2 with an increase in oxygen delivery (8 on nasal cannula, 3 on CPAP, and 2 on mechanical ventilation).

Table 3:

Change in Stress Outcome Measures from Baseline to During or After Non-contact Handheld Optical Coherence Tomography (OCT) and Contact Fundus Photography

	Change from before to during imaging				Change from before to after imaging
Measure	OCT	RetCam	Mean Difference (95%CI)	P *	OCT	RetCam	Mean Difference (95%CI)	P *
Mean Increase in cry score	0.12	1.70	−1.59 (−1.75, −1.42)	<.001	−0.07	0.31	−0.38 (−0.47, −0.28)	<.001
Facial expression worsened, No. (%)*	63 (16.8%)	257 (75.8%)	−59.0% (−65.7%, −52.3%)	<.001	10 (2.6%)	88 (22.5%)	−19.9% (−24.8%, −15.2%)	<.001
Mean Change in heart rate in beats/min	−5.22	10.89	−16.12 (−19.24, −13.00)	<.001	−7.16	18.76	−25.93 (−29.30, −22.55)	<.001
Mean Change in O2 saturation %	0.02	−1.01	1.02 (0.41, 1.63)	<.001	0.19	−1.39	1.58 (0.75, 2.42)	<.001
Mean FiO2	0.05	0.05	−0.01 (−0.15, 0.13)	.91	0.09	0.50	−0.41 (−1.01, 0.19)	.18

^*Generalized estimating equations (GEE) were used to account for repeated visits and measurements. Mean difference was calculated as mean in OCT group minus mean in RetCam group. 95% CI= 95% confidence interval. Rate differences (95% CI) are calculated for percentage of facial expression worsened.

Figure 1:

Change in behavioral and physiological stress during and after imaging with non-contact optical coherence tomography (OCT) or contact fundus photography. Error bars represent 95% confidence interval around the mean (A, C, and D) or proportion (B) for each measure. A) Mean change in cry score during/after contact fundus photography is higher than that of OCT (P<.001 for both). Cry score is measured on a scale of 0–4 (0=no crying, 1=moans or cries minimally, 2=appropriate crying, 3=high pitched crying, 4=inconsolable). B) Mean change in facial expression score (measured on a scale of 0–2 [0=no grimace, 1=grimace alone, 2=grimace and grunt]) during/after contact fundus photography is higher compared with that of OCT (P<.001 for both). C) Mean change in heart rate during/after contact fundus photography was higher than during/after OCT (P<.001 for both). D) Mean change in oxygen (O2) saturation was decreased during/after contact fundus photography compared with during/after OCT imaging (P=.004 during, <.001 after)

Adverse Events during OCT Imaging and Fundus Photography

A total of 32 AEs were recorded in 28 separate imaging sessions, of which two AEs occurred in four imaging sessions (Table 4). There were 14 instances of oxygen desaturation, 12 instances of tachycardia, 4 instances of bradycardia, and 2 instances of emesis. Out of the 28 imaging sessions with AEs, in 18 sessions (64.3%) AEs were transient and imaging resumed after a pause, in 5 sessions (17.9%) imaging was terminated early as a precaution, and in 5 sessions (17.9%) the AE occurred after both imaging modalities were completed. In one case, OCT was not performed per bedside nurse request due to hypoxia following fundus photography, and in a separate case, fundus photography was not performed due to overall infant instability. Sixteen of the 32 AEs (50%) were classified as mild and resolved without intervention, and the other 16 AEs were classified as moderate requiring intervention (increasing FiO2, infant stimulation, repositioning, and/or oral suction). No ocular, unexpected, or severe AEs were identified. After research imaging, there were no additional AEs recorded for at least 30 minutes and for up to 2 hours.

Table 4:

Summary of Adverse Events Across All 399 Imaging Sessions with Non-contact Handheld Optical Coherence Tomography (OCT) and Contact Fundus Photography

Adverse Events	All Imaging N			During or After OCT N (%)			During or After Fundus Photography N (%)
	Total	Mild	Moderate	Total	Mild N (%)	Moderate N (%)	Total	Mild N (%)	Moderate N (%)
Total number	32	16	16	6 (18.2)	1 (16.7)	5 (83.3)	26 (81.3)	11 (42.3)	15 (57.7)
Bradycardia	4	1	3	1 (25.0)	0 (0.0)	1 (100)	3 (75.0)	0 (0.0)	3 (100)
Emesis	2	1	1	2 (100)	1 (50)	1 (50)	0 (0.0)	0 (0.0)	0 (0.0)
Hypoxia	14	4	10	3 (21.4)	0 (0.0)	3 (100)	11 (78.6)	1 (9.1)	10 (90.9)
Tachycardia	12	10	2	0 (0.0)	0 (0.0)	0 (0.0)	12 (100)	10 (83.3)	2 (16.7)

Of the 32 AEs, 6 (18.8%) occurred during or after OCT and 26 (81.3%) occurred during or after imaging with fundus photography. The differences in AEs between imaging modalities are detailed in Table 4. Of the 6 AEs associated with OCT, 1 (16.7%) was mild and 5 (83.3%) were moderate. Of the 26 AEs associated with fundus photography, 11 (42.3%) were mild and 15 (57.7%) were moderate.

Discussion:

To our knowledge, this is the first study published directly evaluating preterm infant stress from handheld non-contact OCT compared with contact fundus photography. Prior studies in the literature have evaluated or compared preterm infant stress or safety (i.e. adverse events) when undergoing BIO, contact fundus photography, and non-contact fundus photography for ROP examination.^{7–9,12,19,20,34} Although OCT was performed for a longer duration compared with fundus photography, it was less stressful based on all metrics evaluated, including behavioral stress indicators (crying and grimacing scores) and physiological measures (heart rate, oxygen saturation) during and after imaging (and their changes from baseline). These results held true regardless of whether OCT or fundus photography was performed first. Furthermore, these changes persisted for 5 minutes after imaging was completed. Preterm infants were more likely to return to their baseline cry score, facial expression score, heart rate, and oxygen saturation after OCT was completed compared with fundus photography. Finally, while the overall incidence of AEs (sustained bradycardia, tachycardia, oxygen desaturation, and emesis) was low, over 80% of AEs occurred during or after contact fundus photography.

Strengths of our study include randomization of imaging order between OCT and fundus photography, direct back-to-back imaging on the same group of infants (thus reducing the risk of confounding differences between infants), and NICU research personnel documenting stress indicators and AEs. Our results expand upon the prior study of Mangalesh et al. which demonstrated that OCT was less stressful compared with BIO using similar measures of behavioral and physiological stress.¹⁹ In that study, both eyes underwent OCT and BIO with order dependent on nursery workflow, whereas in the current study only the randomly selected prespecified study eye was imaged and the imaging modalities were performed in randomized order. While the prior study did not find differences in oxygen saturation between OCT and BIO, this study found a small decrease in oxygen saturation and increased incidence of desaturation events after fundus photography compared with after OCT imaging, even though there were no differences in these measures during imaging. This may be in part because the vital sign abnormalities experienced by the preterm infant when imaged persist even after imaging is completed, in contrast to behavioral signs which are faster to return to baseline. For instance, the mean change in heart rate from baseline to during fundus photography persisted afterwards.

The decreased stress associated with non-contact OCT imaging compared with contact fundus photography is likely secondary to the lack of bright light and anesthetic eye drops and eyelid speculum usage, which was used in all patients undergoing contact fundus photography but is usually not needed in OCT imaging. In fact, other studies such as Prakalapakorn et. al 2018 showed that the use of a non-contact fundus camera (Pictor, Volk, city, state), which did not require an anesthetic eye drop and lid speculum, led to less bradycardia, tachycardia, oxygen desaturation, and prolonged apnea in the 10 minutes after imaging was performed compared with BIO.¹² Our results corroborate other studies which have found contact fundus photography to be undeniably stressful. For instance, Mukherjee et. al showed that heart rate and blood pressure increased during ROP examinations with both BIO or Retcam fundus camera and Dhaliwal showed that the pain scores of infants undergoing Retcam imaging and BIO were similar and tied to placement of the eyelid speculum.^7,34

Limitations

Limitations of this study include the fact that this handheld OCT device is available in research use only. Though multiple handheld swept source OCT devices are in development, none are currently commercially available.^35,36 The clinical relevance or feasibility of handheld OCT in ROP screening examinations is still to be determined. At this time, OCT imaging is not readily available or used by most intensive care nurseries in the US. We also recognize that this sample size of 59 preterm infants may not be representative of the entire population.

Conclusions:

This prospective observational study provides evidence to support the fact that this investigational non-contact handheld OCT is less stressful to preterm infants undergoing ROP examination, both behaviorally and physiologically, and results in fewer AEs compared with contact fundus photography. Further studies are needed on the clinical role that handheld OCT should play in ROP screening examinations and the feasibility of utilizing OCT instead of contact fundus photography for tele-screening purposes.

Key Points.

Question:

Does preterm infant stress from non-contact optical coherence tomography (OCT) imaging differ from that of contact fundus photography during imaging for retinopathy of prematurity (ROP)?

Findings:

This prospective observational study of 397 imaging sessions from 59 infants showed indicators of behavioral (crying and grimacing) and physiological stress (heart rate and oxygen saturation) were lower for OCT, regardless of which imaging modality was performed first. The incidence of adverse events was overall low but greater for fundus photography.

Meaning:

These findings suggest that non-contact OCT is less stressful than contact fundus photography in preterm infants undergoing ROP examination.