Drs. Kemper, Wallace, and Quinn reviewed the application of
telemedicine for retinopathy of prematurity (ROP) screening and concluded
that “insufficient data to recommend that retinal imaging be adopted by
NICUs” based largely on claims that “retinal imaging would almost surely
miss some cases of sight-threatening ROP.”(1) In examining the data, we
come to the opposite conclusion—telemedicine is a highly effective
mechanism for identification of sight-threatening ROP when used in
conjunction with confirmatory bedside binocular indirect ophthalmoscopy
(BIO).
How can it be that two groups come to such disparate conclusions?
The short answer is we and the authors of the present study are hoping to
achieve different goals through telemedicine screening for ROP. Is the
goal of telemedicine screening to identify all ROP present in the eye or
is it to identify ROP that would benefit from bedside BIO by an
ophthalmologist “who has sufficient knowledge and experience to enable
accurate identification and location and sequential retinal changes of
ROP.”(2) We hold to the latter definition. A properly designed screening
test includes a threshold above which an action will be taken. In the case
of telemedicine screening for ROP, the threshold would be referral to an
ophthalmologist experienced in ROP for a BIO examination.
Presently, national guidelines in the U.S.A. have two components on
the actual physical aspects of screening: 1) the technique should be
binocular indirect ophthalmoscopy and 2) the examiner should be an
“ophthalmologist who has sufficient knowledge and experience to enable
accurate identification and location and sequential retinal changes of
ROP.”(2)
In their analysis, Kemper et al preferentially included studies that
included “referral-warranted ROP” as defined by Ells, et al, as the gold
standard.(3) Unfortunately, numerous design flaws were included in the
trials analyzed. The first two papers share following three flaws for
applicability to telemedicine screening for ROP today:
1. They do not include specific identification of referral-warranted
ROP.
2. They use an outmoded contact RetCam 120 hand piece that was never
intended for ROP screening and was incompatible with the specula used,
thereby precluding adequate visualization of the fundus, ably depicted in
Figure 1 (Yen, et al) and Figure 3 (Roth, et al) and discussed in both
papers as a flaw.
3. They only evaluated an infant a maximum of twi time points in
order to make a determination of the presence of ROP and its ability to
predict future advanced disease. (4, 5)
Furthermore, in the paper by Yen, et al, “masked readers” were
utilized, but no mention is made of their specific experience in ROP or
their previous experience in evaluation of ROP images for screening
purposes.(4) The former is a stipulated requirement of the national
screening guidelines,(2) while the second is a recognized learning curve
for the technique of telemedicine.
Most importantly, the authors of the papers included in the analysis
by Kemper et al took the unusual approach of looking at the images in a
vacuum—two non-sequential points in time—to ascertain the presence of any
stage of ROP (not referral-warranted) and for presence of pre-threshold or
threshold.(4, 5) This is not how telemedicine is employed here in the
U.S. and throughout the world, where the technique involves weekly,
sequential, longitudinal viewing to ascertain the presence of disease and
its rate of progression. Frankly, while interesting from a purely
hypothetical perspective, it is akin to allowing a bedside BIO examiner to
make an interpretation at only two time points, which is not in adherence
with the national guidelines regarding disease activity.(2) A careful
reading of the Yen paper reveals that they used two screening points not
“…to test any particular screening strategy itself, but simply to
determine the amount of useful information that could be gleaned from the
images in predicting outcome.”(6) Hence, the data results focused on the
ability to predict pre-threshold and threshold, as opposed to actual
detection. The Roth paper similarly focused on the detection of ROP, as
opposed to referral-warranted ROP.(5) This paper, included 59 eyes of 32
patients for a total of 100 simultaneous BIO and photographic
examinations.(5) This leads to a situation where a maximum of 2
examinations per eye were performed in order to make a determination of
the disease state, something that is not recommended in clinical
practice.(2) This paper does not meet the authors stated goals of
including “referral-warranted ROP”, as no mention is made of the need for
treatment or predictive ability for referral-warranted disease.
Three papers by Chiang, et al, are included for review, all three
involving data obtained from the Jackson Memorial Hospital from January 1,
1999, to December 31, 2000.(7-9) This is one year after the study
reported by Roth, et al, at the same institution and a similar author
(JTF)—and all three papers report using the same equipment at the same
institution that was found to be deficient for evaluating the fundus by
Roth, et al, and Yen, et al.(5-9) Additionally, the central weakness of
these three papers is that the three examiners in each paper admittedly
were inexperienced in ROP screening in general,(7-9) and telemedicine in
particular, again not adhering to national screening guidelines,(2) making
the data difficult to interpret. It is interesting to note that when 3
trained, experienced ROP screeners are utilized, the sensitivity rose to
100% for infants requiring treatment in a follow up study by Chiang, et
al.(10)
Finally, the one paper that performed prospective longitudinal
telemedicine ROP screening in accordance with inclusion criteria stated by
the authors of the present paper reported a sensitivity of 100% and
specificity of 96% for identification of referral-warranted ROP.(3)
The experience of one of the authors herein has been that in 3 years,
no case of referral-warranted or treatment-warranted ROP has been missed
in the Stanford University Network of Retinopathy of Prematurity (SUNDROP)
telemedicine network.(11, 12) These reports are retrospective in nature,
however there has been 100% capture of all infants discharged from the
hospital with office BIO by the same experienced ROP examiner who
performed the telemedicine screening (DMM) with no adverse anatomic or
functional outcomes. For 3 years SUNDROP has provided telemedicine
screening for ROP to 4 separate NICUs served by a central reading center
and has demonstrated the ability of telemedicine to serve as an arbiter of
which babies at risk for ROP need an immediate confirmatory bedside BIO
examination.
Therefore, based upon the studies detailed above that meet the stated
entry criteria in the paper by Kemper et al1, and upon our own extensive
experience (11,12), it is our position that for the stated goal of
identifying ROP patients that would benefit from BIO, telemedicine has met
the requirements of an effective screening mechanism for ROP.
REFERENCES
1. Kemper AR, Wallace DK, Quinn GE. Systematic review of digital
imaging screening strategies for retinopathy of prematurity. Pediatrics
2008;122:825-830.
2. Screening examination of premature infants for retinopathy of
prematurity. Pediatrics 2006;117:572-576.
3. Ells AL, Holmes JM, Astle WF, et al. Telemedicine approach to
screening for severe retinopathy of prematurity: a pilot study.
Ophthalmology 2003;110:2113-2117.
4. Yen KG, Hess D, Burke B, Johnson RA, Feuer WJ, Flynn JT.
Telephotoscreening to detect retinopathy of prematurity: preliminary study
of the optimum time to employ digital fundus camera imaging to detect ROP.
J AAPOS 2002;6:64-70.
5. Roth DB, Morales D, Feuer WJ, Hess D, Johnson RA, Flynn JT.
Screening for retinopathy of prematurity employing the retcam 120:
sensitivity and specificity. Arch Ophthalmol 2001;119:268-272.
6. Yen KG, Hess D, Burke B, Johnson RA, Feuer WJ, Flynn JT. The
optimum time to employ telephotoscreening to detect retinopathy of
prematurity. Trans Am Ophthalmol Soc 2000;98:145-150; discussion 150-141.
7. Chiang MF, Starren J, Du YE, et al. Remote image based retinopathy
of prematurity diagnosis: a receiver operating characteristic analysis of
accuracy. Br J Ophthalmol 2006;90:1292-1296.
8. Chiang MF, Keenan JD, Starren J, et al. Accuracy and reliability
of remote retinopathy of prematurity diagnosis. Arch Ophthalmol
2006;124:322-327.
9. Chiang MF, Keenan JD, Du YE, et al. Assessment of image-based
technology: impact of referral cutoff on accuracy and reliability of
remote retinopathy of prematurity diagnosis. AMIA Annu Symp Proc 2005;126-
130.
10. Chiang MF, Wang L, Busuioc M, et al. Telemedical retinopathy of
prematurity diagnosis: accuracy, reliability, and image quality. Arch
Ophthalmol 2007;125:1531-1538.
11. Murakami Y, Jain A, Silva RA, Lad EM, Gandhi J, Moshfeghi DM.
Stanford University Network for Diagnosis of Retinopathy of Prematurity
(SUNDROP): 12-month experience with telemedicine screening. Br J
Ophthalmol 2008;92:1456-1460.
12. Silva RA, Murakami Y, Jain A, Gandhi J, Lad EM, Moshfeghi DM.
Stanford University Network for Diagnosis of Retinopathy of Prematurity
(SUNDROP): 18-month experience with telemedicine screening. Graefes Arch
Clin Exp Ophthalmol 2009;247:129-136.
Conflict of Interest:
Scientific Advisory Board, Clarity Medical Systems Inc. (Dublin, CA; makers of the RetCam family of widefield digital cameras—DMM (paid Q4 2006-Q1 2008, unpaid Q2 2008-present), MTT
eLetters: ![[Read eLetters]](/icons/misc/sectionDOWN.gif){kind=icon}
