JBUR 5341 No.

of Pages 10

burns xxx (2017) xxx –xxx

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/burns

Prospective comparative evaluation study of Laser

Doppler Imaging and thermal imaging in the
assessment of burn depth

Christopher Wearn a,b,1 , Kwang Chear Lee a,b,1 , Joseph Hardwicke a,b ,
Ammar Allouni a , Amy Bamford a, Peter Nightingale c ,
Naiem Moiemen a,b, *
a
Scar Free Foundation Centre for Burns Research, University Hospital Birmingham NHS Foundation Trust, Queen
Elizabeth Hospital Birmingham, Mindelsohn Way, Edgbaston, Birmingham B15 2WB, UK
b
University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
c
Wellcome Trust Clinical Research Facility Queen Elizabeth Hospital Birmingham, Mindelsohn Way, Edgbaston,
Birmingham B15 2WB, UK

article info abstract

Article history: Introduction: The accurate assessment of burn depth is challenging but crucial for surgical
Accepted 9 August 2017 excision and tissue preservation. Laser Doppler Imaging (LDI) has gained increasing
Available online xxx acceptance as a tool to aid depth assessment but its adoption is hampered by high costs, long
scan times and limited portability. Thermal imaging is touted as a suitable alternative
however few comparison studies have been done.
Keywords:
Methods: Sixteen burn patients with 52 regions of interests were analysed. Burn depth was
Burn depth
determined using four methods LDI, thermal imaging, photographic and real-time clinical
Thermal imaging
evaluation at day 1 and day 3. LDI flux and Delta T values were used for the prediction of
Laser Doppler Imaging
outcomes (wound closure in <21 days). Photographic clinical evaluation of burn depth was
Clinical assessment
performed by 4 blinded burn surgeons.
Results: Accuracy of assessment methods were greater on post burn day 3 compared to day 0.
Accuracies of LDI on post burn day 0 and 3 were 80.8% and 92.3% compared to 55.8% and 71.2%
for thermal imaging and 62.5% and 71.6% for photographic clinical assessment. Real-time
clinical examination had an accuracy of 88.5%. Thermal imaging scan times were
significantly faster compared to LDI.
Discussion: LDI outperforms thermal imaging in terms of diagnostic accuracy of burn depth
likely due to the susceptibility of thermal imaging to environmental factors.
Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.

* Corresponding author at: Scar Free Foundation Centre for Burns Research, University Hospital Birmingham NHS Foundation Trust, Queen
Elizabeth Hospital Birmingham, Mindelsohn Way, Edgbaston, Birmingham B15 2WB, UK.
E-mail addresses: nmoiemen@aol.com, naiem.moiemen@uhb.nhs.uk (N. Moiemen).
1
Joint first authors.
http://dx.doi.org/10.1016/j.burns.2017.08.004
0305-4179/Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10
2 burns xxx (2017) xxx –xxx
1. Introduction
The accurate assessment of burn depth is a critical step in the
management of the burn injured patient, with a key decision
being whether a wound will benefit from surgical manage-
ment. The depth of the burn injury is a key determinant of
wound healing time and a relationship between the time to
wound closure and the incidence of hypertrophic scarring
(HTS) has been clinically demonstrated [1]. Establishing the
difference between superficial dermal burns that will heal
spontaneously within 21days and deep dermal burns which
can result in longer healing time and associated pathological
scarring is challenging, even for experienced burn surgeons [2–
4]. Burn depth classification via visual interpretation has
remained relatively unchanged since the 16th century [5],
however in the last few decades the increasing use of
technology to aid the assessment of burn depth and the
progression of the burn wound has allowed earlier diagnosis of
burns that may benefit from excision, and tissue preservation
of those that will heal spontaneously.
Douglas Jackson, at the Birmingham Accident Hospital,
was the first to identify a link between progression of the
clinical depth of the burn wound and the dermal microvas-
cular blood flow [6]. The link between burn depth and blood
flow in the superficial vascular plexus inspired the develop-
ment of a number of technologies to measure this, either
directly or indirectly. Laser Doppler Imaging (LDI) utilizes a
623nm red diode laser, which superficially penetrates the
skin and is reflected by circulating erythrocytes in the dermal
vasculature. A transducer measures the Doppler shift in light
frequency from the erythrocytes moving relative to the laser
source and provides a semi-quantitative measure of perfu-
sion. LDI has now been validated in a large multi-centre
study over of over 400 burn wounds and shown to have a
technical accuracy of 96% in predicting whether wounds will
be healed by 21 days [7,8]. The National Institute for Clinical
Excellence (NICE) in the UK recommends the use of LDI in
addition to clinical assessment for clinically indeterminate
burn wounds [9].
Thermal imaging provides an indirect measure of cutane-
ous blood flow and was originally described by Lawson in 1961,
to accurately predict the depth of burns in dogs [10]. Mladick
et al. in 1966 found that thermograms correlated very closely
with the pattern of depth of the burn injuries in patients [11]
and later, Hackett concluded that clinical examination of the
burn failed to predict depth in one third of cases compared to
10% failure with thermal imaging [12]. Hardwicke et al.
conducted a pilot evaluation study at our centre utilizing a
current generation high resolution real time thermal imaging
camera and found that it could differentiate between clinically
full thickness, deep partial thickness and superficial partial
thickness burns based on temperature comparisons with non-
burnt skin [13]. The images were high resolution and could be
acquired rapidly suggesting potential advantages over other
imaging modalities such as LDI. We therefore conducted a
comparative evaluation study to compare the accuracy of
high-resolution digital thermal imaging with the current gold
standard, LDI, in assessing wound healing potential in adult
burn injured patients.
Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
2. Materials and methods
2.1. Patient selection
Adult patients admitted to the Birmingham Burns Centre,
University Hospitals NHS Foundation Trust between were
screened for eligibility for prospective recruitment into the
study over a 12 month period from May 2012. The study
included adults aged 18 years and above presenting within 24h
of an acute burn injury <15% of Total body surface area.
Detailed inclusion and exclusion criteria can be seen in Table 1
and Fig. 1. Burn wounds overlying skin tattoos were not
included due to reflectance artifact. The study protocol
received approval by an NHS Research Ethics Committee
(REC reference: 12/SW/0028).
2.2. Study protocol
Patients enrolled into the study underwent study visits for
burn depth assessment on day 1 (within 24h after-injury);
again on day 3 (at 48–96h after-injury) and wound healing
evaluation on Day 14; and Day 21 if not fully healed at Day 14.
Burn depth estimation was performed by four methods: (1)
Laser Doppler Imaging (LDI) with a Moor LDI2-B1 (Moor
Instruments, Axmoor, UK); (2) Thermographic imaging with
the FLIR SC660 thermal imaging camera (FLIR Systems, Inc.,
Wilsonville, USA); (3) Clinical assessment with 2D photogra-
phy; and (4) Real-time clinical assessment. Two dimensional
(2D) clinical photography was undertaken at all visits using a
Nikon D300S DSLR camera (Nikon Corporation, Tokyo, Japan)
by the clinical photography team and these photographs were
then assessed by four consultant burn surgeons blinded to the
outcomes. Real-time clinical assessment at the time of
presentation (day 0) by the on-call consultant was collected
retrospectively. Decisions on need for surgical excision and
skin grafting were made on clinical grounds in obviously full
thickness or deep dermal burns; LDI was used to assist
decision-making in the case of clinically indeterminate burn
wounds. Cases which were found to have a significant area
with a healing potential of >21days underwent excision and
skin grafting, in line with the standard of care prior to
commencement of the study. If the patient had multiple
separate burn wounds, all were included as study wounds and
Table 1 – Inclusion and exclusion criteria for study.
Inclusion criteria Exclusion criteria
Adults (18–99 years) Children (<18years)
Within 24h of burn Unable or refusal to give informed
injury consent
<15% TBSA Suspected burn wound
infection/cellulitis
All burn depths Chemical burns
Incomplete images at day 0 and 3
Excised and grafted before day 21
Skin tattoos
Participation in other interventional
trials
JBUR 5341 No. of Pages 10

burns xxx (2017) xxx –xxx 3

Fig. 1 – CONSORT diagram for the study detailing the number of patients enrolled, excluded from study, and final number
included for analysis.

allocated an identification number. All LDI and thermal
imaging were performed either in a designated clinical
assessment room or in the patient’s own room. Ambient
temperature was standardised to 23  C and humidity was
measured consistently to be around 50%. The LDI and thermal
images were then uploaded and analysed objectively at the
completion of the trial, using their respective software (see
below). Prior to wound assessment and scanning, burn
wounds were checked to ensure adequate cleaning and
debridement and absence of dressing residues. The scanning
times were recorded for each method using a stopwatch and a
patient satisfaction questionnaire was given to each patient
after the first two study visits.
2.3. Thermal imaging
Static thermographic images were recorded at a distance of
70cm. For the purposes of standardization, skin cooling [14]
and occlusive dressings [15] were not used. The base
temperature was adjusted to 30  C and the peak temperature
adjusted to the patient’s core temperature, as measured by
1
tympanic thermometer (Braun Thermoscan , Braun GmbH,
Germany). For comparison with the burn wound temperature,
a control area of non-burnt skin, where possible on the contra-
lateral side, was thermally imaged. If this was not possible (due
to multiple burn sites,), an area of non-burnt skin adjacent to
the burn wound was used.
Image analysis was performed using the FLIR Thermal
researcher Pro V4.0 software (FLIR Systems, Inc., Wilsonville,
USA). Based on the previous work of Hackett [16], the burn
depth diagnosis was determined by the difference between the
mean temperature of the burn wound and the mean
temperature of the control area of non-burnt skin [17] (Delta
T, DT) (Table 2). A full thickness or deep dermal depth burn was
considered to have a healing potential (HP) of >21days and a
superficial partial thickness burn a HP of <21 days.
2.4. LDI scanning
The study team performed all scans, each trained, assessed
and deemed competent in the use of LDI by the manufacturer.
Images were acquired at a distance of 70cm from the burn
wound. Image analysis was performed using the Moor LDI2-BI
V2 software (Moor Instruments, Axmoor, UK). Mean LDI flux

Table 2 – Determination of burn depth diagnosis from static thermal imaging according to criteria used by Hackett15.
DT=Mean temperature of non-burnt skin (TC) — mean temperature of burn wound (TB).
DT value ( C) Burn depth interpretation Healing potential (HP) (days)
<1.5 Superficial partial thickness (SPT) <21
1.5 and <2.5 Deep dermal (DD)/deep partial thickness (DPT) >21
 2.5 Full thickness (FT) >21

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10

4 burns xxx (2017) xxx –xxx

Table 3 – Validated LDI colour palette and healing easier to grade and what features of the wound guided them
potential for each flux range. PU=Perfusion units. to make their decisions.
LDI Flux value LDI palette Healing potential (HP)
(PU) colour (days) 2.6. Real-time clinical analysis

>601 Red <14

441–600 Pink 14
261–440 Yellow 14–21
201–260 Green 21
0–200 Blue >21
values (Perfusion units, PU) were calculated for each burn
wound and the corresponding healing potential (HP) assigned
according to the validated LDI colour palette [7] (Table 3).
Regions of Interest: Burn wounds were divided into regions of
interest (ROI) according to differential healing times identified
at D14 and D21 wound healing assessment. These ROI were
then identified using the 2D clinical photos and an outlined
template was then created from these ROIs and applied to LDI
and thermal image using anatomical and wound reference
points (Fig. 2). The same ROI was analysed for each LDI image
and each thermal image. Healing times were calculated for
each ROI from the time of injury to the time of complete wound
closure.
2.5. Clinical analysis of photographs
2D digital photographs of the regions of interest of the burns
at day 0 and day 3 were shown to 4 consultant burn surgeons
who were asked to grade them as either superficial
(superficial partial thickness or mid-dermal) or deep (deep
dermal or full thickness). Superficial burns were defined as
wounds with a HP of <21 days and deep burns as wounds
with HP of >21 days. The surgeons were given information of
the site of the burn being assessed and the age of the burn
wound. The surgeons were also asked at the end of the
assessment which of the day 0 or day 3 images they found
Data on real-time burn wound depth evaluation at the time of
presentation by an experienced burn surgeon or registrar was
collected retrospectively from patients’ medical records. The
regions evaluated were matched as closely as possible to the
ROIs used in LDI and thermal imaging. For the purpose of
analysis, superficial partial thickness, partial thickness and
mixed depth burns were classified as superficial and deep
dermal and full thickness burns as deep.
2.7. Statistical analysis
Study data was recorded in an electronic Case Report Form
1 1
(CRF) and exported to Microsoft Excel spreadsheet (Micro-
1
soft Corporation, Redmond, WA) before statistical analysis
with PASW Statistics version 18 (SPSS Inc., Chicago, Illinois,
USA), and R version 3.1.1 (The R Foundation for Statistical
Computing, Open Source). Both Laser Doppler Imaging and
thermal imaging were compared with the clinical assessment
of the burn injury. The primary outcome measure of the study
is the ability to predict complete wound closure in 21days or
less. Secondary outcome measures include: comparison of the
accuracy of wound healing potential prediction between day
0 and day 3 assessments, scanning times and patient
satisfaction. The accuracy, sensitivity, specificity, predictive
values and likelihood ratios were calculated for each method
on day 0 and day 3. Confidence intervals were calculated using
exact approaches wherever possible and McNemar’s test was
used to compare the accuracy of methods. The DTComPair
package for the R statistical program was used to calculate the
statistical differences between the positive predictive value
(PPV) and negative predictive value (NPV) [18]. Correlation
between Thermal imaging DT and LDI flux values on Day 0 and

Fig. 2 – a) 2D photograph of burn on antecubital fossa (day 0); b) LDI image of burn (day 0); c) Thermal image of burn (day 0); d) 2D
photograph of burn (day 3); e) LDI image of burn (day 3); f) Thermal image of burn (day 3).

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10

burns xxx (2017) xxx –xxx 5

1 was calculated using the Pearson’s correlation coefficient.

Table 4 – Demographic data for study group.
Scanning times were compared using the non-parametric
Mann–Whitney test. Fleiss’ kappa was used to measure the Demographics Results
inter-rater agreement between the four clinicians on burn n 16
depth assessment via 2D photography [19]. Landis and Koch Male:female 13:3 (4:1)
[20] provided a guide on how the kappa values can be Mean BMI (kg/m2 ) 26.3 (6.4)
interpreted. Mean age (years) 37.5 (16.4)
Mean TBSA (%) 4.1 (3.7)
Mean no. of burn wounds/patient 1.9 (0.7)
Mean no. of ROIs/patient 4.3 (2.3)
3. Results

During the study period, a total of 32 patients were enrolled

into the study, after exclusions this left data from 16 patients
for final accuracy analysis. The demographic data for the study
group is summarized in Table 4. Burn wounds were managed
conservatively until wound closure. Within the study group,
flame and scald burns were the most common mechanism
(38%) followed contact burns (12%), and flash burns (6%). A
total of 20 discrete burn wounds were included, which
following analysis of healing times via 2D clinical photography
were divided into 52 ROIs. Analysis of healing times showed
just over half of the ROIs had healed by 14days (n=27/52), a
quarter healed by 21days (n=16/52) and 17.3% took greater
than 21days to heal (n=9/52).
Comparative analysis of the accuracy of predicting whether
ROIs would take greater than 21days to heal by LDI, thermal
imaging, photographic clinical assessment and real-time
clinical assessment is shown in Table 5.
Accuracy for LDI was higher on both day 0 (p<0.01, chi
square 18.9 with 1 degree of freedom) and day 3 (p<0.01, chi-
square 6.7 with 1degree of freedom) compared to thermal
imaging. There were no statistical differences between the
sensitivity of LDI and thermal imaging on both days 0 and 3.
Specificity for LDI however was statistically better compared
to thermal imaging on both days 0 and 3 (p<0.01). There were
no statistical differences in Positive (PPV) and Negative
Predictive Values (NPV) between LDI and thermal imaging
on day 0. The PPV for LDI on day 3 however was statistically
higher compared to thermal imaging (p<0.01) but not the NPV
(p=0.35).
The effect of the timing of scans on the primary outcomes
was also analysed. LDI had significantly higher accuracy and
PPV on day 3 compared to day 0 (p<0.05, chi square=4.167 with
1 degree of freedom). Sensitivity, specificity and NPV of LDI
were also higher on day 3 but these did not reach significance.
For thermal imaging, compared to day 0, day 3 scans were
more accurate (borderline significance, p=0.0523, chi
square=3.765 with 1 degree of freedom) and had higher
specificity (p<0.05) however no significant difference was
seen in sensitivity, PPV and NPV.
There is a significant correlation between PU and DT on
both day 0 (Pearsons=0.515, p<0.01) and day 3 (Pear-
sons=0.509, p<0.01) (Fig. 3). Thermography had a clear
advantage in terms of time taken for scan, which was
significantly faster by an average of 6.3min on day 0 and
4.4min on day 3 (p<0.01) (Table 6).
3.1. Photographic clinical assessment
52 matched ROIs of the images of the burns on day 0 and 3 were
selected for use to determine the accuracy, sensitivity, PPV and
NPV of clinical assessment of burn depth.
In terms of the primary outcome, LDI was found to out-
perform photographic assessment at both day 0 and 3. LDI was

Table 5 – Comparative analysis of accuracy at predicting healing potential of ROIs to be >21 days. The total number used to
calculate Clinical Day 0 and 3 (photographic) is 4 times higher (208) than LDI, thermal and real-time assessments as the total
results for 4 assessors is used (i.e. 524=208).
Accuracy % Sensitivity % Specificity % PPV % NPV %
(n) (n; 95% CI) (n; 95% CI) (n; 95% CI) (n;95% CI)
LDI day 0 80.8 33.3 90.7 42.9 86.7
(42/52) (3/9; 2.5–64.1) (39/43; 82.0–99.4) (3/7; 6.2–79.5) (39/45; 76.7–96.6)
Thermal day 0 55.8 55.6 55.8 20.8 85.7
(29/52) (5/9; 23.1–88.0) (24/43; 41.0–70.7) (5/24; 4.6–37.1) (24/28; 72.6–98.7)
Clinical day 0 62.5 33.3 68.6 18.2 83.1
(photo, average) (130/208) (12/36; 19.1–51.1) (118/172; 56.8–71.6) (12/66; 9.1–27.3) (118/142; 74.5–88.1)
Real-time clinical day 0 88.5 33.3 100 100 87.8
(46/52) (3/9; 2.5–64.1) (43/43; 100.0–100.0) (3/3; 100.0–100.0) (43/49; 46.8–96.9)
LDI day 3 92.3 55.6 100 100 91.5
(48/52) (5/9; 23.1–88.0) (43/43; 100.0–100.0) (5/5; 100.0–100.0) (43/47; 83.5–99.4)
Thermal day 3 71.2 44.4 76.7 28.6 86.8
(37/52) (4/9; 12.0–76.9) (33/43; 64.1–89.4) (4/14; 4.9–52.2) (33/38; 76.1–97.6)
Clinical day 3 71.6 50.0 76.2 30.5 87.9
(photo, average) (149/208) (18/36; 33.2–66.8) (131/172; 69.0–82.2) (18/59; 19.5–44.0) (131/149; 81.3–92.5)

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10

6 burns xxx (2017) xxx –xxx

Fig. 3 – Relationship between mean flux [PU] and mean DT [ C] on day 0 and day 3.

assessment was higher; 88.4% vs 76.7%, p<0.01) nor PPV and

Table 6 – Scan times of LDI versus Thermography (n=16).
Scan times include time taken for setup of equipment and
patient.
Scan times (min)
Day 1 Day 3
LDI 10.28 7.99
Thermography 3.98 3.57
statistically better in terms of accuracy compared to 2/4 of the
clinical assessors (photographic) on day 0 and 3/4 of the
assessors on day 3 (Tables 7 and 8). Photographic clinical
assessment was also statistically lower in specificity on both
day 0 and day 3, and PPV for day 3 (p<0.01). The average
sensitivity of photographic clinical assessment was lower on
day 3 compared to LDI as well but this did not reach statistical
significance.
On the other hand, thermal imaging accuracy had no
advantage over clinical photographic assessment. There was
no statistical difference seen in terms of accuracy (except for
1 assessor [Assessor 3]), sensitivity and specificity (except for
1 assessor [Assessor 3] in which the specificity of clinical
NPV between thermal imaging and photographic clinical
assessment on both days 0 and 3.
Comparing accuracy of photographic assessment on day
0 and day 3, accuracy of day 3 was higher compared to day 0 but
this was only statistically significant in 1 assessor (Assessor 1).
The agreement of burn depth between raters was also found to
be higher on day 3 compared to day 0. On day 0, there is a
significant variation in the ratings provided by the four raters
with only a slight agreement between raters (Fleiss kap-
pa=0.20). On day 3 however, the agreement improves to 0.44
(moderate agreement). This is demonstrated clearly in the
graphs of the marginal distribution of the depth ratings (Fig. 4)
on day 0 and day 3.
3.2. Real-time clinical assessment
Real-time assessment data was available for all the 52 ROIs on
day 0. On day 0, it was significantly better in terms of accuracy
and PPV compared to the photographic assessment of all
4 assessors (p<0.05). No significant statistical difference was
seen for sensitivity or NPV but the specificity of real-time
assessment was significantly better than photographic as-
sessment of all 4 assessors (p<0.05).

Table 7 – P-values of cross-comparisons of accuracy values between day 0 clinical assessors (photographic) and day 0 LDI,
thermal imaging, real-time clinical assessment, and day 3 clinical assessments (photographic). Gray shaded cells denote
significance. P-values are from McNemar’s test.
Day 0 LDI Thermal Clinical (day 3) Clinical (real-time)
(accuracy%) (80.8%) (53.9%) (various) (88.46%)
Clinical assessor 1 <0.01 0.72 67.3%; 0.04 <0.01
(50.0%)
Clinical assessor 2 0.011 1.00 65.4%; 0.21 <0.01
(53.9%)
Clinical assessor 3 0.58 <0.01 80.8%; 0.58 0.02
(75.0%)
Clinical assessor 4 0.30 0.15 73.8%; 1.00 0.03
(71.5%)

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10

burns xxx (2017) xxx –xxx 7

Table 8 – P-values of cross comparisons of accuracy values comfortable Some patients however felt that the LDI equip-
between day 3 clinical and day 3 LDI and thermal imaging. ment was large and intimidating. A few patients (n=3)
Gray shaded cells denote significance. P-values are from appreciated being able to visualise in colours the superficial
McNemar’s test.
and deep areas of their wounds with both the LDI and thermal
Day 3 LDI Thermal imaging.
(accuracy %) (92.3%) (71.2%)
Clinical assessor 1 <0.01 0.31
(67.3%) 4. Discussion
Clinical assessor 2 <0.01 0.66
(65.4%) Clinical assessment has been shown in the literature to be less
Clinical assessor 3 0.11 0.33 than optimal and accurate in only 60-80% of cases especially in
(80.8%)
burns of intermediate or mixed depth and this can lead to
Clinical assessor 4 0.04 1.00
(73.1%)
unnecessary excision with worsened outcomes and aesthetic
results. Laser Doppler and thermal imaging have emerged as
adjuncts to help increase the accuracy of clinical burn depth
assessment and are the two most commonly used objective
Compared to day 0 LDI outcomes, there were no statistically methods to estimate the depth of burn via the indirect
significant differences in accuracy, sensitivity, specificity, and measurement of the cutaneous blood flow and as a prognostic
NPV. factor to predict healing time [21,22].
Real-time assessment however was found to have signifi- Our study has shown that LDI has a significantly higher
cantly higher accuracy (88.5% vs 55.8%, p<0.01), specificity accuracy on both day 0 and day 3 compared to thermal imaging
(100% versus 55.8%, p<0.01) and PPV (100% versus 20.8%, and photographic clinical assessment. This is mainly due to
p<0.05) compared to day 0 thermal imaging. Thermal imaging the significantly lower specificity (due to a high false positive
(D0) had a slightly higher sensitivity compared to real-time rate) of thermal imaging compared to LDI, leading to a
assessment but this did not reach statistical significance significantly better PPV for LDI versus thermal imaging (on
(55.6% versus 33.3%, p=0.68). day 3). There was no significant difference in NPV between all
Visual clues that clinicians felt were helpful to determine 3 modalities. A limitation of our study however is the wide
burn depth include the colour of the burn, presence of confidence intervals for sensitivity due to the relatively small
blanching and fixed red staining. Other factors deemed number of burn wounds that took longer than 21days to heal
important were the clinical history, anatomical area, age of (n=9). The small number of full thickness burns remaining was
patient, and type of dressings used. Interestingly, despite the due to our standard of care in which most deep dermal or full
better primary outcomes for photographic clinical assessment thickness wounds would be excised during the first few days of
on day 3, half of the clinicians felt that assessment of burn presentation unless they were indeterminate or too small in
depth was easier on day 0, with only 1 clinician saying it was size. Other studies have overcome this limitation by histologi-
easier on day 3 and one reporting that both days were the cally analysing excised burn tissue to determine burn depth
same. The reason given for this was that the visualization of [23]. This method however was not suitable for our study as we
burn wounds on day 3 were more likely to be hindered by defined ROI’s based on time of healing.
discolouration (e.g. by silver containing dressings or topicals,) Photographic clinical assessment of burn wounds in this
or exudates/crusts. study was performed via the analysis of 2D photographs. This
limits the depth indication clues available to the clinician to
3.3. Patient satisfaction questionnaire solely visual appearance and thus is only partly comparable to
real-time assessment [24].
Feedback was available from 11 of the 16 patients in the study. In fact, with the data that was available on real time clinical
The majority (64%, n=7/11) preferred the thermal camera as it assessment on day 0 from reviewing all patients records, there
was quicker compared to the LDI with the rest having no strong is an indication that real-time assessment on day 0 is not only
preference and stated both methods to be quick and superior to 2D photo assessment but thermal imaging day

Fig. 4 – Photographic assessment: Marginal distribution of burn depth assessments across the 4 clinicians on day 0 and day 3.
The graphs show the percentage of images that have been rated as deep or superficial by each clinician.

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10
8 burns xxx (2017) xxx –xxx
0 assessments as well. However it has to be noted that the real-
time assessment in our study had some limitations: the wound
was assessed as a whole rather than for Region of Interest, data
were available only on admission (day 0) and were performed
by different consultants.
Despite other studies that have compared LDI and thermal
imaging showing that the former has a significantly better
accuracy [25], thermal imaging still remains a much re-
searched imaging modality in burns with several papers
published on the topic in the last few years. The main reasons
for this are the relatively lower cost, faster imaging time and
significantly better mobility of the thermal imaging devices.
This is especially true with the release of newer smaller
thermal cameras including models which can be attached to a
mobile phone (FLIR One). Unlike the laser Doppler which
requires a relatively lengthy time to scan the region of interest
(which increases with the size of the burn,), the thermal
camera offers a quicker, more portable and simpler alterna-
tive. The instantaneous scan that thermal imaging produces
allows burns to be assessed objectively quickly which is very
useful in children [26]. The patients in our study also showed a
preference of the thermal camera due to it being faster and less
intimidating compared to the LDI. One study has also shown
that thermal imaging can be used to identify and quantify burn
conversion utilising a modification of the technique [27].
As thermal imaging utilises the infra-red (heat) signals
emanating from the subject in comparison to the laser Doppler
which has employs its own stronger light source, thermal
imaging will invariably be more susceptible to subject and
environmental changes and hence less accurate. Thus various
aspects and steps in thermal imaging have to be standardized
to help make it more reproducible and accurate.
One of these includes standardizing the time of imaging to
day 0 and 3 which has not always been done in other studies.
We have shown that LDI outperformed thermal imaging for
the primary outcome of diagnostic accuracy at predicting
healing >21days on both day 0 and day 3. Furthermore,
thermal imaging was not found to be significantly better than
clinical judgement (both photographic and real-time) in terms
of accuracy for both day 0 and 3 which is in contrast to other
studies [28]. The increase in accuracy of burn depth assess-
ment in all 3 modalities on day 3 compared to day 0 agrees with
the reported literature. Hoeksema et al. reported that day
3 accuracies of both LDI and clinical assessments were higher
on day 3 compared to day 0 (97% versus 54%, and 40.6% versus
61.5% respectively,). Liddington et al. also came to a similar
conclusion which found that thermography of burns, to assess
depth, should be performed within 3days following the injury
[29]. Miccio et al, in an animal study, showed that the
temperature half area (summation of temperatures across
the burn wound) on day 2 had a stronger correlation with scar
depth after healing compared to the measurements on day 1
[30]. Time points beyond 7days showed poor correlation with
scar depth.
Our experience is that the low accuracy of the thermal
camera is due to external environmental influences that were
not controlled adequately (e.g. wound exposure time, evapo-
ration, humidity etc.), which rendered the camera less
accurate than LDI imaging. In our study, we ensured wounds
were dry (although interestingly one paper performed the
Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
opposite and made the whole wound wet instead [31],) and
covering the burn wounds with cling film before measurement
as several studies have suggested covering with a non-
permeable membrane to prevent a falsely low temperature
secondary to evaporative heat loss [14,15,29]. However in our
study, the images that we obtained with the cling film were
inferior to those without (e.g. poorer contrast between
different thermal zones) and contained artefacts due to
creases in the film rendering the images unsuitable for
thermal measurement and were thus not included in the
final analysis. Sites which are being chosen as the controls
have to be selected carefully as well as temperatures have been
shown to vary in different areas of the body [32].
Several authors have tried different techniques to stan-
dardize thermal imaging in hopes of improving its accuracy.
Renkielska et al. [33,34] and Prindeze et al. [35] utilized active
dynamic thermal imaging (ADT, cold excitation) to improve
the accuracy of thermal imaging. However utilizing the
complex protocol required for these methods negates the
most significant advantage of thermal imaging over LDI which
is speed of image acquisition (average time 3.6 versus 7.9min).
This includes equipment set-up time, and the actual time for
image acquisition for the thermal camera is only a few seconds
as it is “point and shoot”. Additionally, it is of note that
guidelines for thermal imaging in other medical specialties
recommend that the patient be allowed time (15–20min) to
equilibrate with the temperature of the room before imaging is
obtained [36,37]. Time is also required for the camera itself to
equilibrate to the room temperature, although no specific
times have been published by FLIR as it is dependent on both
the temperature at which the camera is stored at and the room
in which it is used (direct correspondence with company,).
Cameras usually will have an indicator to show when
equilibrium between the device and room temperature has
been reached.
The main advantage that thermal imaging has in terms of
speed is being challenged with new innovations in other
technologies. Newer, faster models of LDI scanners have now
been introduced onto the market e.g. the Laser Doppler Line
Scanner (LDLS) which only takes 4s to scan an area of 300cm2
with comparable accuracy to traditional LDI imagers (94.2%
versus 94.4%) [38] as well as Laser Speckle Contrast Imaging
which is also able to obtain images within seconds [39] but is
however susceptible to motion artefacts (especially periodic
movements e.g. breathing) which can be partially overcome by
taking a series of scan and choosing images with the lowest
perfusion values [40]. Other technologies include devices
which utilise spectrometry analysis [41,42] such as the
Scanoskin [25] which is also “point and shoot” as well as
extremely mobile. Other promising technologies include
ultrasonography [43,44] especially in combination with real-
time color flow imaging of the microcirculation [45] and laser-
induced indocyanine green fluorescence [46].
5. Conclusions
Further work and innovation in the thermal imaging field
needs to be done in order to standardize the technique and
improve the accuracy of the technology and preserve its place
JBUR 5341 No. of Pages 10

burns xxx (2017) xxx –xxx 9

in the burn depth diagnostic panel. Unlike LDI, which has had
multi-centre trials to validate its results [7,8], the current
colour palette that is utilized in thermal imaging has yet to be
validated against a large cohort of burn patients. If this can be
achieved, along with the creation of a standardised protocol for
imaging to improve its accuracy, then better adoption of the
devices may be possible.
Conflict of interest
The authors have no conflict of interest.
Financial disclosure
Both devices utilized in this trial (Moor LDI2-B1 and FLIR SC660)
were paid for by the Queen Elizabeth Hospital Charity. No other
financial support was used in the production of this article.
REFERENCES
[1] Deitch EA, Wheelahan TM, Rose MP, Clothier J, Cotter J.
Hypertrophic burn scars: analysis of variables. J Trauma
1983;23(10):895–8.
[2] Jaskille AD, Shupp JW, Jordan MH, Jeng JC. Critical review of
burn depth assessment techniques: part I. Historical review. J
Burn Care Res 2009;30(6):937–47.
[3] Jaskille AD, Ramella-Roman JC, Shupp JW, Jordan MH, Jeng JC.
Critical review of burn depth assessment techniques: part II.
Review of laser doppler technology. J Burn Care Res 2010;
31(1):151–7.
[4] Kaiser M, Yafi A, Cinat M, Choi B, Durkin AJ. Noninvasive
assessment of burn wound severity using optical technology:
a review of current and future modalities. Burns 2011;
37(3):377–86.
[5] Lee KC, Joory K, Moiemen NS. History of burns: the past,
present and the future. Burns Trauma 2014;2(4):169–80.
[6] Jackson DM. The treatment of burns: Hunterian lecture
delivered at the Royal College of Surgeons of England on 17th
March, 1953. Ann R Coll Surg Engl 1953;13(4):236–57.
[7] Pape SA, Baker RD, Wilson D, Hoeksema H, Jeng JC, Spence RJ,
et al. Burn wound healing time assessed by laser Doppler
imaging (LDI). Part 1: derivation of a dedicated colour code for
image interpretation. Burns 2012;38(2):187–94.
[8] Monstrey SM, Hoeksema H, Baker RD, Jeng J, Spence RS, Wilson
D, et al. Burn wound healing time assessed by laser Doppler
imaging. Part 2: validation of a dedicated colour code for image
interpretation. Burns 2011;37(2):249–56.
[9] Guidelines N. moorLDI2-BI: a laser Doppler blood flow imager
for burn wound assessment. NICE; 2017. . [updated March
201128/07/2016] http://guidance.nice.org.uk/MTG2.
[10] Lawson RN, Wlodek GD, Webster DR. Thermographic
assessment of burns and frostbite. CMAJ 1961;84:1129–31.
[11] Mladick R, Georgiade N, Thorne F. A clinical evaluation of the
use of thermography in determining degree of burn injury.
Plast Reconstr Surg 1966;38(6):512–8.
[12] Hackett ME. Colour thermography in the diagnosis of burn
depth. Transactions of the fifth international congress in
plastic and reconstructive surgery. Butterworth. p. 813–20.
[13] Hardwicke J, Thomson R, Bamford A, Moiemen N. A pilot
evaluation study of high resolution digital thermal imaging in
the assessment of burn depth. Burns 2013;39(1):76–81.
[14] Anselmo VJ, Zawacki BE. Effect of evaporative surface cooling
on thermographic assessment of burn depth. Radiology
1977;123(2):331–2.
[15] Cole RP, Shakespeare PG, Chissell HG, Jones SG.
Thermographic assessment of burns using a nonpermeable
membrane as wound covering. Burns 1991;17(2):117–22.
[16] Hackett ME. The use of thermography in the assessment of
depth of burn and blood supply of flaps, with preliminary
reports on its use in Dupuytren’s contracture and treatment of
varicose ulcers. Br J Plast Surg 1974;27(4):311–7.
[17] Renkielska A, Nowakowski A, Kaczmarek M, Dobke MK,
Grudzinski J, Karmolinski A, et al. Static thermography
revisited—an adjunct method for determining the depth of the
burn injury. Burns 2005;31(6):768–75.
[18] Stock C, Hielscher T. DTComPair: comparison of binary
diagnostic tests in a paired study design. R package version
1.0.3. 2014.
[19] Geertzen J. Inter-Rater Agreement with multiple raters and
variables. 2012 Available from: https://nlp-ml.io/jg/software/
ira/. [updated June 6, 2017].
[20] Landis JR, Koch GG. The measurement of observer agreement
for categorical data. Biometrics 1977;33(1):159–74.
[21] Mazurek MJ, Frew Q, Sadeghi AM, Tan A, Syed M, Dziewulski P.
A pilot study of a hand-held camera in a busy burn centre:
prediction of patient length of recuperation with wound
temperature. Burns 2016;42(3):614–9.
[22] Monstrey S, Hoeksema H, Verbelen J, Pirayesh A, Blondeel P.
Assessment of burn depth and burn wound healing potential.
Burns 2008;34(6):761–9.
[23] Pape SA, Skouras CA, Byrne PO. An audit of the use of laser
Doppler imaging (LDI) in the assessment of burns of
intermediate depth. Burns 2001;27(3):233–9.
[24] Boccara D, Chaouat M, Uzan C, Lachere A, Mimoun M.
Retrospective analysis of photographic evaluation of burn
depth. Burns 2011;37(1):69–73.
[25] Burke-Smith A, Collier J, Jones I. A comparison of non-invasive
imaging modalities: Infrared thermography,
spectrophotometric intracutaneous analysis and laser
Doppler imaging for the assessment of adult burns. Burns
2015;41(8):1695–707.
[26] Medina-Preciado JD, Kolosovas-Machuca ES, Velez-Gomez E,
Miranda-Altamirano A, Gonzalez FJ. Noninvasive
determination of burn depth in children by digital infrared
thermal imaging. J Biomed Opt 2013;18(6):061204.
[27] Prindeze NJ, Hoffman HA, Ardanuy JG, Zhang J, Carney BC,
Moffatt LT, et al. Active dynamic thermography is a sensitive
method for distinguishing burn wound conversion. J Burn Care
Res 2016;37(6):e559–68.
[28] Singer AJ, Relan P, Beto L, Jones-Koliski L, Sandoval S, Clark RA.
Infrared thermal imaging has the potential to reduce
unnecessary surgery and delays to necessary surgery in burn
patients. J Burn Care Res 2016;37(6):350–5.
[29] Liddington MI, Shakespeare PG. Timing of the thermographic
assessment of burns. Burns 1996;22(1):26–8.
[30] Miccio J, Parikh S, Marinaro X, Prasad A, McClain S, Singer AJ,
et al. Forward-looking infrared imaging predicts ultimate burn
depth in a porcine vertical injury progression model. Burns
2016;42(2):397–404.
[31] Newman P, Pollock M, Reid WH, James WB. A practical
technique for the thermographic estimation of burn depth: a
preliminary report. Burns 1981;8(1):59–63.
[32] Zhu WP, Xin XR. Study on the distribution pattern of skin
temperature in normal Chinese and detection of the depth of
early burn wound by infrared thermography. Ann N Y Acad Sci
1999;888:300–13.
[33] Renkielska A, Nowakowski A, Kaczmarek M, Ruminski J.
Burn depths evaluation based on active dynamic IR thermal
imaging—a preliminary study. Burns 2006;32(7):
867–75.

Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
JBUR 5341 No. of Pages 10
10 burns xxx (2017) xxx –xxx
[34] Renkielska A, Kaczmarek M, Nowakowski A, Grudzinski J,
Czapiewski P, Krajewski A, et al. Active dynamic infrared
thermal imaging in burn depth evaluation. J Burn Care Res [41]
2014;35(5):e294–303.
[35] Prindeze NJ, Fathi P, Mino MJ, Mauskar NA, Travis TE, Paul DW,
et al. Examination of the early diagnostic applicability of active
dynamic thermography for burn wound depth assessment [42]
and concept analysis. J Burn Care Res 2015;36(6):626–35.
[36] Schwartz RG, Elliott R, Goldberg GS, Govindan S, Conwell T,
Hoekstra PP, et al. The American Academy of Thermology. [43]
Guidelines for neuromusculoskeletal thermography. Thermol
Int 2006;16:5–9.
[37] Haddad DS, Brioschi ML, Arita ES. Thermographic and clinical [44]
correlation of myofascial trigger points in the masticatory
muscles. Dentomaxillofac Radiol 2012;41(8):621–9.
[38] Hoeksema H, Baker RD, Holland AJ, Perry T, Jeffery SL, [45]
Verbelen J, et al. A new, fast LDI for assessment of burns: a
multi-centre clinical evaluation. Burns 2014;40(7):1274–82.
[39] Stewart CJ, Frank R, Forrester KR, Tulip J, Lindsay R, Bray RC. A [46]
comparison of two laser-based methods for determination of
burn scar perfusion: laser Doppler versus laser speckle
imaging. Burns 2005;31(6):744–52.
[40] Zötterman J, Mirdell R, Horsten S, Farnebo S, Tesselaar E.
Methodological concerns with laser speckle contrast imaging
Please cite this article in press as: C. Wearn, et al., Prospective comparative evaluation study of Laser Doppler Imaging and thermal
imaging in the assessment of burn depth, Burns (2017), http://dx.doi.org/10.1016/j.burns.2017.08.004
in clinical evaluation of microcirculation. PLoS One 2017;12(3):
e0174703.
Afromowitz MA, Callis JB, Heimbach DM, DeSoto LA, Norton
MK. Multispectral imaging of burn wounds: a new clinical
instrument for evaluating burn depth. IEEE Trans Biomed Eng
1988;35(10):842–50.
Eisenbeiss W, Marotz J, Schrade JP. Reflection-optical
multispectral imaging method for objective determination of
burn depth. Burns 1999;25(8):697–704.
Iraniha S, Cinat ME, VanderKam VM, Boyko A, Lee D, Jones J,
et al. Determination of burn depth with noncontact
ultrasonography. J Burn Care Rehab 2000;21(4):333–8.
Adams TS, Murphy JV, Gillespie PH, Roberts AH. The use of
high frequency ultrasonography in the prediction of burn
depth. J Burn Care Rehab 2001;22(3):261–2.
Goertz DE, Christopher DA, Yu JL, Kerbel RS, Burns PN, Foster
FS. High-frequency color flow imaging of the microcirculation.
Ultrasound Med Biol 2000;26(1):63–71.
Still JM, Law EJ, Klavuhn KG, Island TC, Holtz JZ. Diagnosis of
burn depth using laser-induced indocyanine green
fluorescence: a preliminary clinical trial. Burns 2001;
27(4):364–71.