Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Statistics from Altmetric.com
What is already known?
Infrared thermography has applications in adult medicine, with a number of studies showing it to be an accurate method of recording skin surface temperature.
Many diseases are associated with skin temperature changes that can be detected with infrared thermography.
In the screening for diseases, for example, breast cancer, infrared thermography has been found to have a particularly high sensitivity when compared with current management.
What this study adds?
The use of infrared thermography in children has not been widely explored, with a limited number of relevant studies.
Infrared thermography has been shown to be more accurate in detecting skin temperature changes in children, compared with an adult population.
Infrared thermography (IRT) has been used in medicine for decades, with the first medical use of the technology described in 1959 for imaging arthritic joints.1 Infrared radiation is normally emitted by the human skin, with varying degrees of radiation being recorded from different regions of the body. Using IRT, a unique infrared ‘map’ of the body can be recorded and, through computer processing, can be displayed as a colour image. The clinical use of this technology derives from the changes in blood flow associated with particular diseases, which confer an alteration of the local skin temperature. Thermography has been used in a range of fields in adult medicine, but its usage in children has not been as widely explored.2 Additionally, technological improvements have allowed patients to be imaged with greater detail and accuracy, opening up new scope for research.
Materials and methods
Databases and search strategy
Electronic searches were performed independently by two authors, using the databases of MEDLINE (via Web of Science), the Cochrane Library, CINAHL (EBSCO) and SCOPUS. We also searched Open Grey, Google Scholar and recent conference abstracts. The manufacturers of current IR cameras were also contacted, as were authors of recently published studies. Articles published from 1990 to July 2016 were included.
The search strategy that was used aimed to include articles that covered the topics of IRT and children, widening this search as much as possible, to include any possible synonyms (online supplementary file 1).
Supplementary file 1
After removal of duplicate articles, the authors then independently screened the search results by title and abstract. Studies were selected on the basis of specified inclusion criteria: studies must involve the use of IRT imaging technology in a clinical setting, including one or several samples of living human subjects, with subjects being under the age of 18 years. Disagreement in study inclusion or exclusion was resolved by discussion. All eligible studies were included, except those not archived online or with no available English translation of their abstract.
Resulting studies were then reviewed and included based on relevance to the topic of clinical IRT in children. Finally, the references of included studies were checked for additional studies that may have been overlooked.
The study selection is summarised in online supplementary file 2.
Supplementary file 2
Meta-analysis was not performed due to the marked heterogeneity in applications, study design and outcomes: this is a narrative summary of the available literature.
IRT in children
The field of IRT in medicine is diverse and has a footing in most major specialties; however, the specific use of this technology in paediatrics is not as widely documented: this review, therefore, intends to summarise the available literature on IRT in paediatric medicine.
In an assessment of skin temperature in 25 healthy children, Kolosovas-Machuca et al used IRT to compare the temperature patterns of various regions of the body. They reported a maximum temperature difference of 5.1°C in the y-axis of the body and 0.7°C in the x-axis. Citing a similar study into adult patients, the authors highlighted reduced variability of results in children, which could confer increased precision when applied as a diagnostic tool.3 Symonds et al reported a similar conclusion, in a study of skin temperature response to cold challenge in 26 healthy participants, including adults and children. The largest temperature increase was found in the child cohort (increase of 1107%±365% in children, 33%±300% in individuals aged 13–18 years and 113%±195% in adults, following cold challenge, p<0.05).4 Significantly more publications report on the use of IRT in an adult population but, as has been suggested by these two studies, IRT may have greater precision and better diagnostic outcomes in a paediatric population.
There is insufficient evidence for using IRT in mass population fever screening.5 However, in paediatric patients, applying this technology to individual temperature monitoring (instead of tympanic thermometry) demonstrated more positive outcomes.
Selent et al 6 assessed the application of IRT in fever screening, reporting a range of sensitivity and specificity of 76.4%–83.7% and 79.4%–86.3%, respectively: standard temperature measurement reported a sensitivity of 83.9% and a specificity of 70.8%. In another study, Chan et al 7 reported similar results, with a sensitivity and specificity of 83% and 88% for IRT. This highlighted a benefit of the thermal systems, in that they could more accurately exclude afebrile patients, compared with traditional methods. However, Fortuna et al 8 criticised that IRT tended to overestimate in afebrile patients and underestimate in febrile patients, when compared with rectal temperature readings.
Although with varying consensus on its usage in temperature screening, a benefit of using IRT in the screening of young children for fever is the convenience and speed at which measurements can be taken. With further improvements in IRT technology, it may have applications in the field of emergency paediatrics for rapid high-volume screening.
Monitoring vital signs
Although vital signs are currently monitored using observations conducted at regular intervals, efforts are being made to explore technologies that would allow continuous monitoring.
Heimann et al showed IRT to be an effective tool in reporting fluctuations in neonatal body temperature, with a significant difference reported in preterm infants in different ambient temperatures (p<0.05), while a similar conclusion was reported by Anderson et al, in case studies of two sleeping infants.9 10 This simple application of IRT in paediatric monitoring has been expanded in subsequent studies, in which sophisticated tracking software assists in the assessment of respiratory rate.
Abbas et al found IRT to be promising in neonatal respiratory monitoring, with the possibility of accurate tracking software, but results were limited by a small sample size.11 12 This application of IRT has also been explored in studies by Al-Khalidi et al, Elphick et al and Goldman et al, using IRT to measure the respiratory rate of subjects, comparing results with standard methods of respiratory monitoring. Al-Khalidi et al reported a correlation coefficient of 0.994 between the IRT and the standard methods, while Elphick et al reported similar results, with a correlation coefficient of 0.578–0.999 in the paediatric cohort.13 14 Goldman et al found a high Cronbach’s alpha value of 0.976 (95% CI 0.992 to 0.944) between the IRT and control measurements and successfully identified individuals with respiratory disease, using the time-lag between the ribcage and abdomen (p=0.0125).15 As these studies have suggested, there is potential for further research into non-contact, continuous respiratory monitoring with IRT.
Incorporating a similar method of neonatal monitoring, IRT has been used to screen for necrotising enterocolitis (NEC) in neonates . In a study of 13 neonates at risk of the disease, Rice et al found that those with NEC had a lower abdominal skin temperature (35.3±0.8°C) than those without the disease (36.6±0.9°C; p<0.05).16 17 With larger studies, IRT could be incorporated in the management of this life-threatening neonatal disease.
Trauma and wound healing
Many studies have explored the potential of IRT in detecting temperature changes occurring during traumatic injury or infection.
Sanchis-Sánchez et al explored the use of IRT in ruling out fractures in paediatric trauma patients, reporting a sensitivity of 0.91 and a specificity of 0.88.18 Similarly, Silva et al 19 examined the use of IRT in locating areas of trauma in 51 patients, finding that the technology matched the site of pain in 73% of locations, as well as 7 out of 11 fracture sites. Ćurković et al used IRT in 19 children with forearm fracture: the average temperature of the injured side was 1.17°C higher than that of the healthy side after 1 week and 0.84°C higher after 2 weeks; however, Ćurković et al were unable to produce statistically significant results due to the small sample size.20 Additionally, a recent pilot study has found IRT to be successful in identifying the affected region in patients presenting with acute non-specific limp, with areas of fracture associated with the greatest temperature change, but conclusions were limited by the small sample size.
Saxena et al performed IRT on 483 paediatric surgical patients, over 10 years of clinical practice. In five children with partial amputation, a temperature differential of 2.5±0.3°C was observed following surgical treatment and revascularisation, which reduced to 1.8±0.3°C after 48 hours. Eighteen neonates had surgically implanted skin patches: there was an initial average temperature differential of −4.8±0.6°C following surgery, which increased to 3.4±0.5°C at 30–42 days after surgical intervention, illustrating revascularisation. Furthermore, Saxena et al cited a patient who underwent surgical repair of a thoracic wall abnormality, in which IRT identified a significant temperature increase of 3.7°C, associated with the formation of a sternal wound abscess. Similarly, 42 children were assessed for infection, with areas of abscess showing a temperature increase of 3.6±0.5°C and wound infections also displayed a positive temperature differential. Finally, Saxena et al cited a particular case involving a 2-year-old child with severe gas gangrene of an upper extremity, which required amputation: IRT was used to assess the level of amputation necessary.21–23
Similar to its applications in monitoring wound healing, IRT has also been used to assess burn injuries in children. In a pilot study of 13 children, Medina-Preciado et al reported that the average temperature of superficial dermal burns were 1.7°C higher than the contralateral side, while that of deep dermal burns was 2.3°C lower than the contralateral side (p<0.05). Additionally, when compared with histological results, IRT correctly identified 100% of cases of both superficial and deep burns, while clinical assessment identified 83.33% of superficial and 42.85% of deep burns.24 This outcome was reiterated by Saxena et al, also reporting a 2.8±0.6°C temperature differential across superficial burns.21
These accounts of the surgical applications of IRT illustrate a variety of areas in which thermography can be applied to monitor revascularisation of tissue, as well as screen for signs of infection.
Haemangioma and varicocele
Similar to previous studies, IRT may be used to monitor the progression of haemangioma or varicocele, through the irregular blood flow patterns associated with their formation.
Saxena et al used IRT to image haemangiomas in 102 affected children: 52 patients had a rapidly progressing haemangioma, which showed a temperature differential of 1.5±0.3°C, while those that underwent complete resolution displayed a differential of <0.5°C21. Mohammed et al also reported a decrease in temperature associated with haemangioma resolution, while Garcia-Romero et al displayed how IRT could be used to monitor treatment response, in 10 patients with haemangioma undergoing treatment with systemic beta-blockers.25 26
Saxena et al used IRT in six boys with varicocele, reporting a positive temperature differential of 4.1±0.3°C in the affected side, which reduced significantly with surgical intervention.21 This finding was echoed in a case study by Iwata et al, in which a varicocele repair in a 12-year-old boy resulted in similar temperature change in the affected side.27
These studies into both haemangiomas and varicoceles represent a potential way in which IRT could be incorporated to monitor treatment response in certain diseases.
Many skin conditions involve alterations in the relative thickness of the skin that may confer changes in temperature: a finding that may be quantified by IRT. Exploring the use of IRT in identifying children with localised scleroderma, Martini et al reported a sensitivity and specificity of 92% and 68%, respectively.28 IRT found a similar correlation between disease severity and skin temperature, in another case study involving a patient with localised scleroderma, as well as a patient with psoriasis.29 30 The skin is an accessible organ that lends itself well to imaging with IRT, and these studies illustrate how it could be applied to paediatric dermatology.
The effect of diabetes mellitus on skin perfusion is often only clinically evident after decades of disease. However, Zotter et al found that IRT identified a significant temperature difference between patients affected by diabetes and healthy controls, following cold challenge testing (p<0.05).31 This study highlighted a potential advantage of IRT in diabetic screening, but a larger study is required to reinforce conclusions.
There are a limited number of studies exploring the use of IRT in joint inflammation in children, even though the use of IRT in rheumatoid conditions in adults is well documented. Lasanen et al assessed the application of IRT in the screening of 58 children with signs of joint inflammation, reporting a statistically significant temperature increase in inflamed ankle joints, compared with controls (p<0.05). However, in knee joints, no such difference was shown.32 This study suggests that the efficacy of IRT in screening for joint inflammation may be specific to the area affected, with some joints exhibiting more acute changes in temperature.
IRT may have a variety of applications in neurology. Goetz et al used IRT to monitor hydrocephalus shunt patency, using cold challenge testing to assess temperature differentials, while Zurek et al used IRT to monitor tissue perfusion in a study involving a novel treatment for cerebral palsy.33 34 An innovative study by Coben et al used IRT to record the temperature change in a specific area of the head (named ‘Fpz’), lying over a region of the brain implicated in attention-deficit/hyperactivity disorder. Using the temperature differential to detect patients with disease, a sensitivity of 66% was reported and Coben et al suggested that IRT was superior to the limited alternative diagnostic tests for the disease.35
Kaercher et al explored IRT as a method of ophthalmic examination in 34 patients with X-linked hypohidrotic ectodermal dysplasia (XLHED). In the child cohort, IRT had a sensitivity of 66.7%, but standard methods of diagnosis reported sensitivity values of 72.7%–100%. Although representing no improvement on current practice, IRT illustrated marked temperature differences between children with XLHED and the healthy controls and was shown to be a quick and reliable tool to be used in conjunction with other methods of diagnosis.36
Clark et al studied the use of IRT in detecting food intolerance, in a study of 16 children with known peanut allergy. Following administration of peanut protein via nasal spray, the active group exhibited a higher nasal temperature than the control after 20 min (p<0.05).37 This result was reiterated in a previous study by the same authors, in which egg protein was found to cause a temperature change in those with egg allergy.38 IRT may represent an improvement over alternative methods of allergy screening, with a reduced risk of adverse events associated with nasal challenge over oral challenge.
Cheema et al described a case report of the use of IRT in the assessment of a thoracic epidural block, in which a clearly delineated skin temperature change was found from the dermatomes of T4 to T10, indicating the epidural blockade.39 This case study illustrates the relationship between the peripheral nervous system and skin perfusion, which has relevance in a number of clinical specialties.
Results from this review have been summarised in table 1.
Although a number of studies showed promising clinical applications, with sensitivity and specificity figures similar to that of accepted diagnostic tools, the quality of many included studies was relatively low, with case studies and small pilot studies providing little clinical evidence of efficacy. Conclusions drawn from this review must be considered in this context, but due to the nature of the application of IRT in a clinical environment, true diagnostic accuracy studies are uncommon.
The risk of bias and the applicability of studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies Tool (QUADSA2) (table 2).
As thermal cameras have increasingly higher resolutions, the sensitivity improves and the accuracy with which diseases may be identified increases. IRT has shown to be particularly useful in an emergency setting, with applications into the assessment of both burns and fractures having the potential to change current management pathways.18–24 Also highlighted in this review, studies into respiratory rate monitoring in neonates proved to be successful, with accurate measurement illustrated in a number of publications.11–15 IRT can express greater accuracy in children and, in such a population, non-contact methods of investigation are well-received.3 4 Indeed, there is a need for further research into the application of IRT in paediatrics.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.