Methods to assess area and volume of wounds – a systematic review

Abstract

Wound measurement is important in monitoring the healing process of chronic wounds and in evaluating the effect of treatment. The objective of this systematic review was to evaluate evidence from the literature on accuracy, agreement, reliability and feasibility of wound measurement techniques described since 1994. Studies were identified by searching the electronic databases PubMed, Embase and Cochrane Library. Of the 12 013 studies identified, 43 were included in the review. A total of 30 papers evaluated techniques for measuring wound area and 13 evaluated techniques for measuring wound volume. The six approaches for measuring wound area were simple ruler method (10 papers), mathematical models (5 papers), manual planimetry (10 papers), digital planimetry (16 papers), stereophotogrammetry (2 papers) and digital imaging method (20 papers). Of these studies, 10 evaluated accuracy, 15 agreement, 17 reliability and 25 mentioned feasibility. The number of wounds examined in the studies was highly variable (n = 3–260). Studies evaluating techniques for measuring wound volume included between 1 and 50 wounds and evaluated accuracy (4 studies), agreement (6 studies), reliability (8 studies) and feasibility (12 studies). Digital planimetry and digital imaging were considered the most accurate and reliable methods for area measurement, particularly in larger and irregularly shaped wounds. None of the three‐dimensional technologies have so far had a major impact, because of their low accuracy, high cost and complexity in handling the system set‐up.

Introduction

Measurement of wound size is important in monitoring the wound healing process and in evaluation of the effects of treatment. Wound size and area are useful predictors of the final outcome 1, 2, 3, 4. Flanagan suggested that a 40% reduction in ulcer area after 4 weeks is a good predictor of wound healing 5. The accuracy, agreement, reliability and feasibility of wound assessment methods are therefore central issues in the practical management of wounds and in the development of new treatment options for a large group of individuals.

Various approaches to objective measurement of wound size have been suggested. The focus in the last few decades has been on two‐dimensional methods to measure wound area, which can be divided into contact methods (e.g. manual and digital planimetry) and non‐contact methods (e.g. simple ruler method, mathematical models such as the elliptical method, stereophotogrammetry (SPG) and digital imaging) 6. More recently, three‐dimensional (3D) methods for measuring wound volume have made it possible to evaluate the process of wound healing in relation to all dimensions including depth, and thereby reflect the formation of granulation tissue. It is hypothesised that a 3D approach provides a more accurate evaluation of the biological changes and thus results in more relevant data than a simple two‐dimensional approach. However, none of the newer methods appear to be used on a routine basis, possibly due in part to the absence of valid comparative studies. As a first step towards this, we have performed a systematic review of the available evidence for the accuracy, agreement, reliability and feasibility of wound measurement methods.

Methodology

The objective of this review was to evaluate the techniques used for wound measurement described in the literature between 1 January 1994 and 30 June 2014. Studies including wound surface area and volume measurements were reviewed by describing the evidence for accuracy (the ability to measure the true size of the wound), agreement (the consistency in wound measurement between different methods), reliability (the consistency of wound measurements performed by the same clinician or different clinicians) and feasibility (the practical use of the wound measurement method).

Methods

Search strategy

Studies were identified by searching the electronic databases PubMed, Embase and Cochrane Library. The search terms used in the literature search in the PubMed database were (wound OR ulcer) AND measurement AND (technique OR techniques OR method OR methods). In the Embase database and Cochrane Library, the search terms were ‘wound’ and ‘measurement’.

Study selection

Papers written in English focusing on the evaluation of a wound measurement method were included. Exclusion criteria were evaluation of other aspects of wound assessment (e.g. blood flow, temperature, pH) or wound treatment. Two researchers (LBJ and KBY) first screened the information in the study titles and abstracts and then reviewed the full‐text papers and agreed on study inclusion. In cases of disagreement, a third researcher (JAS) evaluated the paper. The retrieved papers were hand‐searched for additional references.

The resulting 43 studies were subdivided into measurement of surface area (n = 30) and volume determination (n = 13) (Figure 1). Figure 2 illustrates the study selection process.

Figure 1.

Overview of the wound measurement methods.

Figure 2.

Flow diagram of the literature search.

Results

Wound surface area measurement

The 30 papers described a highly variable number of wounds (n = 3–260). Six different methods were evaluated including simple ruler method (10 papers), mathematical models (5 papers), manual planimetry (10 papers), digital planimetry (16 papers), SPG (2 papers) and digital imaging (20 papers). Measurement accuracy was evaluated by 10 studies, agreement by 15 studies, reliability by 17 studies and feasibility by 25 studies. Table 1 summarises the characteristics of these studies.

Table 1.

Wound surface area measurement methods

Studies	Methods	Number of ulcers	Statistical methods	Outcomes	Feasibility
Bhedi et al., India, 2013 24	1. Simple ruler method (length × width) 2. Manual planimetry 3. Digital imaging method	40 patients with different wounds	Agreement: mean values and standard deviation	Simple ruler method was significantly different from manual planimetry and digital imaging (P < 0·05). No significant difference between manual planimetry and digital imaging (P > 0·05)	The manual planimetry was time‐consuming. The digital imaging method was non‐invasive
Bilgin and Günes, Turkey, 2013 25	1. Simple ruler method (length × width) 2. Acetate tracings (manual planimetry) 3. Digital planimetry	80 pressure wounds (56 wounds under 10 cm2) measured three times	Agreement: intra‐class correlation coefficients (ICC)	Strong agreement between the three methods in wounds <10 cm2 (ICC = 0·95). Moderate agreement in wounds >10 cm2 (ICC = 0·70)	Feasibility was not assessed
Bowling et al., UK, 2009 10	1. Digital elliptical measurement (DEM) 2. Hand‐measured elliptical measurement (EM)	36 diabetic foot ulcers in 31 patients	Agreement: Pearson, Spearman and Kendall rank coefficients Reliability: intra‐ and inter‐observer variation	Strong correlation between area measurement by DEM and EM: Pearson 0·961, Spearman 0·929 and Kendall 0·791 (P < 0·001). Intra‐observer variation of wound length and width was <3·0% and that of wound area was <8·0%. Inter‐observer variation of wound length was <6·5% and that of wound width <10·0%.	The digital measurement method was non‐invasive, reducing risk of infection, simple to use and could easily be implemented in routine clinical practice. Feasible for electronic transmission and could have significant potential for telemedicine.
Chang et al., Australia, 2011 23	1. Visitrak® (digital planimetry) 2. Digital imaging method and the software ImageJ to calculate wound area	Four wounds on two pigs were measured once a week for 5 weeks	Agreement: two‐tailed paired t‐test	T‐statistic of 1·285 with a P‐value 0·206 showed no statistical differences in wound area measurement between the two methods	The digital imaging method was a non‐contact method, thus reduced risk of contamination and limited patient discomfort
Flowers et al., Australia, 2008 36	AMWIS software (digital imaging method)	18 wound images measured by four clinicians	Inter‐rater reliability: ICC Pearson correlation coefficient	ICC was 0·958, p < 0·001 Pearson correlation coefficients were 0·710–0·997	Easy to use, useful in remote settings
Foltynski et al., Poland, 2013 32	1. Elliptical method 2. Visitrak® (digital planimetry) 3. TeleDiaFoS® system 4. Silhouette Mobile® device	16 wound templates in different shapes	Accuracy: relative error (RE) Reliability: coefficient of variation (CV)	Overall absolute RE: 1. 13·3%, 2. 6.8%, 3. 2·1%, 4. 2·3% Overall CV: 1. 6·0%, 2. 6·3%, 3. 1·6%, 4. 3·1%	The functionality of the TeleDiaFos® system was limited to the sole of the foot and flat surfaces
Foltynski et al., Poland, 2014 34	1. AreaMe® software 2. Visitrak® (digital planimetry) 3. Silhouette Mobile® device	108 wound shapes were measured five times	Accuracy: RE, compared using Mann–Whitney U‐test Reliability: coefficient of variation (CV), compared using Mann–Whitney U‐test	Overall REs were −3·4% for AreaMe®, −6·3% for Visitrak® and −2·1% for Silhouette Mobile®. Overall CV was 1·6% for AreaMe®, 4·2% for Visitrak® and 1·0% for Silhouette Mobile®.	The largest wounds measured on AreaMe® fitted into a rectangle of 15 × 21 cm. The area measurements using AreaMe® could be stored in the local database.
Gethin and Cowman, Ireland, 2006 14	1. Acetate tracings (manual planimetry) 2. Visitrak® (digital planimetry)	25 wounds under 10 cm2 and 25 wounds over 10 cm2	Agreement: the difference between the two methods analysed by independent t‐test	The difference between the two methods was not significant for wounds < 10 cm2 (t = 0·995, P = 0·330) but significant for wounds > 10 cm2 (t = 2·887, P = 0·008)	Both methods depended on accurate wound tracing. Easy to use.
Haghpanah et al., USA, 2006 31	1. Simple ruler method (length × width) 2. Visitrak® (digital planimetry) 3. VeV MD system (digital imaging method)	40 wounds measured twice by four observers using the three techniques	Accuracy: root mean square error (RMSE) Reliability: intra‐ and inter‐rater variation assessed by two‐way ANOVA	Mean RMSEs were 81 for simple ruler method, 19 for Visitrak and 16 for VeV MD. Intra‐rater variation was non‐significant for all methods. Inter‐rater variation was significant for all three methods (P < 0·010).	Simple ruler method was fast, easy to learn and of low cost. Visitrak was easy to learn and portable, of moderate cost, but did not store data. VeV MD was portable, non‐contact and able to store data, but expensive and time‐consuming.
Hammond and Nixon, USA, 2011 28	Silhouette Mobile® device	five wounds measured by three raters, repeated five times	Reliability: intra‐ and inter‐rater variation; coefficients of variation (CV) Overall reliability: ICC	The CV for intra‐rater variation was 2·6% and 3·2% for inter‐rater variation. Overall reliability ICC was 99·76%.	The device was fast, non‐invasive, easy to use without much training and had low variation
Kantor and Margolis, USA, 1998 9	1. Digital planimetry 2. Simple ruler method (length × width)	260 leg ulcers	Agreement: Pearson, Spearman and Kendall correlation coefficient	For overall area measurement: Pearson = 0·9532 Spearman = 0·9716 Kendall = 0·8828 (P < 0·001 for all). For areas >40 cm2 Pearson = 0·7875 Spearman = 0·8353 Kendall = 0·6667 (P < 0·05)	Feasibility was not assessed
Ladyzynski et al., Poland, 2011 33	1. TeleDiaFoS® system 2. Elliptical method 3. Visitrak® (digital planimetry) 4. Silhouette Mobile® device	36 diabetic foot ulcers	Agreement: linear regression; Bland–Altman analysis	Correlation coefficients between instruments 1 and 2: 0·949, 1 and 3: 0·985, 1 and 4: 0·987	The functionality of the TeleDiaFos® system was limited to the sole of the foot and flat surfaces
Lagan et al., USA, 2000 15	1. Manual planimetry 2. Digital planimetry	11 wounds in seven patients measured three times by one clinician	Intra‐rater reliability: coefficient of variation (CV)	Manual planimetry had significantly larger variability than digital (P = 0·02)	Both methods were dependent on investigator skill, precision and interest
Langemo et al., Canada, 1998 19	1. Simple ruler method (length × width) 2. Manual planimetry 3. Stereophotogrammetry (SPG) L × W 4.Stereophotogrammetry (SPG) area	Three wound models measured by 66 raters	Accuracy: standard error of measurement (SEM) Inter‐rater reliability: ICC Intra‐rater reliability: Pearson correlation coefficient	SEM values were 12·33–17·27 for simple ruler method, 8·67–10·0 for manual planimetry, 5·37–6·45 for SPG. ICCs were 0·98 for simple ruler method and manual planimetry and 0·99 for SPG.	The SPG method was a non‐contact method, but time‐consuming
				Pearson coefficients were 0·48–0·68 for simple ruler method, 0·85–0·92 for manual planimetry and 0·38–0·40 for SPG.
Mayrovitz et al., USA, 1997 11	1. Elliptical method 2. The formula 0.73 × L × W 3. Digital planimetry	81 diabetic neuropathic plantar ulcers	Accuracy: RMSE	RMSEs were 8·7% for the elliptical method and 6·2% for the formula 0·73 × L × W	Formula 0·73 × L × W was fast, easy to use, inexpensive and reasonably accurate in ulcer area estimation
Mayrovitz and Soontupe, USA, 2009 16	Digital planimetry	six wounds	Accuracy: area errors Reliability: coefficient of variation (CV)	Mean area error (± SD) ranged from −2·5% ± 7·01% to 2·32% ± 6·04%. CV <2·5% in four wounds; 6·72–7·56% in two wounds.	Feasibility was not assessed
Miller et al., Australia, 2012 29	Silhouette Mobile® device	14 wounds measured by seven clinicians	Inter‐rater reliability: ICC Intra‐rater reliability: Pillai's trace	ICC was 0·985 for large wounds and 0·624 for small wounds. No significant difference in intra‐rater reliability (P > 0·05).	Feasibility was not assessed
Oien et al., Sweden 2002 13	1. Digital planimetry 2. Mechanical planimetry 3. Grid Tracing (manual planimetry) 4. Simple ruler method (length × width)	20 patients with 50 chronic leg ulcers of various aetiology and sizes	Agreement: Pearson correlation coefficient	Coefficient = 1·00 (P < 0·01) between digital and mechanical planimetry; 1·00 (P < 0·01) between digital planimetry and grid tracing; 0·99 (P < 0·01) between digital planimetry and simple ruler method.	Digital planimetry was fast and practical and thus useful in clinical practice
Rajbhandari et al., UK, 1999 21	1. Graph paper method (manual planimetry) 2. Digital imaging method	30 diabetic foot ulcers in 18 patients	Reliability: coefficient of variation (CV)	CV was 16% for digital imaging method and 27% for graph paper method (P = 0·05)	Digital imaging method was faster and easier to use, and was a non‐contact method
Rogers et al., USA, 2010 7	1. Simple ruler method (length × width) 2. Silhouette Mobile device®	Ten patients with circular or oval‐shaped wounds of varying aetiology	Agreement: percentage of overestimation was calculated by (l × w – TA)/TA)	The simple ruler method overestimated wound area by an average of 41%	The Silhouette Mobile device® was non‐invasive and more accurate compared with the simple ruler method
Samad et al., UK, 2002 22	1. Digital imaging method 2. Acetate tracing (manual planimetry)	11 shapes with known area measured by four raters	Accuracy: area error	Acetate tracing underestimated the area by 3·9% (P < 0·05). Digital imaging showed no significant error.	Digital imaging was faster and more accurate and could be used for telemedical wound care
Santamaria et al., Australia, 2002 35	AMWIS software (digital imaging method)	Three objects with known size were recorded by the same rater. 100 wounds were measured repeatedly (440 images)	Accuracy: measurement error Pearson correlation coefficient	Measurement error 1·28% The correlation coefficient r was 69·7 (P < 0·01)	Easy to use, useful in remote assessment in telemedicine
Shaw et al., Ireland, 2007 18	1. Visitrak® (digital planimetry) 2. Digital imaging method 3. Elliptical method	16 diabetic foot ulcers measured three times with each method	Accuracy: one‐sample t‐test Reliability: CV	Visitrak® measured wounds inaccurately under 25 mm2 (P < 0·001), and the elliptical method tended to underestimate small wounds (P < 0·001). Mean CV was 7·0 for Visitrak, 4·7 for imaging and 8·5 for elliptical method.	The Visitrak® system was fast, easy to use, inexpensive and non‐invasive, but less accurate in measuring small wounds (<25 mm2)
Shetty et al., India, 2012 8	1. Graph method (manual planimetry) 2. Simple ruler method (length × width) 3. Digital imaging method	Ten wounds measured by three clinicians	Agreement: difference between graph method and the other two methods analysed by ANOVA (analysis of variance)	Mean over/under‐ estimation was 28·9–42·9% with the simple ruler method and 1·6–2·4% with the digital imaging method compared with the graph method. The difference between the three methods was not statistical significant (P > 0·001).	The graph method was accurate but cumbersome, time‐consuming and a contact method with risk of wound contamination. Digital imaging was non‐invasive but time‐consuming, the photos needed to be taken from a fixed distance, inaccurate in wounds on a curved limb.
Sprigle et al., USA, 2012 30	1. WMD (wound measurement device) 2. Simple ruler method (length × width)	19 pressure ulcers	Accuracy evaluated by (TA – measured area)/TA Intra‐ and inter‐rater reliability: Intra‐class correlation coefficient (ICC) Agreement: difference in area	Mean errors for WMD were 1·90% at 0°, 1·76% at 5° and 4·28% at 10°. Intra‐rater ICC for WMD >0·975. Inter‐rater ICC was 0·966 in first trial and 0·978 in second trial. On average, WMD measured the area 17·4% higher than simple ruler method.	The WMD was non‐invasive, easy to use and of low cost but had limitations in using a smartphone.
Stockton et al., Australia, 2015 20	1. LifeViz system® 2. Visitrak® (digital planimetry)	25 children with burn injury	Agreement and inter‐rater reliability: Intra‐class correlation coefficients (ICC)	Agreement ICC was 0·994. Inter‐rater ICC was 0·989.	LifeViz system® was fast, easy to use, non‐contact and produced high quality images
Sugama et al., Japan, 2007 17	1. Visitrak® (digital planimetry) 2. Digital imaging method	Ten pressure ulcers in the reliability test for Visitrak® system 30 pressure ulcers used to compare the two methods	Intra‐ and inter‐rater reliability: Intra‐class correlation coefficients (ICC) analysed by ANOVA Agreement: coefficient of variation (CV)	Intra‐rater and inter‐rater ICC were both 0·99 for Visitrak®. CV was 3·1% for Visitrak® and 3·9% for digital imaging (P < 0·001).	Visitrak® was easy to use and faster than digital imaging
Thawer et al., Canada, 2002 26	1. Digital imaging method 2. Digital planimetry	45 patients with chronic wounds (arterial, neuropathic, decubitus and venous)	Intra‐ and inter‐rater reliability: Intra‐class correlation coefficients (ICC) analysed by ANOVA. Standard error of measurement (SEM)	Intra‐rater ICC was 0·99 and SEM was 0·32 for digital planimetry; ICC was 0·99 and SEM was 0·18 for digital imaging. Inter‐rater for single measurement: ICC was 0·91 and SEM was 0·83 for digital planimetry; ICC was 0·94 and SEM was 0·94 for digital imaging.	Digital imaging was more costly but non‐contact, thus lower risk of wound contamination. Digital planimetry was cheaper but a contact method with risk of wound contamination, pain and damage to the wound bed.
Van Poucke et al., Belgium, 2010 27	Two techniques to measure ulcer area from digital images: 1. Freehand drawing (FH) 2. Closed polygon graph algorithm (CP)	16 patients with 20 chronic wounds	Agreement: correlation coefficient Intra‐rater reliability: Kruskal–Wallis test (H‐test) Inter‐rater reliability: coefficient variation (CV)	Correlation between the two methods for area measurement was r = 0·99 (P < 0·001) Intra‐rater reliability showed significant differences for both techniques. Mean inter‐rater CV was 5·87% for FH and 5·81% for CP.	Feasibility was not assessed
Wunderlich et al., USA, 2000 12	1. Digital imaging method 2. Acetate tracing (manual planimetry) 3. Simple ruler method (length × width)	five wounds were measured five times by four clinicians	Agreement: Pearson correlation coefficients Intra‐class correlation coefficient (ICC)	Pearson r was 0·97 between 1 and 2, 0·95 between 1 and 3, 0·93 between 2 and 3. ICC was 0·94 between 1&2, 0·57 between 1 and 3, 0·76 between 2 and 3.	Digital imaging was more expensive but a non‐contact method

TA, true area; l, length; w = width.

Simple ruler method

The simple ruler method consists of multiplying the greatest length and width of the wound to determine the surface area. It is a low cost method and easy to use. However, it is accurate only in perfectly rectangular wounds and does not take into account wound irregularities. It is thus likely to overestimate the area. Rogers et al. 7 found that the simple ruler method overestimated the wound area by an average of 41% compared with digital planimetry. Similarly, Shetty et al. 8 reported 29–43% overestimation by the simple ruler method compared with manual planimetry. Both studies, however, evaluated only a small number of wounds (n = 10). In summary, the simple ruler method is easy to use and is of low cost, but overestimates the wound area.

Mathematical models

The mathematical models are based on the assumption that most wounds are spherical or elliptical in nature, and the area can therefore be calculated using standard mathematical formulae. The most common approach is the elliptical method, in which the area is calculated by multiplying π (π = 3·14) by the shortest and longest radii of the wound.

Kantor and Margolis 9 found a strong positive correlation between area measurements using the ellipse formula compared with digital planimetry, with Pearson, Spearman and Kendall correlation coefficients all greater than 0·88 (P < 0·001). However, the correlation was lower for wounds larger than 40 cm². Bowling et al. 10 investigated 36 diabetic foot ulcers in 31 patients over a period of 12 weeks and found a strong correlation of width, length and area between a hand‐measured elliptical method and a digital imaging system (P < 0·001), although the discrepancy between measurements increased with wound size. Intra‐ and inter‐rater variations using the digital elliptical measurement method were small, indicating a high reliability. Mayrovitz et al. 11 investigated shape and area measurements in 81 diabetic plantar ulcers using a new area formula 0·73 × L × W (L = length, W = width) based on a shape factor, which is an index of wound circularity between 0 and 1 (1 is a perfect circle). This new formula was found to be more accurate than the elliptical model when compared with digital planimetry.

In summary, the mathematical models are fast, easy to use and non‐invasive, but inaccurate in wounds with an irregular shape.

Planimetric measurement

Planimetric methods can be manual or electronic. In the manual method, a transparent film is placed on top of the wound and the margin of the wound is traced with a pen. The tracing is subsequently placed on a metric grid and wound area is determined by counting the number of squares in the grid covered by the traced area. In digital planimetry, the margin of the wound is retraced onto a tablet computer that performs the same calculations 12.

Oien et al. 13 compared four methods of wound area measurement in 20 patients with 50 chronic leg ulcers of various aetiology. The techniques were digital planimetry, mechanical planimetry, grid tracing (manual planimetry) and simple ruler method (width × length). All methods demonstrated a high degree of agreement in wounds less than 10 cm² (P < 0·01), but differences occurred with increasing wound size. Gethin and Cowman 14 measured 50 superficial leg ulcers using acetate tracings (manual planimetry) and Visitrak^® (Smith & Nephew, Hull, UK), which is a variant of digital planimetry. No statistically significant difference in wound area was found between the two methods for wounds below 10 cm² (P = 0·330), but the difference was significant for wounds above 10 cm² (P = 0·008). Lagan et al. 15 found that digital planimetry produced a smaller degree of variability (higher reliability) compared with manual planimetry. Mayrovitz and Soontupe 16 concluded that digital planimetry was fast and provided an accurate and reliable estimate of wound area. Sugama et al. 17 reported high intra‐ and inter‐rater reliability [intra‐class correlation coefficients (ICC 0·99)] for the Visitrak^® system and found a significant positive correlation (P < 0·001) between the Visitrak^® system and the digital imaging method in 30 pressure ulcers. Furthermore, Visitrak^® was faster and easier to use than the digital imaging method. In a three‐way comparison of the Visitrak^® system, digital imaging and the elliptical method, Shaw et al. 18 found Visitrak^® to be the most valid method, showing high reliability in the measurement of wounds larger than 25 mm².

In summary, planimetric methods are relatively easy to learn and are accurate and reliable. A drawback is the need for contact with the wound, which carries a risk of contamination. Digital planimetry is slightly more accurate and reliable than manual planimetry.

Stereophotogrammetry

This is a non‐contact method in which a stereographical camera linked to a computer captures an image of the wound. The image is downloaded to the computer, and the wound perimeter is simply traced by moving the cursor on the monitor. The computer software calculates the wound area, length and width. Wound size can be measured in both two and three dimensions.

Langemo et al. 19 compared four wound measurement methods including simple ruler method, manual planimetry (square counting) and SPG (Vista Medical, Winnipeg, Manitoba, Canada) for measurement of L × W and of area. The SPG area approach was found to be the most accurate with the smallest standard error of measurement. Inter‐rater reliability was equal for the four methods (ICC 0·98), but the SPG system had lower intra‐rater reliability. Stockton et al. 20 reported a high inter‐rater reliability (ICC 0·989) for the SPG system, LifeViz^® (QuantifiCare, Sophia Antipolis, France) and a high agreement with the Visitrak^® system in area measurements (ICC 0·994).

In summary, the SPG method is an accurate, non‐contact method, which reduces the risk of wound contamination, but it is also time‐consuming and expensive.

Digital imaging

In the digital imaging method, an image of the wound is captured and transferred to a computer. The margin of the wound is traced on the screen using a pointing device and the software uses a scale near the wound on the photo to estimate the area.

Rajbhandari et al. 21 compared a graph paper method (manual planimetry) with digital imaging to assess 30 diabetic foot ulcers in 18 patients. Digital imaging had a significantly lower inter‐rater reliability with a coefficient of variation (CV) of 16% and 27% for the digital and graph paper methods, respectively (P = 0·05). Samad et al. 22 compared these two methods and reported similar findings (P < 0·05) and also noted digital imaging to be faster than manual planimetry and useful in telemedical wound care. In contrast, Wunderlich et al. 12 reported a high correlation between the two methods (Pearson r = 0·97).

Chang et al. 23 evaluated digital imaging in combination with ImageJ software and Visitrak^® in measuring four wounds in two pigs. A two‐tailed paired t‐test showed no statistically significant difference between the two methods (P = 0·206). Bhedi et al. 24 and Bilgin and Günes 25 demonstrated that digital imaging and manual planimetry were equally accurate in wound area measurements less than 10 cm². Thawer et al. 26 compared the intra‐ and inter‐rater reliability of digital planimetry and digital imaging method in measuring chronic wounds (n = 45). The techniques were equally reliable indicated by similar ICC and standard errors of measurement.

Van Poucke et al. 27 evaluated two digital image techniques for measuring wound area. In the first technique, the wound margin was drawn freehand with an electronic pointer. The second technique was based on a closed polygon graph algorithm in which the wound margin was drawn by multiple lines that eventually met up. Both techniques had a low reliability and were considered not to be useful in the clinical assessment of wounds.

Hammond and Nixon 28 evaluated the SilhouetteMobile^® system (ARANZ Medical, Christchurch, New Zealand), which is a handheld device designed to measure wound surface area and depth. The system was found to be fast and easy to handle and had high intra‐ and inter‐rater reliability. In contrast, Miller et al. 29 demonstrated low reliability of the SilhouetteMobile^® (ARANZ Medical, Merivale, Christchurch, New Zealand) system in measuring small wounds. Sprigle et al. 30 found a handheld wound measurement device, WMD, based on smartphone technology to be accurate at different distances and angles and fairly accurate even at skewed angles. Haghpanah et al. 31 compared the VeV MD Vista Medical, Winnipeg, Manitoba, Canada (digital imaging) system with Visitrak^® and the simple ruler method. The VeV MD and Visitrak^® were equally accurate (with lowest root mean square error values) and more accurate than the simple ruler method.

Foltynski et al. 32 and Ladyzynski et al. 33 compared a new imaging device system TeleDiaFos^® (Nalecz Institute of Biocybernetics and Biomedical Engineering, Warsaw, Poland) to three reference methods: the elliptical method, the Visitrak^® system and the Silhouette Mobile^® device. The TeleDiaFos^® method is an imaging system that scans the foot and uploads the photo to a server to calculate the area after an operator has drawn the wound outline. Both studies found the TeleDiaFos^® imaging system to be more accurate than the reference methods, but the functionality of the method was limited to the sole of the foot. The same research group evaluated the software AreaMe^® (Institute of Biocybernetics and Biomedical Engineering, Warsaw, Poland) to calculate wound area from digital images. The software was more accurate and reliable than Visitrak^®, but less accurate and reliable than the Silhouette Mobile^® device 34. Santamaria et al. 35 and Flowers et al. 36 reported the software program AMWIS to be accurate and with a high inter‐rater reliability when measuring wound area from digital images (measurement error 1·28%, ICC 0·958). Furthermore, the software program was found to be useful in telemedical wound care.

In summary, digital imaging is a non‐contact method that is equally accurate and reliable as planimetric methods 26. However, it is also time‐consuming and can be affected by the illumination, location and size of the wound.

Wound volume measurements

Of the 13 papers evaluating methods to measure wound volume, 4 studies evaluated accuracy, 6 evaluated agreement, 8 evaluated reliability and 12 studies mentioned feasibility. Table 2 summarises the characteristics of these studies.

Table 2.

Wound volume measurement methods

Studies	Methods	Number and type of ulcers	Statistical methods	Outcomes	Feasibility
Albouy et al., France, 2007 50	1. 3D model including wound classification. 2. Volume measurement by filling wound cavity with water	One wound was measured by 16 pairs of 3D images, and by filling wound cavity with water 20 times	Accuracy: average error and standard deviation	Average error on the volume was 0·4% and the standard deviation was 3·19%	Feasibility was not evaluated
Barone et al., Italy, 2011 39	1. 3D thermoscanner 2. Digital planimetry	Seven leg ulcers	Accuracy: percentage difference between ground truth and wound areas obtained from thermoscanner and digital planimetry	Spread of differences was ±4% on colour images and ±7% on reference method; ±5% on thermal images and ±12% on reference method	The 3D thermoscanner was non‐invasive, expensive
Bowling et al., UK, 2011 51	3D digital optical system	Three wounds. One clinician measured the wounds ten times; four other clinicians measured each wound once.	Reliability: intra‐ and inter‐rater variation due to measurement process and to limited image accuracy	Mean intra‐ and inter‐rater variations due to measurement were 1·6% and 8·4%; due to limited accuracy were 1·9% and 2·7%.	The 3D system was non‐invasive, fast and could be used for remote assessment in telemedicine
Davis et al., USA, February 2013 41	1.SPG system, LifeViz® 3D system 2. Simple ruler method	13 pressure ulcers in 10 patients were measured repeatedly during a 6‐week period	Inter‐rater reliability: intra‐class correlation coefficient (ICC) Agreement: ANOVA to compare the differences between 3D measurements and the simple ruler method	ICC for volume measurement by SPG was 0·9867 (P < 0·001). ANOVA showed a significant difference only for width measurements (P < 0·0001). Surface area, depth and length were not significantly different.	LifeViz® was portable, easy to use, non‐invasive and required minimal training time. Moist wound base and undermined margins could affect the images. Wounds >6 cm in the greatest axis could not be captured.
Davis et al., USA, September 2013 46	1. Laser‐assisted wound measurement device (LAWN) 2. Digital imaging method using ImageJ software 3. Measurement of depth with a ruler 4. Dental paste	12 wounds on the dorsum of a pig	Agreement: student paired t‐test	4 of 12 area measurements with LAWN were statistically different from digital imaging (P < 0·05). All volume measurements with LAWN were artificially lower than digital imaging and dental paste method (P < 0·05)	The LAWN was not mobile in the clinic and was inaccurate for volume measurements
Gardner et al., USA, 2012 52	VeVMD, a digital image analysis system	34 plantar‐surface diabetic foot ulcers	Inter‐rater and intra‐rater reliability	Inter‐rater reliability was 0·745 and intra‐rater reliability was 0·868.	Easy to use, non‐invasive
Kecelj Leskovec et al., Slovenia, 2007 47	1. Laser‐based 3D measuring device 2. Computer planimetry with photography	15 venous leg ulcers in 8 patients	Agreement: The Bland–Altman analysis	Precision in volume 7·5%	The 3D measuring device was fast, small, easy to handle and a non‐contact method. Problems with determining ulcer edges and irregularity of skin around the ulcer.
Langemo et al., Canada, 2001 42	1. Kundin® Device 2. Stereophotogrammetry (SPG)	two wound models were measured twice using both methods by 24 clinicians	Accuracy: standard error of measurement (SEM) Intra‐rater reliability: Pearson correlation coefficient Inter‐rater reliability: Intra‐class correlation coefficient (ICC)	SEM values were 2·83 and 4·56 for Kundin® device and 1·68 and 1·72 for SPG. Pearson coefficients were 0·36 and 0·41 for Kundin® and 0·79 and 0·57 for SPG. ICC for both methods was identical at 0·98.	The SPG was easy to use, more expensive. The Kundin device® was inexpensive, easy to use. No pictorial display is available.
Plassmann and Jones, UK, 1998 38	1. MAVIS (Measurement of Area and Volume Instrument System) 2. Transparency method (manual planimetry) 3. Alginate casts	50 wounds	Agreement: Pearson correlation coefficient r Inter‐rater reliability: Standard deviation (SD) in percentage	For area measurements, r = 0·9446 between transparency tracing and MAVIS; for volume measurements, r = 0·9078 between alginate casts and MAVIS. MAVIS reduced SD by 3–5% in ulcer area compared with transparency tracings and by 5% in ulcer volume compared with alginate casts.	MAVIS was non‐invasive, fast and more accurate than existing methods
Romanelli et al., Italy, 2007 44	3D laser scanner, Derma project	15 patients with venous leg ulcers	Intra‐ and inter‐rater reliability: Intra‐class correlation coefficient (ICC)	ICCs were 0·9832 for intra‐rater reliability and 0·9714 for inter‐rater reliability	Complex to use, expensives
Schubert and Zander, Sweden, 1996 37	Saline gel injection into wound cavity; subsequent measurement of volume required to fill cavity	11 wounds measured twice on four occasions	Intra‐rater and inter‐rater reliability: Coefficient of variation (CV)	CV was 6% for intra‐rater reliability and 19% for inter‐rater reliability	The method was especially good in evaluating ulcers with small opening area or irregular base
Smith et al., USA, 1998 45	1. 3D laser imaging system 2. Tracing method: (manual planimetry) area and perimeter) 3. The alginate casts: volume	Six wound models were measured five times by each method	Agreement: difference in area and volume measurement by the 3D laser system compared with the tracing method and alginate casts analysed by ANOVA	The area was 2·02 ±1·30 cm2 greater for the 3D system than for manual planimetry (P < 0·05). The volume was 1·04 ±0·61 cm3 greater for the 3D system than for alginate casts (P < 0·05).	Non‐contact, expensive

The early volume measurement methods included saline gel injection into a wound cavity. The method is inexpensive and easy to handle but carries the risk of wound contamination. Schubert and Zander 37 reported coefficients of variation of 6% for intra‐rater reliability and 19% for inter‐rater reliability.

During the last decade, various 3D techniques for measuring wound volume have been proposed. Some of these methods assess wound size as well as wound characteristics. Plassmann and Jones 38 described a non‐invasive method using colour‐coded structured light represented by MAVIS (Measurement of Area and Volume Instrument System (University of Glamorgan, Department of Computer Studies, Pontypridd, Mid Glamorgan, UK)). The wound area is lit by parallel light beams at an angle of 45° and a camera above the wound provides serial photos for use in a mathematical algorithm representing the data as a 3D model. MAVIS was compared with transparency tracing (manual planimetry) and alginate casts and was found to be more accurate in measuring both wound area and wound volume. Barone et al. 39 described a 3D optical scanner based on structured light integrated with a thermal imager. Besides measuring wound size, the camera can detect inflammation using thermal imaging. It was found to be useful in detecting the wound boundary, but the thermal imager was expensive.

The high‐resolution digital cameras of SPG systems including MAVIS II 40 is equipped with a specialised lens enabling two digital wound images to be captured simultaneously at different angles. The data are transferred to a software application that creates a 3D reconstruction of the wound. A 2013 study by Davis et al. 41 evaluated the LifeViz^® 3D system (QuantifiCare, San Mateo, CA) and found a high inter‐rater reliability for volume measurements (ICC = 0·9867; P < 0·001). When compared with the simple ruler method, only measurements of width showed a significant difference (P < 0·0001) while surface area, depth and length values were similar. Langemo et al. 42 compared an SPG system with the Kundin^® device (Pacific Technologies and Development Corporation, San Mateo, CA), which is a ruler‐based device using three disposable paper rulers set at orthogonal angles to measure length, width and depth of the wound. Wound volume was calculated by the formula L × W × D × 0·327. The SPG system was more accurate (smallest standard error of measurement) than the Kundin^® device, but the inter‐rater reliability was similar for the two methods (ICC 0·98).

Laser scanners such as the Derma project 43 have also been used in 3D set‐ups. In the Derma project, a 3D laser scanner was developed based on the Minolta Vivid 910^® scanner (Konica Minolta, Osaka, Japan) to measure wound size and classify tissue 43, 44. The ICC was 0·9832 for intra‐rater reliability and 0·9714 for inter‐rater reliability. Smith et al. 45 described and evaluated a system in which a laser line generator illuminates the wound and two video cameras capture an image, which is then reconstructed into a 3D model by computer software. This laser system overestimated the wound area and perimeter, but the volume measurements were not significantly different from those using the reference method, alginate casts. Davis et al. 46 also evaluated a laser‐assisted wound measuring device (LAWN) (Silhouette Star^®, Aranz medical, Christchurch, New Zealand), but found that this 3D device underestimated depth and volume (P < 0·05).

Kecelj Leskovec et al. 47 evaluated a handheld laser‐based 3D measuring device consisting of a laser projector and a digital camera. The laser projector illuminates the wound with light planes and the camera records the wound from different angles. The image is transferred to a computer and reconstructed into a 3D image. The system was found to be fast, small, easy to handle and non‐invasive, but determining wound edges and skin irregularities around a wound presented a challenge. Zvietcovich et al. 48 found that a 3D laser scanner had higher accuracy and reliability in measuring wound volume than a method using gel injection into the wound cavity.

Wannous et al. 49 and Albouy et al. 50 tested a 3D colour imaging method for measuring surface area and volume and classifying wound tissues (e.g. granulation, slough, necrosis). The system was accurate, with an average error on the volume measurement of 0·4% and standard deviation of 3·19%. Bowling et al. 51 described a novel optical system (Eukona^®, Fuel 3D technologies, Oxford, UK) to create 3D photos of diabetic foot ulcers and reported it to be simple and easy to use with potential for use in a telemedical set‐up. When three wounds were measured by five clinicians, the average intra‐ and inter‐rater variations were 1·6% and 8·4%, respectively, due to measurement process and 1·9% and 2·7%, respectively, due to 3D image process. However, the system gave limited details about the wound base and the study had a small sample size.

Gardner et al. 52 evaluated the VeVMD^® system (Vista Medical, Winnipeg, Manitoba, Canada) that measures wound volume using computerised analysis of digital images. On measuring the volume of 33 diabetic foot ulcers, they found an inter‐rater reliability of 0·745 and an intra‐rater reliability of 0·868. However, when four wounds were excluded from the analysis because of difficulties in assessing the wound margins, both inter‐ and intra‐rater reliability improved substantially (0·970 and 0·981, respectively).

In summary, the volume measurement techniques are often immobile, expensive and complex to use and can be inaccurate and unreliable.

Discussion

Several methods for measuring area and volume of wounds have been described in the literature in recent years. When comparing treatment strategies, it is essential that the measurement methods are accurate, reliable and feasible for correct application in routine clinical work and clinical research studies.

One of the factors affecting the accuracy of wound measurement is definition of the wound boundary, which is often difficult to identify. Slight movements can change the appearance of a wound, and localisation on a curved part of the body (e.g. the heel) can make it difficult to correctly estimate wound size. Some wounds are also extensively undermined, making the volume difficult to assess. Finally, wounds situated in areas with a thick layer of soft tissue can pose problems because of physiological contraction or fibrotic scar formation 24, 38.

Wound measurement strategies have traditionally focused on two‐dimensional methods. The simple ruler method often overestimates the wound area and is inappropriate in large wounds with irregular boundaries, but it is fast, easy to use and inexpensive 7, 8. The elliptical method is also easy to use and inexpensive and is useful for measuring ellipse‐shaped wounds 9, 10, 11.

The planimetric method is relatively easy to learn and is an accurate and reliable method that considers body curvature 17. Digital planimetry is slightly more accurate and reliable compared to the manual method 14, 15, but both methods involve contact with the wound and thus carries the risk of wound contamination. The method can also be time‐consuming 24.

The digital imaging method is equally accurate and reliable compared to the planimetric approach 26 and is a non‐contact method, thus eliminating the risk of wound contamination. Digital images can be affected by the illumination, location and size of the wound, and variations in camera angle can lead to underestimation of the wound area 23. The method is also time‐consuming from the instant the image of the wound is captured by the camera until the wound area is estimated by the software.

The methodological quality of the reviewed studies varied. None of the studies mentioned statistical reasons for the choice of sample size, so it was not possible to assess whether sample size was sufficient. In most of the studies evaluating methods for wound area measurement, the number of wounds ranged between 3 and 108, and there was only one larger study including 260 patients. The number of wounds in studies evaluating methods for volume measurement ranged between 1 and 50, but 10 studies included 15 wounds or fewer. When the inter‐rater reliability was estimated, the number of clinicians involved was not always reported. Furthermore, the time interval between intra‐rater measurements was often omitted 10, 15, 19, 21, 28, 32.

Although measurement methods overestimate or underestimate wound area or volume, they might still be useful in the assessment of wound healing. Measurements of wound size need to be reliable, but not necessarily accurate, in the assessment of the healing rate 53. According to Flanagan, the percentage change in wound area is useful in predicting the wound healing 5.

Linear advancement of wound perimeter has been proposed to be a better parameter of wound healing because it is independent of initial wound size and geometry. The method is thereby better at dealing with wounds of varying size 54. However, Santamaria et al argue the opposite that the use of surface area change per day is a more valid method to calculate wound healing 53.

Various 3D techniques for wound measurement have been developed in recent years, which has allowed evaluation of the wound healing process from a volume perspective. Some of these methods are also able to assess wound characteristics and detect inflammation using thermal imaging 39, 49. None of the reviewed technologies have yet had a major impact because of low accuracy, high cost and complexity in handling the system set‐up. The studies typically focus only slightly on feasibility, which is an important element for evaluating the equipment, the demands for training and whether the approach can be disseminated to other clinical set‐ups. Furthermore, the testing of 3D methodologies has not been rigorous. None of the studies have validated the methods sufficiently and most studies have only compared the 3D techniques with two‐dimensional approaches. Thus, it is still uncertain whether the methods are accurate in measuring volume. This is compounded by the absence of a gold standard method for wound area and volume measurement, against which other methods can be tested.

The measurement methods described here have only been tested on a few different types of wounds, and it is therefore not possible to determine whether the methods are suitable for all types of wounds. Only a few of the studies provide data on selection criteria of wounds, including size and location, thus reducing the generalisability of the results.

A major strength of this systematic review is the fact that we have performed a thorough search of the literature for studies evaluating wound measurement methods used during the last 20 years. However, the large heterogeneity in the study methodology has made the interpretation difficult.

Within the last 5 years, telemedicine has been adopted in a number of clinical settings worldwide and this area is expected to play an increasing role in the diagnosis, monitoring and treatment of wounds of varying aetiology. Future wound assessment technologies should therefore be applicable and feasible in a telemedical setting.

Conclusion

Wound measurement methods need to be accurate, reliable and feasible if they are to be useful in evaluating the wound healing process. From the papers included in this review, we found that methods to measure wound area are better validated than methods to measure wound volume. We recommend the use of digital planimetry or digital imaging if a high accuracy is required (e.g. for a clinical trial) in wound area measurements, but the methods can also be time‐consuming for everyday clinical practice. When requiring a fast method, we recommend using a simple area measurement method for circumferential wounds.

Methods that measure wound volume have the potential to assess wound healing using all the dimensions of a wound and are therefore useful for large and deep wounds. However, volume measurement techniques are not yet sufficiently validated, and further studies are needed to confirm their accuracy and reliability.