Histogram Planimetry Method for the Measurement of Irregular Wounds

Introduction. Irregularly shaped wounds or flap borders usually require specified software or devices to measure their area and follow-up wound healing. In this study, an easy way of area measurement called histogram planimetry (HP) for wounds with irregular geometric shapes is defined and compared to conventional millimetric wound measurement. Materials and Methods. Ten irregularly bordered geometric shapes were measured by 4 different individuals working as surgical assistants using both HP and manual millimetric measurement tools. The amount of time for each wound shape calculation as well as  the measurements of the wound areas were noted. All measurements were compared for each method and between each individual using the Wilcoxon signed-rank test. Results. There was no statistically significant difference between 2 measurement methods by means of measured areas; however, measurement time was significantly lower when the HP method was used. There also was no significant difference between the individuals’ measurements and calculation times. These results indicated that HP is useful as a conventional millimetric square wound measurement technique with significantly lower measurement times. Conclusion. Due to the development of photo-editor software technologies, measurements in the surgical field have become more accurate and rapid than conventional manual methods without consuming the time and energy needed for other studies. A future study including comparisons between the presented method and complex computerized measurement methods, in terms of duration and accuracy, may provide additional supportive data for the authors’ method.

Among numerous experimental and clinical wound healing studies in surgical practice, decrease in wound size is one of the most important parameters during the follow-up period. Chronic wounds need periodic measurement of wound size to assess the progression of healing. Flap monitorization in both clinical practice and experimental flap survival studies also requires measurement of necrotic and viable tissue areas or reepithelilaization rates for analytic purposes. The most challenging part of these measurements is the irregular borders of the wound or flap shapes that make precise measurement difficult. 

One of the conventional methods to overcome this problem is transferring the irregular shape by tracing it onto acetate paper and counting millimetric squares from scaled paper that has been inserted under the traced shape.¹ However, although more precise measurements may be achieved by this acetate tracing method, it is tedious and time consuming in larger experimental or clinical studies. 

Other solutions including the multiplication of maximal dimensions of such irregular shapes and comparison of all measurements,² computerized wound tracing,³ laser (including 3-dimensional [3D] wound measurement),⁴ and computerized color analysis of the wound and a perimetric assessment⁵ have also been reported. 

In this study, a simple technology which includes both digital imaging and a photo-editor program is presented. A histogram is a pixel distribution graph that allows the user to identify the color density of any selected area of a photograph. Histogram planimetry (HP) is more accessible and less expensive than automated analysis software programs, and it is based on the pixel count of a selected irregular area which is divided by the pixel count of 1 cm² to find a result in terms of cm² or mm².

Four raters without any visual impairments measured the areas of 10 irregular shapes of various sizes under standard room light. The areas of the shapes ranged between 2 cm and 35 cm, which the authorsdetermined before the study. The raters were given 3-minute intervals between 2  shapes, and 15-minute intervals between 2 methods as a rest period.  Area measurements and measurement time of each shape and the total time of measurement with each method was noted. 

Histogram planimetry method
The raters were chosen from different medical disciplines and were basic computer users with some experience with Microsoft Windows. None of the raters had any experience with the photo-editor software used in this study. Thus, all the raters received a short briefing about the basic use of the program including opening a figure, selecting an area, painting the area, counting the pixels, and measuring the area as cm². The same training was also performed for a manual counting method.

An A4-sized acetate paper was used to draw irregular shapes mimicking the wound healing process which were then scanned and saved to a computer. After the scanning procedure, the picture was opened in Adobe Photoshop CS6 (Adobe, San Jose, CA). There are 3 absolute color tones of the program: red, blue, and green. By changing the numeric tone parameters in the color box, 1 of these colors is obtained. For example, to provide absolute red, the numeric color parameter of red color was changed to 255 while the parameters of blue and green were changed to 0. The user interface of the program and the basics of the HP method are shown in Figure 1. Following the color arrangement, the “magic wand tool” was used to select the borders of the irregular area, and the shape was filled by the selected pure color tone by clicking the mouse button until the vertical histogram line of the selected color was stabilized at the right side of the graph. By moving the cursor on this line, the pixel count of the irregularly bordered shape can be noted from the box over the histogram graph. 

The same procedure is performed for the 1 cm² box. The pixel count of the irregular area was divided by the pixel count of 1 cm² to find the area of the irregularly bordered shape as cm². 

Manual counting
Each of the raters also counted millimetric squares by using a loupe with 2.5x magnification. Borders of the shapes were included in the counting, and partial squares smaller than a half square were counted as half squares while partial squares larger than a half square were counted as 1 square. The area of the irregularly bordered shape is noted as mm². 

Statistical analysis
In order to compare differences in duration and measurement, 2 different methods were used. The Wilcoxon signed-rank test was chosen due to the paired nature of the observations (ie, both methods were employed on the same set of 10 areas, and there was a limited number of observations per participant. Confidence intervals were computed using critical values from a t-distribution at 95% confidence interval as the sample size was small. 

The Kruskal-Wallis test, which is a nonparametric test, was also used to show whether participant characteristics bias the area measured or the duration of the tests which also indicates the inter-rater reliability.

The duration of computing the measurement of each shape using the HP method was significantly shorter than the manual counting (MC) method (P < 0.001). The average duration of measurement for the MC method was 14.32 ± 1.17 minutes, while the same value for the HP method was 2.14 ± 0.17 minutes. For each rater, the difference in duration was independent of any characteristics of the participants that might influence the results, according to the Wilcoxon’s signed- rank test (Table 1). In addition, there was no significant difference in the area measured by the MC and HP methods (P > 0.05) (Table 2). 

In the Kruskal-Wallis test, there were no significant differences in the area measured or the duration taken to employ the MC and HP methods (P > 0.05) (Table 3). 

Wound healing is a 3D process that includes both macroscopic and micromolecular aspects. Woodbury et al⁶ suggested a useful measurement tool not only measures dimensions but can also provide a multifactorial assessment of such wounds. They suggested periodic measurement of the area of the wound which may be helpful during the follow-up period with other parameters that may affect healing such as comorbidities of the patient, the presence and characteristics of the exudate in the wound bed, characteristics of necrotic tissue and the wound itself, the structure of periulcer skin, and presence and course of any edema. 

Computerization of such a measurement became essential for the investigators to save both energy and time. However, there are some doubts as to the accuracy with the computerization of the data, as it may cause overestimation or underestimation. Thawer et al⁷ compared computer-assisted and manual wound measurement methods in both animal and human wounds and noted that both methods were reliable, but the computerized method was significantly faster. In the present study, there was also no significant difference between the MC and HP methods in terms of the measurements. Lagan and colleagues⁸ reported a significantly larger degree of variability in the manual planimetry method (thus, less repeatability) than when using digitail planimetry.

Spectrophotogrammetry is another computerized method used for area measurement. This method is based on a specialized device and software in which an operator traces the wound perimeter by eye using the mouse, and the data collected from this tracing is used for the calculation of the area.^9-11 The tracing is optimized by inserting a gridded plane near the wound to eliminate the depth factor of the photo. A major drawback of this system is its cost, which was obtained from 3 different manufacturers for devices mentioned in the literature,^4,5,9,12and ranged from $1600 to $5000. The HP method does not have such a cost and there is no depth factor as in photographic assessment because the acetate paper tracings are scanned to work on a 1:1 scale tracing. The technique examined in this paper uses a 1cm² square drawn on acetate paper to compare and measure pixel distribution to the irregular shape to be measured. A similar comparison is used in the spectrophotogrammetric or computerized planimetry methods to correct the parallax error due to the photographic distortions caused by the cameras.¹

Digital planimetry methods based on different camera and laser technologies was reported to be more useful than the MC method, because they provide unwrapped and flattened images to better calculate the true area of the wound. However, the technologies were usually compared to a technique that overestimated the wound area based on maximal length and width multiplication.¹³ In the present study, the control group was the MC method, which is a more reliable method with minimal overestimation and depends on counting millimetric squares after transferring the shape over a millimetric paper by tracing.¹⁴ Adding the incomplete squares to the measurement provides more accurate results with less underestimation.¹⁵

Other technologies, such as videometers,⁵ which contain a video or digital camera to take color photographs from the wound bed and a software program that creates a color analysis of the wound and a perimetric assessment,⁵ are also in use. With these methods, the intra-rater reliability of the device changes among studies and area calculation results of the same shape may differ in each repeated measurement. 

The authors of the current study used professional photo editing software. The authors tried several other photo editors, but the chosen program seemed to have more sensitive selection and marking tools. Another useful software for wound measurement is ImageJ, which is a freeware program providing pixel-based measurement tools. Chang and colleagues³ used a square ruler near the shape to be calculated and a special camera with special lens placed parallel to the wound surface to achieve 1:1 dimension of the wound without distortion. After taking the photos, the image data were processed using the freeware program; however, curved body contours that carry the wound, circumferential wounds, or tapering the limbs decrease the accuracy of this method. As a way to prevent this problem, the authors herein used acetate tracings which contact all of the borders of the wound on the curved surfaces as stated in individual studies and these tracings were scanned without any depth in the image.^12,16 Also, a 1- cm² square was drawn to compare pixel density of the irregular shape. 

One limitation of the study is the absence of intra-rater reliability analysis; however, repeated measurements of the same shape seemed unnecessary to the authors because of the insignificant difference of area measurements between the HP and MC methods for the same shape. Thus, inter-rater reliability, which was presented in Table 3, shows no difference among the raters for each shape. 

As in all methods or technologies based on acetate tracings, another factor that influences the area measurement is the variability of the tracing among the investigators. It is important to clearly identify or decide the wound edges before tracing. This means over- or underestimation is possible in the wounds that have ill-defined borders. In more automatized systems, the resultant tracing depends on the decisions of the software or the distribution of laser beams which may lead to poorly estimated results.¹²

Using software programs makes expensive specialized devices unnecessary for the investigators (Table 4). However, a calculator or an automatized mathematical processor such as Microsoft Excel (Microsoft Corp,  Redmond, WA) is needed for more rapid measurement and documentation. The authors of the current study also use software which is a combination of pixel measurement and calculation of the areas to provide investigators a rapid, free online analysis of area measurements. 

Digital single-lens reflex professional cameras, commonly used in clinical photography, contain histogram graphs which provide the user simultaneous color pixel analysis of the area in the frame and may be used in the area measurement. However the color tone varies at the different points of the wound, a mean value of pixel distribution is available in most of the cameras to measure any selected area. 

With the advancement of wound healing strategies and increased investigations on specialized treatment systems, rapid and serial measurement of irregular areas has become essential in larger series of subjects. Evaluating factors such as result estimation, cost effectiveness, and the time and energy of the researchers results in investigators choosing computerized and automatized methods to measure wound healing. The HP method provides reliable and fast measurement of irregularly shaped wound areas with low cost and without the use of an expensive specialized device.

Affiliations: Dr Lütfi Kırdar Kartal Training and Research Hospital, Istanbul, Turkey; Bezmialem Vakıf University School of Medicine, Fatih, Istanbul; Bulent Ecevit University School of Medicine, Zonguldak, Turkey; and Akdeniz University School of Medicine, Antalya, Turkey

Correspondence:
Hakan Sirinoglu, MD
Dr Lütfi Kırdar Kartal Training and Research Hospital
Istanbul, Turkey
drhakansirinoglu@gmail.com

Disclosure: The authors disclose no financial or other conflicts of interest.