Clinical study
Personalized prediction of chronic wound healing: An exponential mixed effects model using stereophotogrammetric measurement

https://doi.org/10.1016/j.jtv.2014.04.001Get rights and content

Abstract

Study aim

Stereophotogrammetric digital imaging enables rapid and accurate detailed 3D wound monitoring. This rich data source was used to develop a statistically validated model to provide personalized predictive healing information for chronic wounds.

Materials

147 valid wound images were obtained from a sample of 13 category III/IV pressure ulcers from 10 individuals with spinal cord injury.

Methods

Statistical comparison of several models indicated the best fit for the clinical data was a personalized mixed-effects exponential model (pMEE), with initial wound size and time as predictors and observed wound size as the response variable. Random effects capture personalized differences.

Results

Other models are only valid when wound size constantly decreases. This is often not achieved for clinical wounds. Our model accommodates this reality. Two criteria to determine effective healing time outcomes are proposed: r-fold wound size reduction time, tr-fold, is defined as the time when wound size reduces to 1/r of initial size. tδ is defined as the time when the rate of the wound healing/size change reduces to a predetermined threshold δ < 0. Healing rate differs from patient to patient. Model development and validation indicates that accurate monitoring of wound geometry can adaptively predict healing progression and that larger wounds heal more rapidly. Accuracy of the prediction curve in the current model improves with each additional evaluation.

Conclusion

Routine assessment of wounds using detailed stereophotogrammetric imaging can provide personalized predictions of wound healing time. Application of a valid model will help the clinical team to determine wound management care pathways.

Introduction

The chronic wound that fails to heal presents a significant burden for both the affected individual and for all those who care for them, from the caregiver to the healthcare system. The goal of clinical management is to promote healing, yet the progress can be slow and intermittent. This makes it very difficult for even the expert wound care practitioner to predict time to healing. This is disheartening not just for the patient but also for the clinician who may feel a need to change therapeutic interventions if healing rates do not appear to improve. Healthcare systems also have to plan care for patients requiring indefinite periods of management.

The multi-stage pathway for acute wound healing is well known, comprising three main phases: 1) Initial haemostatis followed by inflammation, 2) proliferation and closure and 3) maturation and remodeling [1]. The healing process in biological systems has been modeled using a Gompertz-like function to follow an exponential course [2]. This presumes that the rate of change of wound size progressively decreases as the residual wound area approaches total closure. The final closure is very slow, with minute changes in wound size changes occurring over long periods of time.

When considering the chronic wound in the clinical setting, the normal pathway is disrupted and these models do not fit reliably. A major challenge with implementing models based on clinical measurement data is that wound measurement has traditionally been highly unreliable, almost subjective. [3] Accurate wound measurement is essential to provide clinically valid measures. Changes in wound geometry indicate whether a wound is getting smaller, i.e. healing, static or increasing in size. Recent developments in stereophotogrammetric digital imaging have provided readily available technology that can rapidly and accurately obtain a detailed 3D wound image at the bedside. Extensive validation of the LifeViz system using both a static imaging phantom and a dynamic imaging phantom adjustable for different volume quantities has shown the system provides a precise objective dimensional images [4]. We have previously shown that this approach has utility and good inter-observer reliability in the clinical in-patient setting [5].

In the current study we explored the feasibility of applying this rich data source to the development of individualized predictive curves using a statistically validated mixed-effects exponential model to provide predictive information on the healing pathway for chronic wounds.

Section snippets

Study protocol

Wound images were obtained from a sample of 13 category III/IV pressure ulcers (PUs) from 10 individuals with spinal cord injury (SCI), three patients in the study cohort had two PUs. Primary inclusion criteria were that patients were ≥18 years of age, but were not scheduled for surgical wound closure or sharp debridement. The LifeViz 3D system (Quantificare Inc.) was used to obtain wound images weekly over a consecutive 6-week period. The data collection methodology is described in detail in

Predictive model development

A total of 147 valid images were obtained for model development using DermaPix output variables to determine wound geometric properties over time.

Exploratory data analysis (EDA) was first applied to the raw data. The results demonstrated that more severe wounds had steeper healing trajectories than smaller wounds. Six of 13 wounds showed a slightly deteriorating trend for all wound geometry variables in the first one to two weeks before showing steady improvements. This suggests a two-phase

Discussion

We have previously shown that advanced wound assessment using stereophotogrammetric imaging has the potential to provide the clinician with an accurate representation of the wound at the time of assessment. Our preliminary model development now shows that accurate wound geometric data can also be used to predict the wound healing pathway on a personalized iterative basis.

Most current wound healing models utilize complete wound closure as the endpoint and apply analysis similar to survival

Conclusion

Preliminary results suggest that the routine assessment of wounds using detailed stereophotogrammetric imaging has the potential to provide personalized predictive indications of wound healing. The proposed mixed effect exponential model compares well with the existing models. Application of a valid predictive wound healing model will help the clinical team to determine treatment plan efficacy, selection of advanced wound therapeutics and even discharge planning.

Conflict of interest

The authors have no commercial relationships which may lead to a conflict of interests with regard to the information presented in this paper.

Acknowledgments

The authors wish to acknowledge the support of personnel and veterans at the Louis Stokes Cleveland Department of Veterans Affairs Medical Center, Cleveland, Ohio. These individuals provided invaluable assistance in obtaining and analyzing the images used in developing the predictive model.

References (11)

  • S.R. Goldberg et al.

    Wound healing primer

    Surg Clin North Am

    (2010 Dec)
  • S. Wallenstein et al.

    Statistical analysis of wound-healing rates for pressure ulcers

    Am J Surg

    (2004 Jul)
  • S.M. O'Meara et al.

    A systematic review of the performance of instruments designed to measure the dimensions of pressure ulcers

    Wound Repair Regen

    (2012 May–Jun)
  • H. Skvara et al.

    Quantification of skin lesions with a 3D stereovision camera system: validation and clinical applications

    Skin Res Technol

    (2013 Feb)
  • A.J. Davis et al.

    Repeatability and clinical utility in stereophotogrammetric measurements of wounds

    J Wound Care

    (2013 Feb)
There are more references available in the full text version of this article.

Cited by (14)

  • Smart technologies in wound prevention and care

    2019, Innovations and Emerging Technologies in Wound Care
  • Evaluating accuracy and reliability of active stereophotogrammetry using MAVIS III Wound Camera for three-dimensional assessment of hypertrophic scars

    2017, Burns
    Citation Excerpt :

    An ideal measurement tool for assessing hypertrophic burn scars would be one that is valid, reliable, sensitive to change over time, and is capable of determining scar characteristics such as its volume, area, pigmentation, and circumference. The principle of stereo-photogrammetry to generate 3-D reconstructions of scar tissue, and subsequently, determine its volume, has recently gained considerable attention [10,15–17,22]. Stereophotogrammetry works by capturing two images of a target object from different angles [10].

  • Degradation science: Mesoscopic evolution and temporal analytics of photovoltaic energy materials

    2015, Current Opinion in Solid State and Materials Science
    Citation Excerpt :

    These measures are also analogous to descriptors used in image analyses. Individualized longitudinal predictive models that can be updated over time [142] will be extremely important for critical system monitoring. The software packages for effective change point analyses that are applicable to degradation science and monitor system transitions are needed [143,144].

View all citing articles on Scopus
View full text