Deep Learning-Based Assessment and Sophisticated Quantification for Wound Healing and Skin Metrics

The staggering cost of healthcare in America continues to rise, especially for wound healing. In 2014, wound-related healthcare expenses for Medicare beneficiaries were conservatively estimated at $28.1 billion, with over 8 million patients reporting at least one type of wound or infection.1 Improved methods that analyze the complex stages of wound healing have potential to improve patient outcomes, and subsequently, reduce the cost of care. 

Biomaterials have proven their efficacy for wound treatment.  For example, dressings are often supplemented with agents to encourage cell migration and enhance healing.2 Likewise, skin substitutes effectively reduce scar tissue3, and hydrogels can significantly speed healing, compared to traditional intervention.4 

The healing process consists of four phases: hemostasis, inflammation, proliferation, and remodeling (Figure 1). Hemostasis marks the initial response in protecting the skin’s vascular system. Coagulation then sets the stage for the subsequent inflammatory response, followed by proliferation of the epithelial and extracellular matrix (ECM) components. Late proliferation overlaps with the inflammatory phase, leading to re-epithelialization as ECM forms. Remodeling occurs within weeks and can last from months to years; during this dynamic process, a highly organized collagen matrix reconstructs the site, resulting in scar formation.

Figure 1: Schematic overview of the healing process.

Histology endures as the definitive standard to assess wound healing. Histological evaluation often comprises multiple components of the healing process: angiogenesis, inflammation, restoration of connective tissue matrix, wound contraction, remodeling, and re-epithelialization (Table 1). Routine quantification methods include immunohistochemistry (IHC) and traditional light microscopy. Considering the current lack of a standard scoring system, the use of more sophisticated tools may enhance insight into the wound healing process and facilitate a better understanding of wound therapies.

Table 1: Histological skin cell parameters for the assessment of wound healing.5

Reveal Biosciences, a data-powered pathology lab, develops AI-driven digital assays for quantified, histological outcomes. A full-service histology lab at Reveal offers processing, sectioning, and staining biomaterials, including hydrogels, organoids, and spheroids. Our computational approach takes advantage of layered, deep-learning platforms and real data sets to help researchers assess a wide array of  diseases. Each digital assay incorporates rigorous quality controls (QC) that detect and exclude debris, tissue and staining artifacts, and adjust for color variations due to different technical procedures, staining processes, and individual human approaches. 

Analysis of the normal healing process often requires insights derived not only from the immediate site of injury, but also the surrounding tissue. Accordingly, our digital assay portfolio provides enhanced outcomes for skin and wound healing metrics. For example, Reveal Biosciences’ cellular quantification tools may be implemented in proximity to the region of interest, enhancing the view of cellular infiltration within the context of a  specific target location. For complex environments, these enhanced tools yield quantitative outputs including biocompatibility, cellular infiltration, and scar formation.

Biocompatibility and cellular infiltration

The introduction of a foreign material, whether a wound dressing or an implant, will initiate a cascade of physiological responses. Analysing the extent of these mechanisms is complex. For example, to assess migrating cells to the site of interest, the appropriate immune and tissue markers must be identified and reflect the relevancy of the model. Multi-plex IHC or ISH are commonly-used histological approaches to determine cell-by-cell biomarker positivity and intensity. Sophisticated approaches are required for complex analysis in environments containing many cell types or highly dense cellular regions. Our team has developed a nuclear segmentation digital assay that increases accuracy over traditional image analysis techniques. This approach takes into account cellular polarity, cellular clustering, and complex animal models through layers of machine learning tools.

ECM deposition and remodeling

The healing response and remodeling comprise imperative outcomes when assessing the efficacy of skin substitutes and medical devices. Inherent complexities plague colorimetric approaches to distinguish between remodeling ECM and scar tissue. Limitations also frustrate the analysis of tissue-biomaterial interactions on a 2D platform. Developed from extensive experience with a variety of collagen stains, our digital assays accurately distinguish between collagen and scarring. We have also performed 3D tissue analysis through serial sections: each are quantified and analysed, then reconstructed to provide an enhanced view of the tissue and biomaterial integration.

Routine quantification methods are often limited in complex environments, like wound healing and other tissue-biomaterial interactions. Reveal Biosciences provides solutions from any stage of the workflow, from processing and sectioning to customized image analysis. Contact us at [email protected] to speak with our scientific team and learn more about how our services could benefit your research.


  1. Nussbaum, S. R. et al. An Economic Evaluation of the Impact, Cost, and Medicare Policy Implications of Chronic Nonhealing Wounds. Value Health 21, 27–32 (2018).
  2. Piraino, F. & Selimović, Š. A Current View of Functional Biomaterials for Wound Care, Molecular and Cellular Therapies. BioMed Research International vol. 2015 1–10 (2015).
  3. Greaves, N. S. et al. Skin substitute-assisted repair shows reduced dermal fibrosis in acute human wounds validated simultaneously by histology and optical coherence tomography. Wound Repair Regen. 23, 483–494 (2015).
  4. Yuan, L. et al. Study of the use of recombinant human granulocyte-macrophage colony-stimulating factor hydrogel externally to treat residual wounds of extensive deep partial-thickness burn. Burns 41, 1086–1091 (2015).5. Gupta, A. & Kumar, P. Assessment of the histological state of the healing wound. Plastic and Aesthetic Research vol. 2 239 (2015).
  5. Gupta, A. & Kumar, P. Assessment of the histological state of the healing wound. Plastic and Aesthetic Research vol. 2 239 (2015).

Pathology Revealed

Sign up to receive updates on the latest AI-powered pathology breakthroughs, access digital pathology resources, and more.