Computerised analysis of standardised ultrasonographic images to monitor the repair of surgically created core lesions in equine superficial digital flexor tendons following treatment with intratendinous platelet rich plasma or placebo

https://doi.org/10.1016/j.tvjl.2009.10.014Get rights and content

Abstract

The effectiveness of new therapies to treat tendon injuries is difficult to determine and is often based on semi-quantitative methods, such as grey level analysis of ultrasonographic images or subjective pain scores. The alternatives are costly and long-lasting end-stage studies using experimental animals. In this study, a method of ultrasonographic tissue characterisation (UTC), using mathematical analysis of contiguous transverse ultrasonographic images, was used for intra-vital monitoring of the healing trajectory of standardised tendon lesions treated with platelet rich plasma (PRP) or placebo.

Using UTC it was possible to detect significant differences between the groups in the various phases of repair. At end stage, over 80% of pixels showed correct alignment in the PRP group, compared with just over 60% in the placebo group (P < 0.05). UTC also showed significant differences in the course of the healing process between PRP treated and placebo treated animals throughout the experiment. It was concluded that computerised analysis of ultrasonographic images is an excellent tool for objective longitudinal monitoring of the effects of treatments for superficial digital flexor tendon lesions in horses.

Introduction

Tendon injuries are among the most important contributors to wastage in performance horses (Rossdale et al., 1985, Williams et al., 2001) and are therefore a significant economic and welfare issue in the equine industry. The repair process of traumatised tendon tissue basically follows the normal pattern of wound healing, which can be divided into three over-lapping phases: (1) the acute inflammatory/demarcation phase, (2) the proliferation phase and (3) the maturation/remodelling phase. During the inflammatory phase (ca. 10 days post-injury), disrupted tendon fibres are digested by proteolytic enzymes and then removed by phagocytosis. The proliferation phase (ca. 4–45 days post-injury) is characterised by the formation of a fibroproliferative callus. During the maturation or remodelling phase, from 45 days onwards, there is progressive cross-linking and organisation of the collagen fibrils into tendon bundles (Silver et al., 1983, Stashak, 1998). Maturation of the repair tissue, with collagen bundles regaining alignment along the lines of tension, may take up to a year or even longer (Sharma and Maffulli, 2005, Dahlgren, 2007).

Many new therapies have been introduced over the years for the treatment of tendon disorders, although objective and conclusive evidence to support their clinical use can be hard to obtain. In vitro studies are often used and are valuable in predicting potential treatment effects, but these cannot always be directly translated to the living horse. End-stage studies in experimental animals are used to analyse biochemical, biomechanical and histological features of the repair tissue, but such experiments are time-consuming and costly, require the sacrifice of considerable amounts of experimental animals and do not provide information about the different phases of healing. A technique for the objective, non-invasive evaluation of new and existing therapies throughout the healing process would be of great scientific value. Such monitoring of the healing process might lead to earlier conclusions about the effectiveness of treatments and, through insight in the longitudinal sequence of events, might also reveal some of the mechanisms of action of specific treatments.

In clinical trials, evaluation is usually based on qualitative or (semi-) quantitative ultrasonography (Marr et al., 1993a, Marr et al., 1993b, Reef, 2001, Waselau et al., 2008). The echogenicity of tendons is based on the density and highly structured arrangement of the collagenous matrix. Non-injured tendons show a regular echo-pattern, generated at the interface between tendon bundles and endotenon (Martinoli et al., 1993, van Schie and Bakker, 2000). Tendon injuries, characterised by disintegration of tendon bundles, lead to less intense and more irregular echoes, which are visualised as hypoechoic areas (Marr et al., 1993b).

Ultrasonography is far from straightforward, however, as it is a real-time and operator-dependent technique. Semi-quantitative scoring of ultrasonographic (US) images has been practised (Genovese et al., 1990, Palmer et al., 1994) and various methods to quantify grey levels in US images have been developed (Nicoll et al., 1992, Tsukiyama et al., 1996). It has been shown that quantification of grey levels is insufficient to accurately assess repair of tendon injuries (van Schie et al., 1999) and that grey level statistics are not suited for discrimination of the healing stage of the lesion because disintegration of tendon bundles is a three-dimensional (3D) phenomenon that cannot be captured in 2D images (van Schie et al., 2000).

To overcome the problems associated with grey level analysis, a technique for computerised analysis of ultrasound images (UTC) has been developed. This technique stores contiguous transverse US images at regular distances along the tendon´s long axis. A combination of these creates a 3D block of US information. Each US image is the mixed result of structure-related, single reflections that are generated at the interface of tendon bundles and non-structure-related ‘interference’, which results from multiple interfering echoes generated by structures smaller than the resolution of ultrasonography, such as collagen fibrils (Harris et al., 1991, van Schie and Bakker, 2000). Echoes generated by single reflections on tendon bundles are characterised by consistent echogenicity with respect to intensity and distribution of grey levels over contiguous US images (van Schie and Bakker, 2000). Echoes resulting from interference show rapidly changing grey levels in corresponding pixels of contiguous images (van Schie et al., 2001).

Correlating the grey level of corresponding pixels in contiguous transverse US images allows discrimination between structure-related and non-structure-related echoes and to reconstruct longitudinal information from the transverse US images (van Schie and Bakker, 2000, van Schie et al., 2001). Four types of echoes can be discriminated based on their characteristics: (1) echoes generated at fully aligned structures with a size above the limits of resolution, such as intact tendon bundles; (2) echoes generated by single reflections at structures above the size of resolution but with less alignment or integrity, such as inferior repaired tendon; (3) echoes generated by interference at entities below the limits of resolution, such as accumulation of collagen fibrils, not (yet) organised into tendon bundles, and (4) lack of echoes due to absence of reflections, such as homogenous accumulation of cells or fluid (van Schie et al., 2001). It has been shown that UTC provides accurate information about the 3D arrangement of the collagenous matrix and integrity of tendon tissue, allowing distinct identification of different healing pathways (van Schie et al., 2009).

A recently introduced treatment for tendon disorders uses platelet rich plasma (PRP), an autologous concentrate of platelets rich in growth factors that have been shown to influence tendon repair (Dahlgren et al., 2002, Kashiwagi et al., 2004, Thomopoulos et al., 2007). In previous work, we showed that PRP improved end stage biochemical, biomechanical and histological features of experimentally created superficial digital flexor tendon (SDFT) lesions (Bosch et al., 2009). In the present study, the efficacy of UTC for intra-vital detection of the end stage quality of tendon repair was assessed and the usefulness of UTC to detect differences between treatment groups during the early phases of the repair trajectory was determined. The evaluation of tendon tissue obtained from these horses, which is reported elsewhere (Bosch et al., 2009), revealed significant differences between these two treatment groups. We hypothesised that by using UTC it would be possible to identify differences between the two treatment groups earlier than at end stage and to detect time-related differences over the course of the healing process in the relative distribution of tissue types during the various phases of the healing pathway. Such differences might provide additional information about the possible mechanisms underlying any treatment effects.

Section snippets

Horses

Six 3–5 year-old Standardbred horses (5 geldings, 1 mare; bodyweight [BW] 448 ± 31 kg) were used. Clinical examination showed no lameness and no clinical or US signs of tendon disorders. The horses were housed indoors in individual boxes (16 m2) and fed a maintenance ration of silage and concentrates three times daily. Water was freely available. The study was conducted in compliance with the Dutch Act on Animal Experiments and had been approved by the Utrecht University Committee on the Care and

Results

All tendons had developed a core lesion that ultrasonographically appeared very similar to clinical SDFT lesions (Fig. 1, Fig. 2).

Discussion

Over the years many different treatments have been introduced to treat tendon injuries, all aiming at accelerating healing and/or improving quality of repair tissue, but little hard data exist that permit an objective and conclusive judgement on the effectiveness of these new therapies. An important reason for this is the lack of an objective and quantitative intra-vital monitoring technique. In the present study we tested the suitability of a recently developed new quantitative diagnostic

Conclusions

US tissue characterisation is an easy technique to perform and, other than the custom-made software, it does not require sophisticated and costly equipment. We concluded that UTC provides an excellent tool for the objective assessment of the effectiveness of treatments for SDFT lesions in horses. It is especially suitable for the monitoring of the repair trajectory, as it provides an intra-vital method to investigate the quality of repair. It may allow more adequate treatment selection and

Conflict of interest statement

None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.

Acknowledgement

The authors would like to thank the staff of the Department of Pathology, especially Charlotte de Busschere, for preparing and staining the histological sections.

References (30)

  • D. Little et al.

    Ultrasonographic and MRI evaluation of a novel tendonitis model in the horse

    Veterinary Surgery

    (2006)
  • C.M. Marr et al.

    Factors affecting the clinical outcome of injuries to the superficial digital flexor tendon in National Hunt and point-to-point racehorses

    Veterinary Record

    (1993)
  • C.M. Marr et al.

    Ultrasonographic and histopathological findings in equine superficial digital flexor tendon injury

    Equine Veterinary Journal

    (1993)
  • C. Martinoli et al.

    Analysis of echotexture of tendons with US

    Radiology

    (1993)
  • R.G. Nicoll et al.

    Ultrasonographical and pathological studies of equine superficial digital flexor tendons; initial observations, including tissue characterisation by analysis of image grey scale, in a thoroughbred gelding

    Equine Veterinary Journal

    (1992)
  • Cited by (0)

    View full text