Injuries of the SDF tendon are a common problem in horses.1–5 They are often career-threatening events because they are long-lasting injuries and there frequently is an insufficient repair process that results in the formation of biomechanically inferior scar tissue. This repair tissue cannot withstand the repetitive strains of cyclic loading, which in equine flexor tendons are extremely close to the physiologic limits under normal conditions.6–8
The reaction of tendon tissue to macrotrauma essentially follows the general phases of wound healing, with inflammation and demarcation, proliferation, maturation, and remodeling. Remnants of disrupted and necrotic tendon bundles are removed by phagocytosis and lytic enzymes during the inflammatory phase, which lasts until approximately 10 days after an injury. During the proliferation phase, a fibroproliferative callus is formed from 4 to 45 days after injury. During the maturation or remodeling phase (45 to 120 days after injury), there is progressive cross-linking and organization of the collagen fibrils into tendon bundles.1,9–11 It has been stated12 that early bundle formation is evident after day 90 and that the bundles at 120 days after injury have an appearance similar to that for normal tendon.12
Since its introduction, ultrasonography has gradually developed into the most appropriate imaging modality for the examination of the integrity of equine tendons and hence for the diagnosis of lesions.13 It would be of great clinical importance if this technique also could be used for reliable monitoring of the stages of repair after tendon injury and for the objective evaluation of treatments. Use of ultrasonography for these purposes has been reported,14–16 but assessment was mostly based on subjective semiquantitative scores.
Quantitative approaches based on measurement of gray values in transverse images have been reported17, 18 and, to a certain extent, have improved the accuracy and increased the objectivity. However, gray values alone are not sufficient for an accurate assessment of the degree of structural integrity of a tendon.19–21 This indicates that ultrasonography is of questionable value for use in monitoring tendons.22
A method of computerized processing has been developed in an in vitro experimental environment. This method quantifies the stability of the gray value of each corresponding pixel in contiguous transverse ultrasonographic images by means of mathematic routines.13,21,23,24 The correlation routine described in another study13 permits the exclusive discrimination between structure-related and non–structure-related echoes, which in fact provides the transverse ultrasonographic image with longitudinal information about tendon bundles with a size larger than the limits of resolution. The entropy routine quantifies the overall homogeneity of the tendon tissue and provides information about cellular and fibrillar components with a size below the limits of resolution.24 By means of the combination of both routines, this method of computerized image processing permits accurate and unequivocal assessment of the integrity of tendon tissue. Cluster analysis of data from in vitro experiments24 performed to validate the computerized ultrasonography technique for the discrimination of various tissue types reveals distinct differences in the restoration of the fibrillar wave pattern and the formation and organization of tendon bundles in scars. This suggests the existence of different categories for tendon healing, which have been designated as the repair and inferior repair categories.13, 24 The repair category appears to lead to reestablished organization with correct arrangement of fibrillar components in larger tendon bundles and thus to restoration of the structural hierarchy. In the inferior repair category, remodeling appears to be less optimal, which leads to inadequate organization and arrangement of fibrillar components and no or minimal formation of intact, continuous tendon bundles. In the clinical situation, early discrimination between these 2 repair categories would be of great benefit for adequate prognostication and timely intervention.
The study reported here was conducted to evaluate the feasibility and effectiveness of computerized ultrasonography for monitoring the repair process of tendons for clinical conditions. Therefore, a longitudinal study was conducted in which standardized lesions were created in SDF tendons and allowed to heal with no intervention other than the imposition of a regimen of restricted exercise. The lesions were examined on a regular basis by the use of computerized ultrasonography. The ultimate goal was to determine criteria for use of quantitative ultrasonography that would provide objective information on the stage of tendon repair by use of computerized variables. Furthermore, we evaluated whether 2 categories of healing (repair and inferior repair) could be identified early in the healing process on the basis of differences in ultrasonographic variables identified during in vivo monitoring.
Materials and Methods
Horses—Six Dutch Warmblood horses were used in the study. Horses were between 4 and 8 years of age, comprised 5 geldings and 1 mare, and weighed (mean ± SD) 450 ± 38 kg. The horses had no clinical evidence of lameness, had a normal conformation, and had no history of tendon lesions. The SDF tendons had no clinical or ultrasonographic abnormalities. The horses were housed separately in box stalls under identical conditions throughout the study. The study was approved by the Ethical Committee of Utrecht University.
Surgical procedures and postoperative treatment—Standardized lesions were created in the SDF tendons of both forelimbs of the horses by use of a surgical technique adapted from other studies.25, 26 Briefly, horses were anesthetized and positioned in right lateral recumbency. An incision was made through the skin, subcutaneous tissues, and peritendon over the midmetacarpal aspect of the SDF tendon in a proximodistal direction and approximately 1 cm lateral to the sagittal plane. The skin and subcutaneous tissues were retracted to expose the SDF tendon. A V-shaped segment of the SDF tendon was excised in the sagittal plane; the excised segment had a mean ± SD length of 34.3 ± 3.5 mm, mean width of 3.9 ± 0.8 mm, and mean depth of 3.9 ± 0.8 mm. Because the peritendon was not separated from the tendon, part of this structure was excised along with the tendon segment. The defect in the tendon was not closed, but the skin was closed with an intracutaneous suture.60a A pressure bandage was applied for 2 weeks after surgery. Nonsteroidal anti-inflammatory medication60b was provided for 3 days after surgery.
Horses were confined to box stalls from 2 weeks before surgery until 8 weeks after surgery. From 9 until 12 weeks after surgery, horses were hand-walked twice daily (15 min/session) on a hard surface. At 13 weeks after surgery, horses were exercised at a walk in an automatic horse walker (15-m diameter; ground surface consisted of bricks covered with a thin layer of sand) twice daily (20 min/session). At the conclusion of the study, horses were euthanized by administration of an overdose of pentobarbital60c (200 mg/kg, IV) after sedation with detomidine (0.01 mg/kg, IV).d
Experimental configuration for computerized ultrasonography—Ultrasonography was performed with a 7.5-MHz linear-array transducer.e The hardware configuration consisted of an ultrasonographic transducer and an electromagnetic tracking device (Figure 1). Shoes were removed so that they would not interfere with the electromagnetic field of this device. For the exact location of each ultrasonographic image, the inactive receiver of a low-frequency electromagnetic tracking systemf was affixed to the transducer, and the transmitter was placed on the floor 120 to 180 cm from the injured tendon being evaluated. At the beginning of each imaging session, the transducer was placed at the height of the accessory carpal bone for calibration of the zero position of the tracking system. Subsequently, a transverse sweep was made from proximal to distal along the palmar sagittal line of the metacarpus. Special care was taken to ensure scanning was accomplished in a gradual, sliding motion that maintained a correct angle of insonation of the transducer. Ultrasonographic images were digitized and stored in a computer containing two 500-MHz processors. The speed of this system allowed 20 to 24 images/s to be stored. The electromagnetic tracking system provided information about the exact position of the transducer and thus of each ultrasonographic image recorded. Points of reference were the transmitter and the position of the accessory carpal bone.
Schematic depicting the configuration for computerized ultrasonography of the SDF tendon in 6 horses in the study. The transmitter of the electromagnetic tracking system was placed on the floor 120 to 180 cm from the tendon being evaluated. 1 = Accessory carpal bone. 2 = Receiver of the electromagnetic tracking system. 3 = Ultrasound transducer. 4 = Transmitter of the electromagnetic tracking system. 5 = Transmission of positional information. 6 = Transmission of ultrasonographic images. 7 = Data acquisition and digitizing. 8 = Computer.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Ultrasonographic data collection—Custom-designed softwareg was used to collect 240 consecutive transverse ultrasonographic images. Images were collected at intervals of 0.5 mm, beginning at a predetermined distance from the accessory carpal bone. Collection of images was directed by input from the electromagnetic tracking system. The procedure was repeated at least 3 times/limb.
Both limbs of all 6 horses were scanned during 14 separate sessions. The first session was the day before the surgery (time 0). Tendons were scanned 2 days and 1, 2, 3, 4, 6, 8, 10, 12, 14, 16, 18, and 20 weeks after the surgical intervention.
Image processing—Initial processing of the 240 transverse images involved the relocation of contiguous images along the longitudinal axis of the tendon by use of a custom-designed alignment procedure. This created a block of 3-dimensional ultrasonographic information for a length of 12 cm.
This block was used for several procedures. Tomographic reconstructions were made on the basis of original transverse data to create 2 additional planes of view (ie, reconstructed longitudinal and reconstructed coronal [latitudinal]). The combination of these 3 planes permitted measurement of the extent of the lesion and reliable estimation of its volume by use of the best-fitting ellipsoid. Dimensions were measured at 7 time points (immediately after extirpation and subsequently at day 2 and weeks 1, 2, 3, 4, and 6 after surgery). The extirpated tendon specimen was measured by use of conventional measuring techniques in a relaxed state immediately after excision, but the subsequent postsurgical measurements, obtained by means of computerized ultrasonography and tomographic reconstructions, were performed with the tendon in a normal weight-bearing state.
In each ultrasonographic data block, 9 zones of interest (central part of the lesion and proximal and distal to the lesion) were selected on the basis of results for the first imaging session after surgery. Zone 5 corresponded to the center of the lesion and was used as the point of reference; zones 4 and 6 were 15 mm proximal and distal from the center of the lesion, respectively; zones 3 and 7 were 25 mm proximal and distal from the center of the lesion, respectively; zones 2 and 8 were 40 mm proximal and distal from the center of the lesion, respectively; and zones 1 and 9 were 50 mm proximal and distal from the center of the lesion, respectively. For determination of reference values of computerized pixel analysis for normal SDF tendon tissue, in each of these zones, regions of interest of 2,500 pixels were selected in 12 contiguous ultrasonographic images. These representative regions were processed by means of correlation and entropy routines.13,23,24 After processing, means for the entire lesion were calculated, and processed pixels were classified in 1 of 4 categories.24 Category C represented pixels that were exclusively correlated as a consequence of their constantly high gray values; this is indicative of intact, contiguous tendon bundles. Category E represented pixels with exclusively entropy as a consequence of their rapidly changing gray values over contiguous images; this is indicative of accumulations of cells or fibrils. Category B represented pixels that were correlated but that also had some degree of entropy with slight variation (approx 10%) of gray values over contiguous images; this is indicative of wavy tendon bundles or incomplete intratendinous (endotendon) septa. Category N represented pixels that did not correlate and did not have entropy; in most cases, this is indicative of homogenous accumulations of fluid or cells.
Relative distributions of pixels were determined for all imaging sessions. For monitoring of repair and more specifically for comparison of the computerized pixel analysis with the histomorphologic quality of repair, only data collected in 3 zones of interest, namely zone 5 (center of lesion) and zones 4 and 6 (proximal and distal parts of a lesion, respectively), were used. Data obtained before surgical intervention on day 0 served as reference values to which results for the subsequent imaging sessions were compared. Data from the last imaging session before euthanasia were used to calculate aberrations of the C, B, E, and N ratios from the reference values. Ratios were summed and used for ranking of the tendon lesions in accordance with ultrasonographic severity.
Histomorphologic examination of scars—Histomorphologic criteria for the assessment of the stage of tendon repair were based on experiments of tendon healing reported in other studies.21,23–30 Main histomorphologic features used for scoring were overall cellularity, morphology of fibroblasts, arrangement of fibroblasts, morphology and organization of fibrils, morphology and organization of fibers, and fasciculi. Histologic sections from the same 3 zones of interest that had been used for the ultrasonographic characterization (ie, proximal, central, and distal parts of the lesion) were compared against these criteria by use of a semiquantitative scoring system (Appendix). Results from the 3 zones were averaged to determine an overall score of the scar tissue, similar to the situation for the ultrasonographically determined ratios. Scoring was performed independently by 2 of the authors (HTMvS and AMJ) who were not aware of the source of the tissues. The 4 scars with the lowest score were classified as the repair group, whereas the 4 scars with the highest scores were classified as the inferior repair group.
Data analysis—Data from all 12 tendons were used. We deliberately opted to consider both tendons from the same horse as separate entities because it was believed that this decision could provide an indication of the extent to which a good or bad healing pattern could be determined.
The ultrasonographic variables C, B, E, and N from the repair and inferior repair groups were compared by use of a 2-way ANOVA with Bonferroni post hoc tests. Significance was defined as values of P < 0.05. For comparison of computerized ultrasonography ratios and histomorphologic characteristics of each tendon, regression analysis was used to determine the correlation between aberrations from the reference values for C, B, E, and N and the histomorphologic score.
Results
Reference values of computerized pixel analysis for normal SDF tendon tissue—The relative distribution was determined for pixels classified as C, E, B, or N by means of the correlation and entropy routines for all 9 zones at time 0 (before surgical intervention; Figure 2). The C ratio typically had a slightly lower value in the more distal zones, with a concomitant increase in E and N ratios. These patterns apparently did not differ significantly. Values at time 0 were used in this experiment as reference values for normal SDF tendon tissue. Mean ± SD values for each of the ratios for all 9 zones were as follows: C ratio, 0.706 ± 0.063; B ratio, 0.183 ± 0.061; E ratio, 0.032 ± 0.020; and N ratio, 0.080 ±0.048.
Mean ± SD values for C (diamonds), E (circles), B (triangles), and N (squares) ratios before surgical creation of a lesion in both SDF tendons of 6 horses (time 0) for each of 9 zones representing the entire metacarpal length of the SDF tendon. Zone 5 corresponds to the center of the lesion, zones 4 and 6 are 15 mm proximal and distal from the center of the lesion, respectively; zones 3 and 7 are 25 mm proximal and distal from the center of the lesion, respectively; zones 2 and 8 are 40 mm proximal and distal from the center of the lesion, respectively; and zones 1 and 9 are 50 mm proximal and distal from the center of the lesion, respectively.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Extent of the lesion—We detected approximately a 2-fold increase in lesion length during the first 2 days after surgery (from a mean ± SD of 34.3 ± 3.5 mm to 64.1 ± 14.9 mm; Figure 3). After this initial increase, the lesion continued to increase in length until reaching its largest extent at 4 weeks after surgery (mean, 116.1 ± 20.7 mm). Lesion length at that time ranged from 74.3 to 144.0 mm, which was 2 to 4 times the length of the original lesion. After week 4, no further increase was detected.
Mean ± SD length (A), width (B), and volume (C) of the lesion created in 12 SDF tendons (both SDF tendons of 6 horses) from day 0 (day of surgery to create the lesions) until 6 weeks after surgery.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Mean width of the lesion increased as well, but to a lesser extent than the length and in a more gradual manner (Figure 3). The largest width was approximately twice the original width, which was reached 2 weeks after surgery and remained relatively constant until 6 weeks after surgery.
Mean ± SD volume of the extirpated tendon segment was 295.0 ± 136.2 mm3 (range, 140.0 to 510.0 mm3). Because of the increases in length and width, a progressive increase in volume could be detected during the first 3 to 6 weeks after surgery, which ultimately resulted in a mean volume of 3,469.2 ± 221.7 mm3 (range, 1,750.0 to 5,810.0 mm3). Thus, there was a > 10-fold increase in estimated volume during the first 3 to 6 weeks after surgery (Figure 3).
Histomorphologic examination of scars—The histologic sections of the various tendons were ranked (Table 1). The ranking was identical between the 2 authors who examined the sections. The 4 scars with the most favorable histologic features were identified and classified as the repair group, and the 4 specimens with the worst histologic features were classified as the inferior repair group.
Scores for histomorphologic assessment of the scar tissue in the left (L) and right (R) tendons of 6 horses with a surgically created lesion in the SDF tendons.
| Horse 1 | Horse 2 | Horse 3 | Horse 4 | Horse 5 | Horse 6 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | L | R | L | R | L | R | L | R | L | R | L | R |
| Cellularity | 6 | 3 | 5 | 5 | 4 | 6 | 7 | 7 | 5 | 3 | 7 | 6 |
| Proportion of noemal fibroblasts | 6 | 9 | 6 | 9 | 8 | 7 | 7 | 4 | 6 | 5 | 6 | 6 |
| Predominant shape of fibroblasts | 6 | 6 | 9 | 6 | 9 | 6 | 6 | 6 | 5 | 7 | 6 | 7 |
| Arrangement of fibroblasts | 6 | 9 | 3 | 6 | 7 | 7 | 6 | 3 | 4 | 5 | 3 | 5 |
| Proportion of fibroblasts with parallel arrangement | 8 | 9 | 9 | 7 | 9 | 8 | 9 | 5 | 6 | 6 | 5 | 5 |
| Morphology of fibrils | 8 | 9 | 6 | 9 | 8 | 9 | 9 | 8 | 6 | 6 | 6 | 3 |
| Organization and arrangement of fibers and fasciculi | 9 | 9 | 9 | 9 | 9 | 9 | 9 | 8 | 8 | 5 | 6 | 5 |
| Total | 49 | 54* | 47 | 51 | 54* | 52* | 53* | 41 | 40† | 37† | 39† | 37† |
One of 4 tendons with the highest score, which was considered indicative of the worst healing (ie, inferior repair group).
One of 4 tendons with the lowest score, which was considered indicative of the best healing (ie, repair group).
Histomorphologic features of the repair group were characterized (Figure 4). Scars in this group were characterized by advanced organization and alignment of tendon bundles, which were surrounded by thickened, wavy, and partly incomplete intratendinous (endotendon) septa. Furthermore, cellularity was moderately increased, partly with predominantly aligned, enlarged or plump fibroblasts and partly with less organized ovoid or round fibroblasts located in the vicinity of wavy or remodeling septa. Despite the lack of an overall distinct bundle formation, the fibrillar arrangement was unidirectional and there was a distinct regular wave pattern.
Low-magnification (A), medium-magnification (B), and high-magnification (C) photomicrographs of sections of scar tissue obtained from a tendon with the best characteristics of healing (repair group). In panel A, there is advanced alignment of tendon bundles that are surrounded by thickened, wavy, and partly incomplete intratendinous (endotendon) septa. In panel B, notice the moderate increase of cellularity, partly with more aligned fibroblasts with enlarged or plump nuclei and partly with less organized fibroblasts with ovoid or round nuclei located in the vicinity of wavy or remodeling intratendinous septa. In panel C, there is a lack of overall distinct bundle formation, but there is fibrillar unidirectional arrangement with a predominantly regular wave pattern. Notice that nuclei of fibroblasts are enlarged and partly ovoid. H&E stain; bar = 150 Mm, 25 Mm, and 10 Mm for panels A, B, and C, respectively.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Histomorphologic features of the inferior repair group were also characterized (Figure 5). Scars in this group were characterized by a lack of an overall unidirectional arrangement of tendon bundles, which were frequently arranged in an ellipsoid pattern and surrounded by intratendinous septa that were (for the most part) incomplete and contained enlarged vasculature. Furthermore, there was a moderate increase of cellularity, with elongated and ovoid fibroblasts that were arranged predominantly in rows between poorly organized fibrils. There were loosely arranged fibrils of irregular thickness, the wave pattern was only poorly restored, and there was a complete lack of unidirectional fibrillar alignment.
Low-magnification (A), medium-magnification (B), and high-magnification (C) photomicrographs of sections of scar tissue obtained from a tendon with the worst characteristics of healing (inferior repair group). In panel A, notice that there is no unidirectional arrangement of tendon bundles, which are frequently arranged in an ellipsoid pattern and surrounded by intratendinous septa that are incomplete and contain large vasculature. In panel B, cellularity is moderately increased with elongated and ovoid nuclei of fibroblasts that are predominantly arranged in rows between poorly organized fibrils. There is a complete lack of unidirectional fibrillar arrangement. In panel C, there are loosely arranged fibrils of irregular thickness. In addition, there is no regular wave pattern or unidirectional fibrillar alignment. H&E stain; bar = 150 Mm, 25 Mm, and 10 Mm for panels A, B, and C, respectively.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Monitoring of the repair process by use of computerized pixel analysis in ultrasonographic images collected during successive imaging sessions—Each of the pixel ratios was monitored during successive imaging sessions until 20 weeks after surgery (Figure 6). Immediately after surgical creation of the lesion, there was a sharp decrease in structure-related pixels (C and B) and a strong increase in pixels without any correlation (N), which represented an accumulation of fluid. Also, E pixels increased, which was indicative of cellular activity. This pattern continued during the subsequent weeks and reached a maximum at 3 to 4 weeks after surgery. From then on, the number of noncorrelating (N) pixels decreased steadily, as did the number of pixels with a high degree of entropy (E). There was a concomitant increase of pixels with at least some structural relationship (B and C). At 12 weeks after surgery, C and B ratios were almost identical, which was indicative of incomplete structures. Whereas the C ratio in the subsequent weeks revealed a continuous but slow increase, the B ratio continued to increase more rapidly, whereas E and N ratios decreased further toward presurgery values. By 18 and 20 weeks after surgery, E and N ratios had decreased to presurgery values for normal tendon tissues. At this time, the C ratio in the center of the scar (mean ± SD, 0.46 ± 0.057) was relatively constant at a value slightly lower than the presurgical value. However, the B ratio was still increased (mean, 0.448 ± 0.039).
Mean ± SD values for C (diamonds), E (circles), B (triangles), and N (squares) ratios of scars in all 12 SDF tendons (A), the 4 tendons with the best healing (repair group; B), and the 4 tendons with the worst healing (inferior repair group; C) determined during successive imaging sessions conducted before (time 0) and at various time points until 20 weeks after surgical creation of the lesions. *Within a time point, values for a ratio differed significantly (P < 0.05) between the repair and inferior repair groups.
Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.37
Comparison of computerized ultrasonography ratios and histomorphologic scores—Regression analysis of the ultrasonographic characteristics (in data obtained during the last imaging session immediately prior to euthanasia) resulted in quantification of C, B, E, and N ratios as aberrations from the reference values obtained before surgery, and the histomorphologic quality of repair revealed a relatively strong correlation (r2 = 0.37; P = 0.024) between histomorphologic characteristics and the B ratio. For the C and N ratios, there were also significant, albeit less strong correlations (r2 = 0.19 [P = 0.037] and r2 = 0.13 [P = 0.036], respectively). There was not a significant correlation with the E ratio. The predictive value of the ultrasonographic criteria appeared to be good because the 4 scars with the least aberrations from the ultrasonographic reference values appeared to be identical to the ones classified by the histologic criteria as the repair group, whereas the 4 specimens with the most ultrasonographic aberrations revealed poor histomorphologic characteristics (inferior repair group). When regression lines were drawn, there appeared to be a good fit between the histomorphologic scores of all 12 scars and the respective computerized C and B ratios (regression for the C ratio was r = 0.61 [P <0.05] and for the B ratio was r = 0.64 [P = 0.025]).
Comparison of computerized ratios of repair group and inferior repair group patterns—For the repair group, data for the C, B, E, and N ratios were determined (Figure 6). The same data were determined for the inferior repair group. Comparison of the patterns of both groups led to several observations.
The mean C ratio reached its maximum in the inferior repair group (mean ± SD, 0.468 ± 0.031) at 12 weeks after surgery and then decreased to 0.315 ±0.070 at 20 weeks after surgery. In the repair group, the mean C ratio increased to 0.595 ± 0.080 at 18 weeks after surgery and decreased to only 0.500 ± 0.021 at 20 weeks after surgery. The C ratios began to differ significantly between groups beginning at the 18th week after surgery. The mean B ratio had an opposite pattern. In the inferior repair group, the mean ± SD B ratio increased from 0.407 ± 0.026 at 12 weeks after surgery to 0.682 ± 0.072 at 20 weeks after surgery. In the repair group, the mean B ratio decreased from 0.378 ± 0.023 at 12 weeks after surgery to 0.303 ± 0.090, and it then increased moderately to 0.389 ± 0.050 at 20 weeks after surgery. At 14 weeks after surgery, the B ratios began to differ significantly between groups. Patterns for the N and E ratios were grossly comparable over time for both groups.
Discussion
A standardized approach was necessary for development of a timetable of ultrasonographic changes during the process of tendon healing. This requirement prohibited the use of patients with clinical lesions. Experimentally induced tendon lesions have been described and are based on surgical intervention25,26,28 or chemical induction by means of intratendinous application of collagenolytic enzymes.1,27,ref31–34 Lesions induced with collagenolytic enzymes are often characterized by a strong inflammatory response and extensive peritendinous reaction, most probably as a result of leakage of enzymes during intratendinous injection. A disadvantage of a surgical technique is that a segment of normal tendon is extirpated, which bypasses a major part of the naturally occurring processes, such as vascular disruption, necrosis, and demarcation of disrupted fibers. However, because surgical lesions can be standardized and seem to best mimic the natural situation, the choice was made for a lesion created by means of partial tenectomy. The lesion was created in the sagittal plane, but a slightly medial or lateral skin incision was used to avoid effects such as acoustic shadowing and reverberation on the ultrasonographic image as a result of scar tissue formation in the skin.35–37
The duration of the experiment was limited to 20 weeks. The endpoint of the healing process of tendon lesions may be a year or even more after the original trauma. However, it has been reported12 that the maturation and remodeling phase, in which the organization of newly formed type I collagen fibrils into larger tendon bundles and the progressive reorientation of these fibrils along the lines of axial stress take place, lasts from approximately 45 to 120 days after injury. For this reason, 20 weeks (140 days) was chosen as a time point at which it can be assumed that the key developments in the healing process must have taken place and eventual outcome has been largely determined.
The tracking system was essential for the use of computerized ultrasonography in the in vivo situation. There are many potentially confounding factors, such as inconsistencies in transducer handling (eg, slight changes in the angle of insonation and uneven pressure on a standoff) and movements of the horse during imaging. The tracking system and custom-designed relocation software used for optimal alignment of the images succeeded in compensating for these confounding factors. The small SDs in the values for C, B, E, and N ratios (references) as determined for normal tendons, which were based on 12 tendons with at least 3 blocks/tendon and 9 zones/block, confirmed the efficacy of the procedure.
The completely passive receiver of the tracking system was placed on the ultrasound transducer, whereas the transmitter was positioned on the floor 120 to 180 cm from the injured tendon. Thus, any biological effect as a consequence of the low-frequency electromagnetic field was assumed to be negligible.
We detected a 2to 4-fold increase in length of the lesions and an approximately 2-fold increase in width of lesions, which led to an increase in volume of up to 10 times. Although an increase in lesion size in the period immediately after the initial trauma has been described in another report,12 the degree of enlargement was striking. Retraction of wound margins is a commonly observed phenomenon after trauma and may explain the widening of the gap. However, the large expansion of the lesion in a distal direction can be explained only by progressive tearing along the lines of longitudinal arrangement of the fiber bundles under the influence of bearing weight. Enlargement of the lesion continued until 4 weeks after surgery. On the basis of this observation, which was detected in all horses, it would appear advisable that in the case of partial rupture of the SDF tendon, bearing of weight in the period immediately after the injury should be avoided as much as possible. This can best be achieved by applying a cast on the distal portion of the limb for at least the first 4 weeks after the injury.
It has frequently been stated10–12,38 that the inflammatory phase may require 5 to 10 days and that the injured tendon is weakest between 5 and 7 days after injury, presumably as a result of fibrinolysis. The latter statement may be questioned on the basis of the findings in the study reported here, in which the maximum extent of the lesion was not reached until 3 to 4 weeks after surgery. The enormous increase in length and volume of the lesion apparently was not exclusively caused by the activities of macrophages and lytic enzymes, but instead was attributable to biomechanical influences as well. In this early posttraumatic period, the tendon was characterized by a high fraction of N pixels, which reflected the accumulation of fluid in the created tissue gap. The entropy ratio (E pixels) also increased immediately, which was indicative of the strongly increased cellular activity in the region. The C and B ratios decreased to almost zero in the center of the lesion because there was a complete loss of the original structure of the extracellular matrix of the tendon.
The rapid increase of the correlated pixels (C and B) during the early stages of the fibroproliferative phase (r 6 weeks after surgery) might be erroneously explained as early bundle formation. Such a conclusion would be incongruent with results of other investigators who detected the formation of bundles beginning 90 days after surgery.25,26,28 In fact, the increase of the C ratio only indicates the presence of interfaces with a length of approximately 2 mm (because correlation is processed over 5 images at mutual distances of 0.5 mm). The fact that, in contrast to the situation for normal tendon, values for the B ratio were almost as high as for the C ratio led to the conclusion that the interfaces were lacking continuity and integrity. The high values for numbers of pixels classified as E or N indicated the formation of fibrovascular callus, mainly of peritendinous origin, with precursors of intratendinous septa that contained vasculature.
During monitoring, the E and N ratios in the proximal border of the scar gradually decreased to approximately reference values at 12 weeks after surgery, whereas in the distal periphery of the repair tissue, these ratios approached reference values no sooner than 20 weeks after surgery. At 14 to 16 weeks after surgery, a sharp increase of E and eventually N ratios was detected, most probably as a consequence of increasing amounts of exercise beginning at week 13. The prolonged accumulations of fluid or cellular and fibrillar components in the periphery of the lesion may have had various causes. These include increased vascularity, continuous remodeling activities to connect the scar with the surrounding normal part of the tendon, or minor reruptures as a consequence of differences in modulus of elasticity. The fact that the distal transitional zone appeared to lag behind the proximal zone may have been related to the continuing enlargement of the defect in a distal direction during a period of approximately 4 weeks after the initial trauma, whereas there was hardly any progress in the proximal direction during this period.
Histomorphologically, the morphology and organization of fibrils, fibers, and fasciculi (and in most cases also the arrangement of fibroblasts) proved to be the decisive characteristics that permitted the discrimination between the repair and inferior repair categories. Cellularity did not appear to be a major determining factor. Scars in the repair category were characterized by restoration of a regular wave pattern and an advanced ultrastructural organization and alignment. In scars of the inferior repair group, the fibrillar wave pattern had no substantial degree of restoration, and fibrils, fibers, and fasciculi were poorly organized with a lack of unidirectional alignment. Frequently, the collagenous ultrastructure in scars with inferior repair was arranged in an ellipsoid pattern, which suggested the presence of compartments. This may be related to an increase in glycosaminoglycans, which will increase the overall water-binding capacity of the tissue.39 The observation that this elliptical fibrillar arrangement is frequently detected in tendons with chondroid metaplasia and the fact that fibroblasts adapt their cytoskeleton and the production of extracellular matrix components to the local lines of stress40–42 suggested that in these compartments, the lack of uniaxial tension may promote the differentiation of fibroblasts into chondrocytes.
Interestingly, 2 of the 4 scars in the inferior repair group were in the same horse, and the 4 scars in the repair group were in only 2 horses. The chance that 2 specimens from the same animal would be classified within the same group (assuming healing characteristics are randomly distributed) is approximately 30%, but the chance that this would happen twice in the same group is only approximately 3%. This observation apparently indicates that there are individually determined inherent differences in healing tendency. Although we acknowledge that this observation does not qualify as scientific proof, it supports a long-standing clinical impression.
Graphs representing development of the C, B, E, and N ratios for the repair and inferior repair groups were extremely similar for the first 12 weeks after surgery. At approximately 12 weeks, the divergence into 2 pathways became apparent. At that time, N and E ratios had almost reached their lowest values. In the repair group, in which the B ratio had been increasing since approximately week 6 after surgery, the B ratio remained relatively constant. However, in the inferior repair group, the B ratio continued to increase until the horses were euthanized at 20 weeks. The difference between groups was significant at 14 weeks after surgery. With respect to the C ratio, differences between the groups began at approximately 16 weeks after surgery. Similar to results for the B ratio, the C ratio also began increasing at approximately 6 weeks after surgery, and the C ratio continued to increase in the repair group but remained relatively constant (with even a tendency to decrease) in the inferior repair group. For the C ratio, the difference between groups was significant at 18 weeks after surgery. When compared with the original ratios in normal tendon tissue (presurgery reference values), the repair group had a subnormal C ratio (–25%) and an increase in B ratio (approx 100%) at the time of euthanasia, but in the inferior repair group, the C ratio was much lower (–60%) and the B ratio was extremely increased (300%). These differences can be translated in structural terms as a steady increase of tendon bundles with intact intratendinous septa over longer distances in the repair group, whereas in the inferior repair group, the high B ratio indicated incomplete endotendon septa at the time of euthanasia at 20 weeks after surgery. The E and N ratios had no major role during this period. However, although small in absolute terms, there were substantial relative differences between the groups that may have indicated differences in activity of the tissues at the cellular level. The potential exists that these differences are biologically important, but this cannot be quantitatively substantiated on the basis of the study reported here.
On the basis of comparisons of ultrasonographic variables and histomorphologic characteristics of tendon lesions at 20 weeks after surgery, and from the monitoring by computerized ultrasonography of the healing process leading up to that stage, it was concluded that the process of structural realignment along the lines of axial tension is the key factor for the adequate repair of tendon lesions and is crucial for functional recovery. This process of remodeling and restructuring along the long axis of the tendon is reflected in the decrease of the E and N ratios and the increase of the C and B ratios, as determined by the use of computerized ultrasonography. The process starts at approximately 6 weeks after injury and appears to reach a critical phase between 12 and 16 weeks after injury. Then, in the case of undisturbed repair, structural realignment of fibrils over larger distances should begin (reflected by an increase in the C ratio) through the restoration of incomplete endotendon septa (reflected by a decrease in the B ratio). In some cases, this development appears to be hampered, possibly by the precocious formation of mature cross-links,43, 44 which leads to the formation of an inferior quality of repair tissue that can be assumed to be functionally insufficient. This indicates that the period between 12 and 16 weeks after injury can be considered to be crucial in the process of tendon healing and that any attempts at influencing this process (eg, physical exercise, medication, or approaches such as focused shock wave treatments45, 46) should be undertaken before or during that period. It can be concluded that the computerized analysis of ultrasonographic images collected by means of an adequate tracking device provides an excellent tool for the objective assessment and monitoring of lesions in the SDF tendon of horses.
ABBREVIATIONS
| SDF | Superficial digital flexor |
Vicryl, USP 2-0 polyglactin 910, Ethicon GmbH, Norderstedt, Germany.
Quadrisol, Vedaprofen, Intervet, Boxmeer, The Netherlands.
Euthesate, Ceva Sante Animale, Naaldwijk, The Netherlands.
Domosedan, Orion Pharma, Espoo, Finland.
Acoustic Imaging Performa, Dornier Medizintechnik, Munich, Germany.
Fastrak, Polhemus, Colchester, Vt.
RealTimeCorrelator, EchoConsult, Wassenaar, The Netherlands.
References
- 1.
Silver IA, Brown PN, Goodship AE, et al. A clinical and experimental study of tendon injury, healing and treatment in the horse. Equine Vet J Suppl 1983;1:1–49.
- 2.
Rossdale PD, Hopes R, Wingfield-Digby NJ, et al. Epidemiological study of wastage among racehorses 1982 and 1983. Vet Rec 1985;116:66–69.
- 3.
Goodship AE, Birch HL. The pathophysiology of the flexor tendons in the equine athlete, in Proceedings. Dubai Int Equine Symp 1996;1:83–107.
- 4.
Wilson JH, Robinson RA, Jensen R, et al. Equine soft tissue injuries associated with racing: descriptive statistics from American racetracks, in Proceedings. Dubai Int Equine Symp 1996;1:1–21.
- 5.
Olivier A, Nurton JP, Guthrie AJ. An epizoological study of wastage in Thoroughbred racehorses in Gauteng, South Africa. J S Afr Vet Assoc 1997;68:125–129.
- 6.
Stephens PR, Nunamaker DM, Butterweck DM. Application of a Hall-effect transducer for measurement of tendon strains in horses. Am J Vet Res 1989;50:1089–1095.
- 7.
Jansen MO, Savelberg HHCM. Stress and strain of equine tendons of the forelimb at failure. Equine Vet J Suppl 1994;17:57–60.
- 8.
Platt D, Wilson AM, Timbs A, et al. Novel force transducer for measurement of tendon force in vivo. J Biomech 1994;27:1489–1493.
- 9.
Goldin B, Block WD, Pearson JR. Wound healing of tendon—1. Physical, mechanical, and metabolic changes. J Biomech 1980;13:241–256.
- 10.
Enwemeka CS. Inflammation, cellularity, and fibrillogenesis in regenerating tendon: implications for tendon rehabilitation. Phys Ther 1989;69:816–823.
- 11.
Stashak TS. Management of wounds associated with tendons, paratendons, and tendon sheaths. In: Stashak TS, ed. Equine wound management. Philadelphia: Lea & Febiger, 1991;238–257.
- 12.↑
Watkins JP. Treatment principles of tendon disorders. In: Auer JA, ed. Equine surgery. 2nd ed. Philadelphia: Saunders Co, 1992;916–924.
- 13.↑
van Schie HTM, Bakker EM. Structure-related echoes in ultrasonographic images of equine superficial digital flexor tendons. Am J Vet Res 2000;61:202–209.
- 14.
Reef VB, Martin BB, Elser A. Types of tendon and ligament injuries detected with diagnostic ultrasound: description and follow-up, in Proceedings. Annu Meet Am Assoc Equine Pract 1988;34:245–248.
- 15.
Genovese RL, Rantanen NW, Simpson BS, et al. Clinical experience with quantitative analysis of superficial digital flexor tendon injuries in Thoroughbred and Standardbred racehorses. Vet Clin North Am Equine Pract 1990;6:129–145.
- 16.
Palmer SE, Genovese RL, Longo KL, et al. Practical management of superficial digital flexor tendinitis in the performance horse. Vet Clin North Am Equine Pract 1994;10:425–481.
- 17.
Nicoll RG, Wood AKW, Rothwell LTW. Ultrasonographical and pathological studies of equine superficial digital flexor tendons; initial observations, including tissue characterisation by analysis of image gray scale, in a Thoroughbred gelding. Equine Vet J 1992;24:318–320.
- 18.
Tsukiyama K, Acorda JA, Yamada H. Evaluation of superficial digital flexor tendinitis in racing horses through gray scale histogram analysis of tendon ultrasonograms. Vet Radiol Ultrasound 1996;37:46–50.
- 19.
van Schie JTM, Bakker EM. The weird world of speckle and scatter, in Proceedings. Sci Meet Eur Coll Vet Surg 1996;5:70–71.
- 20.
van Schie JTM, Bakker EM, van Weeren PR. Ultrasonographic evaluation of equine tendons: a quantitative in vitro study of the effects of amplifier gain level, transducer-tilt and transducer-displacement. Vet Radiol Ultrasound 1999;40:151–160.
- 21.
van Schie HT, Bakker EM, Jonker AM, et al. Ultrasonographic tissue characterization of equine superficial digital flexor tendons by means of gray level statistics. Am J Vet Res 2000;61:210–219.
- 22.↑
Smith RKW, Birch HL, Patterson-Kane J, et al. A review of the etiopathogenesis, and current proposed strategies for prevention, of superficial digital flexor tendinitis in the horse, in Proceedings. Annu Meet Am Assoc Equine Pract 2000;46:54–58.
- 23.
van Schie HTM, Bakker EM, Jonker AM, et al. Efficacy of computerized discrimination between structure-related and non–structure-related echoes in ultrasonographic images for the quantitative evaluation of the structural integrity of superficial digital flexor tendons in horses. Am J Vet Res 2001;62:1159–1166.
- 24.↑
van Schie HTM, Bakker EM, Jonker AM, et al. Computerized ultrasonographic tissue characterization of equine superficial digital flexor tendons by means of stability quantification of echo patterns in contiguous transverse ultrasonographic images. Am J Vet Res 2003;64:366–375.
- 25.
Watkins JP, Auer JA, Gay S, et al. Healing of surgically created defects in the equine superficial digital flexor tendon: collagentype transformation and tissue morphologic reorganization. Am J Vet Res 1985;46:2091–2096.
- 26.
Watkins JP, Auer JA, Morgan SJ, et al. Healing of surgically created defects in the equine superficial digital flexor tendon: effects of pulsing electromagnetic field therapy on collagen-type transformation and tissue morphologic reorganization. Am J Vet Res 1985;46:2097–2103.
- 27.
Williams IF, McCullagh KG, Goodship AE, et al. Studies on the pathogenesis of equine tendonitis following collagenase injury. Res Vet Sci 1984;36:326–338.
- 28.
Bertone AL, Stashak TS, Smith FW, et al. A comparison of repair methods for gap healing in equine flexor tendon. Vet Surg 1990;19:254–265.
- 29.
Jann HW, Good JK, Morgan SJ, et al. Healing of transected equine superficial digital flexor tendons with and without tenorrhaphy. Vet Surg 1992;21:40–46.
- 30.
Crevier-Denoix N, Collobert C, Pourcelot P, et al. Mechanical properties of pathological equine superficial digital flexor tendons. Equine Vet J Suppl 1997;23:23–26.
- 31.
Henry GA, Patton CS, Goble DO. Ultrasonographic evaluation of iatrogenic injuries of the equine accessory (carpal check) ligament and superficial digital flexor tendon. Vet Radiol 1986;27:132–140.
- 32.
Spurlock GH, Spurlock SL, Parker GA. Ultrasonographic, gross and histologic evaluation of a tendinitis disease model in the horse. Vet Radiol 1989;30:184–188.
- 33.
Keg PR, van den Belt AJ, Merkens HW, et al. The effect of regional nerve blocks on the lameness caused by collagenase induced tendonitis in the midmetacarpal region of the horse: a study using gait analysis, and ultrasonography to determine tendon healing. Zentralbl Veterinarmed [A] 1992;39:349–364.
- 34.
Redding WR, Booth LC, Pool RR. The effects of polysulphated glycosaminoglycan on the healing of collagenase induced tendinitis. Vet Comp Orthop Traumatol 1999;12:48–55.
- 35.
Penninck DG. Imaging artifacts in ultrasound. In: Nyland TG, Mattoon JS, eds. Veterinary diagnostic ultrasound. Philadelphia: WB Saunders Co, 1995;19–29.
- 36.
Kremkau FW, Taylor KJ. Artifacts in ultrasound imaging. J Ultrasound Med 1986;5:227–237.
- 37.
Kremkau FW. Artifacts. In: Diagnostic ultrasound: principles, instruments, and exercises. 6th ed. Philadelphia: WB Saunders Co, 2002;273–318.
- 38.
Leadbetter WB. Cell-matrix response in tendon injury. Clin Sports Med 1992;11:533–578.
- 39.↑
Todhunter RJ. Anatomy and physiology of synovial joints. In: McIlwraith CW, Trotter GW, eds. Joint disease in the horse. Philadelphia: WB Saunders Co, 1996;1–28.
- 40.
Meikle MC, Sellers A, Reynolds JJ. Effect of tensile mechanical stress on the synthesis of metalloproteinases by rabbit coronal sutures in vitro. Calcif Tissue Int 1980;30:77–82.
- 41.
Merrilees MJ, Flint MH. Ultrastructural study of tension and pressure zones in a rabbit flexor tendon. Am J Anat 1980;157:87–106.
- 42.
Clegg PD, Strassburg S, Smith RK. Cell phenotypic variation in normal and damaged tendons. Int J Exp Pathol 2007;88:227–235.
- 43.
Chvapil M. Present status of the pharmacology of fibrosis and scar contractures, in Proceedings. Dubai Int Equine Symp 1996;1:395–406.
- 44.
Reef VB, Genovese RL, Wade Byrd J, et al. Treatment of superficial digital flexor tendon injuries with beta-aminoproprionitrile fumarate (BAPN-F): sonographic evaluation of early tendon healing and remodeling, in Proceedings. Dubai Int Equine Symp 1996;1:423–430.
- 45.
Rompe JD. Dose-dependent effects of extracorporeal shock waves on rabbit Achilles tendon. In: Rompe JD, ed. Shockwave applications in musculoskeletal disorders. Stuttgart, Germany: Thieme Verlag, 2002;7–15.
- 46.
Bosch G, Lin YL, van Schie HT, et al. Effect of extracorporeal shock wave therapy on biochemical composition and metabolic activity of tenocytes in normal tendinous structures in ponies. Equine Vet J 2007;39:226–231.
Appendix
Criteria used for the semiquantitative assessment of histomorphologic characteristics of scar tissue.
| Variable | Grade 0 | Grade 1 | Grade 2 | Grade 3 |
|---|---|---|---|---|
| Cellularity | Normal | Slight increase | Moderate increase | Severe increase |
| Proportion of normal fibroblasts (%) | 100 | 67 to < 100 | 33 to < 67 | < 33 |
| Predominant shape of fibroblasts | Spindle-shaped | Elongated | Plump or ovoid | Ovoid or round |
| Arrangement of fibroblasts | Rows between fibers | Rows between fibers and nests in intratendinous septa | Localized nests in intratendinous septa | Generalized |
| Proportion of fibroblasts with parallel arrangement (%) | 100 | 67 to < 100 | 33 to < 67 | < 33 |
| Morphology of fibrils | Regular thickness and wave | Irregular thickness and regular wave | Irregular thickness and no regular wave | Fine and no regular wave |
| Organization and arrangement of fibers and fasciculi | Distinct and overall parallel | Distinct and locally parallel | Not distinct and locally parallel | Not distinct and not locally parallel (ie, loose and irregular) |
