By continuing to browse this site you agree to us using cookies as described in About Cookies
Wiley Online Library is migrating to a new platform powered by Atypon, the leading provider of scholarly publishing platforms. The new Wiley Online Library will be migrated over the weekend of March 17 & 18. You should not experience any issues or loss of access during this time. For more information, please visit our migration page: http://www.wileyactual.com/WOLMigration/
Osseous resorption of the flexor surface of the distal phalanx of the horse has been identified previously using magnetic resonance (MR) imaging; however, little is known about the prevalence and characteristics of this lesion.
To establish the MR prevalence of resorptive lesions in the flexor surface of the distal phalanx and identify concurrent lesions associated with this finding and associations between the MR and radiographic findings.
Horses with distal extremity MR and radiographs performed within 2 weeks of each other were included in the study. The flexor surface of the distal phalanx was graded independently on both modalities for the presence of osseous resorption. The sensitivity and specificity of radiography for identifying osseous resorption was calculated using MR as the gold standard.
Eighty-two MR studies met the inclusion criteria, 8 of which included osseous resorption of the flexor surface of the distal phalanx. Concurrent injury to the deep digital flexor tendon and the navicular bone was a common finding. An association between MR and radiographic grades was identified. Radiography had a high specificity (0.96) but lower sensitivity (0.45–0.55) for detecting osseous resorption.
Osseous resorption of the flexor surface of the distal phalanx can be observed both on MR and radiographic examinations of horses with lameness localised to the foot. This lesion is most often associated with other pathological changes in the podotrochlear apparatus.
Focal radiolucency of the flexor surface of the distal phalanx is a newly recognised radiographic sign associated with pathological changes of the podotrochlear apparatus.
The increased use of magnetic resonance (MR) imaging has improved our knowledge of lesions contributing to palmar foot pain in the horse [1, 2]. There is now a greater understanding of the close anatomical and functional relationships of various osseous and soft tissue structures within the foot as well as an awareness that injury to any one of these structures rarely occurs in isolation [3-7]. Although radiography has limitations when diagnosing pathological changes in horses presenting for evaluation of palmar foot pain, the expanding knowledge base gained from MR studies can positively influence the interpretation of radiographic studies [8, 9].
Osseous resorption at the insertion of the distal sesamoidean impar ligament (DSIL) on the distal phalanx has been previously identified in a case report using transcuneal ultrasonography . Osseous cyst-like lesions in this region have also been sporadically reported in the more recent MR literature [3, 9, 11] as have less well defined irregularities of the flexor surface of the distal phalanx [2, 3, 7, 12].
Histopathological studies have shown that stress placed upon the podotrochlear apparatus can result in remodelling at the insertions of the DSIL and deep digital flexor tendon (DDFT) on the distal phalanx [13-15]. Normal DSIL and DDFT insertions on the distal phalanx consist of dense connective tissue fibre bundles separated by loose connective tissue septa that contain multiple small arteriovenous complexes, elastic tissue and a vast network of sensory nerve fibres [13, 14, 16]. It has been suggested that degenerative and regenerative changes occurring at this level may result from stress-induced damage to this intricate neurovascular network . This can result in osseous remodelling of the flexor surface of the distal phalanx [7, 16] and/or increased proteoglycan deposition within the DSIL and DDFT at their insertions . Insertional desmitis of the DSIL has also been identified in association with resorptive osseous lesions at this location [9, 10]. The presence of osseous pathological changes suggests that some of these lesions might be visible radiographically.
We hypothesised that osseous resorption of the flexor surface of the distal phalanx would be present in a population of horses undergoing MR examination for evaluation of foot lameness and that some of these changes could be identified radiographically.
The objectives of this study were to: 1) establish the prevalence of resorptive lesions in the flexor surface of the distal phalanx in lame horses presenting for MR imaging (MRI), 2) identify concurrent lesions associated with distal phalangeal osseous resorption 3) assess the association between the MR and radiographic findings and 4) evaluate the radiographic sensitivity and specificity for the detection of these lesions.
Materials and methods
Prevalence of osseous resorption and modality correlation
Medical records of horses with foot lameness undergoing MR examination of the distal extremities of their forelimbs at the Veterinary Medicine Teaching Hospital, University of California Davis between May 2009 and November 2010 were reviewed. Horses were included in the study if they had complete forelimb distal extremity MR and complete digital radiographic examinations performed at our institution within 2 weeks of each other. Age and breed of the horses were obtained from the medical records.
Sensitivity and specificity of radiographs for osseous resorption
For the calculation of the sensitivity and specificity, the radiographs from 3 horses with an osseous resorptive lesion of the flexor surface of the distal phalanx identified via computed tomography or MR examination outside of the inclusion period, were added to the original group. These were included to increase the precision of the statistical analysis by increasing the number of affected cases in the population.
Magnetic resonance assessment
Magnetic resonance imaging was performed under standing sedation using a 0.27 Tesla open MR unit with a dedicated extremity radiofrequency coil (Hallmarq)1. A standard clinical protocol was used including transverse gradient-echo T1-weighted and fast spin-echo T2-weighted sequences, a horizontal transverse spin-echo proton density sequence, dorsal gradient echo T1-weighted, fast spin-echo T2-weighted and spin-echo proton density sequences and sagittal isotropic gradient-echo 3D T1-weighted, fast spin-echo T2-weighted and fast spin echo short tau inversion recovery (STIR) sequences. The transverse images were positioned perpendicular to the DDFT proximal to the navicular bone, the horizontal transverse images were parallel to the sole and dorsal images were perpendicular to the DDFT distal to the navicular bone.
All MR images were evaluated using DICOM viewing software (eFilm)2. The flexor surface of the distal phalanx was assessed on the sagittal isotropic gradient-echo 3D T1-weighted sequence using the following grading system: grade 1 – normal, smooth osseous outline, grade 2 – mildly irregular osseous outline with one or more defects less than 2 mm in width and grade 3 – osseous resorption, characterised by one or more osseous defects greater than 2 mm wide (Fig 1). Each distal phalanx was graded independently by 2 radiologists (A.C.Y. and M.S.) and a consensual grade obtained for the cases where disagreement was present. The distance between the palmar border of the distal phalanx and the palmar and dorsal aspect of the resorptive lesions at the flexor cortex was measured in each of the grade 3 cases, using the sagittal isotropic gradient-echo 3D T1-weighted images and the linear measuring tool of eFilm2. One radiologist (M.S.) graded irregularity of the distal border of the navicular bone and STIR hyperintensity of the distal phalanx at the flexor surface as either present or absent in all of the 82 MR studies that satisfied the inclusion criteria. Any case with a grade 3 MR lesion in the flexor surface of the distal phalanx was also evaluated for concurrent abnormalities in the other structures of the distal extremity by the previously mentioned radiologists. The original MR reports of the rest of the population, generated by board certified radiologists, were reviewed for mention of other lesions in the podotrochlear apparatus, distal interphalangeal joint and DDFT.
Radiographs were obtained using an Eklin EDR33 unit. All horses had a full distal extremity series that included a lateromedial, dorso45°proximal-palmarodistal oblique, a dorso65°proximal-palmarodistal oblique view of the distal phalanx, a dorso65°proximal-palmarodistal oblique view of the navicular bone and an extended palmaro50°proximal-palmarodistal oblique (navicular skyline view) . Feet were cleaned and the sulci of the frog packed with malleable packing material prior to image acquisition.
Radiographs were reviewed using DICOM viewing software (eFilm)2 on a workstation equipped with 3-megapixal monochrome LCD monitors (Totoku ME355i2)4. Two radiologists (A.N.D and S.M.P.) blinded to the results of the MR studies, independently assessed the dorso65°proximal-palmarodistal oblique views of the distal phalanx and navicular bone. Each distal phalanx was assessed for lucency in the region of the flexor surface of the distal phalanx. They were graded as grade 1: no evidence of lucency or small, round, well-defined lucencies <2 mm, grade 2: heterogeneous opacity or poorly defined lucencies <2 mm or grade 3: lucencies >2 mm (Fig 2). The radiographs were reviewed a second time by one of the radiologists 4 months after the initial interpretation and without access to the initial grades in order to assess intraobserver variability.
One of the MRI grade 3 cases was subjected to euthanasia due to poor prognosis. A necropsy was performed and the distal extremity examined both grossly and histopathologically. Serial sagittal en bloc sections of the distal and middle phalanges, the navicular bone and deep digital flexor were obtained at the time of necropsy. Tissues were fixed in 10% buffered formalin for 48 h and subsequently decalcified in nitric acid. The tissues were trimmed to include portions of the distal phalanx, the navicular bone, DDFT and associated soft tissues in a single embedding cassette. Trimmed tissues were routinely processed and embedded in paraffin; 5 μm thick sections were cut and routinely stained with haematoxylin and eosin stains (Fig 3).
A Chi-square test was used to evaluate the association between the breed and the presence of a resorptive lesion between the MR and radiographic grades and between the MR grades and the presence of other lesions. Intra- and interobserver agreement for the radiographic grades was assessed using a Kappa-weighted test. The sensitivity and specificity of radiographic grade 3 for identifying osseous resorption was calculated using the MR grade 3 as the gold standard.
Prevalence of osseous resorption and other lesions identified on MR
During the study period, 181 distal extremity MR examinations were performed, 82 of which met the inclusion criteria. The main reason for exclusion of the other studies was the lack of available radiographs, which was commonly the case with horses initially evaluated elsewhere and referred to our institution specifically for the MR examination. These 82 studies were obtained from 65 lame horses, as 17 horses had bilateral examination for lameness in both limbs. Breeds represented included Warmbloods (25), Quarter Horses (23), Thoroughbreds (6), Paints (2), Irish Sport Horses (2), Morgans (2) and an Arab, Friesian, Tennessee Walker, Paso Fino and pony breed (one each). The average age of horses presenting for examination was 9.7 years (range 3–18 years).
Thirty-six (44%) of the flexor surfaces of the distal phalanges were grade 1, 38 (46%) grade 2 and 8 (10%) grade 3. The average age of the 8 horses with grade 3 lesions was similar to the overall population (9.3; range 4–13 years). Five of the grade 3 lesions were noted in Quarter Horses, 2 in Warmbloods and one in a Thoroughbred. The proportion of Quarter Horses in the MR grade 3 group (5/8 or 62.5%) was higher than the proportion of Quarter Horses in the study population (23/65 or 35.4%); however a statistically significant difference was not detected (P = 0.087). All grade 3 lesions were located between 0.5 and 1.1 cm from the palmar border of the distal phalanx, with the exception of one lesion that extended between 0.2 and 0.9 cm from the palmar border. Seven of the 8 MR grade 3 studies had a concurrent lesion in the DDFT with lesions at one or more of the following locations: distal to the navicular bone (3 cases), at the navicular bone (5 cases) and proximal to the navicular bone (6 cases). Among the MR grade 3 studies, changes within the navicular bone were seen in 7/8 studies including irregularity of the distal margin (7 cases), STIR hyperintensity (5 cases), irregularity of the proximal margin (4 cases), irregularity of the flexor cortex (2 cases) and distal fragmentation (one case). Other findings in MR grade 3 studies included effusion of the distal interphalangeal joint (7 cases), effusion of the navicular bursa (4 cases), effusion of the digital sheath (2 cases), collateral desmitis of the distal interphalangeal joint (one case) and collateral sesamoidean desmitis (one case).
The MR grades were significantly associated with overall navicular bone abnormality (P = 0.004), irregularity of the distal border of the navicular bone (P = 0.005) and distal phalangeal STIR hyperintensity at the attachment of the DSIL (P<0.001). Irregularities of the distal border of the navicular bone were observed in 13/36 (36%) grade 1, 27/38 (71%) grade 2 and 6/8 (75%) grade 3. Short tau inversion recovery hyperintensity of the distal phalanx adjacent to the attachment of the DSIL was present in 3/36 (8%) grade 1, 9/38 (24%) grade 2 and 6/8 (75%) grade 3. No significant associations were detected between the MR grades and the presence of DDFT lesions, location or type of DDFT lesions, lesions of the collateral ligament of the distal interphalangeal joint, lesions of the collateral sesamoidean ligament or effusion of the synovial structures.
Radiologist A identified 38 grade 1, 38 grade 2 and 9 grade 3 studies on the first assessment and 47 grade 1, 29 grade 2 and 9 grade 3 on the second assessment. Radiologist B identified 54 grade 1, 23 grade 2 and 8 grade 3 studies. The kappa-weighted test was 0.61 for intraobserver agreement (Radiologist A) and 0.47 for interobserver agreement.
Correlation MR and radiographic grades
The Chi-square test revealed an association between MR and radiographic grades (P<0.001).
Sensitivity and specificity of radiographs
The sensitivity and specificity of radiographic grade 3 for identifying osseous resorption were 0.55 and 0.96, respectively for one radiologist and 0.45 and 0.96, respectively for the other.
Histopathological examination of bones from a horse with an MR grade 3 lesion revealed an approximately 5 mm wide, irregular cavitation within the flexor surface of the distal phalanx at the site of insertion of the DSIL and DDFT (Fig 3). The bone surface lining the cavitation contained multifocal scalloped indentations, rarely with attached osteoclasts (resorption bays) and was focally proliferative, with attached palisading and hypertrophied osteoblasts. Variably cellular fibrous connective tissue and granulation tissue concentrically arranged in a lamellar pattern filled the cavitation with the granulation tissue extending multifocally into the surrounding osseous resorption bays. Fibroblast cellularity increased in a centripetal manner. Loose eosinophilic acellular debris could also be appreciated within the lesion, extending to the flexor interface with the DDFT. Both the dorsal surface of the DDFT and the DSIL were multifocally architecturally disrupted by hypereosinophilic debris (liquifactive necrosis) and irregular clefting. Both structures had extensive multifocal chondroid metaplasia.
This study explores osseous resorption of the flexor surface of the distal phalanx as identified on radiographs and MRI. Resorption was present in 9.8% of horses in our study population yet it remains unmentioned in multiple studies describing MR lesions of the podotrochlear apparatus associated with foot lameness [2, 4, 5, 15, 18]. Quarter Horses were overrepresented in our group of horses having the resorptive lesion. Although this was not statistically significant when compared with the proportion of Quarter Horses in the study population, this was most likely due to a lack of power in our statistical analysis since the P value of 0.087 still suggests that a strong tendency is present.
In a previous study, cyst-like lesions in the flexor surface of the distal phalanx were present in 2 of 50 limbs  as well as in 3 of 12 limbs in a report evaluating distal phalangeal osseous cyst-like lesions in various locations . In the latter study, osseous cyst-like lesions in the flexor surface were identified almost as frequently as resorption at the insertion of the collateral ligaments of the distal interphalangeal joint (4/12) . These previously reported cyst-like lesions most likely represent the same pathological abnormality as those reported as osseous resorption (MR grade 3) in the current study. The term ‘osseous resorption’ was used in this study in lieu of the term ‘cyst-like lesion’, due to the variation in shape and lack of fluid within the cavitary lesions identified with MRI.
In a previous study, mild irregularity at the insertion of the DSIL on the distal phalanx was reportedly more common in lame horses (65%) than control horses (50%) and irregularity at the insertion of the DDFT was only ever detected in lame horses . In this study, less than half (44%) of the population had a smooth cortical margin along the flexor surface, while a similar number (46%) had a mildly irregular surface (grade 2). The large number of grade 2 flexor surfaces suggests that some irregularity might be a variation of normal and represent normal but prominent vascular channels. Alternatively, the irregularity may represent mild or early pathological osseous resorption. A pathological process is also supported by the observation of a higher prevalence of distal phalangeal STIR hyperintensity and navicular bone irregularity in the MR grade 2 group compared with the MR grade 1 group.
Distal phalangeal osseous resorptive lesions have been most commonly described at the insertion of the collateral ligaments of the distal interphalangeal joint with concurrent desmopathy [9, 19-21]. The grade 3 osseous resorptive lesions in this study were identified on MR images in a consistent location between 0.5 and 1.1 cm from the palmar border of the distal phalanx. This area of the flexor cortex corresponds to the dorsal aspect of the insertion of the DSIL and the palmar aspect of the insertion of the DDFT.
Critical limitations exist in the assessment of DSIL changes by MRI [1-3]. A grading system for DSIL changes was initially included in the design of this study but was found by both radiologists to be difficult to apply. This resulted in very low interobserver agreement. For this reason, DSIL imaging findings were not considered in this report and the association between DSIL lesions and the resorptive lesions on the distal phalanx was therefore not assessed.
Bones from an individual horse with an osseous resorptive lesion in the flexor surface of the distal phalanx underwent histopathological examination in the current study. This examination revealed evidence of concurrent abnormalities within the DSIL and distal aspect of the DDFT. The focal bone loss in the distal phalanx with associated fibrous tissue and necrotic debris in our histopathology specimens may be secondary to the presence of inflammatory mediators and cytokines (e.g. tumour necrosis factor-alpha [TNF-α]) associated with the soft tissues lesions in the adjacent DSIL and the DDFT . This supports the hypothesis that insertional desmopathy or tendonopathy can induce osseous resorptive lesions [3, 9, 11, 17, 19-21, 23]. The involvement of both the DSIL and DDFT seen histopathologically is not only consistent with the location of the osseous resorptive lesions identified on MR but also provides further evidence that these structures often respond to stress placed upon the foot as a single functional unit .
Other proposed causes for osseous resorption in the podotrochlear apparatus have included enlarged vascular channels [2, 14, 24] and synovial invaginations [1, 2, 14, 15, 24, 25]. In this clinical study, histopathological analysis was only possible for one horse. Although osteonecrosis was a prominent finding in this instance, it is possible that this lesion is not representative of all grade 3 imaging cases. Presumed vascular channel enlargement as suggested by hyperintense linear areas on the 3D T1-weighted isotropic MR sequences were seen in several cases. Furthermore, histopathological studies have shown the insertions of the DSIL and DDFT to have a rich, sensory neurovascular supply, characterised by specialised arteriovenous complexes and a large number of small vascular foramina within the distal phalanx [14-16]. Biomechanical stress placed upon the podotrochlear apparatus is proposed to damage this intricate neurovascular network, interrupting the blood supply to this region and triggering degenerative and regenerative changes . Although not confirmed by the histopathology examination reported herein, it is still possible that ischaemic damage plays a role in the development of osseous resorptive lesions of the flexor surface of the distal phalanx.
Enlargement of synovial invaginations has previously been suggested to cause osseous resorption in the distal aspect of the navicular bone [1, 2, 14, 15, 24, 25]. Although a cystic structure lined with synovial cells has previously been identified at the insertion of the DSIL on the distal phalanx , the histopathological sample in this study did not have any evidence of synovial tissue within the cavitary lesion. A thin connective tissue septa exists between the DSIL and the DDFT, making it unlikely that synovium from the navicular bursa would be in direct contact with the bone and be responsible for these resorptive lesions .
All 8 grade 3 MR cases in the present study had evidence of concurrent lesions in regional structures that supports the theory that the close anatomical and functional relationships of the podotrochlear apparatus often results in injury to more than one structure at a time [3-6, 14]. Furthermore, the significant association of the grade 3 MR lesions with navicular bone abnormalities such as irregularity of its distal margin and STIR hyperintensity of the distal phalanx is highly suggestive of increased stress associated with the DSIL .
Due to the limitations of low- and high-field MRI for assessment of the DSIL, a previous study has looked for more readily identifiable abnormalities to serve as indicators of injury to the DSIL itself . The presence of a cystic structure in the distal third of the navicular bone, distal border fragmentation and STIR hyperintensity at the distal phalangeal insertion of the DSIL are all MR findings indicative of DSIL abnormalities . In a similar way, it is possible that the osseous resorption of the flexor surface of the distal phalanx observed in the present study might also be used as an indicator of DSIL injury.
In our clinical experience, the inclusion of a high-resolution sagittal isotropic gradient-echo 3D T1-weighted sequence subjectively improves the ability to detect osseous lesions both on the flexor surface of the distal phalanx and on the distal border of the navicular bone. The isotropic gradient-echo 3D-T1-weighted sequence has a higher spatial resolution with a slice thickness of 0.75 mm, compared with 3 mm for the regular 3D T1-weighted and 5 mm for most of the spin-eco sequences on this system. The use of this sequence and the specific evaluation of the flexor surface of the distal phalanx may have contributed to the greater prevalence of flexor surface irregularity identified in this study when compared with previous reports. Gradient-echo sequences have been shown to less accurately represent the size and shape of osteochondral lesions on high-field MR due to susceptibility artefact  however, in the same study, the T1-weighted gradient-echo sequence was commonly preferred with the low-field system for lesion detection.
The high radiographic specificity identified in this study suggests that radiographic studies of the foot should be carefully assessed for identification of focal radiolucency in the flexor cortex. This radiographic sign can be considered specific for osseous resorption of the flexor surface of the distal phalanx. A negative radiographic study does not, however, rule out the presence of this lesion due to the much lower sensitivity. Magnetic resonance examination or recheck distal extremity radiographs are therefore warranted if clinically indicated.
The fact that radiographic interobserver agreement was only fair also suggests that the flexor surface of the distal phalanx can be a difficult area to interpret and open to subjectivity even when a specific grading scale is applied. The lack of agreement was mostly due to poor differentiation between radiographic grades 1 and 2. Superimposition of the flexor surface of the distal phalanx with the adjacent middle phalanx, solar structures and potential packing artifacts can contribute to difficulty in assessing this area . In our experience, the use of the lateromedial projection to confirm or rule out focal resorption of the flexor surface is often of limited value due to the larger thickness of bone superimposed on the lesion area in this projection.
Along with the underlying aetiology, the clinical significance of the osseous resorptive lesions identified in this study also requires further investigation. The presence of multiple lesions in the foot has been associated with a poorer prognosis [7, 12, 28]. The retrospective nature of this study was a limitation in the assessment of the clinical presentation and outcome of these patients and prospective studies may help provide further information in these areas. The degree to which osseous resorption of the flexor surface of the distal phalanx contributes to the lameness grade would still, however, be complicated by the presence of concurrent lesions in most of these cases.
Osseous resorption of the flexor surface of the distal phalanx might be more common than previously recognised indicating that radiographs should be carefully assessed for this finding in horses presenting for evaluation of palmar foot pain. Further evaluation with cross-sectional imaging is still warranted in negative radiographic studies and in order to evaluate the presence of concurrent abnormalities.
Authors' declaration of interests
No competing interests have been declared.
Sources of funding
No funding was received for this study.
A.C. Young: data collection through medical records, interpretation of MR images, redaction of the manuscript. A.N. Dimock: interpretation of radiographic images, revision of the manuscript. S.M. Puchalski: interpretation of radiographic images, revision of the manuscript. B. Murphy: histology interpretation, contribution to the redaction of the manuscript. M. Spriet: original idea, study design, interpretation of MR images, statistical analysis, contribution to the redaction of the manuscript.