MAGNETIC RESONANCE IMAGING OF THE INITIAL ACTIVE STAGE OF EQUINE LAMINITIS AT 4.7 T

Authors


  • Funded by the American College of Veterinary Radiology and The Ohio State University Intramural Equine Research Fund.

  • Presented at the International Society for Magnetic Resonance in Medicine, Berlin, Germany, May 2007, and in part at the Joint ACVR/IVRA Meeting, Vancouver, Canada, August 2006.

  • Dr. Arble's present address is Center for Veterinary Health Sciences, Oklahoma State University, Stillwater, OK.

  • Dr. Mattoon's present address is Department of Veterinary Clinical Sciences, Washington State University, Pullman, WA.

Address correspondence and reprint requests to Jason B. Arble, at the above address. E-mail: arble@okstate.edu

Abstract

Equine laminitis is a severely debilitating disease. There is a poor understanding of the underlying pathophysiology, and traditional imaging modalities have limited diagnostic capacity. High field strength magnetic resonance (MR) imaging allows direct visualization of the laminae, which other modalities do not. This would prove useful both in assessment of clinical patients and in further investigation into the pathophysiology of the disease. The objective of this study was to characterize the anatomic changes within the equine foot associated with the initial active stage of laminitis. Images obtained using a 4.7 T magnet were compared with digital radiographs using histologic diagnosis as the reference standard. Objective measurements and subjective evaluation for both modalities were evaluated for the ability to predict the histologic diagnosis in horses with clinical signs of laminitis as well as in clinically normal horses and horses that were in a population at risk for developing laminitis. Signal intensity and architectural changes within the corium and laminae were readily seen at 4.7 T, and there was a strong association with the histologic diagnosis of active laminitis. Measurements obtained with MR imaging were more sensitive and specific predictors of laminitis than those obtained radiographically. Subjective evaluation with MR imaging was more sensitive than with radiography and should become more specific with greater understanding of normal anatomy.

Introduction

Laminitis, characterized by loss of the support of the distal phalanx due to separation of the dermal and epidermal laminae, can occur secondary to equine systemic inflammatory disease, dietary change, or carbohydrate overload. The exact pathophysiologic basis of laminitis is unclear.1–3 An imaging modality that allowed direct visualization of the laminae might enhance the understanding of laminitis.

The appearance of chronic laminitis has been described using magnetic resonance (MR) imaging,4,5 but the changes seen with MR imaging during the initial phase are uncharacterized. It was our purpose to identify the specific changes seen in horses during the initial active phase of laminitis using MR imaging and compare the imaging findings with histologic appearance of the laminae. Further, we compared radiographic and MR imaging findings using the histologic appearance as the reference standard. We also assessed whether the changes seen in horses with initial active laminitis may be seen in horses that are at risk for developing acute laminitis but do not yet exhibit clinical signs.

Materials and Methods

Adult, full-sized horses with acute active laminitis, or current disease known to predispose to acute laminitis, having undergone euthanasia were considered for inclusion in this study. All horses were euthanized due to normal clinical course without regard to this study. Horses with a prior history of laminitis in the affected limbs were excluded. The initial active phase of laminitis was defined as foot pain with increased temperature of the dorsal hoof surface, with or without bounding digital pulses, for a duration of less than 15 days.

Twenty-eight horses ranging in age from 2 to 23 years met the criteria for inclusion in the study. If laminitis was detected clinically in the front limbs, then both front limbs and one hind limb were collected. If laminitis was detected only in the hind limbs, then both hind limbs and one front limb were collected. All four extremities were collected if the horse had laminitis in both the front and rear limbs. This resulted in 81 feet being studied.

Horses were divided into three groups based on clinical presentation and possible risk for developing acute laminitis. Group 1 consisted of nine horses that were suffering from an initial active onset of laminitis for less than 14 days at the time of euthanasia. Group 2 consisted of 10 horses that were suffering from diseases known to predispose to acute laminitis but did not have clinical signs of acute laminitis at the time of euthanasia. Diseases in Group 2 horses consisted of gastrointestinal disease in nine and catastrophic skeletal injury in one. The nine horses in Group 3 were used as normal controls. These horses had no prior history of laminitis and were euthanized for causes that are not associated with a predisposition to acute laminitis.

After euthanasia, the shoes were removed if present and the distal extremity was disarticulated from the limb at the metacarpo(tarso)phalangeal joint. The feet were cleaned and all inorganic and metallic material removed. Specimens were stored in a walk-in cooler until imaging could be performed. Specimens that could not be imaged within 4 days were frozen at −20°C. Frozen specimens were allowed to thaw 18–24 h before imaging.

Radiographs of each foot were made with an indirect flat panel digital radiography system.* Lateral-to-medial and dorsal 60° proximal–palmarodistal/plantarodistal projections were made for each foot and stored on a server maintained onsite and dedicated to storage of digital images. Images were stored in DICOM format and evaluated using an eFilm imaging workstation. Any metallic debris detected radiographically was removed before MR imaging.

All distal extremities were imaged on a 4.7 T/40 cm magnet controlled by a Bruker Avance console. After a localizer spin echo sequence, T2*-weighted images were acquired in transverse and sagittal planes using a 3D gradient echo (GRE) sequence. T1 and proton density (PD)-weighted images were acquired in the transverse plane. The sequence parameters are summarized in Table 1.

Table 1.   Magnetic Resonance Imaging Sequence Parameters
 FOV
(cm3/cm2)
MatrixTR
(ms)
TE
(ms)
Thickness/
Gap (mm)
  1. FOV, field of view; TR, time to repetition; TE, time to echo; 3DGRE, three-dimensional gradient recall echo; PD, proton density.

T2*-weighted 3DGRE12 × 12 × 8384 × 384 × 4016.53.71.3 × 1.3
T1 weighted12 × 12256 × 2567004.32.5 × 2.5
PD weighted12 × 12256 × 25630004.32.5 × 2.5

All feet were sectioned within 2 hours of completion of imaging. Using a band saw, sagittal and transverse cuts were made through the dorsal hoof wall and distal phalanx and sections from the mid-portion of the dorsal hoof wall in both planes were harvested. Tissues specimens were fixed in formalin for a minimum of 48 h before being trimmed for paraffin embedding and tissue sectioning. Each slide contained both a sagittal and transverse section that included as much of the corium and keratinized hoof wall as possible and was stained with hemotoxylin and eosin. All slides were evaluated by a board-certified veterinary pathologist (S.E.W.) experienced in equine disease; the pathologist was not aware of the history or imaging findings. Histologic diagnosis of normal or laminitis was made based on laminar disorganization and necrosis, pyknosis, and rhexis of the epidermal laminae.6 All samples were evaluated for separation between the dermal and epidermal laminae and evidence of freezing-induced artifacts and tissue necrosis.

All radiographic and MR images were randomized separately and reviewed by two board-certified veterinary radiologists (J.S.M. and W.T.D.) and one board-certified equine surgeon (J.K.B.). All reviewers were unaware of the clinical status and histologic findings. Measurements included proximal and distal dorsal hoof wall thickness and the dorsal hoof wall-to-distal phalanx ratio.7 Representative transverse T2* 3D GRE images from the proximal, middle, and distal one-third of each foot were identified by dividing the soft tissues dorsal to the distal phalanx into equal thirds and then selecting a slice from the center of each region (Fig. 1). Each image was evaluated for alteration of the following: corium architecture, corium signal intensity, laminar architecture, laminar signal intensity, and laminar separation. Each change was graded normal, mild, moderate, or severe: grades 1–4, respectively. Quantitative measurements of the MR images included: thickness of the corium, laminae, dermal tissue (laminae plus corium), and the ratio of laminae to total dermal tissue (Fig. 2). The laminae-to-dermis ratio was calculated by dividing the thickness of the laminae by the total thickness of the dermal tissue at the same location. All radiographic and MR image measurements were made by a single individual (J.B.A.).

Figure 1.

 Sagittal T2*-weighted image of the equine digit. The transverse lines represent the locations of the three transverse images provided to the observers for evaluation of each foot.

Figure 2.

 Transverse T2*-weighted image illustrating the measurements made of the laminae and corium used to derive the laminae-to-dermis ratio.

Changes noted by one author (J.B.A.) on a pilot study were used as categories to be evaluated by each observer when reviewing the representative MR images from each foot. Pearson correlation was used to evaluate the association between a change noted with MR imaging and a histologic diagnosis of laminitis. Each category was also assigned a numeric qualifier from 1 (normal) to 4 (severe change). A chi-squared analysis was performed for each category at each level (proximal, middle, distal) to detect any significant difference for the number values assigned for each category between normal and laminitic feet. The probability of a change being detected was also calculated with respect to the histologic diagnosis.

A two-sample Student's t-test was used to compare the mean differences in measurements obtained radiographically vs. MR imaging, in normal vs. laminitic feet. Dot plots were constructed to compare the distribution of measurement values and to determine cut points for the laminae-to-dermis ratio in MR images and dorsal hoof wall-to-distal phalanx ratio in radiographs. The sensitivity, specificity, positive predictive value, and negative predictive value for the three different foot levels were calculated using the derived cut point.

The ability to specifically identify laminar separation using MR imaging and radiography was evaluated in terms of sensitivity and specificity using histologically evident separation as the standard. The ability to predict a histologic diagnosis of laminitis based on the observed changes was evaluated by calculating the sensitivity and specificity for the detection of an abnormality in each separate category. Sensitivity and specificity for each abnormality were calculated with two different cut-off points. A cut point of 1 denoted that any abnormality from mild to severe (2–4) was a prediction of disease. A cut point of 2 denoted that mild changes (2) were considered normal while moderate to severe changes (3 or 4) were predictors of laminitis. Sensitivity and specificity were also calculated for each observer for the ability to correctly come to the diagnosis of laminitis or normal for both MR imaging and radiography.

Results

Data from one foot in Group 2 was corrupted beyond retrieval. A single foot from a horse suffering from a nonweight-bearing injury in the contralateral limb was included, as the weight-bearing foot was at risk for support-limb laminitis.

Nineteen of 27 feet in Group 1 horses, and two of 27 feet in Group 2 horses were laminitic. All remaining feet were normal.

The 3D GRE sequence transverse images (Fig. 3) were characterized by the best anatomic detail and contrast resolution of the three MR imaging sequences performed and was therefore exclusively used in the MR evaluations. The normal corium and laminae had greater signal intensity than the hoof wall and the cortical bone of the distal phalanx. The internal portion of the hoof wall and the medullary cavity of the distal phalanx had similar intermediate signal intensity. The peripheral borders of the hoof wall were not detectable due to a lack of signal from this structure. The individual primary laminae were easily identified. Although differentiation between sensitive and insensitive primary laminae was not possible in the MR images, the epidermal laminae typically had a less intense signal which was similar to the interior hoof wall, while the dermal laminae were more hyperintense. The architecture of the corium was readily identifiable with a heterogeneous pattern that was interspersed with hyper- and hypointense blood vessels. The margin between the corium and the cortex of the distal phalanx was distinct and had an undulating pattern. When compared with histologic samples at 4 × magnification (Fig. 4), all of the anatomic architecture was identifiable except the secondary laminae. Ability to see the secondary laminae and consistently differentiate the sensitive from insensitive primary laminae was a limitation in MR anatomic resolution.

Figure 3.

 T2*-weighted transverse image of a normal equine digit.

Figure 4.

 Histological section of the interdigitation of normal laminae. H and E stain at 4 × magnification.

Examining the MR images, five specific changes were identified that were likely associated with the initial active stage of laminitis. Pearson correlation confirmed the association of histologically evident laminitis with all five changes which included loss of normal corium architecture, loss of laminar architecture, increased signal intensity in the corium, increased signal intensity in the laminae, and separation within the laminae (Table 2). When compared with a transverse T2*-weighted MR image from a histologically normal foot (Fig. 5), the above changes were readily identified (Fig. 6).

Table 2.   Pearson Correlation of Magnetic Resonance Changes and Histological Diagnosis of Laminitis
CorrelationsLaminar
Architecture
Laminar
Intensity
Corium
Architecture
Corium
Intensity
Separation
Laminar intensity0.762    
Corium architecture0.7930.672   
Corium intensity0.7380.6430.888  
Separation0.8040.7490.7590.724 
Histological Dx0.7430.5590.7270.6850.653
Figure 5.

 Magnified transverse T2* image of a normal dorsal foot. Note the heterogeneous appearance of the corium and uniform appearance of the laminae. HW, hoof wall; L, lamina; C, corium; P3, distal phalanx.

Figure 6.

 Magnified transverse T2* image of the dorsal foot with initial active laminitis. Note the change in corium intensity and homogeneity compared with Fig. 5. There is disruption and loss of uniformity in the laminae. A line of separation is identified by the arrowheads. HW, hoof wall; L, lamina; C, corium; P3, distal phalanx.

There was marked alteration in the appearance of the corium characterized by the lack of normal vasculature and a homogeneous appearance with overall increased signal intensity. The laminar layer was altered with multiple foci of detail loss and, less commonly, a focal increase in signal intensity. Disruption and distinct separation within the laminae were readily seen beyond merely loss of normal architectural distinction (Fig. 7).

Figure 7.

 T2* transverse MR image (A), lateromedial radiograph (B), and histologic section with H and E stain (C) of a foot with laminitis. The MR image (A) displays the alterations in the corium and lamina. An area of separation is marked with an asterisk (*). An area of increased signal intensity is shown by the white arrow. A focal hypointense signal is also present within the lamina (white arrowhead), likely representing gas or an area of acute focal hemorrhage. The mild degree of separation explains the lack of radiographic findings in the foot (B). The histologic sample shows the separation between the epidermal lamina and the dermal lamina. Note the area of hemorrhage exhibited by accumulation of red blood cells (black arrow). HW, hoof wall; L, lamina; C, corium; P3, distal phalanx; EL, epidermal lamina; DL, dermal lamina.

There was a significant difference (P<0.001) in the means of the laminae-to-dermis ratios between histologically normal vs. laminitic feet for proximal, middle, and distal locations, with the largest difference distally (Table 3). There was also a significant difference (P<0.001) between histologically normal and laminitic feet when comparing the maximum laminae-to-dermis ratio (Table 3). Multiple cut points could be selected which influenced the sensitivity and specificity when using the laminae-to-dermis ratio to predict whether or not a foot will be histologically normal or laminitic (Table 4).

Table 3.   Student's t-Test Results Comparing Means of Normal vs. Laminitic Feet Using MR-Derived Maximum Laminae-to-Dermis Ratio for Proximal, Middle, and Distal Portions of the Foot and Overall Maximum
 Mean (mm)SD (mm)T-valueP-value
Maximum ratio of foot
 Normal0.640.05−7.690.001
 Laminitic0.830.11  
Maximum ratio proximal
 Normal0.60.05−8.590.001
 Laminitic0.780.1  
Maximum ratio middle
 Normal0.610.05−8.120.001
 Laminitic0.80.1  
Maximum ratio distal
 Normal0.620.06−8.510.001
 Laminitic0.820.1  
Table 4.   Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV) for Maximum Laminae-to-Dermis Ratio for Proximal, Middle, and Distal Portions of the Foot and Overall Maximum
 Cut-Off
Value
Sensitivity
(%)
Specificity
(%)
PPV
(%)
NPV
(%)
Maximum ratio of foot0.78193.381.093.3
0.75819894.493.7
0.818110010093.8
Maximum ratio proximal0.7819894.493.7
Maximum ratio middle0.7819894.493.7
Maximum ratio distal0.7819585.093.4
0.75819894.493.7

Mean proximal hoof wall thickness, distal hoof wall thickness, and dorsal hoof wall-to-distal phalanx ratio were significantly different (P<0.001) between histologically normal and laminitic feet (Tables 5 and 6).

Table 5.   Student's t-Test Results Comparing Means of Normal vs. Laminitic Feet Using Dorsal Hoof Wall-to-Distal Phalanx Ratio, Proximal Hoof Wall Thickness, and Distal Hoof Wall Thickness Measured Radiographically
 Mean (mm)SD (mm)T-valueP-value
Dorsal hoof wall-to-distal phalanx ratio
 Normal26.992.13−4.20.001
 Laminitic30.563.69  
Proximal dorsal hoof wall thickness (mm)
 Normal18.381.54−3.240.004
 Laminitic20.763.24  
Distal dorsal hoof wall thickness (mm)
 Normal17.981.52−3.950.001
 Laminitic20.93.27  
Table 6.   Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV) for Radiographic Dorsal Hoof Wall-to-Distal Phalanx Ratio
 Cut-Off
Value
SensitivitySpecificityPPV
(%)
NPV
(%)
Dorsal hoof wall-to-P3 ratio0.311/21=52.4%55/60=91.7%42.3084.60

The categorical observations (1, normal; 2, mild; 3, moderate; 4, severe) for each abnormality detectable on MR images were evaluated for the ability to predict whether or not the foot was histologically laminitic or normal and whether the location within the foot (proximal, middle, distal) had any influence on the observers' interpretation (Figs. 8–12). The difference in the assigned grades (normal, mild, moderate, severe) was statistically significant (P<0.001) between normal and laminitic feet for all categories even when the results for all three observers were summed. Sensitivity and specificity for each abnormality were calculated for the two different cut-off points of 1 (any abnormality) or 2 (only moderate and severe changes) and also with regard to the region of the foot (Tables 7–11). For each abnormality, sensitivity was better in the distal part of the foot, while specificity was better in the proximal portion of the foot.

Figure 8.

 Distribution of total observed MR loss of laminar architecture with respect to location and histological diagnosis. 1–4 on the x-axis denote the observation of 1 (normal) to 4 (severe change) in the laminar architecture for the proximal, middle, and distal images of each foot. The y-axis represents the number of feet that were assigned that grade by the observers (results are summed for all three observers). “Norm” denotes a histologically normal foot and “Lam” signifies that the observed foot was histologically laminitic.

Figure 9.

 Distribution of total observed MR loss of corium architecture with respect to location and histological diagnosis. 1–4 on the x-axis denote the observation of 1 (normal) to 4 (severe change) in the corium architecture for the proximal, middle, and distal images of each foot. The y-axis represents the number of feet that were assigned that grade by the observers (results are summed for all three observers). “Norm” denotes a histologically normal foot and “Lam” signifies that the observed foot was histologically laminitic.

Figure 10.

 Distribution of total observed MR-altered laminar signal intensity with respect to location and histological diagnosis. 1–4 on the x-axis denote the observation of 1 (normal) to 4 (severe change) in laminar signal intensity for the proximal, middle, and distal images of each foot. The y-axis represents the number of feet that were assigned that grade by the observers (results are summed for all three observers). “Norm” denotes a histologically normal foot and “Lam” signifies that the observed foot was histologically laminitic.

Figure 11.

 Distribution of total observed MR-altered corium signal intensity with respect to location and histological diagnosis. 1–4 on the x-axis denote the observation of 1 (normal) to 4 (severe change) in the corium signal intensity for the proximal, middle, and distal images of each foot. The y-axis represents the number of feet that were assigned that grade by the observers (results are summed for all three observers). “Norm” denotes a histologically normal foot and “Lam” signifies that the observed foot was histologically laminitic.

Figure 12.

 Distribution of total observed MR separation with respect to location and histological diagnosis. 1–4 on the x-axis denote the observation of 1 (normal) to 4 (severe change) of separation for the proximal, middle, and distal images of each foot. The y-axis represents the number of feet that were assigned that grade by the observers (results are summed for all three observers). “Norm” denotes a histologically normal foot and “Lam” signifies that the observed foot was histologically laminitic.

Table 7.   Sensitivity and Specificity Using Loss of Laminar Architecture in the Distal, Middle, and Proximal Foot as a Predictor of Histological Laminitis Using a Categorical Cut-Off Value of 1 and 2
Cut PointDistalMiddleProximal
1
 Specificity (%)60.072.282.8
 Sensitivity (%)98.493.792.1
2
 Specificity (%)89.496.798.9
 Sensitivity (%)81.079.463.5
Table 8.   Sensitivity and Specificity Using Loss of Corium Architecture in the Distal, Middle, and Proximal Foot as a Predictor of Histological Laminitis Using a Categorical Cut-Off Value of 1 and 2
Cut PointDistalMiddleProximal
1
 Specificity (%)51.771.790.6
 Sensitivity (%)100.088.982.5
2
 Specificity (%)83.396.7100.0
 Sensitivity (%)87.376.263.5
Table 9.   Sensitivity and Specificity Using Altered Laminar Signal Intensity in the Distal, Middle, and Proximal Foot as a Predictor of Histological Laminitis Using a Categorical Cut-Off Value of 1 and 2
Cut PointDistalMiddleProximal
1
 Specificity (%)52.862.877.8
 Sensitivity (%)87.379.476.2
2
 Specificity (%)85.093.997.8
 Sensitivity (%)69.855.647.6
Table 10.   Sensitivity and Specificity Using Loss of Corium Signal Intensity in the Distal, Middle, and Proximal Foot as a Predictor of Histological Laminitis Using a Categorical Cut-Off Value of 1 and 2
Cut PointDistalMiddleProximal
1
 Specificity (%)43.970.690.6
 Sensitivity (%)100.093.784.1
2
 Specificity (%)80.096.1100.0
 Sensitivity (%)82.566.755.6
Table 11.   Sensitivity and Specificity Using Separation in the Distal, Middle, and Proximal Foot as a Predictor of Histological Laminitis Using a Categorical Cut-Off Value of 1 and 2
Cut PointDistalMiddleProximal
1
 Specificity (%)71.183.395.6
 Sensitivity (%)87.382.573.0
2
 Specificity (%)91.198.398.9
 Sensitivity (%)66.754.033.3

Comparing MR and radiographic images, there was a distinct difference in the sensitivity and specificity when using objective measurements. Both were statistically significant when comparing the means of normal vs. laminitic feet for each test, but there was a large overlap of radiographic measurements between normal and abnormal feet. This was clearly demonstrated when comparing dot plots derived from the radiographic measurements with dot plots of the laminae-to-dermis ratio calculated from the MR images (data not shown).

Comparing the subjective ability to make a diagnosis of a laminitis vs. normal using radiography, the results for all three observers were similar. The sensitivity for detecting active laminitis was low (47.6–57.2%). The specificity, however, was high (80.0–91.7%). The sensitivity for correctly identifying histologically laminitic feet with MR imaging was much higher (95.2–100%), but the specificity was more variable between observers (46.7–93.3%). This indicates a large number of false-negative predictions of laminitis using radiography, while there was a tendency for a larger number of false-positive results with MR imaging.

Discussion

Based on our results, there are definitive changes that occur with the initial active stage of laminitis and are readily detected with MR imaging and these changes cannot be detected with radiography. Fine anatomic detail is readily achievable using high field strength magnets, which makes direct evaluation of the laminae and corium possible.4,8 After subjective assessment of all sequences, it was felt that the T2*-weighted 3D GRE sequence gave the best anatomic detail coupled with the best contrast resolution in both normal and laminitic feet. Distortion artifacts caused by local field inhomogeneities as a result of residual metal in the hoof did not impact the area of interest and were only evident within millimeters of the shavings. The ability to detect disruption within the primary laminae is very important for evaluation of the laminitic horse and makes MR imaging more sensitive in the initial phase of the disease when compared with radiography. The small foci of increased signal intensity which may indicate focal edema or necrosis were only visible with MR imaging. Additionally, the high detail allowed evaluation of adjacent structures such as the corium, which is often overlooked when collecting specimens for histologic evaluation. The decrease in vascular conspicuity and altered signal distribution within the corium is a reminder that the laminar separation may be an end product and not the only pathophysiologic process responsible for the pain and inflammation during the initial phase of the disease. The images obtained using a high field strength magnet allowed more extensive evaluation of the tissues adjacent to and associated with the laminae.

Many studies have focused on the microenvironment of the laminae and on the blood flow of the foot during acute laminitis.9–22 However, seeing these changes in the corium leads us to question the pathophysiologic processes occurring in the tissue supporting and sustaining the laminae. Further investigation into the hemodynamics and possible inflammatory changes specifically within the dermal layer would provide much needed insight into the mechanisms that result in laminar necrosis and separation. Changes in this location may also explain the hemodynamic alterations in the larger vessels feeding the foot. MR imaging provides a method for evaluation of this tissue that is difficult to harvest and often overlooked.

The increased ability to detect changes in MR images before radiographic evidence of physical breakdown of the laminae not only allows avenues for further investigation but also increases the ability of clinical evaluation. The use of laminar separation as criteria for diagnosing acute laminitis had a lower sensitivity and specificity than other categories but may be a better prognostic indicator regarding the outcome of the disease. Horses that had histologic evidence of laminitis but were labeled normal in regard to laminar separation on MR images are likely early in the disease process.

Another potentially useful criterion in the evaluation of the laminitic horse is the use of the laminae-to-dermis ratio. This may have more clinical utility in lower field strength magnets where individual laminae may not be seen but a distinction between the corium and laminae is still possible. Based on our results, a laminae-to-dermis ratio greater than 0.7 is a good indicator of laminitis. It has been shown that the laminae lengthen and attenuate in laminitis.4,5,16 It may be inferred that an increased laminae-to-dermis ratio is likely due to mechanical breakdown at the level of the secondary laminae. Additionally, by using a ratio to evaluate the relative increase in laminae thickness, the cut point is likely to be influenced less than an arbitrary measurement if the angle of the image slice is not perpendicular to the dorsal hoof wall.

We noticed a trend of abnormalities being more evident in the distal region of the foot vs. the middle or proximal regions. The increase in maximum laminae-to-dermis ratio in the toe region of laminitic feet gave a greater disparity between histologically normal vs. laminitic feet, allowing greater sensitivity and specificity. These objective data support the toe region being most affected in chronic laminits.4 This is most evident with the loss of corium architecture and increase in corium signal intensity. This is consistent with the mechanical stresses being greatest in the toe region.

All subjective changes identified were highly correlated with a histologic diagnosis of laminitis. Additionally, despite the variation introduced by using three observers, the grading of each category was significantly different between histologically normal and laminitic feet. Further investigation is necessary to evaluate the specific association of MR changes to the specific histologic change. This process is hampered by the necessity to section the exact tissue in question on the MR image. One limitation of our study was that a single sample from the middle of the foot was obtained for histologic diagnosis. Another fact to consider is that the process of preparing specimens for histologic examination will alter the tissue samples such as the increased sheering stress from the cutting process and the dehydration of the tissue during processing and fixation. The latter impedes full evaluation for the presence of edema which remains a question in regard to the pathophysiologic mechanisms of laminitis. It is unclear whether full histological evaluation of the corium is possible or if the corium is overlooked during postmortem examination. Histologically, acute laminitis should display hyperemia, hemorrhage, marked edema, and occasional thrombosis within the dermis, and progression of the disease leads to laminar disorganization and necrosis of the laminar tips with limited inflammatory cell infiltration.6 This congestion and edema would explain the changes seen in the corium; however, upon histologic examination of the corium, there were no specific differences in vascularity or tissue organization between normal and laminitic feet.

While all radiographic measurements were characterized by a significant difference between normal and laminitic feet, there was a large overlap, illustrated by the fact that sensitivity for the dorsal hoof wall-to-distal phalanx length ratio at a cut-off value of 0.3 has a sensitivity of 52% and a positive predictive value of only 42%. In contrast, the maximum laminae-to-dermis ratio obtained with MR imaging has a much more segregated distribution between normal and laminitic feet with sensitivity of 81% and specificity of 93–98% depending on which portion of the foot is measured.

The results for all three observers were similar with regard to the subjective ability to make a diagnosis of a laminitic vs. normal foot using radiographs. The sensitivity for detecting initial active laminitis was low (47.6–57.2%), because until there is mechanical breakdown of the laminae, there are no distinct radiographic changes. The specimens that did have radiographic changes may have been in transition to chronic laminitis.23,24 Horses with radiographic changes are arguably classified as having chronic laminitis, according to commonly used definitions. For the purpose of our study, however, this was a continuation of the same inciting event and was within the initial active phase. The fact that changes such as laminar separation may be detected earlier with MR imaging before radiographic rotation and distal displacement occur may necessitate revisiting the definition between acute and chronic laminitis. Other than laminar separation, there were no histologic findings to indicate a chronic disease process.

There was a tendency for over-interpreting the MR images in making a subjective diagnosis of laminitis. Two factors need to be considered. First, the observers were only asked to evaluate three out of the 64 total images that comprised an entire study of the foot. The discontinuity made reading of each individual image more objective; however, this prohibited following structures over the entire length of the foot. Second, the observers were only given short tutorials on the specific changes they were detecting, and there are no published observations on what constitutes expected MR imaging findings in horses with acute laminitis.

With regard to the Group 2 horses that were at risk for developing laminitis, only two feet from one horse had histologic changes characteristic of laminitis. Most likely, this horse was misclassified because both histologic and MR imaging changes were diagnostic for laminitis. Unfortunately, using clinical evaluation to group the horses introduced another variable. This horse was endotoxemic which placed it in the at risk category; however, the horse was also being treated for a front foot abscess which has many clinical signs in common with laminitis, including foot pain, increased hoof wall temperature, and altered digital blood flow. An abscess was not identified on necropsy or on MR images, leading us to speculate that the clinical signs of laminitis were misinterpreted. It then follows that subclinical laminitis does not produce changes detectable with MR imaging, or was not present in Group 2 horses. This conclusion is premature given the small sample size and the limited staining and histologic description in our study.

We recognize that the fine anatomic detail seen at 4.7 T is currently not achievable clinically. However, distinction between the corium and laminae is possible in magnets that are used for clinical imaging, and the ratio of the laminae-to-dermal thickness within the foot and the changes seen within the corium can be extrapolated. Further work is necessary to characterize the changes seen with laminitis when using clinically available magnets. Further study into the pathophysiologic mechanisms of laminitis will benefit from studies of live animals, particularly with serial examinations.

In conclusion, alterations in laminar and corium architecture and signal intensity as well as separation within the laminae are readily seen at 4.7 T and are highly correlated with a histologic diagnosis of acute laminitis. The sensitivity of diagnosing laminitis is much higher with high field strength MR imaging than with radiography. Sensitivity and specificity are high for predicting laminitis using a laminae-to-dermis ratio of 0.7. The sensitivity and specificity of quantitative MR imaging measurements are better than those obtained by radiography. The sensitivity of subjective evaluation is higher with MR imaging than with radiography, and the specificity is likely to improve with greater familiarity of the MR appearance of the equine foot. Finally, there was no evidence to confirm the hypothesis that subclinical changes are present in the soft tissues of the digit in horses that are at risk for laminitis.

Footnotes

  1. *Eklin Medical Systems, Santa Clara, CA.

  2. †Merge Healthcare, Milwaukee, WI.

  3. ‡Bruker, Ettlingen, Germany.

Acknowledgments

The authors would like to thank Dr. Schmalbrock from the Department of Radiology, The Ohio State College of Medicine, for her valuable assistance and expertise and Dr. Wharton from Lexington, KY, for her aid in data collection.

Ancillary