Naturally occurring osteochondrosis latens lesions identified by quantitative and morphological 10.5 T MRI in pigs

Abstract

Juvenile osteochondritis dissecans (JOCD) is a pediatric orthopedic disorder that involves the articular–epiphyseal cartilage complex and underlying bone. Clinical disease is often characterized by the presence of radiographically apparent osteochondral flaps and fragments. The existence of early JOCD lesions (osteochondrosis latens [OCL] and osteochondrosis manifesta [OCM]) that precede the development of osteochondral flaps and fragments is also well recognized. However, identification of naturally occurring OCL lesions (confined to cartilage) using noninvasive imaging techniques has not yet been accomplished. We hypothesized that 10.5 T magnetic resonance imaging (MRI) can identify naturally occurring OCL lesions at predilection sites in intact joints of juvenile pigs. Unilateral elbows and knees (stifles) were harvested from three pigs aged 4, 8, and 12 weeks, and scanned in a 10.5 T MRI to obtain morphological 3D DESS images, and quantitative T2 and T1ρ relaxation time maps. Areas with increased T2 and T1ρ relaxation times in the articular–epiphyseal cartilage complex were identified in 1/3 distal femora and 3/3 distal humeri and were considered suspicious for OCL or OCM lesions. Histological assessment confirmed the presence of OCL or OCM lesions at each of these sites and failed to identify additional lesions. Histological findings included necrotic vascular profiles associated with areas of chondronecrosis either confined to the epiphyseal cartilage (OCL, 4- and 8-week-old specimens) or resulting in a delay in endochondral ossification (OCM, 12-week-old specimen). Future studies with clinical MR systems (≤7 T) are needed to determine whether these MRI methods are suitable for the in vivo diagnosis of early JOCD lesions in humans.

1 |. INTRODUCTION

Juvenile osteochondritis dissecans (JOCD) is a pediatric orthopedic disorder affecting the articular–epiphyseal cartilage complex and the underlying subchondral bone. Clinically apparent disease is characterized by pain and gait deficits that are associated with a focal delay of endochondral ossification and, in a subset of patients, development of chondro-osseous flaps and fragments. Lesions are most commonly diagnosed in the knee (distal femur), elbow (distal humerus), and ankle joints of young athletes and, if left untreated, they may progress to early-onset osteoarthritis.¹ Although the exact pathogenesis of JOCD in people remains incompletely understood, the ischemic theory, implicating disruption of endochondral ossification secondary to ischemic necrosis of the epiphyseal cartilage, has recently been gaining traction.^2,3

Much of the current understanding of JOCD in humans stems from extrapolating findings from osteochondrosis (OC) in animals, a disorder that shares the imaging, histologic and clinical features of JOCD.^4,5 Histological sentinel studies conducted at predilection sites in asymptomatic juvenile pigs and horses have identified two early stages of OC.^5,6 The earliest lesions, termed osteochondrosis latens (OCL), are confined to the epiphyseal cartilage and consist of areas of chondrocyte necrosis associated with necrotic cartilage canals, often accompanied by matrix degeneration. Delayed ossification of these necrotic areas of epiphyseal cartilage is known to result in radiographically apparent focal failure of endochondral ossification, consistent with progression to the next stage of the disease, termed osteochondrosis manifesta (OCM).^6–9 Exposure of OCM lesions to biomechanical trauma is believed to trigger formation of chondro-osseous flaps and fragments, which are characteristic features of the end stage disease, termed osteochondrosis dissecans (OCD). Importantly, regions of epiphyseal cartilage necrosis consistent with OCL and OCM have recently been identified in biopsy specimens obtained from JOCD predilection sites in cadavers of children, bolstering support for the theory of a shared pathogenesis between JOCD and OC.¹⁰ Identification of a characteristic, avascular region in the epiphyseal cartilage at the distal femoral JOCD/OC predilection sites using specialized magnetic resonance imaging (MRI) techniques in juvenile humans and pigs also supports the notion of shared pathophysiology across species.^3,11 Indeed, this avascular region was absent in juvenile goats, a species that does not naturally develop OC, further supporting the role of vascular failure in the pathogenesis of JOCD/OC.³

Due to limited access to cadaveric specimens of children for histological evaluation, further understanding of the pathophysiology of JOCD requires the development of a noninvasive approach that can identify and monitor the progression and/or resolution of OCL and OCM lesions in vivo. Implementation of quantitative and morphologic MRI techniques in animal models of orthopedic disorders that are characterized by ischemic necrosis of the epiphyseal cartilage (OC and Legg-Calvé-Perthes Disease) demonstrated the sensitivity of T2 and T1ρ relaxation time maps to acute ischemic injury, suggesting that these are suitable methods to further examine the pathophysiology of JOCD.^12–14 These quantitative cartilage mapping methods have been used to identify surgically induced OCL and OCM lesions in a goat model of OC by focusing on specific areas that were rendered ischemic after transection of cartilage canal vessels.^12,15 The use of these imaging methods to identify naturally occurring OCL and OCM lesions at predilection sites has not previously been accomplished. Doing so would demonstrate their value as research tools and diagnostic methods for children suspected of developing OCL or OCM lesions.

The objective of the current study was to demonstrate the utility of noninvasive MRI methods in identifying naturally occurring OCL and OCM lesions in the distal femoral and humeral predilection sites in cadaveric specimens of intact knee and elbow joints obtained from juvenile pigs. The nearly 100% prevalence of OCL and OCM lesions in both the distal femur and distal humerus in 12-week-old pigs, along with a joint size that is small enough to allow exhaustive histological evaluation for validation of the accuracy of the MRI techniques, supported the selection of these animals for the proposed study. Performing these studies at 10.5 T provided the highest sensitivity for identifying and characterizing these lesions with future studies planned to investigate the translatability of these techniques to in vivo studies at lower, clinical field strengths (≤7 T).

2 |. METHODS

2.1 |. Animals

Unilateral elbows (n = 3) and knees (stifles) (n = 3) were harvested immediately after euthanasia from three male Yorkshire-cross piglets aged 4, 8, and 12 weeks. These ages were chosen to include: (1) an age before that when OCL and OCM lesions at these sites have been reported (4 weeks); and (2) ages in which OCL and OCM lesions would likely be present (8 and 12 weeks, respectively).^16,17 The University of Minnesota Institutional Animal Care and Use Committee approved the reported experiments (IACUC # 2004-38082A).

2.2 |. Imaging

Intact joint specimens were scanned in a whole-body 10.5-T MRI scanner¹⁸ using an 8-channel dipole transceiver head coil.^19,20 Morphological images were acquired with a 3D DESS (double-echo steady-state; repetition time/echo time [TR/TE] = 21/6.2 ms, resolution = 0.3 × 0.3 × 0.3 mm³) using a 19-min scanning time. T2-weighted images were acquired with a 2D multiecho spin echo sequence with fat suppression (FS; TR = 5800 ms, 13 TEs = 10.4–135.2 ms, resolution = 0.3 × 0.3 × 1.5 mm³) and a 37-min scanning time. T1ρ-weighted images were acquired using an adiabatic T1ρ-prepared 3D turbo spin echo sequence (TR = 6000 ms, 6 spin-lock times = 0–80 ms, spin-lock frequency = 500 Hz, resolution = 0.45 × 0.45 × 2.0 mm³) and a scanning time of 25 min. Imaging parameters are listed in Table 1. T2 and T1ρ maps were calculated by fitting a mono-exponential function to MR signal decay using a nonlinear least-square fitting routine in MATLAB (v2019b; Math-Words). Clearly demarcated areas of increased signal intensity in 3D DESS, T2- and T1ρ-weighted images within the articular–epiphyseal cartilage complex (AECC) were considered suspicious for the presence of OCL and/or OCM lesions.

TABLE 1.

10.5 T MRI imaging parameters

Acquisition parameters	3D-DESS	T2 FS	T1ρ
Sequence	Double Echo Steady State	2D Multi Echo Spin Echo	3D Turbo Spin Echo
Repetition time (ms)	21	5800	6000
Echo time (TE)/spin lock (SL) time (ms)	6.2	13 TEs: 10.4–135.2	6 SL times: 0–80
Echo train length	2	13	280
Number of averages	1	1	1
Field of view (mm)	128 × 128	128 × 128	173 × 173
Matrix size	384 × 384	384 × 384	384 × 384
Resolution (mm)	0.33 × 0.33	0.33 × 0.33	0.45 × 0.45
Slices/thickness (mm)	144–288/0.33	26–30/1.5	40/2
Flip angle (degrees)	60	180	180
Pixel bandwidth (Hz/px)	128	250	766
Spin-lock frequency (Hz)	n/a	n/a	500
Scan time (min)	18:42	37:30	24:36

Abbreviation: MRI, magnetic resonance imaging.

2.3 |. Histology

At the conclusion of the MRI sessions, the joints were disarticulated and distal humeri and femora were fixed in 10% neutral buffered formalin, decalcified in 10% ethylenediaminetetracetic acid, and sagittally sectioned into serial 2.0 mm thick slabs that spanned the total thickness of the epiphysis. On gross examination, a focal region accompanied by delay of endochondral ossification was present in the 12-week-old humeral specimen at the trochlea. The other specimens were grossly normal. Individual slabs (approximately 7–9/site) were processed into paraffin blocks for histological evaluation. At least two 5-μm-thick sections were collected from the surface of each slab and stained with hematoxylin & eosin (H&E). Additional serial sections at 50-μm intervals were obtained and stained with H&E to confirm identification of all OCL and OCM lesions present in the evaluated specimens (approximately 4–10 sections per suspected lesion). Histological sections were assessed by a blinded, board-certified veterinary pathologist with experience in musculoskeletal pathology. OCL lesions were defined as areas of chondronecrosis associated with necrotic vascular profiles that were confined to the epiphyseal cartilage.⁸ OCM lesions were defined as areas of chondronecrosis in the epiphyseal cartilage that were accompanied by a delay in endochondral ossification.⁸

2.4 |. MRI data analysis

First, 3D DESS, T2- and T1ρ- weighted images were qualitatively evaluated by a blinded investigator for the presence of discrete areas of increased signal intensity suggestive of the presence of OCL and OCM lesions. Subsequent quantitative evaluation of the MRI images was performed by identifying ROIs in the T2 and T1ρ maps that corresponded with the location of histologically identified OCL (areas of chondronecrosis) and OCM (areas of chondronecrosis with delay of endochondral ossification) lesions. For precise delineation of the ROIs, the position (i.e., distance from articular cartilage surface) and size of the lesions were first measured in ImageJ (Version 1.53a) in the histological sections by a board-certified veterinary pathologist then coregistered in the corresponding MR images. In samples with multiple discrete lesions, the largest lesion was used for MRI assessment and statistical analysis. Control ROIs were designated in comparably sized, histologically unaffected areas adjacent to but discrete from the lesions. Average T2 and T1ρ relaxation times were calculated from the ROIs comprising each OCL and OCM lesion and control regions.

2.5 |. Statistical analysis

Mean T2 and T1ρ relaxation times were both compared between the lesion and control ROIs using paired t-tests. First, the comparison was performed separately for the OCL and for the OCM lesions and their respective control ROIs. Then, comparisons were made across all lesions (OCL and OCM) and control ROIs. Statistical analyses were performed using GraphPad Prism (version 9.2.0 for macOS; GraphPad Software). For all tests, p < 0.05 was considered statistically significant.

3 |. RESULTS

3.1 |. Histological assessment

Histological analysis of sections corresponding with MRI slices containing suspected early JOCD lesions confirmed the presence of OCL and/or OCM lesions at each of those locations (Table 2), which included all distal humeri (3/3) and one distal femur (1/3; 12-week-old) (Figures 1–4). Comprehensive histological evaluation of all distal humeri and femora failed to identify any additional OCL or OCM lesions that were not identified by MRI.

TABLE 2.

Histological characteristics of lesions identified by 10.5 T MRI

Specimen	Joint affected	Site	Maximum widtha (mm)	Maximum thicknessb (mm)	Histologic features	Diagnosis
4-week	Elbow	Humerus—trochlea	2.5	1.5	Epiphyseal chondronecrosis Necrotic vascular profiles	OCL
8-week	Elbow	Humerus—trochlea	3.3	1.5	Epiphyseal chondronecrosis Necrotic vascular profiles	OCL
12-week	Elbow	Humerus—trochlea	6.5	2	Epiphyseal chondronecrosis Necrotic vascular profiles Delay of endochondral ossification Cleft formation without displacement Bone resorption, Myelofibrosis	OCL, OCM
12-week	Knee/stifle	Femur—medial condyle	6.5	5.4	Epiphyseal chondronecrosis Necrotic vascular profiles Delay of endochondral ossification	OCM

Note: Italics denote diagnostic features.

Abbreviations: MRI, magnetic resonance imaging; OCL, osteochondrosis latens; OCM, osteochondrosis manifesta.

^aWidth of lesions were measured in the serial histological section that captured the lesions at their largest point and refers to the greatest extent of the lesion as measured in the plane approximately parallel to the underlying chondrosseous junction.

^bThickness refers to the maximum thickness of affected cartilage as measured in the plane perpendicular to the chondrosseous junction.

FIGURE 1.

OCL lesion at the distal humerus of a 4-week-old pig. (A) A hyperintense region within the cartilage in the 3D DESS image (indicated by white arrow) that corresponds to (B) an area with increased T2 values (orange and yellow) within the articular epiphyseal cartilage complex (black arrow) in the T2 map with (C) corresponding histological section showing associated region of chondrocyte necrosis with matrix pallor (outlined in black), consistent with an OCL lesion in the distal humerus of a 4-week-old pig. White clefts between the cartilage (purple) and marrow and bone of the epiphysis and metaphysis are artifactual. H&E, 0.5× magnification. (D) Corresponding T1ρ map with increased T1ρ values in the OCL lesion (black arrow). (E) Higher magnification image (20×) showing localized area of chondronecrosis centrally (bordered by dotted lines) characterized by matrix pallor and chondrocytes that are pyknotic and eosinophilic with loss of detail (chondrocyte cell death). H&E, 10× magnification; scale bar = 100 μm. (F) High magnification image (40×) of the area indicated by the box bordered by dashed lines in (E) demonstrating necrotic chondrocytes (black arrowheads) with diffuse eosinophilia and condensed morphology as compared to unaffected neighboring chondrocytes (lower left). H&E, 40× magnification; scale bar = 20 μm. H&E, hematoxylin & eosin; OCL, osteochondrosis latens.

FIGURE 4.

OCM lesion at the distal femur of a 12-week-old pig. (A) A hyperintense region (white arrow) within the cartilage in the 3D DESS image (white to pale gray) that corresponds to (B) an area with increased T2 values (orange to yellow) in a wedge shape within the articular epiphyseal cartilage complex (black arrow) in the T2 map with (C) corresponding histological section showing a region of focal delay of endochondral ossification and chondrocyte necrosis with matrix pallor (black circle), consistent with OCM in the distal femur of a 12-week-old pig. H&E, 0.5× magnification; scale bar = 5 mm. (D) Corresponding T1ρ map with increased T1ρ values (orange to yellow) in the OCM lesion (black arrow). (E) High magnification image oriented with the articular surface towards the top of the panel (not pictured) showing focal delay of endochondral ossification (black arrows) and pale matrix and necrotic chondrocytes. H&E, 4× magnification; scale bar = 500 μm. (F) High magnification image of the area indicated by the box in (E) demonstrating necrotic chondrocytes (black arrowheads) with diffuse eosinophilia and pallor of the associated matrix surrounding a necrotic vascular profile (*). H&E, 20× magnification; scale bar = 50 μm. H&E, hematoxylin & eosin; OCM, osteochondrosis manifesta.

Solitary lesions were identified in one humeral (OCL, 4-week-old) and one femoral (OCM, 12-week-old) specimen, while the 8-week-old humeral specimen had three distinct OCL lesions, the largest of which was used for the quantitative comparisons (Figure 2). Evaluation of the 12-week-old humeral specimen demonstrated the presence of an OCM lesion (Figure 3) which was accompanied by an additional, separate, and discrete area of epiphyseal cartilage necrosis that was diagnosed as a small OCL lesion. All humeral lesions were located at the trochlea of the distal humerus, whereas the distal femoral lesion was located at the medial condyle. The OCL lesions (n = 5) were discrete areas within the epiphyseal cartilage characterized by loss of staining (pallor) of the matrix surrounding a central arboriform vascular profile with necrotic endothelium and necrosis of the adjacent chondrocytes. Both OCM lesions (n = 2) were identified in the 12-week-old pig, one in the distal humeral trochlea and one in the medial femoral condyle. The humeral OCM lesion included a localized area of necrosis and clefting at the chondrosseous junction, with expansion of the underlying marrow spaces by granulation tissue and resorption of subchondral bone by osteoclasts, indicative of some degree of chronicity (Figure 3). In contrast, the OCM lesion at the distal femur lacked bone resorption and fibrosis (Figure 4). This lesion was characterized by a focal delay of endochondral ossification accompanied by vascular necrosis affecting a central and multiple regional vascular profiles within the epiphyseal cartilage.

FIGURE 2.

OCL lesions at the distal humerus of an 8-week-old pig. (A) Three distinct neighboring hyperintense (white to pale gray) regions within the cartilage in the 3D DESS image (white arrows) that corresponds to (B) areas with increased T2 values (orange and yellow) within the articular epiphyseal cartilage complex (black arrows) in the T2 map with (C) corresponding histological section showing associated regions of chondrocyte necrosis with matrix pallor (outlined in black), consistent with OCL lesions in the distal humerus of an 8-week-old pig. H&E, 0.5× magnification. (D) Corresponding T1ρ map with increased T1ρ values (orange and yellow) in the OCL lesions (black arrows). H&E, hematoxylin & eosin; OCL, osteochondrosis latens.

FIGURE 3.

OCM lesion at the distal humerus of a 12-week-old pig. (A) A hyperintense region within the cartilage in the 3D DESS image (white arrows) that corresponds to (B) an area with increased T2 values (yellow to green) within the articular epiphyseal cartilage complex (black arrows) in the T2 map with (C) corresponding histological section showing associated region of thickened articular–epiphyseal cartilage complex with delay of endochondral ossification and locally extensive subjacent chondroid matrix necrosis (outlined in black), consistent with OCM in the distal humerus of a 12-week-old pig. H&E, 0.5× magnification. (D) Corresponding T1ρ map with increased T1ρ values (orange and yellow) in the OCM lesion (black arrows). (E) High magnification image oriented with the articular surface towrads the top of the panel (not pictured) showing clefting between the thickened articular–epiphyseal cartilage complex and the subjacent area of necrosis, with fibrosis extending into the subchondral bone (bright pink). H&E, 10× magnification; scale bar = 100 μm. (F) High magnification image (40×) of the area within the box in (E) demonstrating necrotic chondrocytes (black arrowheads), overlying a cleft (*) and accumulation of eosinophilic necrotic debris. H&E, 40× magnification; scale bar = 50 μm. H&E, hematoxylin & eosin; OCM, osteochondrosis manifesta.

The maximum extent of the OCL lesions, measured in histological sections, were 2.5 and 3.3 mm in the 4- and 8-week specimens, respectively. The OCM lesions in the 12-week humeral and femoral specimens both had a maximum extent of 6.5 mm (Table 2).

3.2 |. Morphological and quantitative MRI assessment

The mean T1ρ and T2 relaxation time values were increased by 20% and 26.9% in the 4-week-old humeral OCL ROI and by 43.9% and 33.7% in the 8-week-old humeral OCL ROI compared to their respective control ROIs (Table 3). Values in the 12-week-old specimens were increased by 10.3% (T1ρ) and 5.9% (T2) in the humeral OCM ROI and 15.4% (T1ρ) and 12.6% (T2) in the femoral OCM ROI compared to their respective control ROIs (Table 3). The humeral OCL lesion in the 12-week-old specimen was omitted from the quantitative evaluation due to its small size/indistinct borders. The comparison of all lesion ROIs (OCL and OCM) to their paired control ROIs showed a statistically significant increase in both the mean T2 (n = 4, p = 0.047) and mean T1ρ (n = 4, p = 0.035) relaxation times. The magnitude of this increase was similar for both the T1ρ and T2 values (Table 3). The OCM lesion in the 12-week-old humerus had increased T2 and T1ρ values within the AECC, coupled with small increases in signal extending into the subjacent epiphysis within the area of focal thickening of the epiphyseal cartilage (Figure 3).

TABLE 3.

Comparison of T2 and T1ρ relaxation times between lesion ROIs and control ROIs

Specimen	Control T2 (ms)	Lesion T2 (ms)	T2 increase	Control T1ρ (ms)	Lesion T1ρ (ms)	T1ρ increase
4-week Humerus	70.65	89.65	26.9%	78.75	94.81	20.0%
8-week Humerus	65.21	87.22	33.7%	64.45	92.78	43.9%
12-week Humerus	67.49	71.49	5.9%	76.45	84.44	10.3%
12-week Femur	74.56	83.96	12.6%	77.39	89.34	15.4%
Mean ± SD	69.5 ± 4.1	83.1 ± 8.1	19.8 ± 12.8%	74.3 ± 6.6	90.3 ± 4.5	22.4 ± 14.9%
p-value		0.047			0.035

Note: p-values based on paired t-test between mean lesion and mean control ROI relaxation times.

Abbreviation: SD, standard deviation.

Additionally, qualitative visual assessment of 3D DESS images showed increased signal intensities in the locations corresponding to OCL and OCM lesions and no increased signal was detected in the control regions as demonstrated in Figures 1A–4A.

4 |. DISCUSSION

Our results demonstrate that quantitative MRI methods detect both naturally occurring OCL and OCM lesions at the distal humeral and femoral predilections sites ex vivo in intact joints in juvenile pigs. Blinded, qualitative evaluation of the 3D DESS, T2- and T1ρ- weighted images identified each OCL and OCM lesion present in the specimens with no false positive or false negative findings, as validated by histology. Quantitative evaluation of the relaxation time maps confirmed the initial findings by demonstrating significantly increased T2 and T1ρ relaxation times within the lesion ROIs compared to adjacent control ROIs. These findings suggest that T2 and T1ρ cartilage maps have strong potential as diagnostic tools to identify naturally occurring OCL and OCM lesions, warranting their further evaluation in vivo at a clinically relevant field strength (≤7 T).

The mean T2 and T1ρ relaxation times across all measured OCL and OCM lesions were significantly increased when compared to their respective control regions. The difference in T2 and T1ρ relaxation times when averaged for the 4- and 8-week-old specimens between areas of chondronecrosis (OCL) and controls were 30.3 ± 4.8% and 32 ± 16.9%, respectively. The differences in the present study were somewhat smaller in magnitude than those observed in regions of induced chondronecrosis in a goat model of OC, which had a mean percent difference of 58.4 ± 16.7% and 54.4 ± 30.4% for T2 and T1ρ, respectively.¹⁵ The variation in magnitude of the difference between the relaxation times in lesions and normal regions may be related to surgical induction causing a complete disruption of local blood supply versus a less extensive vascular failure occurring in natural lesions. Another possible cause may be the stage of development in the respective animal models, with the goats undergoing surgery to induce ischemia at 4 days of age while the pigs assessed in this manuscript likely developed ischemia and subsequent lesions at a later timepoint.¹⁵ The different tissue composition of OCM lesions, including retention of necrotic, and in some cases nonnecrotic, cartilage accompanied by remodeling of the adjacent subchondral bone and marrow, coupled with the greater duration of the lesions, is likely responsible for the smaller differences in T2 and T1ρ in the 12-week-old specimens (humeral ΔT2 = 5.9% and ΔT1ρ = 10.3%, femoral ΔT2 = 12.6% and ΔT1ρ = 15.4%) as compared to the differences in OCL lesions in the 4- and 8-week-old specimens. Comparing the magnitude of change in T1ρ and T2 relaxation times across all lesions and their respective controls demonstrated that both of these indices increased by a similar degree, with both techniques identifying increased relaxation times associated with lesions in each specimen.

All humeral lesions were located at the humeral trochlea medial to the capitellum, which is the most frequent predilection site for elbow OC in pigs, while JOCD in children more frequently affects the capitellum.^21,22 Interestingly, a lesion was identified in the 4-week-old humeral specimen, at an earlier age than OCL lesions have been previously reported to develop.⁴ This may suggest that vascular vulnerability at the elbow joint occurs earlier than at the more frequently affected femoral sites, the area examined in the vast majority of JOCD-OC studies. Furthermore, the 12-week-old humeral lesion exhibited resorption of subchondral bone and replacement of marrow spaces by granulation tissue, suggesting ongoing reparative processes. This is consistent with our previous findings that the majority of OCL and OCM lesions undergo repair, with only a small subset progressing to clinical disease in pigs.²³ Taken together, our findings suggest that it is possible that humeral lesions in pigs occur at younger ages and develop chronic changes earlier than femoral lesions. Unsurprisingly, the femoral lesion was found in the medial femoral condyle, the primary predilection site of JOCD/OC in the pelvic limb both in children and piglets.^9,24

Importantly, our study marks the first time that naturally occurring OCL lesions were identified noninvasively in any species and at any location. In pigs, OC is initially characterized by a focal area of chondronecrosis within the growth cartilage (OCL) at regions of vascular vulnerability, which correspond to anatomically similar regions of human vascular vulerabilities.³ In a subset of cases, these regions of chondronecrosis are large enough that they interfere with progression of the ossification front, leading to a focal failure of endochondral ossification (OCM). OCL lesions in humans are also suspected to progress to OCM, which may undergo stages of osseous repair, including peripheral mineralization, progressive ossification, and osseous bridging between the lesion and the parent bone in the course of healing.^8,25,26 This process may fail in some cases, resulting in a “textbook” case of JOCD characterized by articular cartilage clefting and formation of an osteochondral fragment or “loose body” within the joint.²⁷ Similarly, studies investigating the pathophysiology of OC in growing pigs successfully implemented sequential CT scans to monitor the progression and/or resolution of OCM lesions.^17,28 These studies demonstrated the utility of this technique in screening pigs for early lesions of OC and established a 51%–69% spontaneous healing rate for OCM lesions depending on the joint affected.¹⁷ The MRI methods outlined here improve upon these previously reported diagnostic and monitoring methods by not only providing a sensitive and specific means of detecting both OCM and OCL lesions, but by having greater translational potential for pediatric populations due to the absence of ionizing radiation or contrast material. MRI methods such as T2 and T1ρ may provide additional benefits given their quantitative nature and sensitivity to biochemical changes, with studies having shown a correlation of T2 and T1ρ relaxation times to collagen, proteoglycan, and water content as well as to age-related changes in tissue composition and structure.^29–31 3D-DESS allows evaluation of the cartilage morphology and detection of structural defects in the cartilage by providing excellent contrast between the synovial fluid and cartilage.³² While 3D-DESS was used only for the morphological evaluation in the present study, this technique was also capable of identifying regions of chondronecrosis without communication with the synovial fluid, confirming its sensitivity to areas with increased T2 relaxation times within OCL lesions. 3D-DESS is a readily available clinical sequence that does not need extensive postprocessing, suggesting that this technique may be particularly useful for screening pediatric patients.^33,34 While human studies have been performed at 10.5 T under an investigational device exemption and local IRB approval,^35,36 it is currently not approved for pediatric or patient studies. Clinical translation of these sequences for the diagnosis of elbow OCD in human patients will necessitate their further evaluation at lower field strengths (≤7 T) which have the additional benefit of being clinically approved. Prolonged scan times, another important factor in pediatric patients, will also have to be addressed before these methods are implemented in clinical practice.³⁴

The clinical relevance of our findings is underscored by the recent description of lesions at JOCD predilection sites in randomly sampled pediatric cadavers that were morphologically identical to OCL and OCM lesions seen in animal species.¹⁰ Given the observed propensity for healing of OCL and OCM lesions in pigs, it is reasonable to assume that the prevalence of early asymptomatic lesions in children is higher than the prevalence of symptomatic JOCD lesions.¹⁷ This is supported by the findings of a recent in vivo 3 T study of 25 patients with 34 JOCD-affected knees that found JOCD lesions in five asymptomatic knees.²⁵ Furthermore, this 3 T study used T2* relaxation time mapping as a measure of the degree of lesion ossification and was able to distinguish between the different stages of disease progression—from the early, predominantly cartilaginous, to the late and almost entirely ossified JOCD lesions.²⁵ Sensitive and specific noninvasive methods for detecting and monitoring these early lesions may provide insight into factors that contribute to either healing or progression to clinical disease, which could ultimately impact clinical recommendations. Based on our findings, reasonable next investigative steps include the application of 3D-DESS, T2, and T1ρ techniques to pediatric cadaveric specimens to identify ex vivo OCL and OCM lesions with a clinical (≤7 T) MR system, followed by targeted histological assessment to confirm their presence.

The small sample size may be considered a limitation of this study; however, the high prevalence of OCL and OCM lesions in growing pigs (present in four out of six specimens in the present study) and the confirmation of all lesions by exhaustive histological evaluations, considered the gold standard for identification of early-stage JOCD lesions, supports the ability of these MRI methods for the detection of OCL and OCM lesions. Although our study was completed using a 10.5 T whole body research magnet and optimized RF coil to maximize both signal to noise ratio and resolution, there is reason to believe that these techniques would also be effective in identifying OCL and OCM lesions at clinical field strengths, as demonstrated in a previous study that successfully used T2 cartilage maps at 3 T to identify surgically-induced lesions of chondronecrosis in vivo in a goat model.¹² Indeed, during the past 10 years, novel MRI methods have been extensively applied to shed light on the pathogenesis of JOCD/OC and to describe the evolution of JOCD lesions in children.^{3,11,25–27,37}

In conclusion, our results demonstrate that T2 and T1ρ relaxation time mapping methods are highly sensitive and specific to OCL lesions (characterized by chondronecrosis) and OCM lesions (characterized by chondronecrosis accompanied by delayed endochondral ossification). Future clinical implementation of these sequences may allow noninvasive identification and monitoring of early JOCD lesions and determination of risk factors that contribute to their progression in children.