Osteochondritis Dissecans

The phenomenon of osteochondral loose bodies within a joint was termed osteochondritis dissecans (OCD) by König in 1887^1,2. OCD is conventionally defined as a focal idiopathic alteration of subchondral bone with risk of instability and disruption of adjacent articular cartilage that may result in premature osteoarthritis³. The etiology is generally believed to be multifactorial, but in-depth understanding is lacking partially because of the difficulty of investigating OCD before the development of overt clinical symptoms. Treatment is largely based on skeletal maturity and lesion stability, defined by the presence or absence of articular cartilage fracture and subchondral bone separation, with the goals of symptomatic relief and joint preservation⁴. The literature has grown substantially since the 1996 Current Concepts Review article by Schenck and Goodnight⁵. A multicenter group, Research in OsteoChondritis of the Knee (ROCK), was formed in 2010 to investigate the etiology and diagnostic criteria, optimize prevention and treatment strategies, reduce variation in care, and improve quality of life⁶. The present Current Concepts Review article focuses on OCD of the knee, elbow, and talus, recognizing them as a common entity while highlighting important distinctions.

Epidemiology

OCD is divided into juvenile and adult forms, depending on skeletal maturity at diagnosis^5,7. Most adult lesions are likely unresolved juvenile lesions, but de novo adult OCD has been described⁸. Adult OCD has an incidence rate of 3.42 per 100,000 person-years and predominantly affects ankles, followed by knees⁹. Juvenile OCD is more common, affecting 9.5 to 29 of 100,000 knees, 2.2 of 100,000 elbows, and 2 to 4.6 of 100,000 ankles^10–15. Clinically apparent OCD is rare in patients who are <10 or ≥50 years old^10,11,13,14. The risk of juvenile OCD is higher in patients who are ≥12 years old than in those between 6 and 11 years old and is 3 times higher in knees, 22 times higher in elbows, and 7 times higher in ankles in the older children than in the younger group^9,11,13,14. Boys have more risk of developing OCD in the knee (4 times) and elbow (7 times), whereas girls have more risk of OCD developing in the talus (1.5 times)^11,13,14.

OCD predilection sites differ for each joint. The classic site of knee OCD is the posterior-central aspect of the medial condyle (63.6% to 85%) compared with inferior-central aspect of the lateral condyle (15% to 32.5%), inferomedial aspect of the patella (1.5% to 10%), and trochlea (2%)^11,16–20. Elbow OCD typically affects the anterolateral or central aspect of the capitellum (97.5%) rather than the trochlea (2.5%)¹⁴. Most OCD lesions of the talus are posteromedial (71.8%) and are less frequently anterolateral (22.4%) or central (3.5%)¹³. OCD usually involves a single site but can be multifocal within one or multiple joints with lesions in different stages²¹. Bilateralism occurs in knees (7.3% to 29%) and elbows (8.1%), and is rare in ankles^{11,13,14,22,23}.

OCD tends to affect active children and young adults^8,19. Participation in high-level athletics is associated with the development of OCD in the knee (55% to 60%), elbow (84%), and talus (67.4%)^13,14,16,20. Knee and talar OCD cases are commonly associated with soccer, football, and basketball, whereas elbow OCD is often seen in throwing athletes and gymnasts^13,14,24,25. Interestingly, children with extreme obesity have more risk for the development of elbow and talar OCD, whereas moderate obesity increases the risk of knee OCD²⁶. An observed trend of younger onset of OCD may be related to earlier sports participation or specialization and increasing childhood obesity^{11,13,14,19,27–29}.

Etiology

The etiology of OCD is incompletely understood. Despite nomenclature implying an inflammatory process (osteochondritis), minimal supporting evidence exists, suggesting that the condition would be more appropriately named “osteochondrosis,” as it is in veterinary medicine. Growing evidence from the veterinary literature has favored ischemic necrosis of epiphyseal cartilage as a key event^30–32. Indeed, histochemical studies of human specimens have found subchondral bone and epiphyseal cartilage necrosis, matrix metalloproteinase expression consistent with chondrocyte apoptosis, and cartilage degradation reminiscent of osteoarthritis^6,33,34. Understanding has been confounded by similarities to osteochondral fractures and osteonecrosis associated with hemoglobinopathy, steroids, radiation, and chemotherapy^18,35,36. A recent systematic review concluded that the etiopathogenesis of OCD is likely a convergence of biological and mechanical factors³⁷.

Vascular Factors

The function and survival of subarticular epiphyseal cartilage, which determines adult joint geometry, depend on blood supply from cartilage canal vessels. These vessels regress during development as the secondary ossification center expands^38,39. However, premature interruption of these vessels in animals, naturally or surgically, causes epiphyseal cartilage necrosis, termed osteochondrosis latens^32,40,41. This lesion is initially confined to epiphyseal cartilage without articular cartilage or subchondral bone involvement. When the ossification front reaches the lesion, failure of endochondral ossification occurs, and an island of necrotic epiphyseal cartilage is retained within subchondral bone, termed osteochondrosis manifesta, which provides poor mechanical support to overlying cartilage⁴². Animal studies have demonstrated that these subclinical lesions usually heal without intervention; however, exposure to biomechanical trauma in a single event or repeated events may create osteochondral lesions, termed osteochondrosis dissecans (Fig. 1)⁴³.

Fig. 1.

Figs. 1-A through 1-J Diagrams and photomicrographs demonstrating the pathogenesis of osteochondrosis in animals (Figs. 1-A through 1-G) and OCD in human pediatric cadavers (Figs. 1-H, 1-I, and 1-J). (Figures 1-A through 1-G modified, with permission, from: McCoy AM, Toth F, Dolvik NI, Ekman S, Ellermann J, Olstad K, Ytrehus B, Carlson CS. Articular osteochondrosis: a comparison of naturally-occurring human and animal disease. Osteoarthritis Cartilage. 2013 Nov;21[11]:1638–47; Ytrehus B, Carlson CS, Ekman S. Etiology and Pathogenesis of Osteochondrosis. Vet Pathol. 2007 Jul;44[4]:429–48; and Ekman S, Carlson C. The pathophysiology of osteochondrosis. Vet Clin North Am Small Anim Pract. 1998 Jan;28[1]:17–32; and Figures 1-H, 1-I, and 1-J modified from: Tóth F, Tompkins MA, Shea KG, Ellermann JM, Carlson CS. Identification of areas of epiphyseal cartilage necrosis at predilection sites of juvenile osteochondritis dissecans in pediatric cadavers. J Bone Joint Surg Am. 2018 Dec 19;100[24]:2132–39.) Fig. 1-A Normal endochondral ossification. Fig. 1-B Development of osteochondrosis latens because of failure of the cartilage canal blood supply, resulting in epiphyseal cartilage necrosis. Fig. 1-C Osteochondrosis manifesta lesion develops as a delay in the progression of the ossification front. Figs. 1-D and 1-E Healing of an osteochondrosis manifesta lesion by incorporation into subchondral bone. Figs. 1-F and 1-G Development of an OCD lesion as a result of trauma, causing collapse of articular cartilage overlying areas of necrotic epiphyseal cartilage. Fig. 1-H Hematoxylin and eosin staining of articular and epiphyseal cartilage from the central aspect of the medial femoral condyle from a 4-year-old female donor containing an osteochondrosis latens lesion. Arrowheads demarcate necrotic epiphyseal cartilage. Right: High-magnification image demonstrates cartilage canal containing multiple degenerate vessels. Fig. 1-I Hematoxylin and eosin staining of articular and epiphyseal cartilage from the central aspect of the medial femoral condyle from a 9-year-old male donor containing an osteochondrosis manifesta lesion. Right: High-magnification image demonstrates necrotic epiphyseal cartilage partially surrounded by bone. Fig. 1-J Hematoxylin and eosin staining of articular and epiphyseal cartilage from the central aspect of the lateral femoral condyle from a 4-year-old female donor containing a healing osteochondrosis manifesta lesion. Arrow points to a degenerate cartilage canal. Right: High-magnification image demonstrates necrotic epiphyseal cartilage completely surrounded by bone.

Histologic studies have identified lesions identical to osteochondrosis latens and manifesta at predilection sites in human cadaveric specimens⁴². Furthermore, 3-dimensional quantitative magnetic resonance imaging (MRI) demonstrating vascular regression and delayed secondary ossification at predilection sites of OCD in pediatric cadaveric knees supports a vascular etiology (Fig. 2, Video 1)³⁹. Indeed, susceptibility-weighted MRI revealed common vascular patterns in developing distal femora of humans and pigs, species that share identical predilection sites³⁸. Interestingly, goats, which have not been reported to develop distal femoral osteochondrosis, have dense vascular networks that seemingly protect against the condition³⁸. Tóth et al. demonstrated that, even after surgical interruption of distal femoral vasculature, most goats only develop subclinical osteochondrosis⁴⁰. Tenuous epiphyseal cartilage vasculature is a shared feature among predilection sites across species susceptible to OCD^44–46.

Fig. 2.

Figs. 2-A and 2-B Three-dimensional quantitative MRI demonstrating vascular watershed zones and regression. Fig. 2-A Three-dimensional reconstruction of susceptibility-weighted MRI data obtained from a distal femoral specimen from a 3-month-old human cadaveric donor demonstrating axial (central) and abaxial (peripheral) vascular beds with an avascular region in the midsagittal plane. Fig. 2-B Visualizations and density changes in vascular network with age. Maximum-intensity projections of the quantitative susceptibility mapping-post-processed vessel-enhanced (vesselness) maps from 6 selected cadaveric specimens with increasing age are shown. The vasculature is color-coded according to the vascular network (lateral peripheral = green vessels, central = gray, and medial peripheral = red). No central network was observed in the specimens from donors who were ≥4 years old. By 8 years of age, there was a small, residual peripheral vascular network. At 10 years of age (not shown), no substantial epiphyseal cartilage vascularity could be detected. A = anterior, L = lateral, M = medial, and P = posterior. (Modified, with permission, from: Ellermann JM, Ludwig KD, Nissi MJ, Johnson CP, Strupp JP, Wang L, Zbýň Š, Tóth F, Arendt E, Tompkins M, Shea K, Carlson CS. Three-dimensional quantitative magnetic resonance imaging of epiphyseal cartilage vascularity using vessel image features: new insights into juvenile osteochondritis dissecans. JBJS Open Access. 2019 Dec 5;4[4]:e0031. Reproduced with permission.)

Endochondral Ossification-Related Factors

Ribbing initially described separation of an accessory epiphyseal ossification center that progressed to OCD after subsequent trauma⁴⁷. It was later proposed that insult to the secondary subarticular physis, which fails to heal or forms irregular subchondral tissue, predisposes to OCD^5,6,48,49. Laor et al. described chondro-epiphyseal widening and subchondral edema on MRI studies of children with knee OCD, implicating an ossification abnormality⁵⁰. Similarly, an advanced MRI study of juvenile OCD demonstrated an epiphyseal cartilage etiology with delayed endochondral ossification⁵¹. Tóth et al. identified epiphyseal cartilage necrosis adjacent to failed endochondral ossification at OCD predilection sites in pediatric cadaveric knees⁴². In animal models, surgically induced ischemic necrosis of epiphyseal cartilage is associated with delayed endochondral ossification, creating a vulnerable area where articular cartilage is only supported by necrotic tissue⁴⁰.

Biomechanical Factors

A history of mild trauma is commonly reported by patients with OCD in the knee (40%), capitellum (33%), and talus (47% to 80%)^7,13,36,52. OCD secondary to acute trauma, such as ankle sprains frequently associated with lateral talar osteochondral lesions, is likely a different entity although treatment may be similar^52–54. Repetitive loading triggering vascular compromise is a widely proposed theory as there is often an association between OCD and athletic activities (55% to 60%)^8,10,16,20. Capitellar OCD predominantly occurs in the dominant arm of throwing athletes and gymnasts subject to valgus or axial stress³⁶. Baseball catchers tend to develop knee OCD at a younger age and in a more posterior location in the knee compared with noncatchers, likely from loading in hyperflexion⁵⁵. Juvenile OCD of the femoral trochlea is most common in basketball, football, and soccer players exposed to forces across the patellofemoral joint^24,56. Correlation with anatomic variations has also been suggested, including narrower intercondylar width, a larger tibial spine, greater posterior-medial tibial slope, dysmorphic femoral condyles, and more distal femoral attachment of the posterior cruciate ligament^57–62. Studies have associated meniscal injury with the development of OCD in the less common lateral aspect of the femoral condyle^20,63,64. Moreover, correlation has been reported between knee OCD and alignment, in which genu varum predisposed to medial condylar lesions and genu valgum predisposed to lateral condylar lesions, and the mechanical axis deviation has been found to significantly coincide with unstable knee lesions^65–67.

Genetic Factors

Genetic predisposition was initially proposed after Mubarak and Carroll reported the development of OCD in 12 family members over 4 generations⁶⁸. It was disputed by Petrie in his study of 34 patients, in which only 1.2% (1) of 86 first-degree relatives was found to have OCD⁶⁹. Multiple case reports have since described monozygotic and fraternal twins with OCD of the knee, capitellum, and talus^70–77. A recent study found that 14% (14) of 103 children treated for knee OCD had positive family histories⁷⁸. OCD is found in various inherited conditions, including dwarfism, tibia vara, Legg-Calvé-Perthes disease, and Stickler syndrome^35,79–81. It has been considered a subgroup of epiphyseal dysplasia with abnormal accessory ossification nuclei^47,82. Familial OCD is an autosomal dominant syndrome involving the aggrecan gene that manifests as multiple joint OCD, disproportionate short stature, and early osteoarthritis^83–85. Recently, a genome-wide association study in humans identified genetic loci related to juvenile OCD^86,87.

Diagnostic Strategies

History and Physical Examination

Patients with OCD may not recall trauma. Symptoms depend on the lesion location and stage. Stable lesions can cause nonspecific symptoms, including vague or intermittent pain, whereas unstable lesions or loose bodies can cause mechanical symptoms, including catching or locking. For medial femoral condylar lesions, the Wilson test reproduces pain with tibial internal rotation during knee extension from 90° to 30° that is relieved with tibial external rotation⁸⁸. For capitellar lesions, the radiocapitellar compression test reproduces pain with active pronation-supination in full elbow extension⁴⁶. However, these provocative maneuvers have limited diagnostic value⁸⁹. Restricted range of motion and joint effusion may signify unstable lesions or loose bodies. Limb alignment and joint stability should be evaluated as they influence biomechanical stress on joints.

Imaging

Radiography

Radiography is the first-line workup for OCD and for monitoring treatment response^7,18,22,90. Comprehensive views are recommended for knees (anteroposterior, lateral, notch or tunnel, and skyline or sunrise radiographs), elbows (anteroposterior in extension and 45° of flexion, lateral, and external oblique radiographs), and ankles (anteroposterior, lateral, and mortise radiographs) to examine predilection sites^7,90–93. Bilateral and standing alignment radiographs should be made with a low threshold to evaluate suspected bilateral disease and malalignment. Characteristic findings in early lesions are contour abnormality and radiolucency at the articular surface. More advanced lesions display a well-circumscribed, variably ossified fragment (progeny) separated from underlying bone (parent) by a crescent-shaped radiolucent line that may later ossify with healing. Normal variations of accessory epiphyseal ossification nuclei should be noted¹⁹. Radiography can reliably identify lesion location and size but is inaccurate for determining stability and identifying subtle lesions^16,94–99.

MRI

MRI is the standard of care for diagnosing and characterizing OCD^18,100. T1-weighted sequences allow lesion size measurement. T2-weighted sequences provide information on articular cartilage integrity, reactive marrow edema in the parent bone, and fluid or cystic changes at the parent-progeny interface. However, high signal intensity around the lesion may ambiguously represent fluid or granulation tissue^101,102. Postgadolinium fat-suppressed T1-weighted sequences can distinguish fluid from granulation tissue but may not correlate with healing^{100,103–105}. Gadolinium is controversial in children because of observable deposition in the brain, and no data showing long-term clinical consequences are available^106,107. Ellermann et al. described a noncontrast 3-T MRI protocol, called T2* mapping, using the shortest clinical echo time to provide additional information on soft-tissue composition, including osseus versus cartilaginous progeny lesions, fluid versus a fibrous interface, and normal versus sclerotic parent bone (Fig. 3)⁵¹. This protocol also generates a bone window that can obviate the need for computed tomography (CT).

Fig. 3.

Figs. 3-A through 3-E A 23-year-old man with an OCD lesion on the central aspect of the medial femoral condyle evaluated with an advanced clinical 3-T MRI protocol and arthroscopy. Fig. 3-A T2-weighted fat-saturated coronal TSE (turbo-spin-echo) image (TR [repetition time] = 3,600 ms, and TE [echo time] = 59 ms) demonstrating cystic changes and marrow edema of the parent bone. Substantial fluid at the parent-progeny interface is indicative of an unstable lesion. Overlying articular cartilage appears grossly intact. Fig. 3-B Proton density-weighted coronal TSE image (TR = 3,400 ms, and TE = 27 ms), demonstrating a dark progeny fragment suggesting osseous tissue. Fig. 3-C CT-like GRE (gradient-recalled echo) image (TR = 1,000 ms, and TE = 3.15 ms) with inverted contrast (“bone window”), demonstrating parent bone with cystic changes, sclerosis, and marginal mineralization and/or ossification surrounding the cartilaginous progeny lesion. Fig. 3-D T2* map (TR = 1,000 ms and TE = 3.15, 7.5, 11.9, 22.2, and 34.5 ms). Color-coding signal intensities: blue to black = bone and mineralization, green to red = epiphyseal and articular cartilage, and red = fluid at the parent-progeny lesion interface. T2* mapping affirms a cartilaginous OCD lesion (green centrally) with marginal ossification (blue), deemed unstable because of fluid (red) at the interface. Of note is mild overlying articular cartilage edema (red). Fig. 3-E Arthroscopic photographs confirming an unstable lesion. Arthroscopy was unable to determine the presence of bone on the fragment; thus, the advanced clinical 3-T MRI protocol helped to guide intraoperative decision-making. The patient underwent screw fixation, went on to heal, and returned to all activities.

MRI has excellent sensitivity and specificity for detecting OCD but uncertain accuracy in determining stability^108–110. De Smet et al. defined 4 MRI criteria based on T2-weighted sequences that are used to predict stability (Table I)¹¹¹. Evaluating these criteria for knee OCD, Kijowski et al. found 100% sensitivity but only 11% specificity in 36 juvenile lesions compared with 100% sensitivity and specificity in 33 adult lesions¹¹². After implementing 3 secondary MRI criteria (Table I), the sensitivity and specificity of predicting stability improved to 100%.

TABLE I.

Classification Systems for OCD of the Knee^*

Scintigraphy	Radiography	MRI	Lesion Stability	Arthroscopy
Cahill and Berg250 (1983) Stage 0: Normal scintigraphy and radiography Stage I: Normal scintigraphy but radiographic findings of OCD Stage II: Focal avidity at site of OCD defect, and radiographic findings of OCD Stage III: Focal avidity at site of OCD defect and increased activity in the femoral condyle harboring the OCD lesion, as well as radiographic findings of OCD Stage IV: Stage III and increased activity of the juxta-articular tibial plateau, and radiographic findings of OCD	Berndt and Harty251 (1959) Stage 1: Small subchondral compression Stage 2: Partially detached osteochondral fragment Stage 3: Completely detached but nondisplaced fragment Stage 4: Completely detached and displaced fragment (loose body) Scranton and McDermott modification252 (2001) Stage 5: Subchondral cyst	Dipaola et al.253 (1991) Stage I: Thickening of articular cartilage and low-signal changes Stage II: Articular cartilage breached and low-signal rim behind fragment indicating fibrous attachment Stage III: Articular cartilage breached and high-signal changes behind fragment indicating synovial fluid between fragment and underlying subchondral bone Stage IV: Loose body Hefti et al.16 (1999) Stage I: Small change of signal without clear margins of fragment Stage II: Osteochondral fragment with clear margins but without fluid between fragment and underlying bone Stage III: Fluid is visible partially between fragment and underlying bone Stage IV: Fluid is completely surrounding the fragment, but the fragment is still in situ Stage V: Fragment is completely detached and displaced (loose body) Ellermann et al.51 (2017) Type I: Epiphyseal cartilage lesion with necrotic center (no cleft at the interface) Type II: Epiphyseal cartilage lesion with complete or incomplete rim calcification (cleft at the interface) Type III: Partially or completely ossified lesion (varying degrees of osseous bridging and/or clefting at the interface) Type IV: Healed osseous lesion with a linear osseous scar (no cleft at the interface) Type V: Unhealed detached osseous lesion (sequestrum)	De Smet et al.111 (1996) (T2-weighted MRI) Sign 1: A thin line of high-signal-intensity ≥5 mm in length at the interface between the lesion and underlying bone (fibrovascular granulation tissue) Sign 2: A round area of homogeneous high-signal-intensity ≥5 mm in diameter beneath the lesion (cysts) Sign 3: A focal defect with a width of ≥5 mm in the articular surface of the lesion (displacement of the lesion into the joint) Sign 4: A high-signal-intensity line traversing articular cartilage into the lesion (articular fracture) O’Connor et al. modification102 (2002) Instability determined only when accompanied by cartilage breach on T1-weighted MRI Kijowski et al. modification112 (2008) A high T2-signal-intensity rim or cysts surrounding an adult OCD lesion are unequivocal signs of instability A high T2-signal-intensity rim surrounding a juvenile OCD lesion indicates instability only if it has the same signal intensity as adjacent joint fluid, is surrounded by a second outer rim of low T2 signal intensity, or is accompanied by multiple breaks in the subchondral bone plate on T2-weighted MRI Cysts surrounding a juvenile OCD lesion indicate instability only if they are multiple in number or large in size	Guhl254,255 (1982) Stage 1: Intact lesions Stage 2: Lesions showing signs of early separation Stage 3: Partially detached lesions Stage 4: Craters with loose bodies (salvageable or unsalvageable) Brittberg and Winalski256 (2003) ICRS OCD I: Stable lesions with a continuous but softened area covered by intact cartilage ICRS OCD II: Lesions with partial discontinuity that are stable when probed ICRS OCD III: Lesions with a complete discontinuity that are not yet dislocated (“dead in situ”) ICRS OCD IV: Empty defects and defects with a dislocated fragment or loose fragment within the bed ROCK Group (Carey et al.130; 2013) Immobile lesions: Cue ball: No abnormality detectable arthroscopically Shadow: Cartilage is intact and subtly demarcated (possibly under low light) Wrinkle in the rug: Cartilage is demarcated with a fissure, buckle, and/or wrinkle Mobile lesions: Locked door: Cartilage fissuring at periphery and unable to hinge open Trap door: Cartilage fissuring at periphery and able to hinge open Crater: Exposed subchondral bone defect Other classification systems: Ewing and Voto257 (1988) Johnson et al.258 (1990) Dipaola et al.253 (1991) O’Connor et al.102 (2002) Chen et al.259 (2013)

^*OCD = osteochondritis dissecans, ICRS = International Cartilage Repair Society, MRI = magnetic resonance imaging, and ROCK = Research in OsteoChondritis of the Knee.

CT

CT is useful for evaluating lesion size and location, loose bodies, and particularly osseous healing after fixation^1–5. It is commonly used for capitellar and talar OCD. A study evaluating CT and MRI for osteochondral lesions of the talus in 103 patients found comparable sensitivities and specificities¹¹³. Limitations of CT include radiation exposure and inability to obtain soft-tissue contrast^18,114.

Ultrasonography

Ultrasound is most commonly used for capitellar OCD^{91,93,115–120}. Harada et al. tested on-the-field utility of ultrasound for detecting elbow injuries among 153 youth baseball players and found 35 with abnormalities, of which 2 were confirmed capitellar OCD¹¹⁶. Takahara et al. found ultrasound comparable with radiography (85%; 23 of 27 patients), MRI (90%; 9 of 10 patients), and arthroscopy (93%; 14 of 15 patients) in predicting lesion stability using an ultrasound classification system (Table II)¹¹⁵. Ultrasound is heavily operator-dependent and therefore is not the diagnostic modality of choice in most cases^22,91.

TABLE II.

Classification Systems for OCD of the Elbow^*

Radiography	MRI	Ultrasound	Lesion Stability	Arthroscopy
Minami et al.260 (1979) Grade 1: Translucent cystic shadow in the lateral or middle aspect of the capitellum Grade 2: Clear zone or split line between the lesion and the adjacent subchondral bone Grade 3: Loose bodies	Dipaola et al.253 (1991) Stage I: Thickening of articular cartilage and low-signal changes Stage II: Articular cartilage breached and low-signal rim behind fragment indicating fibrous attachment Stage III: Articular cartilage breached and high-signal changes behind fragment indicating synovial fluid between fragment and underlying subchondral bone Stage IV: Loose body	Takahara et al.115 (2000) Type I: Lesion with localized subchondral bone flattening and cartilaginous thickening Type II: Lesion with nondisplaced fragments and an intact articular surface Type III: Lesion with displaced fragment Type IV: Lesion with osteochondral defect	Takahara et al.131 (2007) Stable (ICRS I): open capitellar physis, normal elbow motion, and Minami grade I Unstable (ICRS II-IV): Closed capitellar physis, restricted elbow motion >20°, and Minami grade II or III Kolmodin and Saluan modification261 (2014) Type I (stable): open capitellar physis, Minami grade I, and normal elbow motion Type II (unstable): Closed capitellar physis, Minami grade II or III, and motion restricted >20°, medial to the radial head center line Type IIIa (unstable): Closed capitellar physis, Minami grade II or III, and motion restricted >20°, extending lateral to the radial head center line Type IIIb (unstable): Closed capitellar physis, Minami grade II or III, and motion restricted >20°, extending lateral to the radial head center line including the lateral margin	Brittberg and Winalski256 (2003) ICRS OCD I: Stable lesions with a continuous but softened area covered by intact cartilage ICRS OCD II: Lesions with partial discontinuity that are stable when probed ICRS OCD III: Lesions with a complete discontinuity that are not yet dislocated (“dead in situ”) ICRS OCD IV: Empty defects as well as defects with a dislocated fragment or a loose fragment within the bed

^*OCD = osteochondritis dissecans, MRI = magnetic resonance imaging, and ICRS = International Cartilage Repair Society.

Bone Scintigraphy

Scintigraphy was historically used to detect OCD and assess healing on the basis of perfusion^94,121,122. Despite excellent sensitivity, scintigraphy became obsolete because of the lack of specificity, MRI availability, and radioisotope exposure^{6,18,36,104,123–126}. Recently, a nonradioactive, non-contrast-enhanced 3-T MRI method, called arterial spin labeling, was demonstrated to visualize distal femoral perfusion similarly to scintigraphy in children with knee OCD¹²⁷.

Classification Systems

There are numerous OCD classifications systems based on the joint involved and the diagnostic modality (Tables I, II, and III). Arthroscopy is the gold standard for determining lesion stability¹²⁸. The ROCK arthroscopic classification system for the knee is the most comprehensive, including 6 mutually exclusive categories divided into immobile and mobile lesions but omitting salvageability, which requires considering skeletal maturity, prior treatments, and imaging (Table I)^129,130. Contributing to MRI-based systems, Ellermann et al. devised a classification system for juvenile OCD that stages natural history from necrotic epiphyseal cartilage to lesions that are healed or not healed⁵¹. While no stand-alone preoperative method can definitively determine stability, the Takahara classification system uses radiography and clinical parameters for capitellar OCD, demonstrating the utility of a multifactorial system (Table II)¹³¹.

TABLE III.

Classification Systems for OCD of the Talus^*

Radiography	CT	MRI	Arthroscopy
Berndt and Harty251 (1959) Stage 1: Small subchondral compression Stage 2: Partially detached osteochondral fragment Stage 3: Completely detached but nondisplaced fragment Stage 4: Completely detached and displaced fragment (loose body) Scranton and McDermott modification252 (2001) Stage 5: Subchondral cyst	Ferkel et al.216 (2008) Stage I: Cystic lesion within dome of talus, and intact roof on all views Stage IIA: Cystic lesion with communication to talar dome surface Stage IIB: Open articular surface lesion with overlying nondisplaced fragment Stage III: Nondisplaced lesion with radiolucency Stage IV: Displaced fragment Other classification systems: Loomer et al.124 (1993)	Hepple et al.262 (1999) Stage 1: Articular cartilage damage only Stage 2a: Cartilage injury with underlying fracture and surrounding osseous edema Stage 2b: Stage 2a without surrounding osseous edema Stage 3: Detached but nondisplaced fragment Stage 4: Detached and displaced fragment Stage 5: Subchondral cyst formation	Ferkel et al.216 (2008) Grade A: Smooth, intact, but soft or “ballottable” (springy) Grade B: Rough surface Grade C: Fibrillations and/or fissures Grade D: Flap present or bone exposed Grade E: Loose, nondisplaced fragment Grade F: Displaced fragment Other classification systems: Pritsch et al.263 (1986)

^*CT = computed tomography, and MRI = magnetic resonance imaging.

Treatment Alternatives

The 2010 American Academy of Orthopaedic Surgeons clinical practice guidelines revealed limited existing high-level evidence for guiding OCD management¹³². Skeletal maturity and lesion stability are generally considered the most important information for clinical decision-making^7,18,22,133. General treatment algorithms for knee, elbow, and talus are shown in Figures 4, 5, and 6.

Fig. 4.

General treatment algorithm for OCD of the knee. Dotted arrow indicates alternative treatment after initial failure. ACI = autologous chondrocyte implantation, H&P = history and physical examination, MACI = matrix-assisted autologous chondrocyte implantation, MRI = magnetic resonance imaging, and OAT = osteochondral autograft or allograft transplantation.

Fig. 5.

General treatment algorithm for OCD of the capitellum. Dotted arrow indicates alternative treatment after initial failure. CT = computed tomography, H&P = history and physical examination, MRI = magnetic resonance imaging, and OAT = osteochondral autograft or allograft transplantation.

Fig. 6.

General treatment algorithm for OCD of the talus. Dotted arrow indicates alternative treatment after initial failure. ACI = autologous chondrocyte implantation, CAM = controlled ankle motion, CT = computed tomography, H&P = history and physical examination, MACI = matrix-assisted autologous chondrocyte implantation, MRI = magnetic resonance imaging, and OAT = osteochondral autograft or allograft transplantation.

Nonoperative Treatment

Nonoperative treatment is well accepted for patients with open physes, stable lesions, or minimal symptoms. Duration is typically a minimum of 3 to 6 months followed by gradual return to activity. Protocols on activity modification, weight-bearing restriction, and immobilization vary^19,96. Immediate cessation of athletics and impact activities is usually recommended. Some authors prefer protected weight-bearing for minimizing compression across the lesion, while others have reported success with unloader bracing protecting the lesion^18,96. Controversy over immobilization stems from incomplete understanding of the etiology. Those who believe OCD is localized to articular cartilage advocate maintaining motion for chondral health, whereas those who believe OCD originates from subchondral bone advocate immobilization to promote osseous healing^17,125. Nomograms have been developed to predict the likelihood of healing with nonoperative treatment based on patient age, mechanical symptoms, lesion size, lesion location, and cyst-like lesions^95,96,134.

Operative Treatment

Surgery is primarily indicated for patients with unstable lesions, failed conservative treatment, or poor nomogram predictions. The technique depends on the lesion characteristics. Concomitant pathology, including limb malalignment and joint instability, may require surgical correction (Fig. 7).

Fig. 7.

Figs. 7-A through 7-F A 14-year-old girl who underwent internal fixation with bioabsorbable headless compression screws and proximal tibial medial opening-wedge osteotomy for treatment of an OCD lesion on the medial femoral condyle with associated varus deformity at the knee. Fig. 7-A Preoperative full-length radiograph of the lower limbs, made with the patient standing, demonstrating left genu varum. Fig. 7-B Preoperative anteroposterior and lateral radiographs of the left knee reveal concave parent bone, suggesting a largely cartilaginous OCD lesion. Fig. 7-C Preoperative T2-weighted MRI demonstrating marrow edema and mild fluid at the parent-progeny interface. Fig. 7-D Arthroscopic photographs showing an OCD lesion delineated by unstable articular cartilage but not displaceable with gentle probing, implantation of 3-mm bioabsorbable headless compression screws, and final appearance of the articular surface. Fig. 7-E Fluoroscopy demonstrating a medial opening-wedge high tibial osteotomy. Fig. 7-F Radiographs, made 2 years postoperatively, showing interval healing of the lesion and consolidation of the osteotomy. The patient returned to all activities.

Stable Lesions

Subchondral drilling is well-established for stable lesions to stimulate influx of mesenchymal cells and growth factors^135,136. Techniques include transarticular (Fig. 8) and retroarticular (transepiphyseal) drilling^135,137. Retroarticular drilling avoids cartilage penetration but requires fluoroscopy. A systematic review concluded that there were comparable patient-oriented and radiographic outcomes¹³⁸. Adjunctive bone-grafting has also been described¹³⁹.

Fig. 8.

Figs. 8-A through 8-J An 11-year-old boy who underwent transarticular drilling for treatment of an OCD lesion in the medial femoral condyle of the left knee. Fig. 8-A Preoperative tunnel and lateral radiographs reveal an osseus OCD lesion. Figs. 8-B, 8-C, and 8-D Preoperative images made with advanced clinical 3-T MRI protocol, obtained with same imaging parameters as in Figure 3, including a T2-weighted coronal image demonstrating marrow edema (Fig. 8-B), a CT-like image contrast bone window (Fig. 8-C) depicting a partially ossified progeny lesion without evidence of osseous bridging with the parent bone (insert shows magnification of the lesion), and T2* map (Fig. 8-D), which provides additional information, specifically on cartilage integrity showing intact residual epiphyseal cartilage (green) and articular cartilage (red). Fig. 8-E Arthroscopic photographs made during transarticular drilling with a 0.045-in (0.114-cm) Kirschner wire to penetrate the lesion in a perpendicular fashion through the articular cartilage. Fig. 8-F Fluoroscopy confirming wire placement. Fig. 8-G Radiographs made 3 months postoperatively demonstrating interval healing. Figs. 8-H, 8-I, and 8-J Comprehensive MRI assessment demonstrating mild residual edema in parent bone and lesion in T2-weighted coronal MRI images (Fig. 8-H); solid osseous bridging and further ossification throughout the entire lesion in the CT-like contrast bone window images (insert shows magnification of the lesion) (Fig. 8-I), which was demonstrated on all slices (not shown); and preserved articular cartilage (red) after surgical intervention (see T2* map), which is mostly for further prognostication (Fig. 8-J). The patient returned to all activities.

Unstable Salvageable Lesions

Principles for treating unstable salvageable lesions include articular surface restoration, fracture fixation, and vascular enhancement¹²¹. Fixation devices include headless, partially threaded, variable-pitch compression screws, and bioabsorbable implants (Video 2). Metal screws can provide excellent purchase to help in healing but are dependent on lesion quality and tissue composition (e.g., cartilaginous versus osseous), may need to be removed before full weight-bearing, and interfere with MRI assessment^140–143. Bioabsorbable implants are equally efficacious but can be associated with loosening, synovitis, and cyst formation^144–147. Fixation using bone pegs or osteochondral plugs and hybrid techniques have been described with positive short-term outcomes^148–151. Adjunctive strategies to address nonosseous tissues at the parent-progeny interface include debridement, marrow stimulation, and bone-grafting.

Unstable Unsalvageable Lesions

Lesions may be unsalvageable because of substantial osteochondral fragmentation, subchondral necrosis, or cartilage degeneration. Cartilage salvage and resurfacing techniques include marrow stimulation, autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OAT), and osteochondral allograft transplantation (OCA). Loose body excision alone yields poor outcomes^152,153. Biomechanically inferior fibrocartilage formation is a shortcoming of marrow stimulation and ACI. Newer-generation ACI can treat deeper defects using bilayered collagen membranes interposed with chondrocytes¹⁵⁴. Matrix-induced autologous chondrocyte implantation (MACI) addresses potential problems of uneven cell distribution or leakage by incorporating chondrocytes into collagen scaffolds¹⁵⁵. OAT is biomechanically superior but limited by a maximum lesion size of 2 to 3 cm² and donor site morbidity^156,157. OCA is an alternative for larger lesions^158–161.

Outcomes

Knee

Nonoperative success rates between 49% and 100% have been reported for skeletally immature patients with stable knee lesions^{8,12,16,95,96,101,123,162–165}. Weiss et al. found that 33.5% (69) of 206 patients with juvenile OCD in the knee required surgery, and patients between the ages 12 and 19 years had a 7.4-times greater risk than 6 to 11-year-old patients¹⁶⁶. Conversely, patients with adult OCD often require surgery¹⁷. Sanders et al. calculated that the cumulative rates of osteoarthritis and arthroplasty were 30% and 8%, respectively, at 35 years in 86 patients who had been treated nonoperatively¹⁶⁷. A body mass index (BMI) of >25 kg/m², patellar lesions, and adult onset were risk factors¹⁶⁷. With multifocal lesions, Backes et al. found that only 25.5% (12) of 47 lesions healed conservatively after 12 months²¹. For stable lesions that had failure of nonoperative treatment, subchondral drilling has provided good outcomes^{136,138,168–175}. Transarticular drilling in 23 children yielded radiographic healing at 4.4 months and improved functional outcomes¹⁷⁶. Edmonds et al. performed retroarticular drilling in 59 children who had had failure after at least 6 months of nonoperative treatment of OCD lesions; they were able to return to activities at a mean of 2.8 months and 98.2% had radiographic healing¹⁷⁷. For unstable salvageable lesions, Kocher et al. that found 84.6% (22) of 26 juvenile OCD lesions healed at 6 months after treatment with fixation, independent of the method or lesion location¹⁷⁸. Similarly, Wu et al. reported that lesions in 76% (66) of 87 patients had healed with fixation at ≥2 years postoperatively despite skeletal maturity¹⁷⁹. For unstable unsalvageable lesions, surgical outcomes have been encouraging^180–184. Carey et al. reported that treatment of 55 patients (age range, 14 to 52 years) with ACI yielded 87% and 82% survivorship at 10 and 20 years, respectively, with 2 arthroplasties in the fifth and sixth decades¹⁸⁵. Newer-generation techniques have also shown promise^182,186,187. Roffi et al. described 19 children who were treated with MACI and bone-grafting, which provided functional improvement, but 16% of them had failure at 10 years¹⁵⁵. A randomized clinical trial in which OAT was compared with microfracture in 50 children who had OCD (International Cartilage Repair Society [ICRS] grade III or IV) found superiority with OAT at 4.2 years¹⁸⁸. Similarly, OCA achieved good short-term outcomes in a small cohort of children¹⁵⁸. Emmerson et al. performed OCA in 64 adults (age range, 15 to 54 years), which resulted in good-to-excellent outcomes in 72% after 7.7 years; however, 15% required reoperation, including 3 arthroplasties (in patients who were 30, 36, and 54 years old), at 3 to 5.5 years¹⁵⁹.

Elbow

Nonoperative treatment provides reliable healing for most early and stable capitellar lesions^189–192. Risk factors include advanced skeletal age, larger lesions, cyst-like lesions, and radial head enlargement^190,191,193. Matsuura et al. showed that compliance with discontinuing heavy elbow use for at least 6 months influenced outcomes, with healing in only 22.7% (17) of 75 noncompliant patients versus 78.6% (66) of 84 and 52.9% (9) of 17 compliant patients with Minami stage-I and II lesions, respectively¹⁸⁹. For stable lesions that have failed nonoperative treatment, marrow stimulation has provided satisfactory outcomes^194–198. Bexkens et al. reported good clinical results with microfracture at 1 year, with 62% (44) of 71 patients returning to sports¹⁹⁹. For unstable salvageable lesions, reliable outcomes have been achieved with fragment fixation^{46,91,131,200}. Hennrikus et al. reported that 77% (20) of 26 elbows healed with fixation, and 67% (12) of 18 questionnaire respondents returned to sports at 36 months²⁰¹. For unstable unsalvageable lesions, osteochondral transfer techniques have demonstrated acceptable outcomes^202–205. Bae et al., in a study of 28 patients treated with knee-to-elbow OAT, reported that 73% (19) of 26 patients had pain resolution and 100% (13) of 13 patients with >6 months of follow-up returned to sports at an average of 9 months²⁰⁶. Mirzayan and Lim performed OCA in 9 patients with juvenile OCD, and all had functional improvement and returned to sports after 2 years²⁰⁷. Lateral humeral condyle closing-wedge osteotomy, which is a less common technique, has shown promise^208–210. Koda et al. performed 77 osteotomies for advanced lesions, with concomitant bone-grafting and fixation, yielding good remodeling in 69% (53) of 77 elbows; however, 53% (41) of the 77 elbows had osteoarthritic changes at 9 years²⁰⁹.

Talus

Skeletally immature patients and early talar lesions generally respond to nonoperative treatment^{15,54,211–213}. Nonetheless, Perumal et al. found that 39% (12) of 31 children with OCD lesions (Berndt and Harty stages II, III, or IV) treated nonoperatively for 6 months eventually required surgery with drilling⁵². Similarly, Letts et al. reported that 54% (13) of 24 children with juvenile OCD (Berndt and Harty stages I through IV) had failure of conservative management and required surgery with a combination of drilling, pinning, and excision⁵³. For osteochondral lesions of <150 mm², microfracture has demonstrated positive outcomes^214–220. Choi et al. reported that 165 osteochondral lesions (mean size, 73 mm²) treated with microfracture provided significant improvements in functional scores at 6.7 years²²¹. Subchondral drilling has shown comparable results^222–225. Masquijo et al. performed retroarticular drilling in 6 children with OCD lesions (Berndt and Harty stage I or II), and they showed improvements at 37 months²²⁶. For osteochondral lesions of >150 mm², advanced cartilage restoration techniques are available and feasible^227–232. Kwak et al. described ACI for 29 osteochondral lesions (mean size, 198 mm²) and reported that 79% (23) of 29 patients had good-to-excellent outcomes at 70 months²³³. A systematic review comprising 343 patients with lesions of <250 mm² demonstrated equivalence between MACI and ACI²³⁴. Additionally, a study of 9 patients with osteochondral lesions treated with MACI after failing microfracture found substantial functional improvements at 7 years²³⁵. OAT has also demonstrated favorable outcomes^236–242. Lee et al. reported on 17 patients with OCD (Berndt and Harty stage III or IV) who were managed with knee-to-ankle OAT and had good-to-excellent outcomes at 36 months²⁴³. OCA has provided comparable success^244–248. VanTienderen et al. performed a systematic review comprising 91 osteochondral lesions treated with OCA and demonstrated improvements at 45 months²⁴⁹.

Overview

OCD is a relatively common cause of joint pain and dysfunction that may lead to premature osteoarthritis. The etiology is being elucidated by clinical and cadaveric studies, animal models, advanced imaging, and genetic analyses. Improved understanding will guide treatment and identify risk factors that can be addressed to reduce the prevalence of the disorder. Similar management principles are applicable to OCD of the knee, elbow, and talus. Skeletally immature patients with stable lesions usually improve with nonoperative treatment, whereas those who have failure of conservative management likely respond to arthroscopic surgery. Older patients and those with unstable lesions often require surgical intervention, and various techniques exist to stimulate subchondral revascularization, promote parent-progeny healing, restore articular congruity, and offload the lesion. Newer-generation strategies, including tissue-engineered scaffolds and biologics, are promising. Outcomes have room for improvement, and prospective randomized trials are needed to determine optimal management protocols. The current best evidence on highlighted topics related to OCD is summarized in Table IV.

TABLE IV.

Grades of Recommendation for OCD^*

Recommendation	Grade†
OCD occurs most frequently in the active pediatric and young adult populations, commonly affecting the knee, elbow, or ankle and may lead to premature osteoarthritis.	A
While generally considered an idiopathic phenomenon, various etiopathogenetic theories are being investigated, including local ischemia, aberrant endochondral ossification of the secondary subarticular physis, repetitive microtrauma, and genetic predisposition.	B
Diagnosis is based on the history, physical examination, radiography, and advanced imaging, with elbow ultrasonography and novel MRI protocols potentially enabling early detection and staging.	A
Treatment largely depends on skeletal maturity and lesion stability, defined by the presence or absence of articular cartilage fracture and subchondral bone separation, as determined by imaging and arthroscopy, and is typically nonoperative for stable lesions in skeletally immature patients and operative for those who have failed conservative management or have unstable lesions.	B
Clinical practice guidelines have been limited by a paucity of high-level evidence, but a multicenter effort is ongoing to develop accurate and reliable classification systems and multimodal decision-making algorithms with prognostic value.	A
There is currently no association between gadolinium deposition in the brain and poorer health outcomes in children; however, gadolinium-based contrast agents should be used judiciously and avoided when possible.	I

^*OCD = osteochondritis dissecans, and MRI = magnetic resonance imaging.

^†According to Wright²⁶⁴, grade A indicates good evidence (Level-I studies with consistent findings) for or against recommending intervention; grade B, fair evidence (Level-II or III studies with consistent findings) for or against recommending intervention; grade C, poor-quality evidence (Level-IV or V studies with consistent findings) for or against recommending intervention; and grade I, insufficient or conflicting evidence not allowing a recommendation for or against intervention.

Supplementary Material

Vid 1

Video 1

Three-dimensional reconstruction of susceptibility-weighted MRI data obtained from a distal femoral specimen from a 3-month-old human cadaveric donor demonstrating central and peripheral vascular beds with a watershed region in the midsagittal plane.

Vid 2

Video 2

A 14-year-old female patient undergoing internal fixation with bioabsorbable headless compression screws for treatment of an osteochondritis dissecans lesion on the medial femoral condyle.

• Osteochondritis dissecans occurs most frequently in the active pediatric and young adult populations, commonly affecting the knee, elbow, or ankle, and may lead to premature osteoarthritis.

• While generally considered an idiopathic phenomenon, various etiopathogenetic theories are being investigated, including local ischemia, aberrant endochondral ossification of the secondary subarticular physis, repetitive microtrauma, and genetic predisposition.

• Diagnosis is based on the history, physical examination, radiography, and advanced imaging, with elbow ultrasonography and novel magnetic resonance imaging protocols potentially enabling early detection and in-depth staging.

• Treatment largely depends on skeletal maturity and lesion stability, defined by the presence or absence of articular cartilage fracture and subchondral bone separation, as determined by imaging and arthroscopy, and is typically nonoperative for stable lesions in skeletally immature patients and operative for those who have had failure of conservative management or have unstable lesions.

• Clinical practice guidelines have been limited by a paucity of high-level evidence, but a multicenter effort is ongoing to develop accurate and reliable classification systems and multimodal decision-making algorithms with prognostic value.