Early knee OA definition–what do we know at this stage? An imaging perspective

Abstract

While criteria for early-stage knee osteoarthritis (OA) in a primary care setting have been proposed, the role of imaging has been limited to radiography using the standard Kellgren–Lawrence classification. Standardized imaging and interpretation are critical with radiographs, yet studies have also shown that even early stages of radiographic OA already demonstrate advanced damage to knee joint tissues such as cartilage, menisci, and bone marrow. Morphological magnetic resonance imaging (MRI) shows degenerative damage earlier than radiographs and definitions for OA using MRI have been published though no accepted definition of early OA based on MRI is currently available. The clinical significance of structural abnormalities has also not been well defined, and the differentiation between normal aging and structural OA development remains a challenge. Compositional MRI of cartilage provides information on biochemical, degenerative changes within the cartilage matrix before cartilage defects occur and when cartilage damage is potentially reversible. Studies have shown that cartilage composition can predict cartilage loss and radiographic OA. However, while this technology is most promising for characterizing early OA it has currently limited clinical application. Better standardization of compositional MRI is required, which is currently work in progress. Finally, there has been renewed interest in computed tomography (CT) for assessing early knee OA as new techniques such as weight bearing and spectral CT are available, which may provide information on joint loading, cartilage, and bone and potentially have a role in better characterizing early OA. In conclusion, while imaging may have a limited role in diagnosing early OA in a primary care setting, there are advanced imaging technologies available, which detect early degeneration and may thus significantly alter management as new therapeutic modalities evolve.

Keywords

CT imaging MRI osteoarthritis radiographs

Introduction

Osteoarthritis (OA) is a highly prevalent joint disease affecting 22.9% of individuals over the age of 40 knee¹ and has demonstrated a significant increase in prevalence of 9.3% from 1990 to 2017.² Given our aging population and the increasing obesity epidemic, we are expecting these numbers to grow gradually and lead to even higher rates of disability and health care costs.³ Therapy is limited and mostly consists of lifestyle modification at earlier disease stages and joint replacement at more advanced disease stages. Though there is the perception that there are no efficient treatments available,⁴ it is generally accepted that if detected at early stages, management strategies are available that help effectively to reduce disease burden.⁵ This creates a window of opportunity that should engage physicians to diagnose OA at the earliest stages possible. However, there is significant uncertainty and a lack of consensus of what represents early OA, which also further challenges the execution of rigorous clinical trials to treat OA.⁶

In 2018, classification criteria were proposed for early-stage OA intended for a primary care setting.⁶ This was based on expert opinions and resulted in the recommendation of the following diagnostic criteria: (i) pain, function, and quality of life (using questionnaires such as the Knee injury and Osteoarthritis Outcome Score (KOOS)); (ii) clinical examination including symptoms such as joint line tenderness; and (iii) knee radiographs demonstrating a Kellgren & Lawrence (KL) grade 0 or 1. At this time, no other imaging criteria were recommended. MRI, while considered as highly effective in diagnosing early structural abnormalities, was considered as not sufficiently standardized, and it was felt that there was a too high population prevalence in asymptomatic individuals to use MRI to diagnose early knee OA. In 2021, these draft criteria were tested using longitudinal outcome data from the Osteoarthritis Initiative (OAI) cohort with encouraging performance in terms of enrichment for structural and clinical progression.⁷

Given the ongoing controversy of the role of imaging, we are revisiting the early definition of OA in this review article from an imaging perspective with a focus on the knee as the most clinically significant manifestation. We will evaluate the role of standard radiographs, MRI morphological and compositional techniques, and computed tomography (CT)-based methods based on published data and ongoing initiatives to achieve better standardization of imaging methods.

Standard radiographic techniques

The KL classification to define knee OA was first published in 1957⁸ and yet is still considered as a standard in diagnosing knee OA. According to the previously published classification by Luyten et al.,⁶ criteria for early OA include KL grades 0 and 1, which is inherently problematic as it includes not only early stages of OA but also normal joints without any manifestations of OA. Thus, these criteria should be best used if patients also meet clinical criteria. Earlier definitions included also KL2 (osteophytes only) in addition to KL0 and 1 along with pain.^9,10 Figure 1 demonstrates representative knee radiographs with KL grades 0, 1, and 2. While considered a standard, KL scores have limited reproducibility and include terminology ‘doubtfully’ and possibly as pertains to joint space narrowing, which is inherently imprecise. Also, joint space narrowing is dependent on the imaging technique (weight-bearing extension anteroposterior or flexion posteroanterior (bend) radiographs), which further complicates a reproducible and standardized diagnosis.¹¹ Moreover, it should be noted that even in individuals with KL grade 1 and 0, cartilage damage is not rarely found. Taguchi et al.¹² found cartilage defects, mostly at the medial femoral condyle, in 26% of women aged 50 years and older in their study investigating MRI studies in patients with KL1 and 2, normal leg alignment and no obesity. This study clearly documents advanced and irreversible cartilage damage already in patients with nearly normal radiographs based on KL scores.

Figure 1.

Radiographs of the knee (posterior–anterior, bent projection) demonstrating a normal knee radiograph without osteophytes (KL0) in 35-year-old man (a), a knee radiograph with tiny osteophytes (arrows) of doubtful significance (KL1) in 43-year-old woman (b), and a knee radiograph with definite osteophytes (arrows) in the medial joint compartment (KL2) in a 42-year-old woman (c).

Given the limitations of KL scoring other systems have been suggested such as the atlas-based Osteoarthritis Research Society International (OARSI) classification. Culvenor et al.¹³ compared KL and OARSI atlas criteria in their definition of the presence of knee OA. The investigators used a standard KL grading of 2 and higher to define tibiofemoral OA, while for the OARSI atlas-based grading, OA was diagnosed if there was grade 2 or higher joint space narrowing and/or osteophytes sum score grade 2 and higher and/or grade 1 joint space narrowing with grade 1 osteophyte. Based on the gradings of 1178 knee radiographs, a diagnosis of knee OA was made in 17.3% (n = 203) of knees using the KL score and in 26.2% (n = 309) of knees using the OARSI atlas-based score. Intra- and inter-rater reliability were in a similar range for both scoring systems with kappa values between 0.86 and 0.97 for intra-rater reliability and between 0.63 and 0.73 for inter-rater reliability. These results show that the OARSI atlas criteria identify a larger number of individuals with knee OA, but to date, there is limited information how valid these data are in identifying early stages of disease. Clearly a better standardization is required to define (early) radiographic OA.

Given the inherent limitations of semiquantitative scores with respect to reproducibility, truly quantitative measurements may be more attractive to better define disease burden in knee OA. Researchers have therefore worked on developing methods that would allow an accurate and reproducible assessment of joint space width.^14–16 The additional advantage of measurements of joint space narrowing is their sensitivity to change, which improves assessment of disease progression. However, to date, there are no established cut-off values that would allow the definition of abnormal joint space narrowing. Establishing a threshold would require a reproducible radiographic technique, an adjustment for joint size, and a measurement technique that is accurate and likewise reproducible. But most of all, this cut-off value would need to be validated through longitudinal studies demonstrating that it can in fact predict development of more advanced stages of knee osteoarthritis.

Given the limitations of standard semiquantitative and quantitative assessment of knee radiographs in diagnosing knee OA, researchers have developed computer-aided/machine learning tools to predict knee OA based on radiographs and found encouraging results.^17,18 One of these studies analyzed trabecular bone texture in knee radiographs using convolutional neural networks in knee radiographs from the OAI and Multicenter Osteoarthritis Study (MOST) cohorts.¹⁷ Using progression of joint space narrowing as an outcome, they found better prediction with trabecular bone texture than with KL-based models in the MOST cohort and similar prediction in the OAI cohort.

In summary, while radiographs are an inexpensive imaging technique, a rigorous diagnosis of early OA requires standardized imaging, image readings, and measurement techniques. Still, even with optimized image acquisition and analysis, radiographs are neither sensitive nor specific for early OA disease stages. Using a combination of radiographic (KL0 and 1) and clinical findings as suggested by Luyten et al.⁶ may be a reasonable compromise and was recently longitudinally validated.⁷

Magnetic resonance imaging: morphological imaging

Prevalence of structural pathology in knees without radiographic OA

As already outlined above, it has been acknowledged that a radiographic status of KL grade 0 or 1 may not be equivalent to absence of disease.^19,20 A large population-based observational study reported presence of any MRI-detected OA-related abnormalities in 89% of knees with apparently completely normal radiographs (i.e. KL grade 0).¹⁹ Particularly, MRI-detected cartilage damage and osteophytes seem to be highly prevalent in the medial patellofemoral and medial posterior tibiofemoral joints.²¹ An increasing prevalence of features, including cartilage damage, bone marrow lesions, and meniscal damage, is found in older patients.¹⁹ While the prevalence of MRI pathology did not show significant differences between the patients with and without knee pain for most definitions of pain, the prevalence of ‘any MRI abnormality’ was markedly increased in those with Western Ontario and McMaster Universities OA index (WOMAC) pain compared with those without. However, it needs to be noted that MRI abnormalities were very frequent in participants without pain (up to 86%).¹⁹ Another study based on the Osteoarthritis Initiative (OAI) focused on KL0 knees with and without risk factors (i.e. obesity, family history of total knee joint replacement, repetitive knee bending, previous knee trauma and/or surgery, and presence of Heberden’s nodes) observed that 96% of persons with knee OA risk factors but without radiographic knee OA had at least one MRI finding present and 87% of persons without such risk factors.²² The prevalence of any MRI pathology was higher (prevalence ratio adjusted for age, sex, and BMI 1.3 [CI 1.1–1.6]) in individuals with risk factors for knee OA than in those without risk factors. The prevalence of MRI-defined OA in the tibiofemoral joint using the definition by Hunter et al.²³ was 23% (CI 19–28%) in subjects with OA risk factors and 9% (CI 3–21%) in those without. In addition, MRI-defined structural tissue pathology was far more prevalent in an early OA model of KL0 knees with established contralateral radiographic OA, compared with a healthy reference sample with KL0 in both knees and no risk factors.²⁴ This has been shown for all MRI features, apart from effusion-synovitis. Regarding longitudinal change, incident cartilage damage (defined semiquantitatively) and quantitative cartilage loss (based on 3D segmentation) was more commonly seen in the contralateral OA model group,^24,25 supporting the assumption that patients with a radiographically normal knee but contralateral radiographic OA represent a suitable model for testing disease-modifying osteoarthritis drugs (DMOADs) in clinical trials of early, nonradiographic OA.

Risk of structural OA development

As multiple MRI features of OA are frequently found in a knee with structural disease and increase the risk of progression,²⁶ it would be relevant to understand the time course of events to be able to address specific pathology early in order to avoid or delay progression to more advanced disease. Although cartilage loss is one of the hallmark features of structural disease, the evidence that OA commences with cartilage lesions is sparse.²⁷ Several authors have suggested that incidental meniscal pathology might be one of the main triggers of disease incidence.²⁸ However, as meniscal damage is frequently found in individuals without symptoms, the role of the meniscus in disease onset is still under debate.^29,30 Synovitis, which is reflected on MRI as joint effusion and synovial thickening, increases the risk of cartilage damage and might play an important yet not fully understood role in early disease.^31,32 Furthermore, subchondral bone marrow abnormalities appear to have an impact on disease progression, and animal models have suggested that bone marrow alterations may be the earliest structural reflections of disease onset.³³ Finally, cartilage abnormalities markedly impact risk for progressive disease and may be important triggers for progression on a subregional level and subsequently of the whole joint leading to radiographic OA.^34,35 Eventually, a more or less concomitant presence of these lesions may be possible.³⁶ It has been reported before that knees exhibiting specific structural joint damage on MRI, especially within the 2 years prior to disease incidence, concomitant presence of several features, and occurrence of such features at one or more time points expose the joint at increased risk of developing incident radiographic OA compared with matched knees not showing these features of structural damage.³⁷ Figure 2 gives an example of isolated and concomitant structural MRI features in knees without radiographic OA (KL0).

Figure 2.

Structural tissue damage is highly prevalent and often concomitantly present in knees without radiographic osteoarthritis. (a)–(d) Structural abnormalities in MRI studies that are not visualized in radiographs but represent disease features that have been shown to lead to progressive OA. (a) Coronal intermediate-weighted (IW) MRI shows a superficial focal cartilage lesion at the lateral femur (large arrow) and a minor adjacent subchondral bone marrow lesion (arrowhead) in this KL0 knee. In addition, there is a discrete but definitive horizontal-oblique tear of the body of the medial meniscus reaching the meniscal base and the undersurface (small arrows). (b) Sagittal IW fat-suppressed MRI of another KL0 knee shows a large subchondral bone marrow lesion (arrows) at the lateral femoral trochlea with adjacent superficial cartilage damage (arrowhead). In addition, there is marked Hoffa-synovitis (asterisk). (c) Other findings commonly observed in KL0 knees include effusion-synovitis (asterisk) as shown on this axial reformation of a dual echo at steady state (DESS) MRI, which has been shown to predict longitudinal cartilage loss in knees without radiographic OA. (d) Meniscal extrusion is a risk factor for cartilage damage and is also commonly observed in knees without radiographic OA. Coronal DESS image shows marked extrusion of the medial meniscal body (parallel lines). Note absence of osteophytes at the joint margins in this KL0 knee.

Clinical relevance of pre-radiographic findings

Another study focused on a subset of the OAI and evaluated subjects without radiographic OA at two defined time points, that is, at 12 and 48 months, using cartilage loss and incident knee symptoms as the outcome parameters. This analysis found that structural tissue damage was associated with incident persistent knee symptoms and that more concomitant features were associated with an increased risk of symptom outcomes and incident tibiofemoral cartilage damage.³⁶ From this, the authors concluded that the observed findings should not be considered incidental in individuals with higher risk for OA and may rather reflect early degenerative disease. The relevance of knee pain in early structural knee OA has been emphasized in other studies. Peat et al.³⁸ found that incident radiographic OA is preceded by prodromal symptoms lasting at least 2–3 years. It should also be noted that this early phase of disease constitutes a window of opportunity and management strategies are available that help effectively to reduce disease burden.⁵ These include lifestyle changes including addressing risk factors for disease such as obesity and lack of physical activity.

MRI definitions of OA and early OA

Compared to radiography, MRI is able to visualize the different knee tissues much more comprehensively.³⁷ Figure 2 highlights structural abnormalities in MRI studies that are not visualized in radiographs but represent disease features that have been shown to lead to progressive OA and may in part be considered as signs of early OA. In 2011, an MRI definition of structural knee OA was introduced by an expert group using Delphi methodology.²³ Consequently, this MRI definition of OA includes more joint tissues and pathologies than the radiographic KL definition, suggesting that it may be more sensitive to detect knee OA at an earlier stage, or that it may result in a more valid definition compared with radiography. In this exercise, tibiofemoral OA was defined on MRI as the presence of a definite osteophyte and full-thickness cartilage loss, or a definite osteophyte or full-thickness cartilage loss plus two of the additional features: (1) subchondral bone marrow lesion or cyst not associated with meniscal or ligamentous attachments, (2) meniscal subluxation, maceration or degenerative tearing (such as a horizontal tear), (3) partial-thickness cartilage loss, or (4) bone attrition. In a subsequent validation exercise, Schiphof and colleagues compared KL grading with the MRI definition in the population-based Rotterdam study and showed that the MRI definition detected more cases of structural OA (9%) compared with radiography (4%).²⁰ In addition, the association between the definitions and knee pain at baseline was higher when the MRI definition was included. This was similar for the association between the definitions and persistent knee pain, and between the definitions and BMI.²⁰ This study, however, did not focus particularly on early OA or an early OA definition by MRI.

Magnusson and colleagues investigated participants from the OAI study to study the differences in prevalence of MRI abnormalities over time in those with and without risk factors for knee OA.³⁹ An operational MRI definition of OA was included but not specified further, a radiographic definition was defined as KL ⩾ 1 and a knee pain definition defined as pain on most days of a month in past 12 months. The study reported that the development of structural OA was comparable irrespective of the presence of risk factors for OA. In addition, the overlap with symptoms was quite limited, in particular in the subgroup without risk factors for OA. These findings suggest that structural tissue damage without symptoms should probably not be defined as OA, but as risk factors for development of clinically symptomatic disease and that structural changes may be part of aging, but not necessarily OA.

Previous studies have suggested classification criteria for early knee OA.^9,10 In two of these, early knee OA was defined as a combination of (1) knee pain, (2) KL grade 0 or 1, and (3) arthroscopic findings of cartilage lesions and/or MRI findings demonstrating articular cartilage degeneration and/or meniscal degeneration and/or subchondral BMLs. These criteria were updated after additional consensus meetings, suggesting that MRI should not be part of the criteria for early OA in routine clinical practice or primary care, in light of lack of validated consensus criteria and the high prevalence of structural joint changes detected by this method in the general population.⁶ Others have emphasized the role of MRI in the research setting due to its superior sensitivity to change and validity in the context of early OA but currently not thought appropriate for the routine clinical care setting because of its high cost and potential risk of over-diagnosis.⁴⁰

In summary, currently no accepted definition of early OA based on morphologic MRI is available. Structural features of OA are highly prevalent in knees without radiographic OA, but their clinical importance is still unclear and differentiation between normal aging and structural OA development remains a challenge. As OA is most likely best treated in early and potentially still reversible stages, development of an MRI definition of early OA seems warranted particularly in the context of clinical DMOAD development.

Magnetic resonance imaging: compositional imaging

Cartilage compositional MRI techniques are sensitive to changes in the extracellular matrix of articular cartilage.⁴¹ Compositional MRI in early OA has the potential to detect the initial stages of cartilage degeneration, at a stage where these changes may still be more amenable to lifestyle or pharmaceutical intervention.⁴² This would provide a considerable advance on conventional (morphological/structural) MRI, even with ongoing improvements in MRI acquisition and analysis. Structural cartilage damage implies damage to the collagen matrix; such changes are likely to be irreversible given the limited innate regenerative potential of collagen in adults.⁴³

A number of different compositional MRI techniques have been described over the past two decades (Table 1). The most widely studied of these is T₂ (transversal relaxation time) mapping, followed by T_1ρ (longitudinal relaxation time in the transverse plane in the presence of a radiofrequency [spin-lock] field) mapping. While other techniques offer improved specificity for a particular cartilage component (e.g. delayed gadolinium-enhanced MRI of cartilage [dGEMRIC] is more specific for proteoglycan/glycosaminoglycan [GAG] content than T_1ρ,⁴⁴ although both techniques are correlated to proteoglycan/GAG based on a recent systematic review)⁴¹, the greater accessibility, lack of requirement for gadolinium-based contrast agent, and feasibility at more commonly available field strengths are all factors which have driven the wider uptake of T_1ρ and, in particular, T₂ mapping. Meta-analyses have shown that T₂ and T_1ρ mapping are reliable and have discriminative validity both in the setting of established OA and, importantly, in individuals at risk of OA.^45,46

Table 1.

Compositional MRI techniques.

Technique	Cartilage component assessed	Pros	Cons
T₂ mapping	Collagen orientation, water content	Easily accessible Feasible at 3T	Not sensitive to early proteoglycan loss Susceptible to error due to magic angle effect
T_1ρ mapping	Macromolecular content, water content	Improved dynamic range and sensitivity to proteoglycan changes vs T₂ Feasible at 3T	Not readily available high SAR (energy deposition to tissue) with spin-lock pulses
T₂* mapping	Collagen orientation, water content	Can be combined with UTE imaging to assess deepest layers of cartilage Faster acquisition Feasible at 3T	Less extensively validated than T₂ UTE requires specialist non-Cartesian pulse sequences
dGEMRIC	GAG	GAG specificity	Requires IV gadolinium-based contrast agent Complicated scan protocol
Sodium	GAG	GAG specificity	Difficult at < 7 T Requires multinuclear capability and special coil Long acquisition
gagCEST	GAG	GAG specificity	Not reliable at < 7 T due to low CEST effect Susceptible to B0 inhomogeneity
DWI/DTI	Collagen orientation, proteoglycan content	Combined proteoglycan/collagen assessment	Typically limited spatial resolution and SNR for cartilage. Commercially available DWI/DTI sequences not optimized for MSK applications

CEST, chemical exchange saturation transfer; dGEMRIC, delayed gadolinium-enhanced MRI of cartilage; DTI, diffusion tensor imaging; DWI, diffusion-weighted imaging; GAG, glycosaminoglycan; gagCEST, GAG chemical exchange saturation transfer; MRI, magnetic resonance imaging; MSK, musculoskeletal; SAR, specific absorption rate; SNR, signal-to-noise ratio; T, Tesla; UTE, ultrashort echo time.

Ability of compositional MRI to predict OA incidence & progression

The OAI MRI protocol included T₂ mapping. Given the size, extensive follow-up, and well-characterized nature of this dataset, it is unsurprising that the majority of studies evaluating the ability of compositional MRI to predict OA incidence and worsening have focused on T₂ mapping. A number of approaches to assessing ability of compositional MRI to predict outcome are available; analysis can be done at the level of an individual focal lesion/abnormality, at the level of a cartilage plate (e.g. medial femur) or a defined subregion thereof (e.g. central medial femur), or at the level of a global value for the entire knee. It is encouraging that T₂ mapping has demonstrated predictive validity across these different scales. For example, Kretzschmar et al.⁴⁷ examined cartilage plates from knees, which developed new-onset cartilage lesions over a 4-year period, comparing against cartilage plates from control knees at both the focal lesion and cartilage plate level. The authors showed that, at the local level, cartilage T₂ values were significantly higher in case knees at 1 year prior to lesion onset, and at 2 years prior to onset at plate level (Figure 3). Razmjoo et al.⁴⁸ leveraged advances in automated segmentation and registration techniques to analyze the T₂ values of the entire OAI at a cartilage plate level (4796 knees). The authors demonstrated strong associations between compartmental T₂ values and incident radiographic OA within 2 years for all cartilage plates except the patella.

Figure 3.

Magnified views of the lateral femoral condyle in an individual with an incident cartilage defect. While the cartilage appears normal on the morphological sequence at 12 months prior to lesion onset (a), local T₂ elevation of the cartilage T₂ map (b, white arrow) is demonstrated in the lesion equivalent area 12 months prior to lesion onset (c, white arrow). Reproduced from Kretzschmar et al.⁴⁷ with permission.

Standardization of compositional MRI

Like any quantitative imaging technique, compositional MRI requires careful attention to detail across acquisition and analysis pipeline to ensure that values obtained are accurate, reproducible, and interpretable. Prior work has demonstrated variability across vendors, RF coils and pulse sequences for cartilage compositional MRI.^49–53 Additional sources of variability lie in image analysis approaches, including quantification methods (e.g. exponential decay function fitting for parameter mapping), in particular in situations where the image signal-to-noise ratio (SNR) is low, and manual image processing methods such as cartilage segmentation are used.^54,55

To address these issues, a musculoskeletal subcommittee was formed under Radiology Society of North America (RSNA) – Quantitative Imaging Biomarker Alliance (QIBA) taskforce in April 2017. The first profile of the committee has been focused on standardizing data acquisition techniques for cartilage T₂ and T_1ρ imaging as potential imaging biomarkers for OA.⁵⁶ 3D T_1ρ and T₂ imaging sequences with the same structure (termed as MAPSS) have been developed on three major MR platforms.⁵⁷ With harmonized protocols and centralized data processing, the inter-vendor inter-site variations have been reduced substantially as compared with earlier studies.⁵⁷ These results highlight the importance of standardizing data acquisition methods, including sequence structures and acquisition protocols, and demonstrate the great promise of minimizing inter-site inter-vendor differences by improving sequence performance (e.g. more robust to B₀ and B₁ inhomogeneity) and by standardizing data acquisition.

In addition to data acquisition, standardization and automation of post-processing pipelines are critical for clinical translation of compositional MRI. Significant advances in the field have been made for automatic cartilage segmentation based on deep-learning methods. Further work to standardize post-processing methods, such as constructing T₂/ T_1ρ maps, is needed.⁵⁸

Clinical utility of compositional MRI

Despite the availability of compositional MRI techniques for more than 20 years, they have yet to make a real impact on clinical care for OA. This may partially relate to some of the technical challenges described above (such as standardization) as well as the long acquisition time required. However, the major reason is the lack of disease modifying drug treatments for OA (DMOAD). This means that the identification of early disease has limited impact on patient management. Even in the absence of DMOAD therapy, improved standardization of compositional MRI could see clinical utility being realized for indications such as monitoring of cartilage repair tissue and assessment of suitability for joint preserving surgery (e.g. high tibial osteotomy and unicompartmental knee arthroplasty). Ongoing RSNA-QIBA efforts will continue to improve standardization in readiness for clinical utility.

Computed tomography

The strength of CT modalities in the assessment of OA lies in their ability to acquire more detailed imaging of mineralized joint tissues compared with radiography and MRI.⁵⁹ CT techniques overcome radiographic superposition of joint structures by acquiring a 3D data volume and involve faster and more easily standardizable acquisitions, lower cost, and the ability to acquire reconstructable 3D data sets with greater ease than MRI. Here, we focus on clinical scale techniques in vivo for early diagnosis of degenerative disease, although it should be noted that much relevant research has been undertaken ex vivo with microCT, not just in bone but also the assessment of articular soft tissues such as the meniscus.^60,61

With focus on the knee, a good place to start is the role of CT in the assessment of osteophytes as a cardinal feature of early OA.²³ One cohort study of individuals without radiographic OA showed that osteophytes were the most common finding on MRI in 74% (527/710),¹⁹ raising doubts over the value of radiography in their detection. Segal et al.⁶² confirmed these by showing that cone beam CT (CBCT) was superior to radiography in knee osteophyte detection with a sensitivity of 93% compared with 60% and an accuracy of 95% compared with 79%. Omoumi et al.⁶³ have subsequently demonstrated quantification of osteophytes by volume and spatial distribution around the proximal tibia using CT. Although it seems reasonable to consider CT as the new gold standard for osteophyte detection, no studies have yet confirmed this in place of MRI, likely due to the difficulty in setting a true reference standard.

The precise role of subchondral bone in early osteoarthritis remains unclear; however, animal studies have continued to demonstrate the importance of bone remodeling in OA development, including how early inhibition of bone remodeling can slow post-traumatic OA development.^64,65 In humans, significant differences in subchondral bone mineral density (BMD) have been demonstrated with multidetector CT between severe OA and non-OA knees with higher density in the OA-affected medial compartment and lower density in the lateral compartment in a pattern supported by a biomechanical hypothesis.⁶⁶ In 2020, Shiraishi et al.⁶⁷ used high-resolution peripheral quantitative CT (HRpQCT) to show strong relationships between subchondral bone microstructure at the anterior medial tibia and KL grade, suggesting it may be a useful index in monitoring treatment. Optimum quantitative assessment of BMD (and other relevant morphometric parameters) with CT means that further work needs to be done with clinical CT techniques to define bone behavior more clearly in early knee OA.

CBCT has recently found application at the knee in the context of OA due to the ability to perform weight bearing acquisitions,⁶⁸ comparisons naturally being made with radiography. In 2016, Segal et al.⁶⁹ showed that greater surface area with a low joint space width (JSW) measured on weight bearing CT (WBCT, Figure 4) correlated more strongly with tibial cartilage lesion severity than radiographic minimum JSW. In 2021, Turmezei et al.⁷⁰ subsequently demonstrated that 3D joint space mapping using WBCT yielded a smallest detectable difference in tibiofemoral JSW measurement of +/−0.1 mm for KL 0 and 1 knees, a sensitivity twice that of radiography.⁷¹ WBCT therefore looks to be a logical successor to radiography in the measurement of early joint space narrowing in OA if protocol standardization can be achieved.⁷²

Figure 4.

Bilateral knee weight-bearing cone beam CT in coronal (a) and sagittal (b, right knee) planes. The dotted lines demarcate the limits of the cone beam, like a discus in 3D, outside of which the imaging data reconstruction is unreliable. A substantial advantage to weight bearing is the ability to look at the physiologically loaded joint space in 3D, here controlled by a positioning device (P) to image in 20 degrees of knee joint flexion. Joint space narrowing is seen at both medial tibiofemoral compartments, worse on the left as marked between the white arrowheads. CT is also excellent for the visualization of osteophytes (white arrow), subchondral sclerosis (black arrowhead, also between the white arrowheads), and subchondral cysts (black arrow). In addition to 3D quantification of joint space width, it is also possible to detect meniscal extrusion (asterisk), evaluate patellofemoral disease, and study 3D joint morphology from the same acquisition that is only marginally more dose than traditional radiography. Imaging data courtesy of Professor Neil Segal, Kansas University Medical Center, Kansas City, USA.

One of the most important recent developments across all clinical imaging has been spectral CT, which has been achieved due to the realization of photon counting detector technology in clinical scanner systems.⁷³ In their study on a large animal model of OA with intra-articular contrast, Rajendran et al.⁷⁴ demonstrated simultaneous high-resolution and multi-energy imaging that allowed morphological assessment of cartilage loss as a marker of joint disease. Noncontrast spectral imaging has also been performed on ex vivo human knee specimens suggesting optimal virtual mono-energetic imaging of cartilage and bone simultaneously at 60 to 70 keV,⁷⁵ but further research is required to see how this may enhance our understanding of early human OA in vivo.

In a similar principle to the diffusion of contrast medium into articular cartilage with dGEMRIC, CT arthrography has shown excellent correlation of attenuation values with sulfated glycosaminoglycan content (R² = 0.90, p < 0.0001).⁷⁶ CT arthrography has also been able to depict articular cartilage lesions with superior performance and greater reliability compared with MR arthrography at the ankle,⁷⁷ but wider clinical application still remains limited by the need for an invasive intra-articular injection of contrast medium.

In keeping with its versatility, there are a number of other interesting CT applications with relevance to early OA such as semiquantitative scoring of intra-articular mineralization as a predisposing factor,⁷⁸ whole body scoring of OA load,⁷⁹ and the use of CT as the anatomical framework for metabolic imaging. An example of metabolic imaging with SPECT/CT has shown significantly higher bone tracer uptake in the medial tibiofemoral compartment in the presence of degenerate or torn menisci compared with normal menisci, which may be of value in identifying individuals with early OA.⁸⁰

The notion of identifying early OA with CT has been in researchers’ minds for decades,⁸¹ but it has been recent developments such as WBCT and spectral CT that have pushed it concertedly to the forefront. This has been allied with the advent of lower scan doses that make CT techniques more palatable for investigation in large numbers, in some cases with the dose equivalence of one day of background radiation per scan, just twice the reported dose of equivalent radiography.^82,83

Discussion

Although OA is a clinical diagnosis, symptoms occur relatively late in the overall disease course. Imaging has an important role to play in identifying preclinical pathological processes that have the potential to identify targets for prevention and treatment. In this review, we have evaluated the role of standard radiographs, morphological and compositional MRI techniques, and CT-based methods in the definition of early OA. Figure 5, taken from reference,⁴² well illustrates and summarizes the role of the different imaging techniques in diagnosing OA. We found that these methods show great promise and are likely to have complementary roles in unraveling the complexity of OA and thus providing new therapeutic approaches tailored to a personalized approach. Using novel imaging modalities to better understand the pathophysiology at early OA stages will provide targets for interventions, that can be performed before there is irreversible tissue damage. This includes, for example, joint inflammation/synovitis which can be well visualized with MRI and potentially treated when structural damage is not yet present. This concept has been studied in the evolution of post-traumatic OA^84,85 and may lead to novel treatment approaches.

Figure 5.

Disease stages in OA as defined by imaging.

Radiographs have been in widespread use in OA since the 1950s.⁸ They have provided the cornerstone of clinical and research work in OA. However significant, well described limitations exist in the use of radiographs for defining early OA as, by the time the first radiographic changes occur, significant pathological changes are already present.¹⁹ Furthermore, it is now well accepted that OA is a disease of the whole joint, yet radiography can only directly visualize bone, with other structures such as articular cartilage inferred from measurement of JSW. However, as radiographs are a very good way to assess bone, there has been a rekindling of work refining assessment of trabecular bone texture.¹⁷ This is promising as it has been shown to be a better predictor of progression of joint space narrowing than with KL-based models in a number of studies.¹⁷

MRI has a major advantage over radiographic techniques in that the whole joint and different tissue components can be visualized. However, to date, there is still no accepted definition of early OA based on morphologic MRI and there remains a lack of clarity as to what structural changes constitute early OA. Structural features of OA are common and predate radiographic OA.¹⁹ However, their clinical importance remains unclear and differentiation between normal aging and structural OA development remains a challenge. Understanding this will be very important as these structural changes are likely to identify different phenotypes of knee OA and different joint responses to risk factors. Pathological responses to important risk factors such as obesity are likely to occur in early life and understanding which can be attributed to aging and which are a pathological response will be important if we are to prevent the development of knee OA and treat it at the very early stages.

Compositional imaging adds important information on the health of the knee joint by providing rich data on the composition of cartilage and other tissues (such as menisci) before the development of morphological changes. The promise of cartilage compositional MRI in early OA lies in its ability to demonstrate cartilage abnormality at an earlier stage than conventional, structural MRI.⁴² However, as with any quantitative imaging technique, compositional MRI requires careful attention to detail across acquisition and analysis to ensure that values obtained are accurate, reproducible, and interpretable. There will also be the need for better standardization of post-processing of images and automated segmentation. It is also likely that the full benefit of compositional imaging will be seen when these data are used in concert with morphological and other imaging information so that a picture of the health of the knee joint, and any response to pathological processes, can be captured.

Recent developments such as WBCT and spectral CT have rekindled interest in CT as a method for defining early OA. The strength of CT modalities lies in the ability to acquire more detailed imaging of mineralized joint tissues compared with radiography and MRI.^59,62 The advent of lower scan doses makes it more feasible to use CT in large studies as it is now possible to do CT at doses only twice that of radiography.⁸² CT techniques can also be used to acquire a 3D volume data, but the relative advantage compared with MRI remains unclear.

Conclusion

In a primary care setting, the role of imaging in defining early knee OA is limited to standard radiographs and KL grades 0 and 1. However, radiographs have significant limitations in characterizing early degenerative change. Both morphological and compositional MRI show abnormalities at earlier stages. While these technologies are substantially more sensitive, there is no established definition for early OA using MRI. Compositional MRI has the potential to diagnose cartilage matrix degenerative changes at the earliest stages, yet clinical application of this technology is still limited. New CT techniques may also provide new insights into early knee degeneration. The role of advanced imaging technologies in OA will have increased clinical significance as new therapeutic modalities evolve, which will require an earlier definition of damage of joint tissues and better identification of therapeutic targets.

Footnotes

Acknowledgements

None.

Declarations

ORCID iDs

Frank W. Roemer

Thomas M. Link

References

Cui

Wang

, et al. Global, regional prevalence, incidence and risk factors of knee osteoarthritis in population-based studies. EClinicalMedicine 2020; 29–30: 100587.

Safiri

Kolahi

Smith

, et al. Global, regional and national burden of osteoarthritis 1990–2017: a systematic analysis of the Global Burden of Disease study 2017. Ann Rheum Dis 2020; 79: 819–828.

Hiligsmann

Cooper

Arden

, et al. Health economics in the field of osteoarthritis: an expert’s consensus paper from the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO). Semin Arthritis Rheum 2013; 43: 303–313.

Mahmoudian

Lohmander

Mobasheri

, et al. Early–stage symptomatic osteoarthritis of the knee – time for action. Nat Rev Rheumatol 2021; 17: 621–632.

Roos

Arden

NK.

Strategies for the prevention of knee osteoarthritis. Nat Rev Rheumatol 2016; 12: 92–101.

Luyten

Bierma-Zeinstra

Dell’Accio

, et al. Toward classification criteria for early osteoarthritis of the knee. Semin Arthritis Rheum 2018; 47: 457–463.

Mahmoudian

Lohmander

Jafari

, et al. Towards classification criteria for early-stage knee osteoarthritis: a population-based study to enrich for progressors. Semin Arthritis Rheum 2021; 51: 285–291.

Kellgren

Lawrence

Radiological assessment of osteoarthritis. Ann Rheum Dis 1957; 16: 494–501.

Luyten

Denti

Filardo

, et al. Definition and classification of early osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc 2012; 20: 401–406.

10.

Madry

Kon

Condello

, et al. Early osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc 2016; 24: 1753–1762.

11.

Meyer

Leroux

Levy

, et al. Flexion posteroanterior radiographs affect both enrollment for and outcomes after injection therapy for knee osteoarthritis. Orthop J Sports Med 2017; 5: 2325967117706692.

12.

Taguchi

Chiba

Okazaki

, et al. Characterization of cartilage defects detected by MRI in Kellgren-Lawrence grade 0 or 1 knees. J Orthop Sci 2017; 22: 868–873.

13.

Culvenor

Engen

Øiestad

, et al. Defining the presence of radiographic knee osteoarthritis: a comparison between the Kellgren and Lawrence system and OARSI atlas criteria. Knee Surg Sports Traumatol Arthrosc 2015; 23: 3532–3539.

14.

Mehta

Duryea

Badger

, et al. Comparison of 2 radiographic techniques for measurement of tibiofemoral joint space width. Orthop J Sports Med 2017; 5: 2325967117728675.

15.

Paixao

DiFranco

Ljuhar

, et al. A novel quantitative metric for joint space width: data from the osteoarthritis initiative (OAI). Osteoarthritis Cartilage 2020; 28: 1055–1061.

16.

Vignon

Piperno

Le Graverand

, et al. Measurement of radiographic joint space width in the tibiofemoral compartment of the osteoarthritic knee: comparison of standing anteroposterior and Lyon schuss views. Arthritis Rheum 2003; 48: 378–384.

17.

Almhdie-Imjabbar

Nguyen

Toumi

, et al. Prediction of knee osteoarthritis progression using radiological descriptors obtained from bone texture analysis and Siamese neural networks: data from OAI and MOST cohorts. Arthritis Res Ther 2022; 24: 66.

18.

Shamir

Ling

Scott

, et al. Early detection of radiographic knee osteoarthritis using computer-aided analysis. Osteoarthritis Cartilage 2009; 17: 1307–1312.

19.

Guermazi

Niu

Hayashi

, et al. Prevalence of abnormalities in knees detected by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis study). BMJ 2012; 345: e5339.

20.

Schiphof

Oei

Hofman

, et al. Sensitivity and associations with pain and body weight of an MRI definition of knee osteoarthritis compared with radiographic Kellgren and Lawrence criteria: a population-based study in middle-aged females. Osteoarthritis Cartilage 2014; 22: 440–446.

21.

Hayashi

Felson

Niu

, et al. Pre-radiographic osteoarthritic changes are highly prevalent in the medial patella and medial posterior femur in older persons: Framingham OA study. Osteoarthritis Cartilage 2014; 22: 76–83.

22.

Kumm

Turkiewicz

Zhang

, et al. Structural abnormalities detected by knee magnetic resonance imaging are common in middle-aged subjects with and without risk factors for osteoarthritis. Acta Orthop 2018; 89: 535–540.

23.

Hunter

Arden

Conaghan

, et al. Definition of osteoarthritis on MRI: results of a Delphi exercise. Osteoarthritis Cartilage 2011; 19: 963–969.

24.

Roemer

Eckstein

Duda

, et al. Frequencies of MRI-detected structural pathology in knees without radiographic OA and worsening over three years: how relevant is contralateral radiographic osteoarthritis? Osteoarthr Cartil Open 2020; 1: 100014.

25.

Eckstein

Maschek

Roemer

, et al. Cartilage loss in radiographically normal knees depends on radiographic status of the contralateral knee – data from the osteoarthritis initiative. Osteoarthritis Cartilage 2019; 27: 273–277.

26.

Roemer

Felson

Wang

, et al. Co-localisation of non-cartilaginous articular pathology increases risk of cartilage loss in the tibiofemoral joint –the MOST study. Ann Rheum Dis 2013; 72: 942–948.

27.

Bijlsma

Berenbaum

Lafeber

FP.

Osteoarthritis: an update with relevance for clinical practice. Lancet 2011; 377: 2115–2126.

28.

Englund

Guermazi

Roemer

, et al. Meniscal tear in knees without surgery and the development of radiographic osteoarthritis among middle-aged and elderly persons: the multicenter osteoarthritis study. Arthritis Rheum 2009; 60: 831–839.

29.

Englund

Guermazi

Gale

, et al. Incidental meniscal findings on knee MRI in middle-aged and elderly persons. N Engl J Med 2008; 359: 1108–1115.

30.

Englund

Roemer

Hayashi

, et al. Meniscus pathology, osteoarthritis and the treatment controversy. Nature Rev Rheumatol 2012; 8: 412–419.

31.

Benito

Veale

FitzGerald

, et al. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005; 64: 1263–1267.

32.

Roemer

Guermazi

Felson

, et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011; 70: 1804–1809.

33.

Libicher

Ivancic

Hoffmann

, et al. Early changes in experimental osteoarthritis using the Pond-Nuki dog model: technical procedure and initial results of in vivo MR imaging. Eur Radiol 2005; 15: 390–394.

34.

Guermazi

Hayashi

Roemer

, et al. Brief report: partial- and full-thickness focal cartilage defects contribute equally to development of new cartilage damage in knee osteoarthritis: the multicenter osteoarthritis study. Arthritis Rheumatol 2017; 69: 560–564.

35.

Roemer

Kwoh

Hannon

, et al. What comes first? Multitissue involvement leading to radiographic osteoarthritis: magnetic resonance imaging-based trajectory analysis over four years in the osteoarthritis initiative. Arthritis Rheumatol 2015; 67: 2085–2096.

36.

Sharma

Chmiel

Almagor

, et al. Significance of preradiographic magnetic resonance imaging lesions in persons at increased risk of knee osteoarthritis. Arthritis Rheumatol 2014; 66: 1811–1819.

37.

Roemer

Demehri

Omoumi

, et al. State of the art: imaging of osteoarthritis-revisited 2020. Radiology 2020; 296: 5–21.

38.

Case

Thomas

Clarke

, et al. Prodromal symptoms in knee osteoarthritis: a nested case-control study using data from the osteoarthritis initiative. Osteoarthritis Cartilage 2015; 23: 1083–1089.

39.

Magnusson

Kumm

Turkiewicz

, et al. A naturally aging knee, or development of early knee osteoarthritis? Osteoarthritis Cartilage 2018; 26: 1447–1452.

40.

Emery

Whittaker

Mahmoudian

, et al. Establishing outcome measures in early knee osteoarthritis. Nat Rev Rheumatol 2019; 15: 438–448.

41.

Emanuel

Kellner

Peters

MJM

, et al. The relation between the biochemical composition of knee articular cartilage and quantitative MRI: a systematic review and meta-analysis. Osteoarthritis Cartilage 2022; 30: 650–662.

42.

Kwoh

CK.

Epidemiology of osteoarthritis. In: Newman

Cauley

(eds) The epidemiology of aging. Dordrecht: Springer, 2012, pp. 523–536.

43.

Heinemeier

Schjerling

Heinemeier

, et al. Radiocarbon dating reveals minimal collagen turnover in both healthy and osteoarthritic human cartilage. Sci Transl Med 2016; 8: 346ra90.

44.

van Tiel

Kotek

Reijman

, et al. Is T1rho mapping an alternative to delayed gadolinium-enhanced MR imaging of cartilage in the assessment of sulphated glycosaminoglycan content in human osteoarthritic knees? An in vivo validation study. Radiology 2016; 279: 523–531.

45.

Atkinson

Birmingham

Moyer

, et al. MRI T2 and T1rho relaxation in patients at risk for knee osteoarthritis: a systematic review and meta-analysis. BMC Musculoskelet Disord 2019; 20: 182.

46.

MacKay

Low

SBL

Smith

, et al. Systematic review and meta-analysis of the reliability and discriminative validity of cartilage compositional MRI in knee osteoarthritis. Osteoarthritis Cartilage 2018; 26: 1140–1152.

47.

Kretzschmar

Nevitt

Schwaiger

, et al. Spatial distribution and temporal progression of T₂ relaxation time values in knee cartilage prior to the onset of cartilage lesions – data from the osteoarthritis initiative (OAI). Osteoarthritis Cartilage 2019; 27: 737–745.

48.

Razmjoo

Caliva

Lee

, et al. T₂ analysis of the entire osteoarthritis initiative dataset. J Orthop Res 2020; 39: 74–85.

49.

Balamoody

Williams

Wolstenholme

, et al. Magnetic resonance transverse relaxation time T₂ of knee cartilage in osteoarthritis at 3-T: a cross-sectional multicentre, multivendor reproducibility study. Skeletal Radiol 2013; 42: 511–520.

50.

Dardzinski

Schneider

Radiofrequency (RF) coil impacts the value and reproducibility of cartilage spin-spin (T₂) relaxation time measurements. Osteoarthritis Cartilage 2013; 21: 710–720.

51.

Kim

Mamoto

Lartey

, et al. Multi-vendor multi-site T_1ρ and T₂ quantification of knee cartilage. Osteoarthritis Cartilage 2020; 28: 1539–1550.

52.

Matzat

McWalter

Kogan

, et al. T₂ relaxation time quantitation differs between pulse sequences in articular cartilage. J Magn Reson Imaging 2015; 42: 105–113.

53.

Pai

Majumdar

A comparative study at 3 T of sequence dependence of T2 quantitation in the knee. Magn Reson Imaging 2008; 26: 1215–1220.

54.

Koff

Amrami

Felmlee

, et al. Bias of cartilage T₂ values related to method of calculation. Magn Reson Imaging 2008; 26: 1236–1243.

55.

Raya

Dietrich

Horng

, et al. T₂ measurement in articular cartilage: impact of the fitting method on accuracy and precision at low SNR. Magn Reson Med 2010; 63: 181–193.

56.

Chalian

Guermazi

, et al. The QIBA profile for MRI-based compositional imaging of knee cartilage. Radiology 2021; 301: 423–432.

57.

Han

Busse

, et al. In vivo T(1rho) mapping in cartilage using 3D magnetization-prepared angle-modulated partitioned k-space spoiled gradient echo snapshots (3D MAPSS). Magn Reson Med 2008; 59: 298–307.

58.

Yang

Colak

Chundru

, et al. Automated knee cartilage segmentation for heterogeneous clinical MRI using generative adversarial networks with transfer learning. Quant Imaging Med Surg 2022; 12: 2620–2633.

59.

Bousson

Lowitz

Laouisset

, et al. CT imaging for the investigation of subchondral bone in knee osteoarthritis. Osteoporos Int 2012; 23(Suppl. 8): S861–S865.

60.

Karjalainen

Kestilä

Finnilä

, et al. Quantitative three-dimensional collagen orientation analysis of human meniscus posterior horn in health and osteoarthritis using micro-computed tomography. Osteoarthritis Cartilage 2021; 29: 762–772.

61.

Lahm

Kasch

Spank

, et al. Correlation between 3D microstructural and 2D histomorphometric properties of subchondral bone with healthy and degenerative cartilage of the knee joint. Histol Histopathol 2014; 29: 1477–1488.

62.

Segal

Nevitt

Lynch

, et al. Diagnostic performance of 3D standing CT imaging for detection of knee osteoarthritis features. Phys Sportsmed 2015; 43: 213–220.

63.

Omoumi

Schuler

Babel

, et al. Proximal tibial osteophyte volumes are correlated spatially and with knee alignment: a quantitative analysis suggesting the influence of biochemical and mechanical factors in the development of osteophytes. Osteoarthritis Cartilage 2021; 29: 1691–1700.

64.

Adebayo

Wan

, et al. Role of subchondral bone properties and changes in development of load-induced osteoarthritis in mice. Osteoarthritis Cartilage 2017; 25: 2108–2118.

65.

Ziemian

Witkowski

Wright

, et al. Early inhibition of subchondral bone remodeling slows load-induced posttraumatic osteoarthritis development in mice. J Bone Miner Res 2021; 36: 2027–2038.

66.

Omoumi

Babel

Jolles

, et al. Quantitative regional and sub-regional analysis of femoral and tibial subchondral bone mineral density (sBMD) using computed tomography (CT): comparison of non-osteoarthritic (OA) and severe OA knees. Osteoarthritis Cartilage 2017; 25: 1850–1857.

67.

Shiraishi

Chiba

Okazaki

, et al. In vivo analysis of subchondral trabecular bone in patients with osteoarthritis of the knee using second-generation high-resolution peripheral quantitative computed tomography (HR-pQCT). Bone 2020; 132: 115155.

68.

Lôbo

CFT

Bordalo-Rodrigues

and Weight-Bearing Computed Tomography International Study Group. Weight-bearing cone beam CT scans and its uses in ankle, foot, and knee: an update article. Acta Ortop Bras 2021; 29: 105–110.

69.

Segal

Frick

Duryea

, et al. Correlations of medial joint space width on fixed-flexed standing computed tomography and radiographs with cartilage and meniscal morphology on magnetic resonance imaging. Arthritis Care Res 2016; 68: 1410–1416.

70.

Turmezei

B Low

Rupret

, et al. Quantitative three-dimensional assessment of knee joint space width from weight-bearing CT. Radiology 2021; 299: 649–659.

71.

Hunter

DJ.

Risk stratification for knee osteoarthritis progression: a narrative review. Osteoarthritis Cartilage 2009; 17: 1402–1407.

72.

Roemer

FW.

Weight-bearing CT for knee osteoarthritis assessment: a story unfolds. Radiology 2021; 299: 660–661.

73.

Willemink

Persson

Pourmorteza

, et al. Photon-counting CT: technical principles and clinical prospects. Radiology 2018; 289: 293–312.

74.

Rajendran

Murthy

Frick

, et al. Quantitative knee arthrography in a large animal model of osteoarthritis using photon-counting detector CT. Invest Radiol 2020; 55: 349–356.

75.

Chappard

Abascal

Olivier

, et al. Virtual monoenergetic images from photon-counting spectral computed tomography to assess knee osteoarthritis. Eur Radiol Exp 2022; 6: 10.

76.

Siebelt

van Tiel

Waarsing

, et al. Clinically applied CT arthrography to measure the sulphated glycosaminoglycan content of cartilage. Osteoarthritis Cartilage 2011; 19: 1183–1189.

77.

Schmid

Pfirrmann

Hodler

, et al. Cartilage lesions in the ankle joint: comparison of MR arthrography and CT arthrography. Skeletal Radiol 2003; 32: 259–265.

78.

Guermazi

Jarraya

Lynch

, et al. Reliability of a new scoring system for intraarticular mineralization of the knee: Boston University Calcium Knee Score (BUCKS). Osteoarthritis Cartilage 2020; 28: 802–810.

79.

Gielis

Weinans

Nap

, et al. Scoring osteoarthritis reliably in large joints and the spine using whole-body CT: osteoarthritis computed tomography-score (OACT-score). J Pers Med 2020; 11: 5.

80.

Rechsteiner

Hirschmann

Dordevic

, et al. Meniscal pathologies on MRI correlate with increased bone tracer uptake in SPECT/CT. Eur Radiol 2018; 28: 4696–4704.

81.

Ruegsegger

Munch

Felder

Early detection of osteoarthritis by 3D computed tomography. Technol Health Care 1993; 1: 53–66.

82.

Kokkonen

Suomalainen

Joukainen

, et al. In vivo diagnostics of human knee cartilage lesions using delayed CBCT arthrography. J Orthop Res 2014; 32: 403–412.

83.

Mettler

Jr Huda

Yoshizumi

, et al. Effective doses in radiology and diagnostic nuclear medicine: a catalog. Radiology 2008; 248: 254–263.

84.

Heilmeier

Mamoto

Amano

, et al. Infrapatellar fat pad abnormalities are associated with a higher inflammatory synovial fluid cytokine profile in young adults following ACL tear. Osteoarthritis Cartilage 2020; 28: 82–91.

85.

Theeuwes

van den Bosch

MHJ

Thurlings

, et al. The role of inflammation in mesenchymal stromal cell therapy in osteoarthritis, perspectives for post-traumatic osteoarthritis: a review. Rheumatology 2021; 60: 1042–1053.