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)

Journal Pre-proof 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)

Kazuteru Shiraishi, Ko Chiba, Narihiro Okazaki, Kazuaki Yokota, Yusuke Nakazoe, Kenichi Kidera, Akihiko Yonekura, Masato Tomita, Makoto Osaki PII:

S8756-3282(19)30449-1

DOI:

https://doi.org/10.1016/j.bone.2019.115155

Reference:

BON 115155

To appear in:

Bone

Received date:

17 March 2019

Revised date:

11 November 2019

Accepted date:

12 November 2019

Please cite this article as: K. Shiraishi, K. Chiba, N. Okazaki, et al., In vivo analysis of subchondral trabecular bone in patients with osteoarthritis of the knee using secondgeneration high-resolution peripheral quantitative computed tomography (HR-pQCT), Bone(2018), https://doi.org/10.1016/j.bone.2019.115155

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2018 Published by Elsevier.

Journal Pre-proof 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)

Kazuteru Shiraishi MD, Ko Chiba MD, PhD, Narihiro Okazaki MD, PhD, Kazuaki Yokota MD, Yusuke Nakazoe MD, Kenichi Kidera MD, PhD, Akihiko Yonekura MD, PhD, Masato Tomita MD,

ro of

PhD, and Makoto Osaki MD, PhD

Department of Orthopedic Surgery, Nagasaki University Graduate School of Biomedical Sciences,

re

-p

Nagasaki, Japan

lP

Corresponding author Ko Chiba

+81-95-819-7321

Jo

ur

[email protected]

na

1-7-1 Sakamoto, Nagasaki 852-8501, Japan

COI The authors declare no conflicts of interest associated with this manuscript.

Journal Pre-proof 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)

Abstract Objective Subchondral bone plays an important role in the pathological mechanisms of knee osteoarthritis (OA). High-resolution peripheral quantitative computed tomography (HR-pQCT) is an imaging modality allowing noninvasive microstructural

ro of

analysis of human bone, and the second generation enables scanning of the knee.

The purpose of this study was to perform in vivo analysis of subchondral trabecular bone in patients with medial knee OA, to elucidate features of bone microstructure in medial knee OA, and to investigate relationships between bone

re

-p

microstructure and both stage of disease and lower limb alignment.

Methods

lP

Subjects were 20 women, including both patients with medial knee OA (Kellgren-Lawrence (KL) grade 2, n=5, KL grade 3, n=7, and KL grade 4, n=4; mean age: 63.0 years; body mass index (BMI): 23.8 kg/m2) and volunteers without knee OA

na

(KL grade 1, n=4, mean age: 66.0 years; BMI: 23.8 kg/m2). The proximal tibia (20-mm length) was scanned by second-generation HR-pQCT at a voxel size of 60.7 µm. A subchondral trabecular bone volume of 5 mm length was

ur

extracted from the medial and lateral plateaus. They were then divided into 4 regions: anterior, central, medial or lateral,

Jo

and posterior. Finally, subchondral bone microstructure parameters were analyzed and compared, between each plateau and region. Relationships between microstructural parameters and disease stage (KL grade, minimum joint space width), and between those parameters and lower limb alignment (femorotibial angle: FTA, mechanical axis deviation: MAD) were also investigated.

Results In the medial plateau, volumetric bone mineral density (vBMD), bone volume fraction (BV/TV), and trabecular thickness were significantly higher and structure model index (SMI) was significantly lower than in the lateral plateau, particularly in the anterior, central, and medial regions (p < 0.01 each). In the anterior region of the medial plateau, vBMD, BV/TV, and connectivity density showed strong positive correlations with KL grade, FTA, and MAD (r-range: 0.61 to 0.83), while trabecular separation and SMI exhibited strong negative

Journal Pre-proof correlations with KL grade, FTA, and MAD (r-range: -0.60 to -0.83).

Conclusions Higher bone volume, trabecular thickness, and a more plate-like structure were observed in the medial tibial plateau than in the lateral. Subchondral bone microstructure at the anterior region in the medial plateau showed strong relationships with KL grade and lower limb alignment. These results indicate that subchondral bone microstructure in this region may provide representative indices, particularly in medial knee OA. Although this study involved a specifically Asian cohort with a lower BMI distribution than other ethnic groups, the technique presented may be useful in studying the pathogenesis of OA

-p

ro of

or evaluating treatment effects.

re

Keywords: HR-pQCT; Osteoarthritis; Knee; Subchondral bone; Bone microstructure

Jo

ur

na

lP

Running Title: Subchondral trabecular bone analysis of knee OA using HR-pQCT

Journal Pre-proof 1. Introduction Knee osteoarthritis (OA) is considered a joint disease of elderly people, causing degeneration of articular cartilage and subchondral bone, chronic pain, and walking disability. In Japan, the prevalence of radiographic knee OA at  40 years old was estimated as 54.6% (42.0% in men, 61.5% in women), and the number of patients with knee OA was approximately 25 million [1]. Knee OA is thus regarded as a serious social issue. OA also causes irreversible changes to the joint with progression. Symptomatic treatment with analgesics remains the mainstay of treatment, and patients with progressive knee OA have no choices other than surgical treatment. Accordingly, clarification of the pathological mechanisms underlying OA is important. However, this is currently not

ro of

simple, and increasing evidence suggest that OA is a whole-joint disease involving complex etiologies, such as cartilage destruction, abnormal metabolism of subchondral bone, and inflammation of the synovial membrane [2-5]. Among these factors, subchondral bone has been considered to play an important role in OA pathogenesis, and is currently expected to

-p

offer a target in the treatment of OA [6, 7].

re

Many studies have demonstrated that subchondral bone resorption increases and trabecular bone becomes thinner in early-stage OA based on OA animal models, such as rats [8], guinea pigs [9, 10], and dogs [11]. Another study reported that

lP

bone resorption inhibitors can prevent cartilage attrition [12]. On the other hand, using histology on human tissue, many studies have demonstrated that subchondral bone volume increases and trabecular bone becomes thicker with degeneration

na

of the articular cartilage in late-stage OA [13,14]. Consequently, the pathological mechanisms in histological terms remain unsolved. In particular, evaluating human histology in early-stage OA is difficult for ethical reasons.

ur

Many studies have reported imaging analyses of knee OA. However, dual-energy X-ray absorptiometry (DXA) can

Jo

assess only areal bone mineral density (aBMD) [15, 16], and quantitative computed tomography (QCT) can assess only volumetric bone mineral density (vBMD) [17, 18]. Although micro-computed tomography (micro-CT) offers high spatial resolution and the capability to assess bone microstructure in detail, many studies have only reported on ex vivo human specimens retrieved from patients who had undergone knee arthroplasty [19-23]. Although some studies tried to assess human bone microstructure in vivo using MRI, some degree of inaccuracy needs to be considered because MRI detects signals mainly from the bone marrow, synovial fluid, or fat, not from trabecular bone, and in addition, chemical-shift artifacts may occur [24, 25]. As described above, extensive in vivo analyses related to human knee OA have not been provided in existing imaging analysis studies. High-resolution peripheral quantitative computed tomography (HR-pQCT) provides in vivo analysis of bone microarchitecture at human peripheral sites with high spatial resolution and low exposure to radiation [26, 27]. Second-generation HR-pQCT has a longer gantry, enabling provision of more proximal images, such as the knee or elbow

Journal Pre-proof [28]. Consequently, several in vivo studies on the knee joint using HR-pQCT have already been reported [28-31]. However, no evaluations of early to end-stage knee OA using HR-pQCT have yet been reported. We therefore performed an analysis of subchondral bone microstructure using cross-sectional HR-pQCT in early to end-stage medial knee OA. The purpose of this exploratory study was to investigate: 1) characteristics of subchondral bone microstructure in medial knee OA; and 2) relationships between subchondral bone microstructure and both stage of disease

Jo

ur

na

lP

re

-p

ro of

and lower limb alignment.

Journal Pre-proof 2. Subjects and methods Subjects Participants in this study were 20 Japanese women, comprising 16 patients with medial knee OA and 4 volunteers without knee OA. Medial knee OA is defined as medial joint space narrowing and osteophytes on the knee radiograph with Kellgren-Lawrence (KL) grade 2 or more in this study [32]. Women with a history of trauma or inflammatory arthritis, such as rheumatoid arthritis, were excluded. These volunteers did not use drugs that affect bone metabolism, such as anti-osteoporosis drugs. This study protocol was approved by Nagasaki University Hospital Clinical Research Ethics Committee (registration

ro of

number: 16103116) and complied with the principles of the Declaration of Helsinki of 1975, revised in 2000. Informed consent was obtained from all participants prior to enrolment.

-p

Imaging

re

(a) HR-pQCT

Imaging of the knee was performed using a second-generation HR-pQCT system (XtremeCT II; Scanco Medical,

lP

Brüttisellen, Switzerland). The knee of the subject was fixed in full extension in a custom leg holder specifically adjustable for imaging of the knee, then inserted into the gantry with the subject in a sitting position (Fig. 1). The scan region was 20

na

mm in length at the knee joint, to include the proximal end of the tibia. Scan settings were as follows: voltage 68 kVp, current 1470 µA, integration time 100 ms, number of projections 900, field of view 140 mm, matrix 2304×2304, and voxel

ur

size 60.7 µm. Total number of slices was 336, scan time was 5.9 min. The radiation dose was 23.7 mGy as the volume

Jo

computed tomography dose index (vCTDI), 48.3 mGycm as the dose-length product (DLP), and 22 μSv as the effective dose. They were calculated based on the measurements using a dosimetry phantom for HR-pQCT.

(b) Plain X-ray Frontal radiographs of the full-length lower extremities were obtained using the X-ray equipment in a standing position, keeping the knee extended.

Measurements (a) HR-pQCT Using bone microstructure measurement software (TRI/3D-BON; Ratoc System Engineering, Tokyo, Japan), registration of subchondral trabecular bone, segmentation of regions of interest (ROIs), and measurement of subchondral bone

Journal Pre-proof microstructural parameters were performed. As shown in Figure 2 and 3, analyses were performed separately for the medial and lateral tibial plateaus. For each plateau, on the coronal and sagittal views, a subchondral trabecular bone volume of 5 mm length was extracted. This volume was 2 mm below and parallel to the joint surface (Fig. 2A and D). A binarized image was then created with a threshold value of 320 mg/cm3 (Fig. 3A and B) [33], and the entire bone region was obtained after filling all bone marrow spaces of the binarized image (Fig. 3C). Subchondral trabecular bone volume was extracted by deleting a 1-mm external margin of the entire bone region on the axial view to exclude cortical bone (Fig. 3D). ROI selection was performed as follows on a slice-by-slice basis for each subchondral trabecular bone compartment.

ro of

Based on the maximum length from medial to lateral, the subchondral trabecular volume was divided into three compartments of equal length on the axial view (Fig. 3E). The compartments of the medial one-third and lateral one-third were defined as the medial plateau (Med) and lateral plateau (Lat), respectively (Fig. 2B and E).

-p

Each plateau (Med and Lat) was then further divided into four regions (anterior, central, posterior, and medial or lateral).

re

This process was also performed on a slice-by-slice basis. The central regions were defined as rectangles with one side equal to one-third of the anterior-posterior length and the other side equal to half of the medial-lateral length in each plateau

lP

(Fig. 3F). To define other regions, two lines were drawn outward at oblique angles (45°) from the two outer vertices of the rectangle defining the central regions (Fig. 3F). Finally, the medial plateau (Med) was divided into four regions: anterior

na

(medA), central (medC), medial (medM), and posterior (medP) regions (Fig. 2C). Similarly, the lateral plateau (Lat) was divided into four regions: anterior (latA), central (latC), lateral (latL), and posterior (latP) regions (Fig. 2F).

ur

The following trabecular microstructural parameters were measured in the volumes of interest (VOIs) of the 5-mm length

Jo

inferior-superiorly established as above: vBMD (mg/cm3), bone volume fraction (BV/TV) (%), trabecular thickness (Tb.Th) (μm), trabecular number (Tb.N) (1/mm), trabecular separation (Tb.Sp) (μm), structure model index (SMI), connectivity density (Conn.D) (1/mm3), and degree of anisotropy (DA) [34]. The vBMD was calculated from X-ray attenuation values using a calibration curve that was created by imaging a reference phantom comprising five hydroxyapatite rods (0, 100, 200, 400, 800 mgHA/cm3). BV/TV was calculated by a voxel-based measurement on binarized images created with a fixed threshold value (320 mg/cm3) [33]. Tb.Th, Tb.N, and Tb.Sp were measured directly. SMI is an index in which a trabecular bone structure showing a value of 0 indicates a plate-like structure, and that showing a value of 3 indicates a rod-like structure. Conn.D is a parameter to qualify trabecular connectivity, with a higher value indicating greater connectivity. DA is a parameter to qualify trabecular structure directions, with a higher value indicating higher anisotropy.

Journal Pre-proof (b) Plain X-ray The following parameters were measured on frontal radiographs of the full-length lower extremities in a standing position (SYNAPSE; Fujifilm, Tokyo, Japan). KL grade was classified into 1 to 4 based on medial joint space narrowing and osteophytes [32]. Minimum joint space width (mJSW) is the minimum width in the medial joint space. Femorotibial angle (FTA) is defined as a laterally opened angle formed by the axes of the femoral and tibial diaphyses [35]. Varus alignment is defined as FTA > 175° [35]. Mechanical axis is the line connecting the center of the femoral head to the center of the ankle joint, while mechanical axis deviation (MAD) is the distance from the center of the knee joint to the mechanical axis [35].

ro of

Varus alignment is defined as medial deviation > 15 mm [35].

Statistical analysis

Regarding characteristics of subchondral bone microstructure in medial knee OA, the statistical significances of

-p

differences between subchondral bone microstructural parameters in the medial and lateral plateaus were determined using

re

the paired t-test after checking for normality and homoscedasticity. The statistical significance of differences among subchondral bone microstructural parameters among all eight regions (28 comparisons) was determined using the paired

lP

t-test with Bonferroni adjustment after analysis of variance (ANOVA) for repeated measures. Regarding relationships between subchondral bone microstructural parameters and either disease stage (KL grade and

na

mJSW) or lower limb alignment (FTA and MAD), Spearman's correlation was used (correlation coefficient = r). A scatter

medP.

ur

plot with linear regression analysis was also added for the relationships of “BV/TV vs. FTA” in medA, medC, medM, and

Jo

SPSS version 16.0 (SPSS, Chicago, IL) was used for all statistical analyses. Because of the relatively small number of subjects in this study, values of p < 0.01 were considered as significant to show more reliable results.

Journal Pre-proof 3. Results The characteristics of subjects are shown in Table 1. Mean age of the 20 participants was 63.6 ± 7.9 years (range, 51–81 years). Mean height was 1.53 ± 0.05 m, mean weight was 56.4 ± 10.3 kg, and mean body mass index (BMI) was 23.8 ± 3.5 kg/m2.

(1) Subchondral bone microstructural characteristics of medial knee OA Bone microstructural parameters of medial knee OA patients (KL grade ≥ 2, n = 16) are shown in Table 2. Significant differences in bone microstructure were identified between the medial plateau (Med) and lateral plateau (Lat).

ro of

The subchondral bone microstructure in the medial plateau was formed by denser, thicker trabeculae, with more plate-like, more anisotropic structures than in the lateral plateau (vBMD: 119 and 70 mg/cm3; BV/TV: 18.0 and 10.4%; Tb.Th: 293 and 231 μm; SMI: 1.81 and 2.20; DA: 1.45 and 1.27, medial and lateral, respectively, p < 0.01 each).

-p

Significant differences were also identified in bone microstructure among all regions within the medial and lateral

re

plateaus. Subchondral bone microstructure in medA, medC, and medM consisted of denser and more plate-like trabeculae compared to medP and the regions in the lateral plateau (vBMD: 124, 124, and 155 mg/cm3; BV/TV: 18.9, 19.3, and

lP

23.9%; SMI: 1.74, 1.76, and 1.52, medA, medC, and medM, respectively, corrected p < 0.01 each). In particular, subchondral bone microstructure in medM consisted of thicker, denser, and more connected trabeculae compared to the

na

other regions (Tb.Th: 306 µm, Tb.N: 0.67/mm, Conn.D: 2.30, Tb.Sp: 495 µm, corrected p < 0.01 each). In contrast, the subchondral bone microstructure in the medP had lower density and a more rod-like structure than the medA, medC, and

Jo

ur

medM.

(2) Relationships between bone microstructure parameters, stages of disease, and lower limb alignment The results of lower limb alignment are shown in Table 1. Mean FTA was 179.6° (175–190°) and mean MAD was 16.0 mm of medial deviation (0–59.2 mm). These results indicated that lower limb alignment in the present study ranged between neutral and varus. Correlations between subchondral bone microstructural parameters and both stage of disease and lower limb alignment (KL grade, mJSW, FTA, and MAD) in all knees (n = 20) are shown in Table 3. While vBMD, BV/TV, Tb.Th, and Conn.D in the Med showed positive correlations with KL grade and FTA (r-range: 0.63 to 0.79), Tb.Sp and SMI in the Med displayed negative correlations with KL grade and FTA (r-range: -0.58 to -0.70). Both vBMD and BVTV in the Med showed positive correlations with MAD (r = 0.56 and 0.59), and Tb.Sp in the Med displayed a negative correlation with MAD (r = -0.56). In addition, vBMD, BV/TV, and Tb.Th had negative correlations

Journal Pre-proof with mJSW (r = -0.56, -0.59 and -0.58, respectively), and Tb.Sp had a positive correlation with mJSW (r = 0.60). On the other hand, only a correlation between Tb.N in the Lat and mJSW was evident in the lateral plateau. In all regions, bone microstructural parameters at medA showed the highest correlations with stage of disease and lower limb alignment. That is, vBMD, BV/TV, Tb.Th, Tb.N, and Conn.D at medA showed positive correlations with KL grade and FTA (r-range: 0.63 to 0.83), whereas Tb.Sp and SMI had negative correlations with KL grade and FTA (r-range: -0.71 to -0.83). In addition, vBMD, BV/TV, and Conn.D at medA showed positive correlations with MAD (r = 0.63, 0.69, and 0.61, respectively) and Tb.Sp and SMI had negative correlations with MAD (r = -0.60 and -0.71). Conversely, vBMD, BV/TV, Tb.Th, and Conn.D at medA displayed negative correlations with mJSW (r-range: -0.56 to -0.63), and Tb.Sp and

ro of

SMI had positive correlations with mJSW (r = 0.69 and 0.68, respectively).

Scatter plots of FTA and BV/TV in the medial plateau are shown in Figure 4. Significant positive correlations were shown in the medA, medM, and medP. In particular, BV/TV at medA showed the strongest correlation with FTA.

-p

Regional distributions of BV/TV are shown in Figure 5, with subject groups subdivided by KL grade. No statistical test

re

was performed when stratifying the subjects by KL grade, due to the limited sample size. For KL grade 1, no marked differences were evident among all eight regions in the medial and lateral plateaus (A). For KL grade 2, BV/TV was the

lP

highest in medM compared to the other regions (B). For KL grade 3, BV/TV at medA, medC, and medM were higher than the others (C). For KL grade 4, BV/TV at medM and medA were markedly higher than the others (D). On the other hand,

Jo

ur

na

BV/TV at medP was lower than at the other three medial regions (C and D).

Journal Pre-proof 4. Discussion The purpose of this study was to investigate subchondral bone microstructure in medial knee OA using HR-pQCT. Higher bone volume, trabecular thickness, and more plate-like structures were evident in medial plateau (particularly in anterior, central, and medial regions) than in lateral. In addition, bone microstructural parameters at the anterior region in the medial plateau showed the highest correlation with stage of disease and lower limb alignment. As shown in Table 2, vBMD, BV/TV, Tb.Th, and DA were significantly higher and SMI was significantly lower in the medial plateau compared with those in the lateral plateau in medial knee OA. This result indicates that subchondral trabecular bone in the medial tibial plateau not only possesses higher bone volume and trabecular thickness, but also

ro of

comprises a more plate-like structure with higher anisotropy in medial knee OA.

The primary factor in this result is estimated to be the heterogeneous load distribution associated with malalignment causing high load on the medial tibial plateau [21, 36]. Consequently, subchondral bone structural changes occur as

-p

adaptations [20, 21].

re

Previous studies have used DXA and QCT to show that BMD was higher in the medial plateau than in the lateral plateau [15-17]. Our previous study using MRI reported that as cartilage area decreased in the medial plateau, BV/TV and Tb.Th in

lP

the medial tibial plateau increased [25]. Some micro-CT studies have analyzed specimens retrieved from patients with late-stage knee OA who underwent total knee arthroplasty. Roberts et al. reported that BV/TV in the medial condyle was

na

consistently higher than lateral condyle in the varus group [19]. Chen et al. reported that increasing BV/TV and decreasing SMI were found beneath damaged cartilage [20]. Cox et al. reported that BV/TV, Tb.N, and Tb.Th were increased and

ur

Tb.Sp was decreased in severe OA [21]. Finnilä et al. also reported that BV/TV, Tb.N, and Tb.Th were increased, and

Jo

Tb.Sp and SMI were decreased in late-stage OA [22]. Our study supports these previous studies in terms of denser, thicker trabecular bone and more plate-like structures in the medial plateau. The heterogeneity of bone microstructural parameters among regions is also shown in Table 2. The vBMD and BV/TV were higher and SMI was lower, particularly at medA, medC, and medM. These results indicate that subchondral trabecular bone structure within the medial plateau (particularly in the medA, medC and medM) has higher subchondral bone volume and a more plate-like structure than other regions. In particular, subchondral trabecular bone at medM shows the highest density, thickness, and connectivity (the highest BV/TV, vBMD, Tb.Th, Tb.N, and Conn.D, and the lowest Tb.Sp and SMI) compared to other regions. These results are consistent with the contention that a higher load concentration exists on anterior, central, and medial regions of the medial plateau in medial OA, as suggested by other studies [19]. On the other hand, subchondral bone microstructure at the posterior region in the medial plateau had lower density and a more rod-like structure than the other three regions in the medial plateau. This result may be attributable to the kinematics

Journal Pre-proof of the femur and tibia. Although the femur rotates externally in flexion and internally in extension relative to the tibia, that range of movement is smaller at the tibial medial plateau than at the lateral plateau. That is, this movement is called a medial pivot motion [37, 38]. In these kinematics, high load may not be concentrated on the posterior region of the medial tibial plateau even from the early stage. Moreover, several studies have reported that the normal kinematics could not be observed in knee OA [39, 40], and a systematic review also showed that the medial contact pattern in the tibial plateau for OA was usually more anterior to that in the healthy knee [41]. These factors may support the heterogeneity of subchondral bone parameters in the medial tibial plateau shown in this study. In previous human imaging studies, similar results concerning differences in bone volume in different regions have been reported with QCT [17] and micro-CT [19, 23], even

ro of

with differences in placing ROIs.

As shown in Table 3, significant correlations were found between microstructural parameters of subchondral bone in the medial plateau and each of KL grade, mJSW, FTA, and MAD. These results may indicate that stage of disease and lower

-p

limb alignment cause higher load concentrations on the medial tibial plateau and result in bone microstructural changes

re

such as denser, thicker trabeculae, and a more plate-like structure in the medial plateau than in the lateral [42]. These results also support studies in which varus deformity was a risk factor for the onset or progression of medial knee OA [43, 44].

lP

In particular, bone microstructural parameters in the anterior region of the medial plateau showed the highest correlations with KL grade, mJSW, FTA, and MAD in all regions. As shown in Figure 4, BV/TV in the anterior region had the highest

na

correlations with FTA and a scatter plot graph showed the strongest positive linear association without significant outliners.

OA.

ur

These results indicate that bone microstructure in the anterior region may offer representative indices, particularly in medial

Jo

As shown in Figure 5, although BV/TV did not differ markedly among the eight regions in KL grade 1, BV/TV was very high in the late stage of disease not only at the medial region but also at the anterior region of the medial plateau. In this regard, representative cases are also shown in Figure 6. A previous kinematic study reported that the contact pattern of the medial plateau became more anterior in OA knees compared with healthy knees [41], which might shift the load distribution to the anterior part. Subchondral bone microstructural analysis using HR-pQCT could contribute in providing bone imaging markers for knee OA. This technique could be useful to elucidate the pathogenesis of OA and to develop or evaluate the effects of medical treatments. This technique could also be useful in selecting or evaluating the effects of surgical methods such as osteotomy. Several limitations must be considered when interpreting the present results. First, the number of subjects was small in this study. We regard this investigation as a pilot study that yielded significant results. Second, this study used a cross-sectional design. A longitudinal study is therefore warranted in the future to clarify details of the establishment of OA. We believe

Journal Pre-proof that HR-pQCT serves as a unique modality to enable such studies. Third, details such as duration of disease, symptoms (pain score, range of motion, walking ability), cartilage evaluation by MRI, and BMD as measured by DXA are missing from this study. Fourth, practical limitations exist on this technique, in that some patients with obesity, short length of the lower limbs, or restricted range of knee motion cannot be scanned by HR-pQCT because of the small size of the scanner gantry. However, no patients encountered difficulty with scanning of the knee in this study. Finally, mean BMI in this study (23.8 ± 3.5 kg/m2) differed from that in Caucasian cohorts. A previous epidemiological study among 1094 Japanese women also showed that mean BMI was 23.6 kg/m2 in the OA group and 22.0 kg/m2 in the control group [45]. In addition, a study among 2200 Japanese people showed that the odds ratio for BMI in relation to OA prevalence was relatively low, at

ro of

1.14 [46]. Mean BMI in this study thus might have been lower than in Caucasian cohorts, but this was an Asian study in which extreme obesity was not relatively notable, similar to other Asian cohorts [47, 48].

-p

Conclusions

re

We performed in vivo analysis of the tibial subchondral bone microstructure in medial knee OA using HR-pQCT. Higher bone volume, higher trabecular thickness, and a more plate-like structure were observed in the medial plateau than in the

lP

lateral plateau. In particular, higher bone volume and a more plate-like structure were found at anterior, central, and medial regions of the medial plateau, compared with the posterior region and the lateral plateau. In addition, bone microstructural

na

parameters at the anterior region in the medial plateau showed the highest correlation with stage of disease and lower limb alignment. Subchondral bone microstructure in that region may thus offer representative indices, particularly for medial

ur

OA. We believe that HR-pQCT serves as a promising modality to specifically examine patients with knee OA. The

Jo

technique presented here may be useful in studying the pathogenesis of OA or evaluating the effects of treatment.

Author Contributions

Study design: KC; Data acquisition: KC, NO, and KY; Data analysis: KS and KC; Data interpretation: KS and KC; Drafting of the manuscript: KS; Revision of manuscript content: KS, KC and MO; Approval of the final version of the manuscript: KS, KC, NO, KY, YN, KK, AY, MT, and MO.

Conflict of interest The authors declare that they have no conflicts of interest.

Acknowledgments

Journal Pre-proof

Jo

ur

na

lP

re

-p

ro of

We would like to thank Shuntaro Sato, a specialist in biostatistics, for providing advice on the statistical analyses.

Journal Pre-proof References

[1] Yoshimura N, Muraki S, Oka H, et al. Prevalence of knee osteoarthritis, lumbar spondylosis, and osteoporosis in Japanese men and women: The research on osteoarthritis/osteoporosis against disability study. J Bone Miner Metab. 2009;27(5):620-628. doi:10.1007/s00774-009-0080-8

[2] Kapoor M, Martel-Pelletier J, Lajeunesse D, Pelletier JP, Fahmi H. Role of proinflammatory cytokines in the

ro of

pathophysiology of osteoarthritis. Nat Rev Rheumatol. 2011;7(1):33-42. doi:10.1038/nrrheum.2010.196

[3] Brandt KD, Radin EL, Dieppe PA, van de Putte L. Yet more evidence that osteoarthritis is not a cartilage disease. Ann

-p

Rheum Dis. 2006;65:1261-1264. doi: 10.1136/ard.2006.058347

re

[4] Lories RJ, Luyten FP. The bone-cartilage unit in osteoarthritis. Nat Rev Rheumatol. 2011;7(1):43-49.

lP

doi:10.1038/nrrheum.2010.197

[5] Loeser RF, Goldring SR, Scanzello CR, et al. Osteoarthritis: a disease of the joint as an organ. Arthritis Rheum.

na

2012 ;64(6):1697-1707. doi:10.1002/art.34453

ur

[6] Cucchiarini M, de Girolamo L, Filardo G, et al. Basic science of osteoarthritis. J Exp Orthop. 2016;3(1):22.

Jo

doi:10.1186/s40634-016-0060-6

[7] Lajeunesse D. The role of bone in the treatment of osteoarthritis. Osteoarthr Cartil. 2004;12: S34-S38. doi:10.1016/j.joca.2003.09.013

[8] Hayami T, Pickarski M, Zhuo Y, Wesolowski GA, Rodan GA, Duong LT. Characterization of articular cartilage and subchondral bone changes in the rat anterior cruciate ligament transection and meniscectomized models of osteoarthritis. Bone. 2006;38(2):234-243. doi:10.1016/j.bone.2005.08.007

[9] Muraoka T, Hagino H, Okano T, Enokida M, Teshima R. Role of subchondral bone in osteoarthritis development: A comparative study of two strains of guinea pigs with and without spontaneously occurring osteoarthritis. Arthritis Rheum.

Journal Pre-proof 2007;56(10):3366-3374. doi:10.1002/art.22921

[10] Ding M, Danielsen CC, Hvid I. Age-related three-dimensional microarchitectural adaptations of subchondral bone tissues in guinea pig primary osteoarthrosis. CalcifTissue Int. 2006;78:113-122. doi:10.1007/s00223-005-0028-5

[11] Intema F, Hazewinkel HAW, Gouwens D, et al. In early OA, thinning of the subchondral plate is directly related to cartilage damage: Results from a canine ACLT-meniscectomy model. Osteoarthr Cartil. 2010;18(5):691-698.

ro of

doi:10.1016/j.joca.2010.01.004

[12] Hayami T, Pickarski M, Wesolowski GA, et al. The Role of Subchondral Bone Remodeling in Osteoarthritis: Reduction of Cartilage Degeneration and Prevention of Osteophyte Formation by Alendronate in the Rat Anterior Cruciate

re

-p

Ligament Transection Model. Arthritis Rheum. 2004;50(4):1193-1206. doi:10.1002/art.20124

[13] Aho OM, Finnilä M, Thevenot J, Saarakkala S, Lehenkari P. Subchondral bone histology and grading in osteoarthritis.

lP

PLoS One. 2017;12(3):1-16. doi:10.1371/journal.pone.0173726

na

[14] Li G, Yin J, Gao J, et al. Subchondral bone in osteoarthritis: Insight into risk factors and microstructural changes.

ur

Arthritis Res Ther. 2013;15(6). doi:10.1186/ar4405

Jo

[15] Hulet C, Sabatier JP, Souquet D, Locker B, Marcelli C, Vielpeau C. Distribution of bone mineral density at the proximal tibia in knee osteoarthritis. Calcif Tissue Int. 2002;71(4):315-322. doi:10.1007/s00223-001-2112-9

[16] Lo GH, Merchant MG, Driban JB, et al. Knee Alignment Is Quantitatively Related to Periarticular Bone Morphometry and Density, Especially in Patients with Osteoarthritis. Arthritis Rheumatol. 2018;70(2):212-221. doi:10.1002/art.40325

[17] Bennell KL, Creaby MW, Wrigley T V., Hunter DJ. Tibial subchondral trabecular volumetric bone density in medial knee joint osteoarthritis using peripheral quantitative computed tomography technology. Arthritis Rheum. 2008;58(9):2776-2785. doi:10.1002/art.23795

Journal Pre-proof [18] Johnston JD, Masri BA, Wilson DR. Computed tomography topographic mapping of subchondral density (CT-TOMASD) in osteoarthritic and normal knees: methodological development and preliminary findings. Osteoarthr Cartil. 2009;17(10):1319-1326. doi:10.1016/j.joca.2009.04.013

[19] Roberts BC, Thewlis D, Solomon LB, Mercer G, Reynolds KJ, Perilli E. Systematic mapping of the subchondral bone 3D microarchitecture in the human tibial plateau: Variations with joint alignment. J Orthop Res. 2017;35(9):1927-1941. doi:10.1002/jor.23474

ro of

[20] Chen Y, Hu Y, Yu YE, et al. Subchondral Trabecular Rod Loss and Plate Thickening in the Development of Osteoarthritis. J Bone Miner Res. 2018;33(2):316-327. doi:10.1002/jbmr.3313

-p

[21] Cox LGE, Van Donkelaar CC, van Rietbergen B, Emans PJ, Ito K. Alterations to the subchondral bone architecture

re

during osteoarthritis: Bone adaptation vs endochondral bone formation. Osteoarthr Cartil. 2013;21(2):331-338.

lP

doi:10.1016/j.joca.2012.10.013

[22] Finnilä MAJ, Thevenot J, Aho OM, et al. Association between subchondral bone structure and osteoarthritis

na

histopathological grade. J Orthop Res. 2017;35(4):785-792. doi:10.1002/jor.23312

ur

[23] Roberts BC, Solomon LB, Mercer G, et al. Relationships between in vivo dynamic knee joint loading, static alignment

Jo

and tibial subchondral bone microarchitecture in end-stage knee osteoarthritis. Osteoarthr Cartil. 2018;26(4):547-556. doi: 10.1016/j.joca.2018.01.014

[24] Bolbos RI, Zuo J, Banerjee S, et al. Relationship between trabecular bone structure and articular cartilage morphology and relaxation times in early OA of the knee joint using parallel MRI at 3T. Osteoarthr Cartil. 2008;16(10):1150-9. doi: 10.1016/j.joca.2008.02.018

[25] Chiba K, Uetani M, Kido Y, et al. Osteoporotic changes of subchondral trabecular bone in osteoarthritis of the knee: A 3-T MRI study. Osteoporos Int. 2012;23(2):589-597. doi:10.1007/s00198-011-1585-2

[26] Burghardt AJ, Link TM, Majumdar S. High-resolution computed tomography for clinical imaging of bone

Journal Pre-proof microarchitecture. Clin Orthop Relat Res. 2011;469(8):2179-2193. doi:10.1007/s11999-010-1766-x

[27] Nishiyama KK, Shane E. Clinical imaging of bone microarchitecture with HR-pQCT. Curr Osteoporos Rep. 2013;11(2): 147–155. doi:10.1007/s11914-013-0142-7

[28] Kroker A, Zhu Y, Manske SL, Barber R, Mohtadi N, Boyd SK. Quantitative in vivo assessment of bone microarchitecture in the human knee using HR-pQCT. Bone. 2017;97:43-48. doi:10.1016/j.bone.2016.12.015

ro of

[29] Kroker A, Bhatla JL, Emery CA, Manske SL, Boyd SK. Subchondral bone microarchitecture in ACL reconstructed knees of young women: A comparison with contralateral and uninjured control knees. Bone. 2018;111:1-8.

-p

doi:10.1016/j.bone.2018.03.006

re

[30] Bhatla JL, Kroker A, Manske SL, Emery CA, Boyd SK. Differences in subchondral bone plate and cartilage thickness between women with anterior cruciate ligament reconstructions and uninjured controls. Osteoarthr Cartil.

lP

2018;26(7):929-939. doi:10.1016/j.joca.2018.04.006

na

[31] Krause M, Hubert J, Deymann S, et al. Bone microarchitecture of the tibial plateau in skeletal health and osteoporosis.

ur

Knee. 2018;25(4):559-567. doi:10.1016/j.knee.2018.04.012

Jo

[32] Kellgren, J.H., & Lawrence, J.S. Radiological assessment of osteo-arthrosis. Annals of the Rheumatic Diseases. 1957;16(4):494–502. doi.org/10.1136/ard.16.4.494

[33] Manske SL, Zhu Y, Sandino C, Boyd SK. Human trabecular bone microarchitecture can be assessed independently of density with second generation HR-pQCT. Bone. 2015;79:213-221. doi:10.1016/j.bone.2015.06.006

[34] Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Müller R. Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res. 2010;25(7):1468-1486. doi:10.1002/jbmr.141

[35] Lobenhoffer P, Heerwaarden RJ van, Staubli AE et al. Osteotomies around the knee. Indications-planning – surgical

Journal Pre-proof techniques using plate fixators. 2008; AO Foundation.

[36] Kumar D, Manal KT, Rudolph KS. Knee joint loading during gait in healthy controls and individuals with knee osteoarthritis. Osteoarthr Cartil. 2013;21(2):298-305. doi:10.1016/j.joca.2012.11.008

[37] Moro-Oka TA, Hamai S, Miura H, et al. Dynamic activity dependence of in vivo normal knee kinematics. J Orthop Res. 2008;26(4):428-434. doi:10.1002/jor.20488

ro of

[38] Tanifuji O, Sato T, Kobayashi K, et al. Three-dimensional in vivo motion analysis of normal knees using single-plane fluoroscopy. J Orthop Sci. 2011;16(6):710-718. doi:10.1007/s00776-011-0149-9

-p

[39] Hamai S, Moro-oka TA, Miura H, et al. Knee kinematics in medial osteoarthritis during in vivo weight-bearing

re

activities. J Orthop Res. 2009;27(12):1555-1561. doi:10.1002/jor.20928

lP

[40] Saari T, Carlsson L, Karlsson J, Kärrholm J. Knee kinematics in medial arthrosis. Dynamic radiostereometry during

na

active extension and weight-bearing. J Biomech. 2005;38(2):285-292. doi:10.1016/j.jbiomech.2004.02.009

[41] Scarvell JM, Galvin CR, Perriman DM, Lynch JT, van Deursen RWM. Kinematics of knees with osteoarthritis show

ur

reduced lateral femoral roll-back and maintain an adducted position. A systematic review of research using medical

Jo

imaging. J Biomech. 2018;75:108-122. doi:10.1016/j.jbiomech.2018.05.007

[42] Heijink A, Gomoll AH, Madry H, et al. Biomechanical considerations in the pathogenesis of osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc. 2012;20(3):423-35. doi: 10.1007/s00167-011-1818-0

[43] Sharma L, Song J, Felson DT, Cahue S, Shamiyeh E, Dunlop DD. The role of knee alignment in disease progression and functional decline in knee osteoarthritis. J Am Med Assoc. 2001;286(2):188-195. doi:10.1001/jama.286.2.188

[44] Brouwer GM, Van Tol AW, Bergink AP, et al. Association between valgus and varus alignment and the development and progression of radiographic osteoarthritis of the knee. Arthritis Rheum. 2007;56(4):1204-1211. doi:10.1002/art.22515

Journal Pre-proof [45] Yoshimura N, Muraki S, Oka H, Kawaguchi H, Nakamura K, Akune T. Association of knee osteoarthritis with the accumulation of metabolic risk factors such as overweight, hypertension, dyslipidemia, and impaired glucose tolerance in Japanese men and women: The ROAD study. J Rheumatol. 2011;38(5):921-930. doi:10.3899/jrheum.100569

[46] Muraki S, Oka H, Akune T, et al. Prevalence of radiographic knee osteoarthritis and its association with knee pain in the elderly of Japanese population-based cohorts: The ROAD study. Osteoarthr Cartil. 2009;17(9):1137-1143. doi: 10.1016/j.joca.2009.04.005.

ro of

[47] Zhou M, Chen J, Wang D, Zhu C, Wang Y, Chen W. Combined effects of reproductive and hormone factors and obesity on the prevalence of knee osteoarthritis and knee pain among middle-aged or older Chinese women: A

-p

cross-sectional study. BMC Public Health. 2018;18(1):1-9. doi:10.1186/s12889-018-6114-1

re

[48] Shin D. Association between metabolic syndrome, radiographic knee osteoarthritis, and intensity of knee pain: Results

Jo

ur

na

lP

of a national survey. J Clin Endocrinol Metab. 2014;99(9):3177-3183. doi:10.1210/jc.2014-1043

Journal Pre-proof Figure legends

Figure 1. (A) Custom leg holder is attached to the manufacturer’s original leg holder. (B) The knee is inserted into the gantry in a sitting position. The opposite leg is positioned resting on a stool.

Figure 2. The ROI selection was performed separately for the medial (A, B, and C) and lateral (D, E, and F) tibial plateaus (shown in

ro of

the case of a left knee).

(A and D) In the coronal (upper) and sagittal (lower) views, a subchondral trabecular bone volume of 5 mm length was extracted. This volume was 2 mm below and parallel to the joint surface.

re

(Med) and the lateral plateau (Lat), respectively.

-p

(B and E) The medial 1/3 and the lateral 1/3 subchondral trabecular bone compartment were defined as the medial plateau

(C and F) In addition, Med was divided into the anterior (medA), central (medC), medial (medM), and posterior (medP)

na

Figure 3.

lP

regions. Lat was divided into the anterior (latA), central (latC), lateral (latL), and posterior (latP) regions.

The algorithm for 3D segmentation is shown in the case of a medial plateau of a left knee

ur

(A) A subchondral bone was shown on the axial view.

Jo

(B) A binarized image was created with a fixed threshold value (320 mg/cm3). (C) The entire bone region was obtained by filling all bone marrow spaces in the binarized image. (D) The subchondral trabecular bone volume was extracted by deleting a 1-mm external margin of the entire bone region to exclude cortical bone. (E) The subchondral trabecular bone volume was divided into 3 compartments of equal length. (F) Further, the subchondral trabecular compartment in each plateau was divided into 4 regions using a rectangle and two oblique lines. The rectangle was drawn with the lines dividing a an antero-posterior length into 1/3 and medio-lateral length into 1/2. Two lines were drawn outward at oblique angles (45°) from the two outer vertices of the rectangle.

Figure 4. Correlations between FTA and BV/TV at the four regions in the medial plateau. Significant positive correlations were

Journal Pre-proof observed at medA, medM, and medP. In particular, BV/TV at medA showed the highest correlation with FTA.

Figure 5. The average BV/TV at the eight regions in the medial and lateral plateau are shown separately in each KL grade. (A) For KL grade 1, no marked differences were evident among all eight regions. (B) For KL grade 2, BV/TV was the highest in medM compared to the other regions. (C) For KL grade 3, BV/TV at medA, medC, and medM were higher than the others. (D) For KL grade 4, BV/TV at medM and medA were markedly higher than the others.

ro of

On the other hand, BV/TV at the medP was lower than the other three medial regions.

(Qualitative comparison, no statistical test was performed when stratifying subjects by KL-grade, due to limited sample

-p

size.)

re

Figure 6.

HR-pQCT images and measurement data of representative two women with KL grade 1 and 4 knees. The patient with KL

na

Table 1 Subject characteristics

lP

grade 4 knee had higher vBMD, BV/TV and Tb.Th particularly at the medA and the medM than KL grade 1 knee.

KL 1

KL 2

KL 3

KL 4

20

4

5

7

4

63.6 ± 7.9 (51 – 81)

66.0

67.4

59.1

64.3

1.53 ± 0.05 (1.48 – 1.62)

1.50

1.55

1.53

1.56

56.4 ± 10.3 (43.6 – 81.5)

54.0

51.8

53.8

69.1

BMI (kg/m²)

23.8 ± 3.5 (18.7 – 31.1)

23.8

21.7

22.8

28.3

mJSW (mm)

1.7 ± 1.3 (0 – 3.5)

3.2

2.5

1.5

0.3

179.6 ± 3.8 (175 – 190)

176.5

177.2

179.4

184.8

16.0 ± 15.8 (0 – 59.2)

5.7

8.2

12.9

41.6

Age (years) Height (m) Weight (kg)

FTA (°) MAD (mm)

Jo

Number of knees

ur

Total

BMI, body mass index; KL grade, Kellgren-Lawrence grade; mJSW, minimum joint space width; FTA, femorotibial angle; MAD, mechanical axis deviation. Left side data are shown as mean ± standard deviation (minimum – maximum). Right side data are shown as mean for patients at each KL grade.

Journal Pre-proof

Table 2 Subchondral bone microstructure in each plateau and region (KL grade ≥2) vBMD

119 ± 51

BV/TV †

18.0 ± 7.8

Tb.Th †

Tb.N

293 ± 73

†

Tb.Sp 577

SMI 1.81 ± 0.32

0.55 ± 0.11

†

± 108

Med

582 124 ± 77 *

18.9 ± 11.0 *

274 ± 80

560

19.3 ± 6.8 *

273 ± 47

medC

± 87

(latA, latL, medP)

re

(latA, latL)

23.9 ± 11.5

0.67 ±

306 ± 88 *

* medM (latA, latC, latL, latP,

medP)

medP)

70 ± 24 Lat

233 ± 32

10.4 ± 3.8

231 ± 23

± 0.27 * (latC, latP,

(latA, medP)

medC, medP)

1.76 ±

2.04 ±

0.37 *

0.58 *

(latA, medP)

(latA)

1.52 ±

0.17 *

1.31 ±

2.30 ±

0.19

1.79

0.68 *

0.96 *

(latA, latC, latL,

(latA, latL,

latP, medP)

medP)

± 0.35 * (latC, latP,

(latA)

(latA, medP)

medP)

ur

medP

9.8 ± 4.3

Jo

66 ± 30

1.72

495 ± 133 *

(latA, latC, latL, latP,

na

(latA, latC, latL, latP,

lP

155 ± 77 *

†

0.62 ± 0.10

-p

124 ± 45 *

0.20

0.63 *

ro of

(latA, latL, medP)

1.45 ±

1.75 ± 0.78 ± 137

(latA, latL)

DA

1.86 ± 0.74

1.74 ±

0.57 ± 0.14

medA

Conn.D

medC, medP)

1.43 ± 0.48 ± 0.16

649 ± 118

2.38 ± 0.31

1.15 ± 0.95 0.18

1.27 ± 0.51 ± 0.11

624 ± 103

2.20 ± 0.26

1.37 ± 0.69 0.16

1.51 ± latA

59 ± 25

8.7 ± 4.0

224 ± 26

0.46 ± 0.18

662 ± 130

2.41 ± 0.28

0.98 ± 0.67 0.15

1.25 ± latC

84 ± 24

12.6 ± 3.8

236 ± 27

0.58 ± 0.16

562 ± 100

2.20 ± 0.24

1.68 ± 0.68 0.13

1.69 60 ± 25 latL

8.9 ± 3.9

216 ± 30

0.52 ± 0.13

639 ± 105

2.28 ± 0.32

1.11 ± 0.73 ± 0.22 * (latC, latP,

Journal Pre-proof medC)

1.37 ± 82 ± 35

12.4 ± 5.7

234 ± 27

0.56 ± 0.17

574 ± 108

2.14 ± 0.39

1.61 ± 0.84 0.20

latP

Med, medial plateau; Lat, lateral plateau; medA, medC, medM, and medP, anterior, central, medial, and posterior regions of the medial plateau, respectively; latA, latC, latL, and latP, anterior, central, lateral, and posterior regions of the lateral plateau,

ro of

respectively. vBMD, volumetric bone mineral density; BV/TV, bone volume fraction; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation;

-p

SMI, structure model index; Conn.D, connectivity density; DA, degree of anisotropy.

re

All data are shown as mean ± standard deviation.

lP

Bold characters indicate parameters showing significant differences, with parentheses below indicating regions with which comparisons show significant differences.

na

Paired t-test was used for comparisons of mean values between Med and Lat. (†p < 0.01)

Jo

corrected p< 0.01)

ur

Paired t-test with Bonferroni adjustment was used for comparisons of mean values among all regions. ( *

Table 3 Correlation coefficients between subchondral bone microstructure and KL grade, mJSW, FTA, and MAD

KL grade

vBMD

BV/TV

Tb.Th

Tb.N

Tb.Sp

SMI

Conn.D

DA

Med

0.76**

0.79**

0.76**

0.54

-0.70**

-0.67*

0.69**

-0.38

medA

0.81**

0.83**

0.66*

0.68*

-0.83**

-0.79**

0.82**

-0.41

medC

0.63*

0.64*

0.53

0.51

-0.70**

-0.61*

0.69*

-0.40

medM

0.68*

0.71**

0.79**

0.40

-0.71**

-0.56*

0.67*

-0.21

medP

0.57*

0.58*

0.68*

0.38

-0.48

-0.30

0.51

-0.20

Lat

0.25

0.28

0.37

0.42

-0.43

-0.03

0.22

-0.33

latA

0.27

0.29

0.36

0.15

-0.33

-0.10

0.23

-0.24

latC

0.33

0.32

0.24

0.46

-0.49

-0.10

0.35

-0.32

Journal Pre-proof 0.29

-0.30

0.01

0.12

-0.02

latP

0.14

0.15

0.31

0.06

-0.31

0.02

0.11

-0.11

Med

-0.56*

-0.59*

-0.58*

-0.32

0.60*

0.41

-0.46

0.46

medA

-0.63*

-0.62*

-0.56*

-0.41

0.69**

0.68*

-0.60*

0.42

medC

-0.54

-0.49

-0.51

-0.25

0.62*

0.49

-0.35

0.48

medM

-0.39

-0.41

-0.61*

-0.18

0.54

0.25

-0.43

0.25

medP

-0.55

-0.52

-0.57*

-0.36

0.53

0.35

-0.50

0.13

Lat

-0.48

-0.49

-0.53

-0.60*

0.45

0.28

-0.37

0.15

latA

-0.47

-0.52

-0.46

-0.41

0.50

0.39

-0.37

0.28

latC

-0.57*

-0.54

-0.39

-0.54

0.46

0.47

-0.46

0.32

latL

-0.33

-0.34

-0.45

-0.38

0.37

0.21

-0.30

0.00

latP

-0.28

-0.27

-0.46

-0.23

0.34

0.16

-0.20

-0.15

Med

0.69**

0.71**

0.63*

0.52

-0.70**

-0.58*

0.70**

-0.13

medA

0.74**

0.76**

0.63*

medC

0.52

0.51

0.38

medM

0.60*

0.60*

medP

0.62*

0.61*

Lat

0.22

0.23

latA

0.35

0.32

latC

0.24

latL

0.14

latP

0.13

Med

-p

ro of

0.23

-0.73**

-0.71**

0.77**

-0.38

0.43

-0.67*

-0.46

0.64*

-0.10

0.63*

0.43

-0.71**

-0.46

0.66*

0.05

0.67*

0.44

-0.57*

-0.26

0.54

0.03

0.22

0.51

-0.43

-0.01

0.30

-0.25

0.33

0.21

-0.45

-0.04

0.36

-0.29

0.20

0.01

0.27

-0.49

0.03

0.33

-0.16

0.19

0.07

0.32

-0.40

-0.01

0.19

-0.22

0.10

0.08

0.21

-0.40

0.05

0.24

0.03

0.56*

0.59*

0.52

0.29

-0.56*

-0.36

0.46

-0.16

0.63*

0.69**

0.55

0.52

-0.60*

-0.71**

0.61*

-0.33

0.41

0.42

0.34

0.18

-0.55

-0.27

0.35

-0.14

medM

0.51

0.53

0.55

0.22

-0.71**

-0.32

0.46

-0.02

medP

0.31

0.29

0.55

0.13

-0.27

0.09

0.22

-0.23

Lat

-0.10

-0.09

-0.10

0.28

-0.26

0.20

0.07

-0.13

latA

0.22

0.22

0.25

0.15

-0.31

-0.12

0.27

-0.04

latC

-0.06

-0.06

-0.34

0.30

-0.36

0.09

0.20

-0.24

latL

-0.25

-0.26

-0.23

0.03

-0.16

0.34

-0.12

-0.46

latP

-0.21

-0.20

-0.20

-0.04

-0.25

0.26

0.03

0.08

medA medC

na

re

0.63*

lP

MAD

0.19

ur

FTA

0.13

Jo

mJSW

latL

Med, medial plateau; Lat, lateral plateau;

Journal Pre-proof medA, medC, medM, and medP, anterior, central, medial, and posterior regions of the medial plateau, respectively; latA, latC, latL, and latP, anterior, central, lateral, and posterior regions of the lateral plateau, respectively. vBMD, volumetric bone mineral density; BV/TV, bone volume fraction; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation; SMI, structure model index; Conn.D, connectivity density; DA, degree of anisotropy. KL grades, Kellgren-Lawrence grade; mJSW, minimum joint space width; FTA, femorotibial angle; MAD, mechanical axis deviation. Statistical significance was analyzed using correlation coefficient by Spearman’s rank test (*p < 0.01, **p < 0.001). Bold characters indicate significant coefficients. Significant positive and negative correlations are highlighted in yellow and green,

ro of

respectively.

HighlightS

Subchondral trabecular bone microstructure in medial knee osteoarthritis was analyzed using second-generation

-p



HR-pQCT.

Higher bone volume, trabecular thickness, and more plate-like structure were observed in medial plateau than in

re



lateral.

lP

Subchondral bone microstructure at the anterior region in the medial plateau correlated strongly with

ur

na

Kellgren-Lawrence grades and lower limb alignment.

Jo



Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6