Micro-computed tomographical and histopathological analyses of medial tibial plateaus from patients with subchondral insufficiency fracture or osteoarthritis of the knee

Abstract

Background

Subchondral insufficiency fracture of the knee (SIFK) is a rare and rapidly progressing knee joint disease. The typical diagnostic finding of SIFK by magnetic resonance imaging (MRI) is a subchondral hypointense line around the bone marrow lesions (BMLs), which are also commonly observed in knee osteoarthritis (OA). Subchondral bone fracture and its repair reactions are implicated for SIFK. However, pathological changes of SIFK and BMLs and the mechanism by which SIFK is induced remain elusive. We aimed to examine characteristics of SIFK and OA by micro-computed tomography (micro-CT) and histology together with biomarker analysis.

Methods

Nineteen patients with femoral or tibial SIFK (10 F-SIFK and 9 T-SIFK patients) and 24 knee OA patients, who were diagnosed by radiography and MRI and underwent unicompartmental knee arthroplasty, were enrolled for this study, and their medial tibial plateaus were examined by micro-CT and pathology. Serum and urine biomarkers were also analyzed.

Results

All the T-SIFK patients showed BMLs with a subchondral hypointense line by MRI. Micro-CT analysis revealed that the T-SIFK lesion comprises multiple subchondral bone fragments covered with articular cartilage. Histologically, the lesion was composed of articular cartilage-covered subchondral bone fragments, debris of bone and bone marrow, fibrogranulation tissue, cartilage and woven bone. The medial tibial plateaus from OA patients frequently exhibited eburnation, which was commonly accompanied by microfracture. All the BMLs observed in T-SIFK and OA were associated with fat necrosis, which was characterized by disrupted fat cells and foamy macrophage infiltration. The posterior tibial slope angle (12.84 ± 2.34° vs 9.58 ± 2.84°) and the rate of medial meniscal posterior root tears (68.4% vs 25.0%) were significantly higher in T-SIFK than OA. Numbers of grade 2 subchondral bone resorption pits in uninvolved areas of the medial tibial plateaus (2.32 ± 1.43 vs 0.72 ± 0.66) and the femoral condyles (5.53 ± 3.73 vs 1.50 ± 2.10) were significantly higher in T-SIFK than OA.

Conclusions

Our data demonstrate that T-SIFK is generated by subchondral bone fracture and its repair reaction and suggest that fat necrosis of the bone marrow is involved in BML formation in T-SIFK and OA.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13075-026-03733-w.

Introduction

Subchondral insufficiency fracture of the knee (SIFK) is a rare knee joint disease with an acute onset of severe knee pain. SIFK occurs mainly in females above 60 years of age without any specific history of metabolic disorder or therapeutic intervention [1]. Historically, Ahlback et al. [2] first reported spontaneous osteonecrosis of the knee (SONK) and considered that this disease is caused by osteonecrosis due to ischemia. However, the subsequent histological study on femoral condyles from SONK patients disclosed that the primary event leading to osteonecrosis in SONK is a subchondral insufficiency fracture and there are two types of lesions: Subchondral fracture without osteonecrosis, early-stage lesion without collapse, and subchondral fracture complicated with osteonecrosis, more advanced-stage lesion [3]. Further histological studies have confirmed the concept [4–7]. In recent years, SONK is categorized as an advanced SIFK lesion, and the term of SONK is abandoned [8]. Accordingly, SIFK may be classified into SIFK without osteonecrosis and SIFK with osteonecrosis.

Diagnosis of SIFK is commonly made, besides clinical history, by the characteristic finding of magnetic resonance imaging (MRI), i.e. a linear hypointense line with the surrounding bone marrow lesions (BMLs). Osteoporosis [9–11], meniscal problems including medial meniscal posterior root tears [12, 13] and osteoarthritis (OA) [14] are suggested as risk factors, but how these factors are involved in the induction and progression of SIFK remains unknown. In SIFK, the medial femoral condyle is most commonly affected in approximately 65% of cases, followed by the medial tibial plateau and lateral femoral condyle [1]. Thus, previous histological studies have focused on the medial femoral condyles [3–7], and little or no information is available for the morphological characteristics of the medial tibial plateau in SIFK patients. Although pathological events are known to occur in the subchondral bone plate and trabeculae, to the best of our knowledge, there are no reports on the bone changes in SIFK analyzed by microcomputed tomography (micro-CT), which is an excellent tool for the analysis of pathological bone changes. In addition, information about biomarkers for SIFK is still limited.

Osteoarthritis (OA) is the most prevalent joint disease in the elderly and characterized by destruction of articular cartilage caused primarily by activated chondrocytes with abnormal metabolism. However, several factors including osteophyte formation, synovitis and subchondral bone changes are implicated for induction and/or progression of OA [15–18]. Among them, the impairment of subchondral bone is commonly associated with cartilage defects and subchondral bone sclerosis is a hallmark of OA [19]. Bone remodeling is known to be increased at early stage and reduced at late stage of OA [17, 20]. Accumulating evidence shows that abnormal subchondral bone remodeling together with angiogenesis and sensory nerve innervation contributes to cartilage destruction and knee pain [20]. The subchondral bone changes in OA joints are commonly detected as BMLs on MRI [20]. However, BMLs are also seen in many pathological conditions including SIFK, trauma, inflammation and tumor of the bone. In knee OA, BMLs are associated with loss of cartilage and bone sclerosis in the medial tibial plateaus [21] and longitudinal studies on large cohorts have indicated that although BMLs are associated with loss of articular cartilage in the knee OA joints [22–24], some BMLs show a decrease in size or even complete resolution during a follow-up period, indicating that they are reversible [23, 25]. Only sparse data are available on the histology of MRI-detected BMLs in various knee joint diseases [25]. Although histological scoring system for BMLs in OA joints has been proposed by focusing on the seven domains of pathological subchondral changes including cysts, fibrosis, hypervascularity, cartilage islands, thickened trabeculae, breaching of the tidemark, and inflammatory cell infiltration [26], no specific and/or commonly occurring pathological changes associated with BMLs have been reported. Therefore, pathological characteristics of BMLs observed in OA and SIFK joints remain elusive.

In the present study, we obtained the medial tibial plateaus at unicompartmental knee arthroplasty (UKA) from the knee joints of patients with SIFK or OA and examined them by micro-CT and histology. Our data have shown that in a stark contrast to OA, the tibial plateaus from SIFK comprise many fractured subchondral bone fragments covered with articular cartilage, while those from OA joints frequently exhibit eburnation. In addition, we have demonstrated that all the BMLs observed in SIFK and OA cases are associated with fat necrosis.

Methods

Patients

This study involved 43 patients who underwent medial UKA. By manual examination, MRI and arthroscopy, all the patients were free of ligamentous instability such as anterior cruciate ligament deficiency. The 35 patients (81.4%) were female and aged between 61 and 82 years, and the 8 patients (18.6%) were male and aged between 62 and 81 years (Table 1). The inclusion criteria were cases diagnosed as SIFK involving the femoral condyle and/or the tibial plateau or knee OA. The exclusion criteria were secondary knee OA due to trauma, after osteotomy, or after arthroscopic surgery such as meniscectomy, and cases without MRI imaging. The clinical data of each patient were recorded, including age, gender, and body mass index (BMI). All surgeries were performed using the Oxford Partial Knee with mobile bearing (Zimmer Biomet, IN). Arthroscopic surgery was first performed to evaluate the anterior cruciate ligament (ACL) and meniscus tear. Medial meniscal posterior root tear was evaluated according to the LaPrade classification [27]. Then, UKA was performed by a medial parapatellar approach according to Zimmer Biomet Oxford Partial Knee Surgical technique. The medial tibial plateau and posterior femoral condyle were collected during surgery and used for further analyses according to the hospital ethics guidelines. Written consent was obtained from all patients. The study protocol complied with the principles outlined in the Declaration of Helsinki of 1975 was approved by the Ethics Committee for Medical Research of Juntendo University School of Medicine (Ethics Committee Research Subject No E22-0352).

Table 1.

Characteristics of the patients with SIFK or OA

Case No	Sex/ Age	Clinical diagnosis	Pain duration (months)	BMI	K/L	BM	MRI findings of MTP				Histology of MTP	mCT of MTP
							BML	Fracture line	SBA	SBC
1	F/82	T, F-SIFK	4	24.7	3	(-)	(+)	(+)	(+)	(+)	SIFK	(-)
2	F/69	T-SIFK	40	30.0	3	(-)	(+)	(+)	(-)	(+)	SIFK	(-)
3	M/67	T-SIFK	270	22.3	3	(-)	(+)	(+)	(-)	(+)	SIFK	(-)
4	M/71	T-SIFK	26	31.8	3	(-)	(+)	(+)	(-)	(-)	SIFK	(-)
5	M/75	T-SIFK	12	28.9	3	(+)	(+)	(+)	(+)	(+)	SIFK	(+)
6	F/73	T-SIFK	8	27.5	3	(+)	(+)	(+)	(+)	(+)	SIFK	(+)
7	F/74	T-SIFK	10	26.7	3	(+)	(+)	(+)	(+)	(+)	SIFK	(+)
8	F/58	T-SIFK	10	28.1	2	(+)	(+)	(+)	(+)	(+)	SIFK	(+)
9	F/70	T-SIFK	12	29.4	3	(+)	(+)	(+)	(+)	(+)	SIFK	(+)
10	F/61	F-SIFK	6	24.4	2	(-)	(-)	(-)	(+)	(-)	OA	(-)
11	F/69	F-SIFK	39	27.9	2	(+)	(-)	(-)	(-)	(-)	OA	(-)
12	F/71	F-SIFK	13	24.4	2	(+)	(-)	(-)	(+)	(-)	OA	(+)
13	M/66	F-SIFK	10	23.1	3	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
14	M/81	F-SIFK	9	22.1	2	(+)	(+)	(-)	(+)	(-)	OA + ebur	(+)
15	F/65	F-SIFK	3	30.8	3	(+)	(+)	(-)	(+)	(-)	OA	(+)
16	F/76	F-SIFK	12	23.7	3	(+)	(+)	(-)	(-)	(-)	OA	(+)
17	F/76	F-SIFK	6	20.3	2	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
18	M/62	F-SIFK	64	25.8	4	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
19	F/65	F-SIFK	4	35.0	3	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
20	F/82	OA	54	23.0	4	(-)	(+)	(-)	(-)	(-)	OA + ebur + mf	(-)
21	F/78	OA	49	20.5	3	(-)	(+)	(-)	(-)	(-)	OA + ebur	(-)
22	F/77	OA	78	27.0	3	(-)	(+)	(-)	(-)	(+)	OA	(-)
23	F/80	OA	205	26.1	4	(-)	(+)	(-)	(-)	(-)	OA + ebur	(-)
24	F/75	OA	140	32.1	3	(-)	(+)	(-)	(-)	(+)	OA + cyst + mf	(-)
25	F/70	OA	35	31.2	3	(-)	(+)	(-)	(-)	(-)	OA + ebur + mf	(-)
26	M/72	OA	34	23.5	3	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
27	F/74	OA	50	21.5	3	(+)	(+)	(-)	(-)	(+)	OA + ebur + mf	(+)
28	F/80	OA	164	20.3	4	(+)	(+)	(-)	(-)	(+)	OA + ebur + mf	(+)
29	M/77	OA	73	26.9	3	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
30	F/64	OA	100	23.4	3	(+)	(+)	(-)	(+)	(-)	OA + ebur + mf	(+)
31	F/69	OA	17	23.6	3	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
32	F/70	OA	20	23.7	3	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
33	F/78	OA	66	24.4	3	(+)	(+)	(-)	(+)	(-)	OA + ebur	(+)
34	F/76	OA	104	26.2	3	(+)	(+)	(-)	(+)	(-)	OA + ebur	(+)
35	F/79	OA	124	26.1	4	(+)	(+)	(-)	(-)	(+)	OA + cyst	(+)
36	F/72	OA	65	19.6	3	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
37	F/66	OA	109	26.6	3	(+)	(+)	(-)	(+)	(-)	OA + ebur + mf	(+)
38	F/72	OA	66	22.7	4	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
39	F/75	OA	9	18.0	3	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
40	F/72	OA	185	19.1	3	(+)	(+)	(-)	(-)	(-)	OA + ebur + mf	(+)
41	F/73	OA	195	22.4	4	(+)	(+)	(-)	(-)	(-)	OA + ebur	(+)
42	F/67	OA	82	23.5	4	(+)	(+)	(-)	(+)	(-)	OA + ebur	(+)
43	F/82	OA	42	20.9	3	(+)	(+)	(-)	(+)	(+)	OA + ebur + mf	(+)

BMI body mass index, K/L Kellgren-Lawrence, BM Biomarkers’ test, MTP Medial tibial plateau, BML Bone marrow lesion, SBA Subchondral bone attrition, SBC Subchondral bone cyst, m micro, T-SIFK SIFK involving the medial tibial plateau, F-SIFK SIFK involving the medial femoral condyle, SIFK subchondral insufficiency fracture of the knee, OA Osteoarthritis, ebur eburnation, mf microfracture

The radiographic evaluation of knee joints

Before surgery, X-rays were taken in standing positions of the anteroposterior and lateral extended view or the 45° of flexion view (Rosenberg view), and the lower limb alignment was measured based on the full-limb standing radiographs by a stereoscopic radiography (sterEOS system, Paris, France). The OA severity was evaluated according to the Kellgren-Lawrence (K/L) classification [28], which is the current standard grading system for OA. Lower limb alignment was assessed by evaluating posterior tibial slope (PTS), femoro-tibial angle (FTA), mechanical axis (%ma), medial proximal tibial angle (MPTA) and lateral distal femoral angle (LDFA). PTS was determined by measuring the angle between the line connecting the highest anterior and posterior points of the tibial plateau and the line perpendicular to the tibial proximal anatomic axis [29].

MRI analysis

The knee joints were analyzed using both one of the fat suppressed MRI sequences, which included proton density-weighted fat-suppressed (PDFS), T2-weighted spectral adiabatic inversion-recovery (T2WI-SPAIR) and T2-weighted fat-suppressed (T2WI-FS) MRI, and proton density-weighted imaging (PDWI) MRI sequences. Subchondral bone lesions including BMLs, subchondral bone cyst (SBC) and subchondral bone attrition (SBA) in the medial tibial plateau were evaluated according to the Whole Organ Magnetic Resonance Imaging Score (WORMS) system [30]: BMLs, SBC, and SBA were defined as poorly marginated aeras of increased signal intensity in the fatty epiphyseal marrow, foci of markedly increased signal in the subarticular bone with sharply defined, rounded margins, and flattening or depression of the articular surfaces, respectively. In the present study we assessed these lesions as either present or absent without scoring. The widths of meniscal extrusion and osteophytes were measured according to previous studies [22, 31, 32]. Because osteophytes are composed of the cartilage and bone parts, widths of the bone part and cartilage part of osteophytes and total width of osteophytes were measured. Meniscal tears were also evaluated according to the MRI Osteoarthtitis Knee Score (MOAKS) [31]. MOAKS grades 2 to 4 were defined as meniscus tears, and grade 5 as medial meniscus posterior root tears.

Micro-CT analysis of the tibial samples

Medial tibial plateau and posterior femoral condyle were fixed in 10% buffered formalin for 3—4 days and analyzed by micro-CT (ScanXmate-L090HH, Comscantechno Co., Ltd., Japan) with an energy of 68 kV and a current of 89 μA over 300 s, producing a resolution of 53.09 μm³ voxel size. Three-dimensional (3D) microstructural images were reconstructed, and bone volume (BV), trabecular thickness (Tb.Th), trabecular number (Tb.N), trabecular separation (Tb.Sp), bone mineral density (BMD), node number (N.Nd) and node node number (N.NdNd) were calculated using TRI/3D-BON-FCS64 software (Ratoc System Engineering Co., Ltd., Japan). The coronal view was adjusted to match the histology slice. Areas of SIFK and eburnation lesions were individually traced on 3D reconstructed micro-CT images and superimposed for each group with WinROOF2018 software (Mitani Corp, Tokyo, Japan) as previously reported by Olah et al. [33]. The parameters of the uninvolved areas in the medial tibial plateaus obtained from SIFK and OA patients were analyzed by trimming a 6 mm × 6 mm × 3 mm area. In addition, the parameters of the areas beneath the eburnated areas with or without microfracture in the OA medial tibial plateaus were analyzed by trimming a 3 mm × 3 mm × 1.5 mm area.

Histological examination

Because of the surgical procedures for UKA, the medial femoral condyle except for the posterior part and the medial meniscus were fragmented and unsuitable for analyses by micro-CT and histology. Thus, the medial tibial plateaus and the posterior parts of the femoral condyles removed at arthroplasty were subjected to histological examination after micro-CT analysis. After decalcification in 10% EDTA solution, the coronal slices containing the lesions detected by MRI were made and embedded in paraffin, and then serial sections. (4 μm in thickness) were prepared by a microtome. They were stained with hematoxylin and eosin and Safranin O/Fast blue staining.

Histochemistry

Tartrate-resistant acid phosphatase (TRAP) staining was performed with deparaffinized sections of the medial tibial plateau using L-(+)-Tartaric acid and New Fuchsin alkaline phosphatase substrate kit (Nichirei Biosciences, Tokyo, Japan). For immunohistochemistry, deparaffinized sections were subjected to immunohistochemistry for type X collagen, CD68, CD31, and perilipin-1 after antigen retrieval by treating with Target Retrieval Solution, pH 6.0 (DAKO Cytomation, Tokyo, Japan) at 100 °C for 10 min and reactions with 3% H₂O₂ and Avidin/Biotin Blocking Kit (ab64212; abcam, Cambridge, UK) and Blocking One Histo (Nacali Tesque, Inc., Tokyo, Japan). Additional antigen retrieval was carried out for type X collagen immunostaining by incubation with Proteinase K (× 1000 dilution; #19,131; QIAGEN, Hilden, Germany) for 5 min at room temperature. They were reacted with mouse anti-Collagen X antibody (X53, #14–9771-82; Thermo Fisher Scientific, Waltham, MA), mouse anti-CD68 antibody (PG-M1, #MA5-12,407; Thermo Fisher Scientific), mouse anti-CD31 antibody (PECAM-1, JC70A, #NCL-L-CD31-607; Leica Biosystems, Nussloch, Germany), or rabbit anti-Perilipin-1 antibody (D418, #3470; Cell Signaling Technology Inc, Danvers, MA). Then they underwent the reactions with corresponding secondary antibodies, followed by the ABC method (Vector Laboratories, Burlingame, CA). Color was developed using Pierce™ DAB Substrate Kit (Thermo Fisher Scientific), followed by counterstaining with hematoxylin.

Analyses of biomarkers

The blood and urine tests were performed using the samples obtained in the early morning while fasting. The blood tests included those for general biochemistry, glucose metabolism, and bone metabolism markers, and the urine tests included those for bone metabolism markers. Blood samples were also collected on the same day, and sclerostin was measured by ELISA (Sclerostin, Human, ELISA Kit; Funakoshi, Tokyo, Japan).

Statistical analysis

All data were expressed as means ± SD and analyzed using IBM SPSS Statistics 29.0 software program. Results between the two independent groups were determined by the two-sided Student’s t-test and Fisher's exact test. P values less than 0.05 were considered significant.

Results

Characteristics of the patients

Table 1 summarizes the characteristics of the patients with SIFK involving the medial tibial plateau and/or medial femoral condyle (T-SIFK and/or F-SIFK) or knee OA. Mean ages of the patients with T-SIFK and/or F-SIFK (n = 19) or knee OA (n = 24) were 70.1 ± 6.3 (range, 61–82 years) and 74.2 ± 4.7 (range, 64–82 years) years of age, respectively, and were significantly younger in the SIFK group (P = 0.02). BMI, which was 26.7 ± 3.7 kg/m² and 23.8 ± 3.5 kg/m² for SIFK and OA patients, respectively, was significantly higher in the SIFK patients (P = 0.02). Duration between the onset of pain and surgery was significantly shorter in the SIFK patients (29.4 ± 60.4 months) than the knee OA patients (86.1 ± 57.0 months) (P = 0.003) (Table 1).

Imaging and histological data of SIFK

Radiograph of the knee joints of T-SIFK showed occasionally (6/9 cases) a horizontal band parallel to the medial tibial plateau or negligible changes (Fig. 1A). However, all the medial tibial plateaus exhibited BMLs comprising a diffuse high signal intensity on fat suppressed MRI (PDFS, T2WI-SPAIR or T2WI-FS) sequences (Fig. 1B) and a low signal intensity on PDWI MRI (Fig. 1C). Importantly, a low signal intensity line (so called a fracture line) within the BML on fat suppressed MRI and PDWI MRI was observed in all the tibial plateaus of T-SIFK (9/9 cases) (Fig. 1B and C). Subchondral bone cyst (SBC) (88.9%, 8/9 cases) was frequently detected in the medial tibial plateau of these cases (Table 1). Subchondral bone attrition (SBA) was observed in 66.7% of the cases (6/9 cases). In the 10 patients with a diagnosis of F-SIFK (Table 1), radiological analysis of these knee joints showed subchondral collapse in the medial femoral condyles (Supplemental Fig. 1A) and MRI analysis demonstrated a low intensity line within the BML in all the medial femoral condyles (Supplemental Fig. 1B and C). However, because of surgical procedures, femoral condyle except for the posterior part became osteo-cartilaginous fragments during the operation of UKA, and thus the medial tibial plateaus and the posterior parts of femoral condyles were subjected to further analysis.

Fig. 1.

Radiographic and MRI findings of the representative case of T-SIFK and macroscopic and micro-CT observations of the surgically removed medial tibial plateau from case No 6 in Table 1. A, Radiographic image of the patient’s right knee joint. Note a faint horizontal band parallel to the medial tibial plateau. Boxed area is magnified below. B and C, Images of the patient’s knee joint observed by proton density-weighted fat-suppressed (PDFS) MRI (B) and proton density-weighted imaging (PDWI) MRI (C). Note that a low signal intensity line within the high signal intensity BMLs on PDFS sequences. Boxed areas are magnified to the right. D, Macroscopic observation of the unfixed medial tibial plateau, which was obtained by UKA. Medial meniscus is removed. E, Micro-CT findings of the medial tibial plateau at the horizontal and vertical axes. After fixation with formalin, the tissue was subjected to micro-CT analysis. Dotted horizontal and vertical lines divide the medial tibial plateau into four quadrants. The red lines in D and E indicate the plane that the tibial plateau is spliced for the analysis by micro-CT and histology. The green lines in D and E indicate the plane for histological analysis of Supplemental Fig. 2. F, Schematic map of the SIFK localization in the medial tibial plateau prepared from micro-CT images of the 5 T-SIFK cases. Ant, anterior; Post, posterior; Med, medial; Lat, lateral part

Macroscopically, all the medial tibial plateaus obtained from the T-SIFK patients at UKA showed the ill-defined lesion with irregular surfaces (Fig. 1D). When the medial tibial plateau was divided into four quadrants, i.e. anteromedial, anterolateral, posteromedial, and posterolateral, the lesion appeared to be present around the anteromedial and posteromedial parts of the tibial plateau (Fig. 1D). Intriguingly, micro-CT analysis clearly demonstrated that the lesion has a distorted oval shape in the anteromedial and posteromedial parts of the tibial plateau, and it is composed of numerous irregularly shaped subchondral bone fragments covered with articular cartilage (Fig. 1E). When the area of SIFK lesion was mapped on 3D reconstructed micro-CT images from 5 SIFK cases, the lesion was located in the central medial part of the medial tibial plateau and extended in the anterior–posterior direction (Fig. 1F).

Histological analysis of representative slices of the medial tibial plateau indicated that articular cartilage is maintained in all areas and the lesion detected on micro-CT comprises the articular cartilage-covered subchondral bone plate fragments, which were detached from the subchondral bone trabeculae (Fig. 2A, B, J and K, and Supplemental Fig. 2A). By closer examination, debris of bone and bone marrow together with serous exudates were present around and below the articular cartilage-covered subchondral bone plate fragments (Fig. 2B and C, and Supplemental Fig. 2A and B). Bone resorption by TRAP-positive osteoclasts was also commonly observed (Supplemental Fig. 2C and D). In addition, the formation of woven bone, cartilage and fibrovascular granulation tissues was seen around the subchondral bone fragments (Fig. 2C, D, L and M, and Supplemental Fig. 2C and F). Immunohistochemistry of CD31 and type X collagen demonstrated increased number of microblood vessels in the fibrogranulation tissue and production of type X collagen by hypertrophic chondrocytes in the newly formed cartilage (Supplemental Fig. 2E and G). The subchondral bone of the fragments showed empty lacunae lacking osteocytes, whereas articular cartilage on the subchondral bone plate was viable, retaining chondrocyte nuclei (Fig. 2E and F). All these data strongly suggest that the lesion represents the subchondral bone fracture and repair, and that the line showing detachment of the subchondral bone plate fragments from the granulation and osteo-cartilaginous tissues containing exudation and bone debris represents a fracture line.

Fig. 2.

Histology and immunohistochemistry of the medial tibial plateau obtained from the T-SIFK patient shown in Fig. 1. Serial sections were stained with HE (A, B, C, D, E, F and G) and Safranin O (J, K, L, and M) and also subjected to immunohistochemistry (H and I). Boxed areas labeled by small letters in A and B (b, c, d, e, f and g) and J (k, l and m) are shown at higher magnifications to the places marked by corresponding capital letters (B, C, D, E, F, G, K, L and M). H and I, Immunostaining for perilipin and CD68, respectively. Arrowheads and arrows indicate perilipin-negative fat cells and CD68-positve macrophages, respectively. Scale bars: 5 mm for A and J, 200 μm for B, C and K, 100 μm for D, L, M and 50 μm for E, F, G, H and I

Imaging and histological data of tibial OA

All 24 patients with knee OA were in advanced stage with K/L grades of 3 or 4 (Table 1) and radiographs of the knee joints showed a narrowing in space between the bones of the medial femoral-tibial joint and osteophyte formation on the joint margins (Fig. 3A). By fat suppressed MRI and PDWI MRI, BMLs were observed in the medial tibial plateaus of all the OA cases (Fig. 3B and C). SBA and SBC were detected by MRI in 25% (6/24 cases) and 25% (6/24 cases), respectively (Table 1). However, a low signal intensity line (a fracture line) within the BML on PDFS sequences was not observed in these OA cases (Fig. 3B and C and Supplemental Fig. 3A and B).

Fig. 3.

Radiographic and MRI findings of the representative case of knee OA and macroscopic and micro-CT observations of the surgically removed medial tibial plateau from case No 30 in Table 1. A, Radiographic image of the patient’s right knee joint. Note a narrowing in space between medial femoral-tibial joint and osteophyte formation. Boxed area is magnified below. B and C, Images of the patient’s knee joint observed by T2-weighted spectral adiabatic inversion-recovery (T2-SPAIR) MRI (B) and proton density-weighted imaging (PDWI) MRI (C). Boxed areas are magnified to the right. D, Macroscopic observation of the unfixed medial tibial plateau, which was obtained by UKA. Dotted horizontal and vertical lines divide the medial tibial plateau into four quadrants. Note that eburnated lesion is detected in the anteromedial part of the medial tibial plateau. E, Micro-CT findings of the medial tibial plateau at the horizontal and vertical axes. After fixation with formalin, the tissue was subjected to micro-CT analysis. The red lines indicate the plane that the tibial plateau is spliced for the analysis by micro-CT and histology. F, Schematic map of the eburnation’s localization in the medial tibial plateau prepared from micro-CT images of the 10 OA cases with eburnation and microfracture. Ant, anterior; Post, posterior; Med, medial; Lat, lateral part

Macroscopical observation of all the medial tibial plateaus removed from the OA knee joints showed articular surface irregularity and 91.7% of the cases (22 of 24 cases) had eburnated lesions (Table 1), which were recognized as focal depressed areas lacking articular cartilage mainly in the anteromedial part of the medial tibial plateaus (Fig. 3D and Supplemental Fig. 3C). By micro-CT analysis, the macroscopically detected eburnated lesions were shown as the areas with sclerotic appearance of trabecular bone and loss of articular cartilage on horizontal and vertical sections (Fig. 3E and Supplemental Fig. 3D). 3D reconstructed micro-CT images showed that the eburnated lesions are located primarily in the anteromedial part of the medial tibial plateau and extended anterolaterally and posteromedially (Fig. 3F and Supplemental Fig. 3E).

Histologically, representative sections of the medial tibial plateau in all 24 cases showed OA changes such as fibrillation, laceration and depletion of Safranin O-stained proteoglycans together with chondrocyte cluster formation in articular cartilage, and the eburnated regions were denuded and osteosclerotic without coverage of articular cartilage (Fig. 4A, B, H and I and Supplemental Fig. 3F, G, K and L). Importantly, 59.1% of the eburnated lesions (13 of 22 cases) were accompanied with microfracture, which was characterized by the presence of dead bone trabeculae and formation of fibrovascular granulation, cartilage, and woven bone tissues (Fig. 4C, D, J and K). In the remaining 40.9% of the eburnated cases without microfracture (9 of 22 cases), the subchondral bone plate and trabeculae in the eburnated area were composed of thickened plate-like structures (Supplemental Fig. 3F, G, K and L), which appeared to be different from the irregular-shaped rod-like trabeculae shown in the eburnated lesions with microfracture (Fig. 4A, B, H and I).

Fig. 4.

Histology and immunohistochemistry of the medial tibial plateau obtained from the knee OA patient shown in Fig. 3. Serial sections were stained with HE (A, B, C, D and E) and Safranin O (H, I, J, and K) and also subjected to immunohistochemistry (F and G). Boxed areas labeled by small letters in A (b, c, d and e) and H (i, j and k) are shown at higher magnifications to the places marked by corresponding capital letters (B, C, D, E, I, J and K). F and G, Immunostaining for perilipin and CD68, respectively. Arrowheads and arrows indicate perilipin-negative fat cells and CD68-positve macrophages, respectively. Scale bars: 5 mm for A and H, 500 μm for B and I, 100 μm for D and K and 50 μm for C, E, F, G and J

To further examine the morphological changes, we carried out micro-CT analysis of the eburnated areas in the medial tibial plateaus and compared the parameters between eburnated cases with microfracture (n = 10) and those without microfracture (n = 7) (Supplemental Fig. 4A). There was no difference in Tb.N (2.2 ± 0.8 vs 1.7 ± 0.2, P = 0.09) between the two groups. However, the eburnated cases with microfracture as compared to the eburnated cases without microfracture showed significantly lower BV/TV (83.8 ± 8.1 vs 93.4 ± 5.6, P = 0.016), Tb.Th (404.6 ± 110.7 vs 565.7 ± 83.1, P = 0.005) and BMD (877.1 ± 104.0 vs 974.8 ± 71.1, P = 0.048). On the other hand, Tb.Sp (74.1 ± 33.7), N.Nd (42.2 ± 17.7), and N.NdNd (70.0 ± 27.4) in the eburnated cases with microfracture were significantly higher than those without microfracture (37.0 ± 27.1, P = 0.029; 20.9 ± 8.6, P = 0.01 and 32.9 ± 14.7, P = 0.005, respectively). These data confirm the contrasting histological findings that the eburnated lesions without microfracture comprise thickened plate-like subchondral bone structure, whereas the eburnated lesions with microfracture are composed of thinner rod-like subchondral bone. In contrast, when micro-CT analysis was performed in the uninvolved areas in the medial tibial plateaus between cases of T-SIFK and F-SIFK (n = 12) and OA (n = 17), all the parameters showed no significant differences (Supplemental Fig. 4B).

Implications of fat necrosis for BMLs in the medial tibial plateaus from SIFK and OA joints

All the knee joints of SIFK and OA patients showed BMLs in the medial tibial plateaus (Table 1). In the SIFK cases, the BMLs, which were characterized by a diffuse high signal intensity with a low signal intensity line on PDFS MRI, appeared to be larger than those in the OA cases. Histologically, the subchondral bone areas corresponding to the BMLs were composed of various pathological changes such as exudative materials, necrotic bone fragments, fibrovascular granulation, cartilaginous tissues, and woven bone (Fig. 2). Interestingly, in the peripheral part of the BMLs, bone marrow fat cells, which were positively immunostained for perilipin, were disrupted, forming small vesicular structures and/or microcystic changes, both of which were negative for perilipin immunostaining (Fig. 2G and H). This lesion was accompanied by infiltration of CD68-positive foamy macrophages and lymphoid cells (Fig. 2I). These histopathological findings indicate that the BML in the peripheral part of the SIFK lesion corresponds to bone marrow fat necrosis.

Lesions of the medial tibial plateaus in OA cases, which included cases from the F-SIFK patients (n = 10) as well as the OA patients (n = 24) (Table 1), could be divided histologically into three types, i.e. OA with eburnation plus microfracture (Fig. 4), OA with eburnation alone (Supplemental Fig. 3), and OA without eburnation (Supplemental Fig. 5). In OA with eburnation plus microfracture, the histology of the subchondral bone areas of the tibial plateau corresponding to the BMLs on MRI showed fibrovascular granulation, cartilaginous tissue, and woven bone together with a small amount of necrotic bone fragments (Fig. 4). In the peripheral part of the lesion, there were foci of fat necrosis, which were characterized by damaged perilipin-negative fat cells and infiltrated CD68-positive foamy macrophages and lymphoid cells (Fig. 4E, F and G). On the other hand, the medial tibial plateaus showing OA with eburnation alone demonstrated that fat necrosis composed of disrupted fat cells and infiltrated foamy macrophages and lymphoid cells are present in the subchondral bone areas, which correspond to the BMLs (Supplemental Fig. 3H, I and J). In addition, histological analysis on the tibial samples of BML-positive OA without eburnation revealed that fat necrosis characterized by disrupted fat cells and infiltrated foamy macrophages and lymphoid cells is dominantly present to the subchondral bone areas corresponding to the BMLs (Supplemental Fig. 5E, F and G). In contrast, the subchondral bone areas of the medial tibial plateaus in the BML-negative OA (OA without eburnation) lacked fat necrosis, showing normal bone marrow fat tissue (Supplemental Fig. 6E, F and G). All these data suggest that fat necrosis contributes at least in part to BMLs on MRI in SIFK and OA knee joint pathology.

Fig. 5.

Comparison of radiological and MRI-based parameters, histology of subchondral bone resorption pits (SBRP), and comparison of SBRP between SIFK and knee OA patients. A, The radiological and MRI-based knee-related parameters. Upper panel shows radiological parameters including Kellgren-Lawrence classification (K/L), posterior tibial slope (PTS), femoro-tibial angle (FTA), mechanical axis (%ma), medial proximal tibial angle (MPTA), and lateral distal femoral angle (LDFA). The lower panel indicates MRI-based knee-related parameters including medial meniscal extrusion (MME), bone-part size of osteophyte (Bp), full-length size of osteophyte (Op), medial meniscal tear (MM tear), and medial meniscal posterior root tear (MMPRT). B, Representative histology of SBRP grade 1 and grade 2. SBRP grade 1 is restricted to the calcified cartilage and grade 2 is beyond the tidemark showing the lesion within articular cartilage. Scale bar, 50 μm. C, Comparison of the numbers of SBRP grade 1 and grade 2 in the uninvolved areas of the medial tibial plateau and the posterior femoral condyle between SIFK and OA patients. Data are expressed as means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001

Fig. 6.

Biomarkers in the samples from SIFK and OA patients. Blood and urine samples were obtained for the patients as described in the Materials and Methods and used for analyses of total cholesterol (Cho), hemoglobin A1c (HbA1c), blood sugar (BS), creatinine (Cre), estimated glomerular filtration rate (eGFR), albumin (Alb), calcium (Ca), phosphorus (P), 25 hydroxy vitamin D (25OHVitD), intact parathyroid hormone (intPTH), sclerostin, type 1 procollagen-N-propeptide (P1NP), undercarboxylated osteocalcin (ucOC), tartrate-resistant acid phosphatase-5b (TRACP-5b), deoxypyridinoline, urine (DPD), type I collagen cross-linked N-telopeptides, urine (NTX), homocysteine (Homo), and pentosidine (Pen). Data are expressed as means ± SD. *, P < 0.05

Comparisons of the parameters obtained by imaging and histological analyses of the SIFK and OA patients

When parameters obtained by radiographical and MRI analyses were compared between the SIFK and OA patient groups, K/L grade of the knee joints was significantly higher in OA group than SIFK group (2.73 ± 0.56 vs 3.21 ± 0.59, P = 0.011) (Fig. 5A). Importantly, the posterior tibial slope angle (PTS) measured by radiograph was significantly larger in SIFK patient group (12.84 ± 2.34°) than OA group (9.58 ± 2.84°) (P = 0.0002), but there were no significant differences in FTA, %ma, MPTA or LDFA (Fig. 5A). Although the degree of MME in SIFK and OA patients was not different between the two groups (6.13 ± 1.70 mm vs 6.50 ± 1.88 mm, P = 0.517, respectively), the bone-part size of osteophyte (Bp) and the full-length size of osteophyte (Op; sum of the bone and cartilage parts) around the medial tibial plateau were significantly lower in SIFK (2.72 ± 0.98 and 3.94 ± 1.13 mm) than OA (3.70 ± 1.53 and 5.24 ± 1.32 mm) (P = 0.001 and P = 0.020, respectively). In addition, the ratio of MME width to osteophyte width was significantly greater in SIFK (1.59 ± 0.35) than in OA (1.26 ± 0.30) (P = 0.002) (Fig. 5A). The occurrence rate of medial meniscal (MM) tears including medial meniscal posterior root tears (MMPRT) in the SIFK and OA patients was more than 80% of each group. Although horizontal tear was a main type of MM tears (10/24 cases; 41.7%) in OA group, rate of MMPRT was significantly higher in SIFK patients (13/19 cases; 68.4%) than OA patients (6/24 cases; 25.0%) (P = 0.004) (Fig. 5A).

In addition, we examined the subchondral bone resorption pit (SBRP) with samples of the medial tibial plateaus and the posterior parts of the medial femoral condyles. As shown in Fig. 5B, SBRP was classified into two types, i.e. grade 1 and grade 2, which correspond to the subchondral bone resorption pits below the tidemark and beyond the tidemark, respectively, according to Shibakawa et al. [34]. When the numbers of grade 1 and grade 2 of SBRP were counted in areas distant from the SIFK or eburnated lesions of the tibial plateaus, grade 2, but not grade 1, was significantly higher in SIFK group than OA group (2.32 ± 1.43 vs 0.72 ± 0.66, P = 0.001). Similarly, the number of grade 2, but not grade 1, evaluated with the femoral condyle samples was significantly higher in SIFK group than OA group (5.87 ± 3.73 vs 1.72 ± 2.39, P = 0.004) (Fig. 5C).

Differences in biological markers between SIFK and OA patients

We also examined biomarkers for lipid (total cholesterol) and glucose metabolism (HbA1C, blood sugar levels), kidney function (Cre, eGFR), bone metabolism (albumin, Ca, P, 25OHVitD, intPTH, sclerostin, P1NP, ucOC, TRACP-5b, NTX), and bone quality (homocysteine, pentosidine), but found no significant differences in any of them. However, a bone resorption marker, urinary deoxypyridinoline (DPD), was significantly higher in the SIFK group than the OA group (6.6 ± 1.9 vs. 5.3 ± 1.0, P = 0.03) (Fig. 6).

Discussion

In the present study, to the best of our knowledge, we have demonstrated for the first time that SIFK lesion is clearly identified by micro-CT analysis as a distorted oval-shaped area composed of multiple subchondral bone plate fragments in the central medial part of the medial tibial plateau. Histopathology has indicated that the SIFK lesion is formed by fracture of subchondral bone plate and its repair reactions. One of the characteristics of the SIFK lesion is that viable articular cartilage is lying on the dead subchondral bone fragments, and this contrasts to the eburnated lesion without articular cartilage in OA joint.

Previous MRI study without pathological analysis on femoral condyle insufficiency fracture, i.e. F-SIFK, showed that full-thickness articular cartilage is lost in the majority of the cases (75.7%, 53 of 70 cases) [35]. Subsequently, case reports on F-SIFK, which was examined by both MRI and histology, disclosed that the lesions of F-SIFK are covered with articular cartilage in early stage of the disease [36] and detachment of the lesion occasionally occurs, leading to formation of the loose body composed of articular cartilage and a thin necrotic subchondral bone plate [6]. In contrast to F-SIFK, due to the anatomical location of the tibia and gravity during walking, it seems likely that in T-SIFK lesions, the cartilage-covered bone fragments are preserved even if the lesion is detached at the fracture line below the subchondral bone plate of the medial tibial plateau. Therefore, our data together with previous studies [1, 6, 13, 36, 37] suggest that SIFK starts primarily from impairment of the subchondral bone, and the SIFK lesion is caused by fracture of the subchondral bone plate covered with full-thickness articular cartilage.

The association of meniscal abnormalities such as meniscal tears and posterior root tears in SIFK patients is well known [12, 13], and the meniscal problems are always associated with MME [9]. In the present study, 94.7% and 68.4% of the SIFK patients showed medial meniscal tears and medial meniscal posterior root tears (MMPRT), respectively, confirming the previous studies [12, 13, 38] and suggesting a possible implication of MMPRT for MME in SIFK patients. Our previous MRI studies of knee OA have suggested that osteophytes generated at the periphery of the medial tibial plateau promote MME through displacement of the coronary ligament by osteophytes and play a key role in OA progression by enhancing MME [32, 39–41]. In the present study, bone-part and full-length sizes of osteophytes in SIFK patients were smaller compared with those in OA patients but MME/Op ratio was significantly higher in SIFK than OA. Thus, the data suggests that MMPRT, but not osteophytes, may play a dual role in MME formation and induction of SIFK.

One interesting finding among the radiographical data in the present study was an increase in posterior tibial slope (PTS) in the SIFK patients (12.84 ± 2.34°) compared to the OA patients (9.58 ± 2.84°). A previous study also reported that the level of PTS is higher in SIFK patients than OA patients [38]. Since the shear force to the anterior tibial plateau is linearly increased with increasing PTS [42], higher PTS allows the medial femoral condyle and medial tibial plateau to generate high-level mechanical forces during joint movement and may act as a risk factor for development of SIFK. Previous studies have indicated that when PTS increases, anterior tibial translation, tibial shear force and forces on the anterior cruciate ligament increase [42, 43]. Therefore, it seems likely that SIFK lesion in the central medial part of the medial tibial plateau, being located more posteriorly than OA eburnation, may be ascribed to the increased mechanical forces loading to the area due to increased PTS. Although higher PTS and MMPRT are closely associated with SIFK as shown in the present and previous studies [38], their causative relationship is unknown. However, since a recent study has shown that increased PTS results in increased compression and anterior shear forces at the posterior medial meniscus root [44], it is possible to speculate that increased PTS induces MMPRT, which subsequently leads to the development of SIFK. However, experimental studies are needed to prove this hypothesis.

Several factors such as changes in bone mass, bone quality and bone repair rates are supposed to be implicated for insufficiency fractures [45]. In fact, osteoporosis and decreased mineral bone density were suggested to be linked to SIFK [9–11], but only 16% of SIFK patients reportedly had osteoporosis [14]. In the present study, however, the bone parameters obtained by micro-CT analysis of the uninvolved areas of the medial tibial plateaus and the biomarkers for bone metabolism derived from SIFK or OA patients showed no differences in both groups and provided no evidence of osteoporosis in SIFK patients. Therefore, induction of subchondral bone fracture in SIFK patients is not explained by local and/or general osteoporosis and thus osteoporosis may not be an essential factor for SIFK development. Although diabetes mellitus is known to contribute to the increased fracture risk through multiple mechanisms such as decreased bone turnover and alterations of the bone properties and microstructure [46], biomarkers for glucose metabolism or bone quality showed no significant differences between the SIFK and OA groups in the present study.

One of the intriguing findings in the present study was that the numbers of grade 2 SBRP in uninvolved areas of the medial tibial plateaus and femoral condyles are significantly increased in SIFK cases than OA cases. In addition, the level of DPD, a bone resorption biomarker, was significantly increased in SIFK patients compared to OA patients. SBRP grade 2, which is characterized by osteoclastic bone resorption and vascular invasion of the subchondral bone plate, is suggested to be one of the mechanisms of subchondral bone remodeling [34]. Thus, it seems plausible to speculate that SBRP grade 2 may contribute to fragility of the subchondral bone plate, leading to insufficiency subchondral fracture in response to enhanced mechanical stress. Although the regulation mechanism of SBRP formation remains unknown, mechanical stress is suggested to stimulate the pit formation [34]. As discussed above, SIFK patients had higher PTS, severe MME and MMPRT as well as higher BMI. Since all these factors produce enhanced mechanical forces in both medial tibial plateaus and medial femoral condyles, they might be involved in the SIFK induction via increased pit formation. However, further studies on the regulation mechanisms of SBRP formation by developing experimental SIFK models are needed.

In a stark contrast to SIFK, articular cartilage in advanced-stage knee OA commonly disappears due to complete destruction of articular cartilage and articular surface exhibits eburnation. In the present study, the majority of OA cases (91.7%, 22 of 24 cases) showed eburnation, and the eburnated lesions were classified into two forms with or without microfracture, which correspond to grade 6 (deformation) and grade 5 (denudation), respectively, according to the OARSI cartilage OA histopathological grading system [47]. To the best of our knowledge, no information is available for prevalence of these lesions in advanced knee OA patients who underwent UKA, but our study suggests the possibility that the eburnated lesion is commonly accompanied by microfracture, since microfracture was observed in 59.1% of our cases. Although the mechanism of the microfracture formation is unclear, histology and micro-CT have shown that the eburnated lesion with microfracture is characterized by the irregular-shaped rod-like trabeculae, increased Tb.Sp, N.Nd and N.NdNd together with decreased BV/TV, Tb.Th and BMD, as compared with the thickened plate-like subchondral bone plate in the eburnated lesion without microfracture. Thus, it seems likely that when the eburnated areas were not endowed with well-developed thickened subchondral bone plate, microfracture may be induced in the areas by increased mechanical stress. However, further studies are needed to demonstrate this hypothesis.

BMLs are frequently observed in both SIFK and OA knee joints [26, 35, 48]. Previous histological studies on the BMLs in OA knee joints reported that regions within the tibial plateaus demonstrating BMLs consisted of bone marrow necrosis, necrotic or remodeled trabeculae, bone marrow fibrosis, bone marrow edema and bone marrow bleeding in only less than 30% of the cases and the remaining cases had normal histology [49]. Another study showed that subchondral regions associated with BMLs comprise cysts, fibrosis, hypervascularity, cartilage islands, trabecular thickening, loss of tidemark integrity and inflammatory cell infiltration [26]. Since BMLs are detected as water signals by MRI, pathological conditions accompanied with hypervascularity including increased bone remodeling activity are considered to contribute to formation of BMLs [50]. Therefore, it is plausible that BMLs observed in SIFK and OA with eburnation plus microfracture in the present study correspond mainly to pathological changes related to fracture, such as destruction of bone and bone marrow tissues, exudation, fibrous granulation, cartilage and bone formation, all of which are accompanied by local increased vascularity and water signals.

Importantly, however, we found that fat necrosis, which was characterized by disrupted adipocytes and infiltration of foamy macrophage, existed not only at the peripheral parts of BMLs in SIFK and OA with eburnation plus microfracture but also in the BMLs of the medial tibial plateaus showing eburnation without microfracture or damaged articular cartilage without eburnation. These data are, to the best of our knowledge, the first observation that fat necrosis is closely associated with all the BMLs examined in both SIFK and OA. Fat necrosis is caused by various factors, which include trauma, ischemia, lipolytic enzymes, autoimmune, and infection. Pathologically, it can be classified into lipogranuloma type fat necrosis, coagulation-like necrosis type fat necrosis, enzymic fat necrosis, and lipomembranous fat necrosis [51]. Among them, fat necrosis observed in the present study was the lipogranuloma type, which appears to be caused by several factors including trauma, ischemia, and inflammation [51]. Altogether, it seems likely that fat necrosis in the epiphyseal subchondral bone area is a common, possibly early, event in BML formation, and may be caused by increased mechanical stress on the subchondral bone underlying the defective articular cartilage in SIFK and OA. In addition, since small localized fat necrosis is readily repaired, this may explain why BMLs in some cases is spontaneously recovered during the follow-up period [23, 25].

The present study has potential limitations that should be considered. First, while this study disclosed the characteristics of T-SIFK lesion by micro-CT and histology, our study contained a small number of patients with T-SIFK (n = 9) and we cannot deny the possibility that the present study may have introduced certain bias into the results. Second, due to UKA the medial femoral condyle tissues containing the F-SIFK lesion were collected as osteochondral tissue fragments, which were unsuitable for pathological examination, and therefore it was not possible to evaluate the similarities and/or differences between T-SIFK and F-SIFK lesions. Further study with a larger number of T-SIFK and/or F-SIFK patients by combined analyses using micro-CT and histopathology is needed to confirm our data in the present study and further clarify the different aspects of the SIFK lesions of the tibia and femur. Lastly, we observed histologically microfractures in approximately 60% of the eburnated lesions of OA knee joints and provided data suggesting that subchondral bone plate of such the lesions lacks well-developed thickened bone plate. However, the microfractures were not detected by MRI or micro-CT prior to histological analysis. Therefore, the development of clinical high-resolution imaging modalities capable of detecting changes in the subchondral bone structure is required to diagnose microfractures within the lesion and provide better treatment to patients.

Conclusions

We have demonstrated by micro-CT and histology that the SIFK lesion in the medial tibial plateau is composed of multiple irregularly shaped subchondral bone fragments covered with viable articular cartilage, debris of bone and bone marrow, fibro-granulation, and formation of cartilage and woven bone, indicating subchondral bone fracture and its repair. In addition, our data have shown that all the BMLs in the medial tibial plateaus from SIFK and OA are associated with fat necrosis of the bone marrow.

Abbreviations

ACL

Anterior cruciate ligament

BMD

Bone mineral density

BMI

Body mass index

BMLs

Bone marrow lesions

BV

Bone volume

DPD

Urinal deoxypyridinoline

F-SIFK

SIFK involving the medial femoral condyle

FTA

Femoro-tibial angle

K/L

Kellgren-Lawrence

LDFA

Lateral distal femoral angle

micro-CT

Microcomputed tomography

MME

Medial meniscal extrusion

MME/Op

Ratio of MME to osteophyte

MMPRT

Medial meniscal posterior root tears

MPTA

Medial proximal tibial angle

MRI

Magnetic resonance imaging

N.Nd

Node number

N.NdNd

Node node number

OA

Osteoarthritis

PDFS

Proton density-weighted fat-depressed

PDWI

Proton density-weighted imaging

%ma

Mechanical axis

SBA

Subchondral bone attrition

SBC

Subchondral bone cyst

SBRP

Subchondral bone resorption pit

SIFK

Subchondral insufficiency fracture of the knee

PTS

Posterior tibial slope

SONK

Spontaneous osteonecrosis of the knee

T-SIFK

SIFK involving the medial tibial plateau

T2WI-FS

T2-weighted fat-suppressed

T2WI-SPAIR

T2-weighted spectral adiabatic inversion-recovery

Tb.N

Trabecular number

Tb.Sp

Trabecular separation

Tb.Th

Trabecular thickness

3D

Three-dimensional

UKA

Unicompartmental knee arthroplasty

Authors’ contributions

TY had a role in acquisition, analysis, and interpretation of the data, and drafting the article. JT and HK had a role in acquisition, analysis, and interpretation of the data. YK and CY assisted in the experiments and had a role in analysis of the data. TN and MI had a role in conception and design and drafting the article. YO had a role in conception and design and drafting the article, had a full access to all the data in the study, and takes responsibility for the integrity of the data and accuracy of the data analysis. All authors read and approved of the final manuscript.

Funding

This study was supported in part by Grant-in-Aid for Scientific Research from the Japanese Society for the Promotion of Science (JSPS) to MI (15K10494 and 18K09082), TN (19K17209) and YO (16H05454 and 19H03788). This study was also funded in part by a JOA (Japanese Orthopaedic Association)-Subsidized Science Project Research 2022–2024; by a High Technology Research Center Grant and the Program for the Strategic Research Foundation at Private Universities (2014–2019) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT); by a Center of Innovation (COI) program, one of the main funding programs under the Center of Innovation Science and Technology based Radical Innovation and Entrepreneurship Program (COI STREAM), which was launched in 2013 by the MEXT and operated by Japan Science and Technology Agency (JST), to Juntendo University.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study protocol, which complied with the principles outlined in the Declaration of Helsinki, was approved by the Ethical Committee Review Board at Juntendo University (Ethics Committee Research Subject No E22-0352). Written informed consent was obtained from all patients.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.