How Do MRI-Detected Subchondral Bone Marrow Lesions (BMLs) on Two Different MRI Sequences Correlate with Clinically Important Outcomes?

Deep learning-based automated detection and segmentation of bone and traumatic bone marrow lesions from MRI following an acute ACL tear

INTRODUCTION

Traumatic bone marrow lesions (BML) are frequently identified on knee MRI scans in patients following an acute full-thickness, complete ACL tear. BMLs coincide with regions of elevated localized bone loss, and studies suggest these may act as a precursor to the development of post-traumatic osteoarthritis. This study addresses the labour-intensive manual assessment of BMLs by using a 3D U-Net for automated identification and segmentation from MRI scans.

METHODS

A multi-task learning approach was used to segment both bone and BML from T2 fat-suppressed (FS) fast spin echo (FSE) MRI sequences for BML assessment. Training and testing utilized datasets from individuals with complete ACL tears, employing a five-fold cross-validation approach and pre-processing involved image intensity normalization and data augmentation. A post-processing algorithm was developed to improve segmentation and remove outliers. Training and testing datasets were acquired from different studies with similar imaging protocol to assess the model's performance robustness across different populations and acquisition conditions.

RESULTS

The 3D U-Net model exhibited effectiveness in semantic segmentation, while post-processing enhanced segmentation accuracy and precision through morphological operations. The trained model with post-processing achieved a Dice similarity coefficient (DSC) of 0.75 ± 0.08 (mean ± std) and a precision of 0.87 ± 0.07 for BML segmentation on testing data. Additionally, the trained model with post-processing achieved a DSC of 0.93 ± 0.02 and a precision of 0.92 ± 0.02 for bone segmentation on testing data. This demonstrates the approach's high accuracy for capturing true positives and effectively minimizing false positives in the identification and segmentation of bone structures.

CONCLUSION

Automated segmentation methods are a valuable tool for clinicians and researchers, streamlining the assessment of BMLs and allowing for longitudinal assessments. This study presents a model with promising clinical efficacy and provides a quantitative approach for bone-related pathology research and diagnostics.

Bone marrow lesions in osteoarthritis: biomarker or treatment target? A narrative review

Osteoarthritis (OA) is a leading cause of pain, functional impairment, and disability in older adults. However, there are no effective treatments to delay and reverse OA. Magnetic resonance imaging (MRI) can assess structural abnormalities of OA by directly visualizing damage and inflammatory reactions within the tissues and detecting abnormal signals in the subchondral bone marrow region. While some studies have shown that bone marrow lesions (BMLs) are one of the early signs of the development of OA and predict structural and symptomatic progression of OA, others claimed that BMLs are prevalent in the general population and have no role in the progression of OA. In this narrative review, we screened and summarized studies with different designs that evaluated the association of BMLs with joint symptoms and structural abnormalities of OA. We also discussed whether BMLs may serve as an imaging biomarker and a treatment target for OA based on existing clinical trials.

[Assessing the Correlation between the Radiological, Macroscopic and Histological Examination of Degenerative Changes of Articular Surfaces in Knee Osteoarthritis with Varus Deformity]

PURPOSE OF THE STUDY

Our study aims to compare the results of preoperative radiography and intraoperative visual assessment of the cartilage with histological assessment of joint surfaces of the medial and lateral compartments resected in patients during the total knee replacement.

MATERIAL AND METHODS

The cohort included 20 patients (9 men and 11 women) with the mean age of 66.6 (±7.0) years who met the inclusion criteria of the study. Degenerative changes of the knee joint seen on a preoperative weight-bearing anteroposterior X-ray were evaluated according to the Kellgren-Lawrence grading system separately for the medial and lateral compartment. Based on the visual appearance, the condition of articular surfaces was assessed using the International Cartilage Repair Society Score (ICRS Grade). The histological assessment of degenerative changes was conducted by a pathologist with the use of the Osteoarthritis Research Society International Osteoarthritis Cartilage Histopathology Assessment System based on six grades of articular cartilage degeneration.

RESULTS

The mean degree of degenerative changes based on the radiological classification was assessed as 3.5 (±0.6) for the medial compartment and 2.1 (±0.4) for the lateral compartment. The visually assessed chondropathy according to the ICRS Grade was 3.7 (±0.6) for the medial femoral condyle and 1.8 (±1.0) for the lateral femoral condyle. The histological score obtained using the Osteoarthritis Research Society International Osteoarthritis Cartilage Histopathology Assessment was 4.9 (±1.1) for the medial femoral condyle and 2.4 (±0.7) for the lateral femoral condyle. In respect of the medial compartment, there was no statistically significant parametric correlation between the intraoperative visual assessment of the cartilage degeneration and the preoperative radiological grade r = 0.45. The histological assessment showed a statistically significant concordance both with the degree of chondropathy r = 0.76 and the radiological grade r = 0.64. In the lateral compartment, the parametric test showed a statistically significant concordance only between the radiological grade and the histological score r = 0.72. The correlation between the visual assessment of chondropathy and the radiological grade r = 0.27 as well as the histological score r = 0.24 was very low.

DISCUSSION

In our cohort assessing the early degenerative changes of the lateral compartment as well as the more advanced degenerative changes of the medial compartment, the correlation between the intraoperative assessment of cartilage degeneration as a diagnostic method to examine the lateral compartment and the preoperative radiological grade was not confirmed. Our results failed to confirm a better reporting value of the visual cartilage degeneration assessment of the lateral compartment as against the preoperative X-ray. The space width without narrowing on an X-ray has no reporting value for this compartment in case of varus deformity.

CONCLUSIONS

The results clearly indicate that the assessment of macroscopic appearance of the cartilage degeneration during arthroscopy does not necessarily guarantee good long-term clinical outcomes after high tibial osteotomy. The respective degrees of cartilage degeneration identified during the intraoperative visual assessment and the radiological grading of osteoarthritic changes did not correlate in either compartment. In the lateral compartment, the initial radiological and histological findings preceded the visually detectable cartilage changes.

KEY WORDS

knee, cartilage, osteoarthritis, radiology, histology, arthroscopy, osteotomy.

Dual-energy CT virtual non-calcium: an accurate method for detection of knee osteoarthritis-related edema-like marrow signal intensity

OBJECTIVES

To evaluate the performance of a dual-energy computed tomography (DECT) virtual non-calcium (VNCa) technique in the detection of edema-like marrow signal intensity (ELMSI) in patients with knee joint osteoarthritis (OA) compared to magnetic resonance imaging (MRI).

METHODS

The study received local ethics board approval, and written informed consent was obtained. DECT and MRI were used to examine 28 knees in 24 patients with OA. VNCa images were generated by dual-energy subtraction of calcium. The knee joint was divided into 15 regions for ELMSI grading, performed independently by two musculoskeletal radiologists, with MRI as the reference standard. We also analyzed CT numbers through receiver operating characteristics and calculated cut-off values.

RESULTS

For the qualitative analysis, we obtained CT sensitivity (Readers 1, 2 = 83.7%, 89.8%), specificity (Readers 1, 2 = 99.5%, 99.5%), positive predictive value (Readers 1, 2 = 95.3%, 95.7%), and negative predictive value (Readers 1, 2 = 97.9%, 98.7%) for ELMSI. The interobserver agreement was excellent (κ = 0.92). The area under the curve for Reader 1 and Reader 2 was 0.961 (95% CI 0.93, 0.99) and 0.992 (95% CI 0.98, 1.00), respectively. CT numbers obtained from the VNCa images were significantly different between regions with and without ELMSI (p < .001).

CONCLUSIONS

VNCa images have good diagnostic performance for the qualitative and quantitative analysis of knee osteoarthritis-related ELMSI.

Associations of osteoclastogenesis and nerve growth in subchondral bone marrow lesions with clinical symptoms in knee osteoarthritis

Background/objective

Subchondral bone marrow lesions (BMLs) are common magnetic resonance imaging (MRI) features in joints affected by osteoarthritis (OA), however, their clinical impacts and mechanisms remain controversial. Thus, we aimed to investigate subchondral BMLs in knee OA patients who underwent total knee arthroplasty (TKA), then evaluate the associations of osteoclastogenesis and nerve growth in subchondral BMLs with clinical symptoms.

Methods

Total 70 patients with primary symptomatic knee OA were involved, then separated into three groups based on MRI (without BMLs group, n ​= ​14; BMLs without cyst group, n ​= ​37; BMLs with cyst group, n ​= ​19). Volume of BMLs and cyst-like lesions was calculated via the OsiriX system. The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaire was used to assess clinical symptoms. Histology and immunohistochemistry were deployed to assess subchondral osteoclastogenesis and nerve distribution. Pearson's correlation coefficient was used to evaluate the associations between volume of BMLs and joint symptoms, and to assess the associations of osteoclastogenesis and nerve growth in subchondral BMLs with joint symptoms.

Results

In BMLs combined with cyst group, patients exhibited increased osteoclastogenesis and nerve distribution in subchondral bone, as shown by increased expression of tartrate resistant acid phosphatase (TRAP) and protein gene product 9.5 (PGP9.5). Volume of subchondral cyst-like component was associated with joint pain (p ​< ​0.05). Subchondral osteoclastogenesis and nerve distribution were positively associated with joint pain in BMLs with cyst group (p ​< ​0.05).

Conclusion

The subchondral cyst-like lesion was an independent factor for inducing pain in OA patients; osteoclastogenesis and nerve growth in subchondral cyst-like lesions could account for this joint pain.

The translational potential of this article

Our results indicated that the increased osteoclastogenesis and nerve growth in subchondral cyst-like lesions could account for the pain of OA joints. These findings may provide valuable basis for the treatment of OA.

Machine Learning-Based Individualized Survival Prediction Model for Total Knee Replacement in Osteoarthritis: Data From the Osteoarthritis Initiative

OBJECTIVE

By using machine learning, our study aimed to build a model to predict risk and time to total knee replacement (TKR) of an osteoarthritic knee.

METHODS

Features were from the Osteoarthritis Initiative (OAI) cohort at baseline. Using the lasso method for variable selection in the Cox regression model, we identified the 10 most important characteristics among 1,107 features. The prognostic power of the selected features was assessed by the Kaplan-Meier method and applied to 7 machine learning methods: Cox, DeepSurv, random forests algorithm, linear/kernel support vector machine (SVM), and linear/neural multi-task logistic regression models. As some of the 10 first-found features included similar radiographic measurements, we further looked at using the least number of features without compromising the accuracy of the model. Prediction performance was assessed by the concordance index, Brier score, and time-dependent area under the curve (AUC).

RESULTS

Ten features were identified and included radiographs, bone marrow lesions of the medial condyle on magnetic resonance imaging, hyaluronic acid injection, performance measure, medical history, and knee-related symptoms. The methodologies Cox, DeepSurv, and linear SVM demonstrated the highest accuracy (concordance index scores of 0.85, Brier score of 0.02, and an AUC of 0.87). DeepSurv was chosen to build the prediction model to estimate the time to TKR for a given knee. Moreover, we were able to decrease the features to only 3 and maintain the high accuracy (concordance index of 0.85, Brier score of 0.02, and AUC of 0.86), which included bone marrow lesions, Kellgren/Lawrence grade, and knee-related symptoms, to predict risk and time of a TKR event.

CONCLUSION

For the first time, we developed a model using the OAI cohort to predict with high accuracy if a given osteoarthritic knee would require TKR, when a TKR would be required, and who would likely progress fast toward this event.