A rig for acquisition of standardized trabecular bone radiographs

Bone Data Lake: A storage platform for bone texture analysis

Trabecular bone (TB) texture regions selected on hand and knee X-ray images can be used to detect and predict osteoarthritis (OA). However, the analysis has been impeded by increasing data volume and diversification of data formats. To address this problem, a novel storage platform, called Bone Data Lake (BDL) is proposed for the collection and retention of large numbers of images, TB texture regions and parameters, regardless of their structure, size and source. BDL consists of three components, i.e.: a raw data storage, a processed data storage, and a data reference system. The performance of the BDL was evaluated using 20,000 knee and hand X-ray images of various formats (DICOM, PNG, JPEG, BMP, and compressed TIFF) and sizes (from 0.3 to 66.7 MB). The images were uploaded into BDL and automatically converted into a standardized 8-bit grayscale uncompressed TIFF format. TB regions of interest were then selected on the standardized images, and a data catalog containing metadata information about the regions was constructed. Next, TB texture parameters were calculated for the regions using Variance Orientation Transform (VOT) and Augmented VOT (AVOT) methods and stored in XLSX files. The files were uploaded into BDL, and then transformed into CSV files and cataloged. Results showed that the BDL efficiently transforms images and catalogs bone regions and texture parameters. BDL can serve as the foundation of a reliable, secure and collaborative system for OA detection and prediction based on radiographs and TB texture.

A dissimilarity-based multiple classifier system for trabecular bone texture in detection and prediction of progression of knee osteoarthritis

There is a growing need for classification systems that can accurately detect and predict knee osteoarthritis (OA) from plain radiographs. For this purpose, a system based on a support vector machine (SVM) classifier and distances measured between trabecular bone (TB) texture images was developed and tested in previous work. Unlike other systems, it allows an image classification without the calculation and selection of numerous texture features, and it is invariant to a range of imaging conditions encountered in a routine X-ray screening of knees. Although the system exhibited 85.4% classification accuracy in OA detection, which was higher than those obtained from other systems, its performance could be further improved. To achieve this, a dissimilarity-based multiple classifier (DMC) system is developed in this study. The system measures distances between TB texture images and generates a diverse ensemble of classifiers using prototype selection, bootstrapping of training set and heterogeneous classifiers. A measure of competence is used to select accurate (i.e. better-than-random) classifiers from the ensemble, which are then combined through the majority voting rule. To evaluate the newly developed system in OA detection (prediction of OA progression), TB texture images selected on standardised radiographs of healthy and OA (non-progressive and progressive OA) knees were used. The results obtained showed that the DMC system has higher classification accuracies for the detection (90.51% with 87.65% specificity and 93.33% sensitivity) and prediction (80% with 82.00% specificity and 77.97% sensitivity) than other systems, indicating its potential as a decision-support tool for the assessment of radiographic knee OA.

A signature dissimilarity measure for trabecular bone texture in knee radiographs

PURPOSE

The purpose of this study is to develop a dissimilarity measure for the classification of trabecular bone (TB) texture in knee radiographs. Problems associated with the traditional extraction and selection of texture features and with the invariance to imaging conditions such as image size, anisotropy, noise, blur, exposure, magnification, and projection angle were addressed.

METHODS

In the method developed, called a signature dissimilarity measure (SDM), a sum of earth mover's distances calculated for roughness and orientation signatures is used to quantify dissimilarities between textures. Scale-space theory was used to ensure scale and rotation invariance. The effects of image size, anisotropy, noise, and blur on the SDM developed were studied using computer generated fractal texture images. The invariance of the measure to image exposure, magnification, and projection angle was studied using x-ray images of human tibia head. For the studies, Mann-Whitney tests with significance level of 0.01 were used. A comparison study between the performances of a SDM based classification system and other two systems in the classification of Brodatz textures and the detection of knee osteoarthritis (OA) were conducted. The other systems are based on weighted neighbor distance using compound hierarchy of algorithms representing morphology (WND-CHARM) and local binary patterns (LBP).

RESULTS

Results obtained indicate that the SDM developed is invariant to image exposure (2.5-30 mA s), magnification (x1.00 - x1.35), noise associated with film graininess and quantum mottle (< 25%), blur generated by a sharp film screen, and image size (> 64 x 64 pixels). However, the measure is sensitive to changes in projection angle (> 5 degrees), image anisotropy (> 30 degrees), and blur generated by a regular film screen. For the classification of Brodatz textures, the SDM based system produced comparable results to the LBP system. For the detection of knee OA, the SDM based system achieved 78.8% classification accuracy and outperformed the WND-CHARM system (64.2%).

CONCLUSIONS

The SDM is well suited for the classification of TB texture images in knee OA detection and may be useful for the texture classification of medical images in general.

Directional fractal signature analysis of trabecular bone: evaluation of different methods to detect early osteoarthritis in knee radiographs

There is ongoing research directed towards the development of cheap and reliable decision support systems for the detection and prediction of osteoarthritis (OA) in knee joints. Fractal analysis of trabecular bone texture X-ray images is one of the most promising approaches. It is cheap and non-invasive. However, difficulties arise when the fractal signature methods are used to quantify bone roughness and anisotropy on individual scales. This is because the fractal methods are able to quantify bone texture only in the vertical and horizontal directions, and previous studies showed that OA bone changes can occur in any direction. To address these difficulties, three directional fractal signature methods were developed in this study, i.e. a fractal signature Hurst orientation transform (FSHOT) method, a variance orientation transform (VOT) method, and a blanket with rotating-grid (BRG) method. These methods were tested and the best performing method was selected. Unlike other methods, the newly developed techniques are able to calculate fractal dimensions (FDs) on individual scales (i.e. fractal signature) in all possible directions. The accuracy of the methods developed in measuring texture roughness and anisotropy on individual scales was evaluated. The effects of imaging conditions such as image noise, blur, exposure, magnification, and projection angle and the effects of translation of the bone region of interest on texture parameters were also evaluated. Computer-generated fractal surface images with known FDs and X-ray images obtained for a human tibia head were used. Results obtained show that the VOT method performs better than the FSHOT and BRG methods.

Knee joint biomechanics following arthroscopic partial meniscectomy

We investigated spatiotemporal data, joint kinematics, and joint kinetics during gait in a group of subjects who had recently undergone arthroscopic partial meniscectomy and compared the results to those of healthy controls. Gait analysis was performed on 105 pain-free meniscectomy patients and 47 controls, walking at a self-selected speed. The meniscectomy population was comparable to controls in spatiotemporal parameters and knee kinematics. However, they had reduced range of motion (ROM) and lower peak moments in the sagittal plane on the operated limb compared to the nonoperated limb. Compared to controls, the meniscectomy patients had significantly larger knee adduction moments over stance, even after accounting for their greater body weight. These differences likely increase articular loads on the medial compartment of the tibiofemoral joint and may contribute to the high risk of knee osteoarthritis following arthroscopic meniscal surgery.

Differences in trabecular bone texture between knees with and without radiographic osteoarthritis detected by fractal methods

OBJECTIVE

To develop an accurate method for quantifying differences in the trabecular structure in the tibial bone between subjects with and without knee osteoarthritis (OA).

METHODS

Standard knee radiographs were taken from 26 subjects (seven women) with meniscectomy and radiographic OA Kellgren & Lawrence grade 2 or worse in the medial compartment. Each case knee was individually matched by sex, age, body mass index and medial or lateral compartment with a control knee. A newly developed augmented Hurst orientation transform (HOT) method was used to calculate texture parameters for regions selected in X-ray images of non-OA and OA tibial bones. This method produces a mean value of fractal dimensions (FD MEAN), FDs in the vertical (FDV) and horizontal (FDH) directions and along a direction of the roughest part of the tibial bone (FDSta), fractal signatures and a texture aspect ratio (Str). The ratio determines a degree of the bone texture anisotropy. Reproducibility was calculated using an intraclass correlation coefficient (ICC). Comparisons between cases and controls were made with paired t tests. The performance of the HOT method was evaluated against a benchmark fractal signature analysis (FSA) method.

RESULTS

Compared with controls, trabecular bone in OA knees showed significantly lower FD MEAN, FDV, FDH and FDSta and higher Str at trabecular image sizes 0.2-1.1mm (P<0.05, HOT). The reproducibility of all parameters was very good (ICC>0.8). In the medial compartment, fractal signatures calculated for OA horizontal and vertical trabeculae were significantly lower at sizes 0.3-0.55 mm (P<0.05, HOT) and 0.3-0.65 mm (P<0.001, FSA). In the lateral compartment, FDs calculated for OA trabeculae were lower than controls (horizontal: 0.3-0.55 mm (P<0.05, HOT) and 0.3-0.65 mm (P<0.001, FSA); vertical: 0.3-0.4mm (P<0.05, HOT) and 0.3-0.35 mm (P<0.001, FSA).

CONCLUSION

The augmented HOT method produces fractal signatures that are comparable to those obtained from the benchmark FSA method. The HOT method provides a more detailed description of OA changes in bone anisotropy than the FSA method. This includes a degree of bone anisotropy measured using data from all possible directions and a texture roughness calculated for the roughest part of the bone. It appears that the augmented HOT method is well suited to quantify OA changes in the tibial bone structure.

Automated classification of articular cartilage surfaces based on surface texture

In this study the automated classification system previously developed by the authors was used to classify articular cartilage surfaces with different degrees of wear. This automated system classifies surfaces based on their texture. Plug samples of sheep cartilage (pins) were run on stainless steel discs under various conditions using a pin-on-disc tribometer. Testing conditions were specifically designed to produce different severities of cartilage damage due to wear. Environmental scanning electron microscope (SEM) (ESEM) images of cartilage surfaces, that formed a database for pattern recognition analysis, were acquired. The ESEM images of cartilage were divided into five groups (classes), each class representing different wear conditions or wear severity. Each class was first examined and assessed visually. Next, the automated classification system (pattern recognition) was applied to all classes. The results of the automated surface texture classification were compared to those based on visual assessment of surface morphology. It was shown that the texture-based automated classification system was an efficient and accurate method of distinguishing between various cartilage surfaces generated under different wear conditions. It appears that the texture-based classification method has potential to become a useful tool in medical diagnostics.