Analysis and multiclass classification of pathological knee joints using vibroarthrographic signals

BACKGROUND AND OBJECTIVE

Vibroarthrography (VAG) is a method developed for sensitive and objective assessment of articular function. Although the VAG method is still in development, it shows high accuracy, sensitivity and specificity when comparing results obtained from controls and the non-specific, knee-related disorder group. However, the multiclass classification remains practically unknown. Therefore the aim of this study was to extend the VAG method classification to 5 classes, according to different disorders of the patellofemoral joint.

METHODS

We assessed 121 knees of patients (95 knees with grade I-III chondromalacia patellae, 26 with osteoarthritis) and 66 knees from 33 healthy controls. The vibroarthrographic signals were collected during knee flexion/extension motion using an acceleration sensor. The genetic search algorithm was chosen to select the most relevant features of the VAG signal for classification. Four different algorithms were used for classification of selected features: logistic regression with automatic attribute selection (SimpleLogistic in Weka), multilayer perceptron with sigmoid activation function (MultilayerPerceptron), John Platt's sequential minimal optimization algorithm implementation of support vector classifier (SMO) and random forest tree (RandomForest). The generalization error of classification algorithms was evaluated by stratified 10-fold cross-validation.

RESULTS

We obtained levels of accuracy and AUC metrics over 90%, more than 93% sensitivity and more than 84% specificity for the logistic regression-based method (SimpleLogistic) for a 2-class classification. For the 5-class method, we obtained 69% and 90% accuracy and AUC respectively, and sensitivity and specificity over 91% and 69%.

CONCLUSIONS

The results of this study confirm the high usefulness of quantitative analysis of VAG signals based on classification techniques into normal and pathological knees and as a promising tool in classifying signals of various knee joint disorders and their stages.

T2 mapping in patellar chondromalacia

OBJECTIVE

To study the correlation between the T2 relaxation times of the patellar cartilage and morphological MRI findings of chondromalacia.

METHODS

This prospective study comprises 50 patients, 27 men and 23 women suffering of anterior knee pain (mean age: 29.7, SD 8.3 years; range: 16-45 years). MRI of 97 knees were performed in these patients at 1.5T magnet including sagittal T1, coronal intermediate, axial intermediate fat sat and T2 mapping. Chondromalacia was assessed using a modified version of Noyes classification. The relaxation time, T2, was studied segmenting the full thickness of the patellar cartilage in 12 areas: 4 proximal (external facet-proximal-lateral (EPL), external facet-proximal-central (EPC), internal facet-proximal-central (IPC), internal facet-proximal-medial (IPM), 4 in the middle section (external facet-middle-lateral (EML), external facet-middle-central (EMC), internal facet-middle-central (IMC), internal facet-middle-medial (IMM) and 4 distal (external facet-distal-lateral (EDL), external facet-distal-central (EDC), internal facet-distal-central (IDC), internal facet-distal-medial (IDM).

RESULTS

T2 values showed a significant increase in mild chondromalacia regarding normal cartilage in most of the cartilage areas (p<0.05), except in the internal distal facet (IDC and IDM), EPC, EDL, and IMM. Severe chondromalacia was characterized by a fall of T2 relaxation times with loss of statistical significant differences in comparison with normal cartilage, except in EMC and IMC, where similar values as mild chondromalacia were maintained (p<0.05).

CONCLUSIONS

Steepest increase in T2 values of patellar cartilage occurs in early stages of patellar cartilage degeneration. Progression of morphologic changes of chondromalacia to more severe degrees is associated to a new drop of T2 relaxation times approaching basal values in most of the areas of the patellar cartilage, except in the central area of the middle section, where T2 values remain increased.

Evaluation of magnetic resonance signal modification induced by hyaluronic acid therapy in chondromalacia patellae: a preliminary study

Hyaluronic Acid (HA) is an alternative method for the treatment of osteoarthritis (OA), which acts on pain through a double action: anti-inflammatory and synovial fluid (SF) visco-supplementation. Magnetic Resonance Imaging (MRI), utilizing specific sequences, is a valid method for studying the initial phase of chondral damage. The analysis of the data, obtained through the intensity of values taken by positioning Region of Interest (ROIs) within the lesion, determining the differences before and after treatment with HA injected into the knee. The results obtained after six months and one year from the injection were statistically different in respect to those taken before, immediately and after three months of treatment. MRI represents a valid tool to evaluate the grade of chondromalacia patellae and also to follow the cartilage modification induced by HA therapy.

Retropatellar chondromalacia associated with medial osteoarthritis after meniscus injury. One year of observations in sheep

BACKGROUND/AIMS

In an ovine meniscal repair model, the patellofemoral (PF) osteoarthritis due to a non-sutured tear or failed repair was investigated.

METHODS

A radial meniscus tear was either sutured with polydioxanone (PDS), with a slow degrading polylactide long-term suture(LTS) or left without treatment. Knee joint cartilage in the PF and medial compartment was evaluated compared to normal knees (healthy controls).

RESULTS

Retropatellar osteoarthritis in the non-sutured and sutured animals was intense in contrast to the control knees after 6 months in all groups (p < 0.001), and after 12 months in the PDS group (p < 0.001), LTS group and non-sutured animals (p < 0.05).

CONCLUSION

Non-sutured meniscus tears and failed repair lead fast to intense PF osteoarthritis corresponding with tibial damage of the injured compartment.

In vivo study on the short-term effect of radiofrequency energy on chondromalacic patellar cartilage and its correlation with calcified cartilage pathology in an equine model

Chondromalacia can cause joint pain and synovial effusion with the potential for developing into osteoarthritis. Thermal chondroplasty using radiofrequency energy (RFE) has been reported to be superior to mechanical debridement for treating chondromalacia. We compared short-term changes in biomechanical properties of articular cartilage after treatment with monopolar (mRFE) or bipolar RFE (bRFE) or mechanical debridement (MD) on experimentally created grade II chondromalacia patellae. Chondromalacia patellae was created arthroscopically in both patellae of 15 ponies. Ten months after surgery, each patella was randomly assigned to one of four experimental groups: sham operated, untreated control; MD; bRFE; and mRFE. Animals were euthanized 6 months after treatment and fresh osteochondral sections were collected from the treated area, the border of the chondromalacic and nonchondromalacic area, and from two untreated areas for analysis of mechanical properties. The same areas were harvested from an additional six untreated ponies. The aggregate modulus (H(A)), Poisson's ratio (nu(s)), and permeability (k) were determined for each area under creep indentation, and cartilage thickness was measured with a needle probe. The relation between zone of calcified cartilage (ZCC) and mechanical properties of hyaline cartilage (HC) was assessed histomorphometrically. Treated areas of all four groups had inferior mechanical properties compared at the same location. The treated and border areas had significantly lower H(A) values than the untreated areas. Permeability values showed significant differences between bRFE and other treated groups. Chondromalacic areas showed thinning of cartilage compared to nonchondromalacic areas. Biomechanical properties of the injured cartilage were inferior to nonchondromalacic cartilage regardless of the treatment type. mRFE had the highest stiffness value compared to other treatments and significantly higher values than MD. A significant correlation was observed between the mechanical properties of HC and ZCC thickness.

Second-look arthroscopy of cartilage changes of the patellofemoral joint, especially the patella, following acute and recurrent patellar dislocation

OBJECTIVE

The purpose of this study was to clarify some pathogenesis of the cartilage changes of the patellofemoral joint in patellar dislocation knees.

METHODS

A first arthroscopy was performed in a total of 60 knees, including 30 knees of acute patellar dislocation (APD) and 30 knees of recurrent patellar dislocation (RPD). At the time of the first arthroscopy, 58 knees sustained open medial patellofemoral ligament repair/reconstruction, and 2 knees had only arthroscopy examination. In APD knees, the second-look arthroscopy was performed 16.7 months after the first arthroscopy, and in RPD knees, 25.2 months after the first arthroscopy.

RESULTS

In most APD knees, cracking primarily on the central dome of the patella did not have a remarkable change but in several knees that had a high number of cracking or interlaced cracking, cracking became worse with fibrillation. Osteochondral defect site primarily on the medial facet showed fibrillation and/or ulceration (erosion). In most RPD knees, fissuring primarily on the central dome did not have a remarkable change, and fibrillation and/or erosion mainly on the medial facet also did not have a remarkable change. However, in four knees with the continuation of patellar dislocation, the patellar cartilage changes had definitely worsened. On the other hand, in the first and second-look arthroscopy, most APD and RPD knees had no remarkable cartilage lesions or cartilage changes of the femoro-trochlear aspect.

CONCLUSIONS

One of the pathogenesis of fissuring in RPD knees seems to be cracking occurring at the time of APD. Two of the pathogenesis of fibrillation may result from the reparative reaction of the cartilage in osteochondral defect site and the cartilage change due to a high degree of cracking damage mainly on the central dome in APD injury. The continuation of patellar dislocation definitely makes the patellar cartilage lesions worse.

Welche Faktoren beeinflussen das Ergebnis der ACT bei Überschreiten der Indikationsgrenzen?

AIM

The present study evaluates the results of ACT at the knee joint with transgression of the usual spectrum of indications and points out the possible limits of the method.

METHODS

In 20 of a total of 71 patients the indication limits of ACT were exceeded. The clinical follow-up over two years was related to the localisation, size, adjacent cartilage, previous and additional operations and the quality of the transplanted chondrocytes.

RESULTS

The average patient age at time of operation was 41.1 years (24 to 55 years). The average defect size was 7.9 cm (2) (3.75 to 14 cm (2)). In the analysed group outside of the indication limits of ACT half of the patients showed good or excellent results at clinical follow-up after 2 years. 10 patients, however, assessed the clinical situation as fair or poor. Within this group 5 patients showed a minimum of two lesions. In 6 patients the treated defect was not completely surrounded by an adjacent cartilage shoulder. In 7 cases the meniscus at the site of the lesion had been partially removed in a previous operation and osteoarthrosis or kissing lesions were detected in 4 cases. The persistent good cell quality had no negative influence on the postoperative results.

CONCLUSION

The results after ACT are influenced by the adjacent cartilage shoulder, kissing lesions, previous meniscectomy and osteoarthrosis. This may limit the indications of the method.

Evaluation of patellar chondromalacia with MR: comparison between T2-weighted FSE SPIR and GE MTC

PURPOSE

To compare two different MR sequences to tissue signal suppression in the study of patellar cartilage abnormalities.

MATERIALS AND METHODS

We examined 26 patients with magnetic resonance (MR) imaging: sequences included spectral presaturation with inversion recovery (SPIR), with fat suppression and T2-weighted images, magnetization transfer contrast (MTC) sequences, T1-weighted and T2-weighted spin-echo sequences. All patients underwent conventional knee arthroscopy and in all patients a hyaline cartilage lesion was assessed in three articular zones: the patellar medial facet, the lateral facet and the patellar crista. Was assessed 78 articular facets. The lesions were classified using a standard arthroscopic grading system adapted to MR imaging: normal cartilage that corresponds to the grade 0 according to the Noyes grading system, low grade lesions that correspond to the grade I e IIa and high grade lesions that correspond to grades IIb and III. The arthroscopic results were compared with MR images. We assessed the MR diagnostic accuracy, sensitivity, specificity and MR positive predictive value and negative predictive value of the two sequences taking into consideration total lesions, and high-grade and low grade lesions separately.

RESULTS

Twenty-four low grade lesions (16 grade I e 8 grade IIa) and 18 high grade lesions (10 grade IIb e 8 grade III) were diagnosed by arthroscopy. Regarding low grade and high-grade lesions together, the accuracy was 77% for MTC sequences and 90% for SPIR sequences. In identifying low-grade lesions, the sensitivity was 88% for SPIR sequence and 42% for MTC sequences. Specificity for the detection of all lesions was 89% for the SPIR sequences and 94% for the MTC sequences. The SPIR sequence visualised water content abnormalities in degenerating cartilage, which are representative of low-grade lesions. The sensitivity of the sequence enabled us to obtain improved contrast for detecting cartilage surface irregularities. The MTC sequences allowed us to grade high-grade lesions susceptible to surgery and small cartilage defects in the presence of joint fluid. The MTC sequences were insufficient in the diagnosis of early stages of chondromalacia because the suppression of the signal of bonded water reduced the contrast among areas of articular cartilage with different water content. For this reason cartilage oedema and early superficial fibrillation were not identified.

CONCLUSIONS

In our experience the SPIR sequence proved superior to the MTC sequence in the identification of low grade lesions of the patellar cartilage. The overall value of such sequences in the study of articular pathology also needs to be assessed in the others sites where the articular cartilage is thinner and surfaces more curvilinear.