The Experts below are selected from a list of 62229 Experts worldwide ranked by ideXlab platform
David T Felson - One of the best experts on this subject based on the ideXlab platform.
-
assessment of Knee Pain from mr imaging using a convolutional siamese network
European Radiology, 2020Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Using 10-fold cross-validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans. • Our article is the first to leverage a deep learning framework to associate MR images of the Knee with Knee Pain. • We developed a convolutional Siamese network that had the ability to fuse information from multiple two-dimensional (2D) MRI slices from the Knee with Pain and the contralateral Knee of the same individual without Pain to predict unilateral Knee Pain. • Our model achieved an area under curve (AUC) value of 0.808. When individuals who had WOMAC Pain scores that were not discordant for Knees (Pain discordance < 3) were excluded, model performance increased to 0.853.
-
pairwise learning of mri scans using a convolutional siamese network for prediction of Knee Pain
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
assessment of Knee Pain from mri scans using a convolutional siamese network
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize Pain in Knee joints with risk of osteoarthritis solely by radiographic findings. We sought to understand if advanced machine learning methods such as deep neural networks can be used to predict and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing their Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions that were most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by a radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
assessment of bilateral Knee Pain from mr imaging using deep neural networks
bioRxiv, 2019Co-Authors: Gary H Chang, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Background and objective It remains difficult to characterize Pain in Knee joints with osteoarthritis solely by radiographic findings. We sought to understand how advanced machine learning methods such as deep neural networks can be used to analyze raw MRI scans and predict bilateral Knee Pain, independent of other risk factors. Methods We developed a deep learning framework to associate information from MRI slices taken from the left and right Knees of subjects from the Osteoarthritis Initiative with bilateral Knee Pain. Model training was performed by first extracting features from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices. The extracted features from all the 2D slices were subsequently combined to directly associate using a fused deep neural network with the output of interest as a binary classification problem. Results The deep learning model resulted in predicting bilateral Knee Pain on test data with 70.1% mean accuracy, 51.3% mean sensitivity, and 81.6% mean specificity. Systematic analysis of the predictions on the test data revealed that the model performance was consistent across subjects of different Kellgren-Lawrence grades. Conclusion The study demonstrates a proof of principle that a machine learning approach can be applied to associate MR images with bilateral Knee Pain. SIGNIFICANCE AND INNOVATION Knee Pain is typically considered as an early indicator of osteoarthritis (OA) risk. Emerging evidence suggests that MRI changes are linked to pre-clinical OA, thus underscoring the need for building image-based models to predict Knee Pain. We leveraged a state-of-the-art machine learning approach to associate raw MR images with bilateral Knee Pain, independent of other risk factors.
-
predicting bilateral Knee Pain from mr imaging using deep neural networks
bioRxiv, 2018Co-Authors: Gary H Chang, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:Background and objective: It remains difficult to characterize Pain in Knee joints with osteoarthritis solely by radiographic findings. We sought to understand how advanced machine learning methods such as deep neural networks can be used to analyze MRI scans and predict bilateral Knee Pain. Methods: We developed a deep learning framework to associate information from MRI slices taken from the left and right Knees of subjects from the Osteoarthritis Initiative with bilateral Knee Pain. Model training was performed by first extracting features from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices. The extracted features from the 2D slices were subsequently combined to directly associate using a fused deep neural network with the output of interest as a binary classification problem. Results: The deep learning model resulted in predicting bilateral Knee Pain on test data with 70.1% mean accuracy, 51.3% mean sensitivity, and 81.6% mean specificity. Systematic analysis of the predictions on the test data revealed that the model performance was consistent across subjects of various Kellgren-Lawrence grades. Conclusion: The study demonstrates a proof of principle that a machine learning approach can be applied to associate MR images with bilateral Knee Pain.
Michael C Nevitt - One of the best experts on this subject based on the ideXlab platform.
-
appendicular lean mass grip strength and the development of Knee osteoarthritis and Knee Pain among older adults
ACR Open Rheumatology, 2021Co-Authors: James S Andrews, Michael C Nevitt, Laura S Gold, Patrick J Heagerty, Peggy M CawthonAbstract:Author(s): Andrews, James S; Gold, Laura S; Nevitt, Michael; Heagerty, Patrick J; Cawthon, Peggy M | Abstract: ObjectiveThe association of sarcopenia with development of Knee osteoarthritis (OA) or Knee Pain in older adults is uncertain. We examined the relationship of grip strength and appendicular lean mass (ALM) with the likelihood of developing Knee OA and Knee Pain in older adults in the Health ABC (Health, Aging, and Body Composition) Study.MethodsALM and grip strength were assessed at baseline by dual-energy x-ray absorptiometry and handheld dynamometry, respectively. Incident clinically diagnosed, symptomatic Knee OA, defined as new participant report of physician-diagnosed Knee OA and concurrent frequent Knee Pain, and incident frequent Knee Pain over 5 years of follow-up were examined. Separate regression analyses, stratified by sex, modeled associations of baseline ALM and grip strength with the likelihood of incident clinically diagnosed, symptomatic Knee OA and incident Knee Pain over follow-up, adjusting for covariates.ResultsAmong the 2779 subjects without OA at baseline, 95 men (6.9%) and 158 women (11.3%) developed clinically diagnosed, symptomatic Knee OA, and, among the 2182 subjects without Knee Pain at baseline, 315 men (28.3%) and 385 women (36.1%) developed Knee Pain over follow-up. Among men only, each SD decrement of ALM was associated with decreasing likelihood of incident Knee OA (odds ratio [OR] per SD decrement: 0.68; 95% confidence interval [CI]: 0.47-0.97), and each SD decrement of grip strength was associated with increasing likelihood of incident Knee Pain (OR per SD decrement: 1.20; 95% CI: 1.01-1.42).ConclusionIn older men, ALM and grip strength may be associated with the development of Knee OA and Knee Pain, respectively.
-
association of intermittent and constant Knee Pain patterns with Knee Pain severity and with radiographic Knee osteoarthritis duration and severity
Arthritis Care and Research, 2021Co-Authors: Lisa C Carlesso, Michael C Nevitt, Gillian A Hawker, James C Torner, Cora E Lewis, Tuhina NeogiAbstract:OBJECTIVE To examine the relation of Knee Pain patterns to Pain severity and to radiographic osteoarthritis (OA) severity and duration. METHODS The Multicenter Osteoarthritis Study is a longitudinal cohort of older adults with or at risk of Knee OA. Participants' Intermittent and Constant Osteoarthritis Pain (ICOAP) scores were characterized as 1) no intermittent or constant Pain, 2) intermittent Pain only, 3) constant Pain only, and 4) a combination of constant and intermittent Pain. Knee Pain severity was assessed using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) Pain subscale and a visual analog scale (VAS). Radiographic Knee OA (ROA) severity was defined as Kellgren/Lawrence grade ≥2, and ROA duration was defined according to the clinic visit at which ROA was first noted. We assessed the relation of ICOAP Pain patterns to Knee Pain severity, ROA severity, and ROA duration using regression models with generalized estimating equations. RESULTS There were 2,322 participants (mean age 68.8 years, body mass index 31.0 kg/m2 , 60% female). Higher ICOAP Pain patterns, i.e., a mix of constant and intermittent Pain, were associated with greater WOMAC Pain severity compared with those patients without either Pain pattern (odds ratio [OR] 43.2 [95% confidence interval (95% CI) 26.4-61.3]). Results were similar for the VAS (OR 71.2 [95% CI 45.7-110.9]). Those patients with more severe and longer duration of ROA were more likely to have a mix of constant and intermittent Pain compared with those without either Pain (OR 3.7 [95% CI 3.1-4.6] and OR 2.9 [95% CI 2.5-3.5], respectively). CONCLUSION Knee Pain patterns are associated with radiographic disease stage and duration, as well as Pain severity, highlighting the fact that Pain patterns are important for understanding symptomatic disease progression.
-
relation of patellofemoral joint alignment morphology and radiographic osteoarthritis to frequent anterior Knee Pain data from the multicenter osteoarthritis study
Arthritis Care and Research, 2020Co-Authors: Erin M Macri, Michael C Nevitt, James C Torner, Cora E Lewis, Tuhina Neogi, I Tolstykh, Rafael Widjajahakim, Michael Roux, Joshua J StefanikAbstract:Objective. Patellofemoral (PF) alignment and trochlear morphology are associated with PF osteoarthritis (OA) and Knee Pain, but whether they are associated with localized anterior Knee Pain is unknown, which is believed to be a symptom specific to PF joint pathology. We therefore aimed to evaluate the relation of PF alignment and morphology, as well as PFOA and tibiofemoral OA, to anterior Knee Pain. Methods. The Multicenter Osteoarthritis Study is a cohort study of individuals with, or at risk for, Knee OA. We evaluated cross-sectional associations of PF alignment, trochlear morphology, and PF and tibiofemoral radiographic OA, with localized anterior Knee Pain (defined with a Pain map). We used 2 approaches: a within-person Kneematched evaluation of participants with unilateral anterior Knee Pain (conditional logistic regression), and a cohort approach comparing those with anterior Knee Pain to those without (binomial regression). Results. With the within-person Knee-matched approach (n = 110; 64% women, mean age 70 years, body mass index [BMI] 30.9), PF alignment, morphology, and tibiofemoral OA were not associated with unilateral anterior Knee Pain. Radiographic PFOA was associated with Pain, odds ratio 5.3 (95% confidence interval [95% CI] 1.6–18.3). Using the cohort approach (n = 1,818; 7% of Knees with anterior Knee Pain, 59% women, mean age 68 years, BMI 30.4), results were similar: only PFOA was associated with Pain, with a prevalence ratio of 2.2 (95% CI 1.4–3.4). Conclusion. PF alignment and trochlear morphology were not associated with anterior Knee Pain in individuals with, or at risk for, Knee OA. Radiographic PFOA, however, was associated with Pain, suggesting that features of OA, more so than mechanical features, may contribute to localized symptoms.
-
the relation of mri detected structural damage in the medial and lateral patellofemoral joint to Knee Pain the multicenter and framingham osteoarthritis studies
Osteoarthritis and Cartilage, 2015Co-Authors: Joshua J Stefanik, Ali Guermazi, David T Felson, Cora E Lewis, Frank W Roemer, K D Gross, Yuqing Zhang, Neil A Segal, Michael C NevittAbstract:Summary Objective To examine the relation of cartilage loss and bone marrow lesions (BMLs) in the medial and lateral patellofemoral joint (PFJ) to Knee Pain. Methods We categorized the location of full-thickness cartilage loss and BMLs in the PFJ on Knee magnetic resonance imaging (MRIs) from the Multicenter Osteoarthritis (MOST) and Framingham Osteoarthritis (FOA) Studies as no damage, isolated medial, isolated lateral, or both medial and lateral (mixed). We determined the relation of MRI lesions in each PFJ region to prevalent Knee Pain. Differences in Knee Pain severity were compared among categories of PFJ full-thickness cartilage loss and BMLs using quantile regression. Results In MOST (n = 1137 Knees), compared with Knees without full-thickness cartilage loss, Knees with isolated lateral or mixed PFJ full-thickness cartilage loss had 1.9 (1.3, 2.8) and 1.9 (1.2, 2.9) times the odds of Knee Pain, respectively, while isolated medial cartilage loss had no association with Knee Pain. BMLs in both the medial and lateral PFJ had 1.5 (1.1, 2.0) times the odds of Knee Pain compared with Knees without BMLs. Knee Pain severity was lowest in Knees with isolated medial PFJ cartilage loss or BMLs. In FOA (n = 934 Knees), neither isolated medial nor lateral cartilage loss was associated with Knee Pain, whereas isolated BMLs in either region were associated with Pain. Conclusions Results were not completely concordant but suggest that Knee Pain risk and severity is greatest with cartilage loss isolated to (MOST) or inclusive of (MOST and FOA) the lateral PFJ. While BMLs in either the medial or lateral PFJ are related to Pain.
-
relation of synovitis to Knee Pain using contrast enhanced mris
Annals of the Rheumatic Diseases, 2010Co-Authors: Kristin Baker, Ali Guermazi, Jingbo Niu, M D Crema, Andrew J Grainger, Margaret Clancy, Laura B Hughes, Joseph A Buckwalter, A Wooley, Michael C NevittAbstract:Background It has been suggested that synovitis causes joint Pain. On non-contrast-enhanced MRIs synovial thickening cannot be assessed and on these images synovitis has been inconsistently associated with Pain. Objective To assess synovial thickening in relation to Knee Pain severity among subjects in the Multicenter Osteoarthritis Study (MOST) using contrast-enhanced (CE) MRI. Methods MOST is a cohort study of people who have, or are at high risk of, Knee osteoarthritis (OA). An unselected subset of 535 participants who volunteered underwent CE 1.5 T MRI of one Knee. Synovitis was scored in six compartments and a summary score was created. Knee Pain severity was assessed using the maximum item score on the Western Ontario and McMaster Osteoarthritis Index (WOMAC) Pain scale. The association between synovitis and Pain severity was examined using a logistic regression model adjusting for age, sex, body mass index (BMI), MRI bone marrow lesions and effusions in the whole sample and in a subgroup without radiographic OA. Results 454 of the 535 subjects undergoing CE MRI had complete data on synovitis and WOMAC Pain. Mean age was 59 years, mean BMI 30 and 48% were women. In Knees with moderate Pain, 80% had synovitis. For Knee Pain, synovitis conferred a 9.2-fold increased odds compared with those without synovitis. In Knees without radiographic OA (n=329), there was also an association of synovitis with an increased prevalence of Pain. Conclusion Synovitis has a strong relation with Knee Pain severity, an association detected more clearly with CE MRI than suggested by previous studies using non-CE MRI measures of synovitis.
Ali Guermazi - One of the best experts on this subject based on the ideXlab platform.
-
assessment of Knee Pain from mr imaging using a convolutional siamese network
European Radiology, 2020Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Using 10-fold cross-validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans. • Our article is the first to leverage a deep learning framework to associate MR images of the Knee with Knee Pain. • We developed a convolutional Siamese network that had the ability to fuse information from multiple two-dimensional (2D) MRI slices from the Knee with Pain and the contralateral Knee of the same individual without Pain to predict unilateral Knee Pain. • Our model achieved an area under curve (AUC) value of 0.808. When individuals who had WOMAC Pain scores that were not discordant for Knees (Pain discordance < 3) were excluded, model performance increased to 0.853.
-
pairwise learning of mri scans using a convolutional siamese network for prediction of Knee Pain
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
assessment of Knee Pain from mri scans using a convolutional siamese network
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize Pain in Knee joints with risk of osteoarthritis solely by radiographic findings. We sought to understand if advanced machine learning methods such as deep neural networks can be used to predict and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing their Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions that were most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by a radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
the relation of mri detected structural damage in the medial and lateral patellofemoral joint to Knee Pain the multicenter and framingham osteoarthritis studies
Osteoarthritis and Cartilage, 2015Co-Authors: Joshua J Stefanik, Ali Guermazi, David T Felson, Cora E Lewis, Frank W Roemer, K D Gross, Yuqing Zhang, Neil A Segal, Michael C NevittAbstract:Summary Objective To examine the relation of cartilage loss and bone marrow lesions (BMLs) in the medial and lateral patellofemoral joint (PFJ) to Knee Pain. Methods We categorized the location of full-thickness cartilage loss and BMLs in the PFJ on Knee magnetic resonance imaging (MRIs) from the Multicenter Osteoarthritis (MOST) and Framingham Osteoarthritis (FOA) Studies as no damage, isolated medial, isolated lateral, or both medial and lateral (mixed). We determined the relation of MRI lesions in each PFJ region to prevalent Knee Pain. Differences in Knee Pain severity were compared among categories of PFJ full-thickness cartilage loss and BMLs using quantile regression. Results In MOST (n = 1137 Knees), compared with Knees without full-thickness cartilage loss, Knees with isolated lateral or mixed PFJ full-thickness cartilage loss had 1.9 (1.3, 2.8) and 1.9 (1.2, 2.9) times the odds of Knee Pain, respectively, while isolated medial cartilage loss had no association with Knee Pain. BMLs in both the medial and lateral PFJ had 1.5 (1.1, 2.0) times the odds of Knee Pain compared with Knees without BMLs. Knee Pain severity was lowest in Knees with isolated medial PFJ cartilage loss or BMLs. In FOA (n = 934 Knees), neither isolated medial nor lateral cartilage loss was associated with Knee Pain, whereas isolated BMLs in either region were associated with Pain. Conclusions Results were not completely concordant but suggest that Knee Pain risk and severity is greatest with cartilage loss isolated to (MOST) or inclusive of (MOST and FOA) the lateral PFJ. While BMLs in either the medial or lateral PFJ are related to Pain.
-
association between bone marrow lesions detected by magnetic resonance imaging and Knee Pain in community residents in korea
Osteoarthritis and Cartilage, 2013Co-Authors: Inje Kim, Ali Guermazi, M D Crema, Donghyun Kim, J Y Jung, Yeong Wook Song, David J Hunter, Hyun Ah KimAbstract:Summary Objective To describe the frequency of bone marrow lesions (BMLs) detected by magnetic resonance imaging (MRI), and to examine the association of BMLs with Knee Pain severity in community residents in Korea. Methods Participants were randomly chosen from the population-based Hallym Aging Study, irrespective of whether they had Knee osteoarthritis (OA) or Pain. Demographic and Knee Pain data were obtained by questionnaire. Radiographic evaluations consisted of weight-bearing Knee anteroposterior radiographs and 1.5-T MRI scans. MRI was performed in the dominant Knees of subjects without Knee Pain and in the more symptomatic Knees of subjects with Knee Pain. BMLs were graded according to the whole-organ MRI score. Results The mean age of the 358 study subjects was 71.8 years, and 34.5% of subjects had radiographically detected Knee OA. The prevalences of BMLs and large BMLs in the tibiofemoral compartments were 80.3% and 40.4%, respectively. After adjusting for age, sex, and body mass index, total and medial compartment BML scores were significantly associated with the presence of Knee Pain, and the association was stronger as the summary score for BML increased. In proportional regression analysis, Knee Pain severity increased with BML severity in any compartment and in the medial compartment. Conclusion BMLs detected by MRI were highly prevalent in this elderly Asian population. BMLs were significantly linked to Knee Pain, and BML severity correlated with Knee Pain severity. BMLs may be important surrogate targets for monitoring Pain and structure modification in OA therapeutics.
G Isik - One of the best experts on this subject based on the ideXlab platform.
-
bipolar versus unipolar intraarticular pulsed radiofrequency thermocoagulation in chronic Knee Pain treatment a prospective randomized trial
Pain Physician, 2017Co-Authors: Ersel Gulec, Hayri Ozbek, Sinan Pektas, G IsikAbstract:Background Chronic Knee Pain is a major widespread problem causing significant impairment of daily function. Pulsed radiofrequency has been shown to reduce severe chronic joint Pain as a non-pharmacological and less invasive treatment method. Objective We aimed to compare the effectiveness of unipolar and bipolar intraarticular pulsed radiofrequency methods in chronic Knee Pain control. Study design Prospective, randomized, double-blind study. Setting Pain clinic in Cukurova University Faculty of Medicine. Methods One hundred patients, aged 20 - 70 years with grade 2 or 3 Knee osteoarthritis were included in this study. Patients were randomly allocated into 2 groups to receive either unipolar (group U, n = 50) or bipolar (group B, n = 50) intraarticular pulsed radiofrequency (IAPRF) with a 45 V voltage, 2 Hz frequency, 42° C temperature, 10 msec pulse width, and 10 minute duration. We recorded visual analog scale (VAS) and Western Ontario and McMaster Universities Osteoarthritis Index LK 3.1WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index LK 3.1) scores of patients at baseline and one, 4, and 12 weeks after the procedure. The primary outcome was the percentage of patients with ≥ 50% reduction in Knee Pain at 12 weeks after the procedure. Results There was a significant difference between the groups according to VAS scores at all post-intervention time points. In group B, 84% of patients, and in the group U, 50% of patients achieved at least 50% Knee Pain relief from the baseline to 3 months. In group B, WOMAC scores were significantly lower than the group U at one and 3 months. Limitations Lack of long-term clinical results and supportive laboratory tests. Conclusion Bipolar IAPRF is more advantageous in reducing chronic Knee Pain and functional recovery compared with unipolar IAPRF. Further studies with longer follow-up times, laboratory-based tests, and different generator settings are required to establish the clinical importance and well-defined mechanism of action of PRF. This study protocol was registered at clinicaltrials.gov (identifier: NCT02141529), on May 15, 2014. Institutional Review Board (IRB) approval date: January 16, 2014, and number: 26/9Key words: Chronic Pain, intraarticular, Knee joint, Knee osteoarthritis, Pain management, pulsed radiofrequency treatment, quality of life, recovery of function.
Vijaya B Kolachalama - One of the best experts on this subject based on the ideXlab platform.
-
assessment of Knee Pain from mr imaging using a convolutional siamese network
European Radiology, 2020Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Using 10-fold cross-validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans. • Our article is the first to leverage a deep learning framework to associate MR images of the Knee with Knee Pain. • We developed a convolutional Siamese network that had the ability to fuse information from multiple two-dimensional (2D) MRI slices from the Knee with Pain and the contralateral Knee of the same individual without Pain to predict unilateral Knee Pain. • Our model achieved an area under curve (AUC) value of 0.808. When individuals who had WOMAC Pain scores that were not discordant for Knees (Pain discordance < 3) were excluded, model performance increased to 0.853.
-
pairwise learning of mri scans using a convolutional siamese network for prediction of Knee Pain
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize the source of Pain in Knee joints either using radiographs or magnetic resonance imaging (MRI). We sought to determine if advanced machine learning methods such as deep neural networks could distinguish Knees with Pain from those without it and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing the Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by an expert radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
assessment of Knee Pain from mri scans using a convolutional siamese network
bioRxiv, 2019Co-Authors: Gary H Chang, Ali Guermazi, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Objectives It remains difficult to characterize Pain in Knee joints with risk of osteoarthritis solely by radiographic findings. We sought to understand if advanced machine learning methods such as deep neural networks can be used to predict and identify the structural features that are associated with Knee Pain. Methods We constructed a convolutional Siamese network to associate MRI scans obtained on subjects from the Osteoarthritis Initiative (OAI) with frequent unilateral Knee Pain (n=1,529) comparing their Knee with frequent Pain to the contralateral Knee without Pain. The Siamese network architecture enabled pairwise learning of information from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices obtained from similar locations on both Knees. Class activation mapping (CAM) was utilized to create saliency maps, which highlighted the regions that were most associated with Knee Pain. The MRI scans and the CAMs of each subject were reviewed by a radiologist to identify the presence of abnormalities within the model-predicted regions of high association. Results Using 10-fold cross validation, our model achieved an area under curve (AUC) value of 0.808. When individuals whose Knee WOMAC Pain scores were not discordant were excluded, model performance increased to 0.853. The radiologist review revealed that about 86% of the cases that were predicted correctly had effusion-synovitis within the regions that were most associated with Pain. Conclusions This study demonstrates a proof of principle that deep learning can be applied to assess Knee Pain from MRI scans.
-
assessment of bilateral Knee Pain from mr imaging using deep neural networks
bioRxiv, 2019Co-Authors: Gary H Chang, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:ABSTRACT Background and objective It remains difficult to characterize Pain in Knee joints with osteoarthritis solely by radiographic findings. We sought to understand how advanced machine learning methods such as deep neural networks can be used to analyze raw MRI scans and predict bilateral Knee Pain, independent of other risk factors. Methods We developed a deep learning framework to associate information from MRI slices taken from the left and right Knees of subjects from the Osteoarthritis Initiative with bilateral Knee Pain. Model training was performed by first extracting features from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices. The extracted features from all the 2D slices were subsequently combined to directly associate using a fused deep neural network with the output of interest as a binary classification problem. Results The deep learning model resulted in predicting bilateral Knee Pain on test data with 70.1% mean accuracy, 51.3% mean sensitivity, and 81.6% mean specificity. Systematic analysis of the predictions on the test data revealed that the model performance was consistent across subjects of different Kellgren-Lawrence grades. Conclusion The study demonstrates a proof of principle that a machine learning approach can be applied to associate MR images with bilateral Knee Pain. SIGNIFICANCE AND INNOVATION Knee Pain is typically considered as an early indicator of osteoarthritis (OA) risk. Emerging evidence suggests that MRI changes are linked to pre-clinical OA, thus underscoring the need for building image-based models to predict Knee Pain. We leveraged a state-of-the-art machine learning approach to associate raw MR images with bilateral Knee Pain, independent of other risk factors.
-
predicting bilateral Knee Pain from mr imaging using deep neural networks
bioRxiv, 2018Co-Authors: Gary H Chang, David T Felson, Shangran Qiu, Terence D Capellini, Vijaya B KolachalamaAbstract:Background and objective: It remains difficult to characterize Pain in Knee joints with osteoarthritis solely by radiographic findings. We sought to understand how advanced machine learning methods such as deep neural networks can be used to analyze MRI scans and predict bilateral Knee Pain. Methods: We developed a deep learning framework to associate information from MRI slices taken from the left and right Knees of subjects from the Osteoarthritis Initiative with bilateral Knee Pain. Model training was performed by first extracting features from two-dimensional (2D) sagittal intermediate-weighted turbo spin echo slices. The extracted features from the 2D slices were subsequently combined to directly associate using a fused deep neural network with the output of interest as a binary classification problem. Results: The deep learning model resulted in predicting bilateral Knee Pain on test data with 70.1% mean accuracy, 51.3% mean sensitivity, and 81.6% mean specificity. Systematic analysis of the predictions on the test data revealed that the model performance was consistent across subjects of various Kellgren-Lawrence grades. Conclusion: The study demonstrates a proof of principle that a machine learning approach can be applied to associate MR images with bilateral Knee Pain.
