Dataset Information

Deep Learning-Based Muscle Segmentation and Quantification of Full-Leg Plain Radiograph for Sarcopenia Screening in Patients Undergoing Total Knee Arthroplasty.

ABSTRACT: Sarcopenia, an age-related loss of skeletal muscle mass and function, is correlated with adverse outcomes after some surgeries. Here, we present a deep-learning-based model for automatic muscle segmentation and quantification of full-leg plain radiographs. We illustrated the potential of the model to predict sarcopenia in patients undergoing total knee arthroplasty (TKA). A U-Net-based deep learning model for automatic muscle segmentation was developed, trained and validated on the plain radiographs of 227 healthy volunteers. The radiographs of 403 patients scheduled for primary TKA were reviewed to test the developed model and explore its potential to predict sarcopenia. The proposed deep learning model achieved mean IoU values of 0.959 (95% CI 0.959-0.960) and 0.926 (95% CI 0.920-0.931) in the training set and test set, respectively. The fivefold AUC value of the sarcopenia classification model was 0.988 (95% CI 0.986-0.989). Of seven key predictors included in the model, the predicted muscle volume (PMV) was the most important of these features in the decision process. In the preoperative clinical setting, wherein laboratory tests and radiographic imaging are available, the proposed deep-learning-based model can be used to screen for sarcopenia in patients with knee osteoarthritis undergoing TKA with high sarcopenia screening performance.

SUBMITTER: Hwang D

PROVIDER: S-EPMC9267147 | biostudies-literature | 2022 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep Learning-Based Muscle Segmentation and Quantification of Full-Leg Plain Radiograph for Sarcopenia Screening in Patients Undergoing Total Knee Arthroplasty.

Hwang Doohyun D Ahn Sungho S Park Yong-Beom YB Kim Seong Hwan SH Han Hyuk-Soo HS Lee Myung Chul MC Ro Du Hyun DH

Journal of clinical medicine 20220622 13

Sarcopenia, an age-related loss of skeletal muscle mass and function, is correlated with adverse outcomes after some surgeries. Here, we present a deep-learning-based model for automatic muscle segmentation and quantification of full-leg plain radiographs. We illustrated the potential of the model to predict sarcopenia in patients undergoing total knee arthroplasty (TKA). A U-Net-based deep learning model for automatic muscle segmentation was developed, trained and validated on the plain radiogr ...[more]

PMID: 35806895

Similar Datasets

Project description:Arthrofibrosis is characterized by excessive extracellular matrix (ECM) deposition that results in restricted joint motion after total knee arthroplasty (TKA). Current surgical and pharmacologic treatment options are limited. Therefore, an in vitro model for joint myofibroblastogenesis is valuable to investigate the arthrofibrotic process and identify diagnostic biomarkers and treatment options. In this study, we obtained intraoperative posterior capsule (PC), quadriceps tendon (QT), and suprapatellar pouch (SP) tissue from knees of four patients undergoing primary TKA for osteoarthritis and characterized primary outgrowth cells from these tissues in the absence and presence of transforming growth factor beta 1 (TGFβ1), a pro-myofibroblastic cytokine. Light microscopy of knee outgrowth cells revealed spindle-shaped cells while immunofluorescence (IF) established staining for the fibroblast-specific antigen TE-7 and Vimentin, which are characteristics of fibroblasts. These fibroblasts differentiate readily into myofibroblasts as highlighted by enhanced alpha smooth muscle actin (ACTA2) mRNA and protein expression and increased collagen mRNA (i.e., collagen type 1 (COL1A1) and collagen type 3 (COL3A1)) expression and collagenous matrix deposition in the presence of TGFβ1. Of note, these studies also revealed that knee-derived fibroblasts are more sensitive to TGFβ1-mediated myofibroblastogenesis than adipose-derived mesenchymal stem cells. Importantly, while outgrowth fibroblasts isolated from four patients and three anatomical regions exhibit similar gene expression profiles, these knee fibroblasts form a unique gene expression cluster within the fibroblast niche as revealed by RNA-sequencing analysis. In conclusion, our study provides a fibroblast/myofibroblast model of outgrowth knee cells derived from patients undergoing primary TKA that can be employed to assess myofibroblast-related processes and test novel pharmacological strategies in vitro for arthrofibrosis.

Project description:ImportanceIn 2018, Medicare removed total knee arthroplasty from the list of inpatient-only procedures, resulting in a new pool of patients eligible for outpatient total knee arthroplasty. How this change was associated with the characteristics of patients undergoing outpatient knee arthroplasty at hospital-owned surgery centers (HOSCs) vs freestanding ambulatory surgery centers (FASCs) is unknown.ObjectivesTo describe the characteristics of patients undergoing outpatient, elective total and partial knee arthroplasty in 2017 and 2018 and to compare the cohorts receiving treatment at FASCs and HOSCs.Design, setting, and participantsThis observational retrospective cohort study included 5657 patients having elective, outpatient partial and total knee arthroplasty in the Florida and Wisconsin State Ambulatory Surgery Databases in 2017 and 2018. Prior admissions were identified in the State Inpatient Database. Statistical analysis was performed from March to June 2022.Main outcomes and measuresCharacteristics of patients undergoing surgery at a FASC vs a HOSC in 2017 and 2018 were compared.ResultsA total of 5657 patients (mean [SD] age, 64.2 [9.9] years; 2907 women [51.4%]) were included in the study. Outpatient knee arthroplasties increased from 1910 in 2017 to 3747 in 2018 and were associated with an increase in total knee arthroplasties (474 in 2017 vs 2065 in 2018). The influx of patients undergoing outpatient knee arthroplasty was associated with an amplification of differences between the patients treated at FASCs and the patients treated at HOSCs. Patients with private payer insurance seen at FASCs increased from 63.4% in 2017 (550 of 867) to 72.7% in 2018 (1272 of 1749) (P < .001), while the percentage of patients with private payer insurance seen at HOSCs increased, but to a lesser extent (41.6% [427 of 1027] in 2017 vs 46.4% [625 of 1346] in 2018; P < .001). In 2017, the percentages of White patients seen at FASCs and HOSCs were similar (85.0% [737 of 867] vs 88.2% [906 of 1027], respectively); in 2018, the percentage of White patients seen at FASCs had increased and was significantly different from the percentage of White patients seen at HOSCs (90.6% [1585 of 1749] vs 87.9% [1183 of 1346]; P = .01). Both types of facilities saw an increase from 2017 to 2018 in the percentage of patients from communities of low social vulnerability, but this increase was greater for FASCs (FASCs: 6.7% [58 of 867] in 2017 vs 33.9% [593 of 1749] in 2018; HOSCs: 7.6% [78 of 1027] in 2017 vs 21.2% [285 of 1346] in 2018). Finally, while FASCs and HOSCs had cared for a similar portion of patients with prior admissions in 2017 (7.8% [68 of 867] vs 9.4% [97 of 1027], respectively; P = .25), in 2018, FASCs cared for fewer patients with prior admissions than HOSCs (4.0% [70 of 1749] vs 8.1% [109 of 1346]; P < .001).ConclusionsThis study suggests that the increase in the number of patients undergoing outpatient knee arthroplasty in 2018 corresponded to FASCs treating a greater share of patients who were White, covered by private payer insurance, and healthier. These findings raise a concern that as more operations transition to the outpatient setting, variability in access to FASCs may increase, leaving hospital-owned centers to bear a greater share of the burden of caring for more vulnerable patients with more severe illness.

Project description:BackgroundIn patients with knee osteoarthritis (KOA) undergoing knee arthroplasty (KA), lower-limb motor function tests are commonly measured during peri-surgical rehabilitation. To clarify their sources of variation and determine reference intervals (RIs), a multicenter study was performed in Japan.MethodsWe enrolled 545 KOA patients (127 men; 418 women; mean age 74.2 years) who underwent KA and followed a normal recovery course. The surgical modes included total KA (TKA), minimally invasive TKA (MIS-TKA), and unicompartmental KA (UKA). Motor functions measured twice before and two weeks after surgery included timed up-and-go (TUG), maximum walking speed (MWS), extensor and flexor muscle strength (MS), and knee range of motion (ROM). Multiple regression analysis was performed to evaluate their sources of variation including sex, age, BMI, and surgical mode. Magnitude of between-subgroup differences was expressed as SD ratio (SDR) based on 3-level nested ANOVA. SDR≥0.4 was set as the threshold for requiring RIs specific for each subgroup.ResultsBefore surgery, age-related changes exceeding the threshold were observed for TUG and MWS. Between-sex difference was noted for extensor and flexor MS, but extension and flexion ROMs were not influenced by sex or age. After surgery, in addition to similar influences of sex and age on test results, surgical modes of UKA and MIS-TKA generally had a favorable influence on MWS, extensor MS, and flexion ROM. All motor function test results showed a variable degree of skewness in distribution, and thus RIs were basically derived by the parametric method after Gaussian transformation of test results.ConclusionsThis is the first study to determine RIs for knee motor functions specific to KOA patients after careful consideration of their sources of variation and distribution shapes. These RIs facilitate objective implementation of peri-surgical rehabilitation and allow detection of patients who deviate from the normal course of recovery.

Dataset Information

Deep Learning-Based Muscle Segmentation and Quantification of Full-Leg Plain Radiograph for Sarcopenia Screening in Patients Undergoing Total Knee Arthroplasty.

Publications

Deep Learning-Based Muscle Segmentation and Quantification of Full-Leg Plain Radiograph for Sarcopenia Screening in Patients Undergoing Total Knee Arthroplasty.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets