1
|
Shimizu H, Enda K, Koyano H, Shimizu T, Shimodan S, Sato K, Ogawa T, Tanaka S, Iwasaki N, Takahashi D. Bimodal machine learning model for unstable hips in infants: integration of radiographic images with automatically-generated clinical measurements. Sci Rep 2024; 14:17826. [PMID: 39090235 PMCID: PMC11294347 DOI: 10.1038/s41598-024-68484-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/25/2023] [Accepted: 07/24/2024] [Indexed: 08/04/2024] Open
Abstract
Bimodal convolutional neural networks (CNNs) are frequently combined with patient information or several medical images to enhance the diagnostic performance. However, the technologies that integrate automatically generated clinical measurements within the images are scarce. Hence, we developed a bimodal model that produced automatic algorithm for clinical measurement (aaCM) from radiographic images and integrated the model with CNNs. In this multicenter research project, the diagnostic performance of the model was investigated with 813 radiographic hip images of infants at risk of developmental dysplasia of the hips (232 and 581 images of unstable and stable hips, respectively), with the ground truth defined by provocative examinations. The results indicated that the accuracy of aaCM was equal or higher than that of specialists, and the bimodal model showed better diagnostic performance than LightGBM, XGBoost, SVM, and single CNN models. aaCM can provide expert's knowledge in a high level, and our proposed bimodal model has better performance than the state-of-art models.
Collapse
Affiliation(s)
- Hirokazu Shimizu
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
- Department of Cancer Pathology, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Ken Enda
- Department of Cancer Pathology, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Hidenori Koyano
- Graduate School of Biomedical Science and Engineering, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Tomohiro Shimizu
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Shun Shimodan
- Department of Orthopaedic Surgery, Kushiro City General Hospital, Kushiro, Hokkaido, Japan
| | - Komei Sato
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Takuya Ogawa
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Shinya Tanaka
- Department of Cancer Pathology, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
- Institute for Chemical Reaction Design and Discovery (WPI-ICReDD), Hokkaido University, Sapporo, Hokkaido, Japan
| | - Norimasa Iwasaki
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan
| | - Daisuke Takahashi
- Department of Orthopaedic Surgery, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan.
| |
Collapse
|
2
|
Nair A, Yatsonsky D, Liu J. Comparison of outcomes of different Graf grades of developmental dysplasia of the hip in infants treated with Tubingen splint versus Pavlik harness - A systematic review. J Orthop 2024; 49:68-74. [PMID: 38075458 PMCID: PMC10701362 DOI: 10.1016/j.jor.2023.11.060] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 10/15/2023] [Revised: 11/19/2023] [Accepted: 11/21/2023] [Indexed: 02/29/2024] Open
Abstract
This systematic review was designed to compare the outcomes of the two braces against each other classified by the Graf method. The databases sources included PubMed, Embase, and Google Scholar. The keywords included "DDH Tubingen versus Pavlik" and Tubingen and Pavlik separately. Included papers provided specific data regarding success and failure rate, avascular necrosis (AVN), duration, and age of intervention. The excluded studies discussed surgeries, diagnosis and mechanism, and ones that weren't in English. Total of 20 papers were included, resulting in 1243 Tubingen and 420 Pavlik samples. It was seen that the Tubingen splint had a statistically significant greater success rate and lower failure rate for Graf 2, D, and 3 hips, while both braces were not very successful for Graf 4 at success rates less than 60 %. Tubingen also had a lower incidence of AVN. Both braces shared similar ages of intervention, duration, and time per day. Both braces are very comparable to each other, each having better success rates for lower Graf grades, which points to the importance of bracing earlier to improve the success rates. The Tubingen splint had a higher success rate, lower failure rate, and lower AVN rate compared to the Pavlik harness. This points to the Tubingen splint potentially being the preferred option for bracing in infants.
Collapse
Affiliation(s)
- Ajay Nair
- Department of Orthopedic Surgery, University of Toledo Medical Center, Toledo, OH, 43614, United States
| | - David Yatsonsky
- Department of Orthopedic Surgery, University of Toledo Medical Center, Toledo, OH, 43614, United States
| | - Jiayong Liu
- Department of Orthopedic Surgery, University of Toledo Medical Center, Toledo, OH, 43614, United States
| |
Collapse
|
3
|
Hockett C, Mayfield LM, Gill CS, Kim HKW, Sucato DJ, Podeszwa DA, Jo CH, Morris WZ. Does Screening Ultrasound Timing in Developmental Dysplasia of the Hip Need to be Adjusted for Moderate Preterm and Near-term Infants: A Prospective Study. J Pediatr Orthop 2024; 44:e25-e29. [PMID: 37773040 DOI: 10.1097/bpo.0000000000002540] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 09/30/2023]
Abstract
INTRODUCTION An initial screening ultrasound is essential for patients at higher risk of developmental dysplasia of the hip (DDH) due to breech presentation or a family history of DDH. The American Academy of Pediatrics recommends screening ultrasounds to be performed after 6 weeks of age to reduce the rate of false positives. However, there is limited evidence regarding whether these screening ultrasounds need to be adjusted for gestational age in prematurity. The purpose of this study was to evaluate the influence of moderate preterm and near-term births on screening hip ultrasounds for high-risk DDH populations. METHODS We identified all prospectively enrolled patients in a single-center database referred for screening hip ultrasound for DDH. We included those hips referred for risk factors of DDH, including breech presentation, family history of DDH, or hip click, and excluded those with known dysplasia or referral for hip instability. Each ultrasound was measured by a pediatric radiologist to determine the alpha angle and femoral head coverage. Patients were classified as "premature" if born at <37 weeks gestation or "full term" if born at ≥37 weeks gestation. All patients underwent screening hip ultrasound between 5 and 8 weeks of age. Sonographic markers of dysplasia and the incidences of abnormal ultrasound and Pavlik harness treatment were compared between cohorts. Significance was set at P <0.05. RESULTS A total of 244 hips in 122 patients were included, 58 hips in the premature cohort and 186 hips in the full-term cohort. The premature cohort had a significantly decreased gestational age compared with the full-term cohort (35.4 ± 1.1 vs 38.5 ± 1.1 wk, respectively, P < 0.001). However, there was no difference between premature and full-term cohorts in sex distribution (69% vs 75%, females, P = 0.39), unadjusted age at the time of ultrasound (6.6 ± 0.7 vs 6.8±0.7 wk, respectively, P = 0.07), or referral reason ( P = 0.14). On hip ultrasound, there was no difference between premature and full-term cohorts with respect to alpha angle (62.6 ± 3.3 vs 62.2 ± 5.3 degrees, P = 0.41), femoral head coverage (54.9 ± 6.3 vs 55.1 ± 10.6, P = 0.19), rate of abnormal ultrasound (18.3% vs 20.7%, respectively, P = 0.68), or the rate of Pavlik harness treatment (0% vs 5.3%, respectively, P = 0.12). DISCUSSION There was no significant difference in alpha angle or femoral head coverage between premature and full-term patients at 5 to 8 weeks of unadjusted age. This preliminary data suggests that screening ultrasounds can be performed without adjusting for prematurity. LEVEL OF EVIDENCE Level II, prognostic study.
Collapse
Affiliation(s)
- Claire Hockett
- Department of Orthopaedic Surgery, Scottish Rite for Children
| | | | - Corey S Gill
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| | - Harry K W Kim
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| | - Daniel J Sucato
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| | - David A Podeszwa
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| | - Chan-Hee Jo
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| | - William Z Morris
- Department of Orthopaedic Surgery, Scottish Rite for Children
- Department of Orthopaedic Surgery, University of Texas, Southwestern Medical Center, Dallas, TX
| |
Collapse
|
4
|
Ghanem I, Karam G, Ghanem D, Saliba I. Femoral nerve palsy during Pavlik harness treatment for developmental dysplasia of the hip is not an indication for harness abandonment. J Child Orthop 2023; 17:205-211. [PMID: 37288053 PMCID: PMC10242367 DOI: 10.1177/18632521231172986] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 12/23/2022] [Accepted: 04/13/2023] [Indexed: 06/09/2023] Open
Abstract
Objectives The aim of this study was to report the incidence of femoral nerve palsy in developmental dysplasia of the hip children treated with Pavlik harness, to identify any possible associated risk factors, and to evaluate its outcome without any specific strap release. Methods A retrospective chart review was conducted on all cases of femoral nerve palsy in a consecutive series of children who underwent Pavlik harness treatment for developmental dysplasia of the hip. In unilateral cases, the developmental dysplasia of the hip was compared to the contralateral side. All hips with femoral nerve palsy were compared to the remaining hips of the series and any possible risk factor for paralysis was recorded. Results In total, 53 cases of femoral nerve palsy of various severity were identified from a group of 473 children with 527 hips treated for developmental dysplasia of the hip at an average age of 3.9 months. However, 93% occurred during the first 2 weeks of treatment. Femoral nerve palsy was more common in older and larger children with the most severe Tonnis type, and a hip flexion angle in the harness above 90° (p < 0.03 for all). All of them resolved spontaneously before completion of treatment without any specific measures. We found no correlation between the presence of femoral nerve palsy or the time taken for spontaneous resolution and treatment failure using the harness. Conclusion Femoral nerve palsy is most observed with higher Tonnis types and high hip flexion angles in the harness, but its presence by itself is not predictive of treatment failure. It resolves spontaneously before completion of treatment and does not require any strap release or harness discontinuation. Level of evidence Level III.
Collapse
Affiliation(s)
- Ismat Ghanem
- Department of Orthopedic Surgery, Hôtel-Dieu de France Hospital, Saint Joseph University, Beirut, Lebanon
| | - Geoffrey Karam
- Department of Orthopedic Surgery, Hôtel-Dieu de France Hospital, Saint Joseph University, Beirut, Lebanon
| | - Diane Ghanem
- Faculty of Medicine, American University of Beirut, Beirut, Lebanon
| | - Ibrahim Saliba
- Department of Orthopedic Surgery, Hôtel-Dieu de France Hospital, Saint Joseph University, Beirut, Lebanon
| |
Collapse
|
5
|
Bakti K, Lankinen V, Helminen M, Välipakka J, Laivuori H, Hyvärinen A. Clinical and sonographic improvement of developmental dysplasia of the hip: analysis of 948 patients. J Orthop Surg Res 2022; 17:538. [PMID: 36510263 PMCID: PMC9743506 DOI: 10.1186/s13018-022-03432-7] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 11/07/2022] [Accepted: 12/04/2022] [Indexed: 12/14/2022] Open
Abstract
BACKGROUND Developmental dysplasia of the hip is a common condition, which varies in severity. Abduction treatment is widely used to correct the development of the hips, but mild forms of DDH can also recover spontaneously. The purpose of this study was to evaluate factors affecting the rate of improvement of developmental dysplasia of the hip, and evaluate any risk factors slowing the process. MATERIAL AND METHODS The study population consisted of patients diagnosed with DDH in Tampere University hospital in the years 1998-2018. Data were retrospectively collected, and associations between clinical variables and rate of improvement were analyzed. Alpha angles were assessed monthly, and associations between risk factors and improvement of alpha angles were studied. A total of 948 patients were included in the analysis. RESULTS More severe first status of the hips was associated with faster improvement in dynamic ultrasound compared to milder DDH in univariate design in first 3 months of age; in the multivariable design, Ortolani positivity was conversely associated with lower alpha angles in 1-month follow-up. Immediate abduction treatment was associated with faster recovery rate compared to delayed abduction or watchful waiting. Female sex and positive family history were associated with slower rate of improvement and lower alpha angles. In multivariable design, female sex, positive family history and treatment strategy remained statistically significant as initiation time of the treatment explained the first found association of clinical hip status and the recovery rate after 2 months of age. CONCLUSION Female sex and positive family history might be independent risk factors for slower recovery in DDH before 6 months of age. These children might need special attention in their follow-up plans and abduction treatment.
Collapse
Affiliation(s)
- Karim Bakti
- grid.502801.e0000 0001 2314 6254Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland
| | - Vilma Lankinen
- grid.502801.e0000 0001 2314 6254Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland ,grid.410552.70000 0004 0628 215XDepartment of Pediatric Surgery, Turku University Hospital, Savitehtaankatu 5, 20520 Turku, Finland
| | - Mika Helminen
- grid.502801.e0000 0001 2314 6254Faculty of Social Sciences, Health Sciences, Tampere University, Tampere, Finland ,grid.412330.70000 0004 0628 2985Tays Research Services, Tampere University Hospital, Tampere, Finland
| | | | - Hannele Laivuori
- grid.502801.e0000 0001 2314 6254Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland ,grid.502801.e0000 0001 2314 6254Center for Child, Adolescent, and Maternal Health Research, Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland ,grid.412330.70000 0004 0628 2985Department of Obstetrics and Gynecology, Tampere University Hospital, Tampere, Finland ,grid.7737.40000 0004 0410 2071Institute for Molecular Medicine Finland (FIMM), Helsinki Institute of Life Science, University of Helsinki, Helsinki, Finland ,grid.7737.40000 0004 0410 2071Medical and Clinical Genetics, University of Helsinki, Helsinki University Hospital, Helsinki, Finland
| | - Anna Hyvärinen
- grid.502801.e0000 0001 2314 6254Faculty of Medicine and Health Technology, Tampere University, Tampere, Finland ,grid.412330.70000 0004 0628 2985Tays Research Services, Tampere University Hospital, Tampere, Finland ,Department of Surgery, Mehiläinen Länsi-Pohja Oy, Kemi, Finland ,grid.412326.00000 0004 4685 4917Department of Otorhinolaryngology and Head and Neck Surgery, Oulu University Hospital, Oulu, Finland ,grid.10858.340000 0001 0941 4873Clinical Medicine Research Unit, Medical Research Center, University of Oulu, Oulu, Finland
| |
Collapse
|
6
|
Dwan K, Kirkham J, Paton RW, Morley E, Newton AW, Perry DC. Splinting for the non-operative management of developmental dysplasia of the hip (DDH) in children under six months of age. Cochrane Database Syst Rev 2022; 10:CD012717. [PMID: 36214650 PMCID: PMC9549867 DOI: 10.1002/14651858.cd012717.pub2] [Citation(s) in RCA: 6] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/05/2022]
Abstract
BACKGROUND Developmental dysplasia of the hip (DDH) describes the abnormal development of a hip in childhood, ranging from complete dislocation of the hip joint to subtle immaturity of a hip that is enlocated and stable within the socket. DDH occurs in around 10 per 1000 live births, though only one per 1000 are completely dislocated. There is variation in treatment pathways for DDH, which differs between hospitals and even between clinicians within the same hospital. The variation is related to the severity of dysplasia that is believed to require treatment, and the techniques used to treat dysplasia. OBJECTIVES To determine the effectiveness of splinting and the optimal treatment strategy for the non-operative management of DDH in babies under six months of age. SEARCH METHODS We searched CENTRAL, MEDLINE, Embase, seven other electronic databases, and two trials registers up to November 2021. We also checked reference lists, contacted study authors, and handsearched relevant meetings abstracts. SELECTION CRITERIA Randomised controlled trials (RCTs), including quasi-RCTs, as well as non-RCTs and cohort studies conducted after 1980 were included. Participants were babies with all severities of DDH who were under six months of age. Interventions included dynamic splints, static splints or double nappies (diapers), compared to no splinting or delayed splinting. DATA COLLECTION AND ANALYSIS Two review authors independently selected studies, extracted data and performed risk of bias and GRADE assessments. The primary outcomes were: measurement of acetabular index at years one, two and five, as determined by radiographs (angle): the need for operative intervention to achieve reduction and to address dysplasia; and complications. We also investigated other outcomes highlighted by parents as important, including the bond between parent and child and the ability of mothers to breastfeed. MAIN RESULTS We included six RCTs or quasi-RCTs (576 babies). These were supported by 16 non-RCTs (8237 babies). Five studies had non-commercial funding, three studies stated 'no funding' and 14 studies did not state funding source. The RCTs were generally at unclear risk of bias, although we judged three RCTs to be at high risk of bias for incomplete outcome data. The non-RCTs were of moderate and critical risk of bias. We did not undertake meta-analysis due to methodological and clinical differences between studies; instead, we have summarised the results narratively. Dynamic splinting versus delayed or no splinting Four RCTs and nine non-RCTs compared immediate dynamic splinting and delayed dynamic splinting or no splinting. Of the RCTs, two considered stable hips and one considered unstable (dislocatable) hips and one jointly considered unstable and stable hips. No studies considered only dislocated hips. Two RCTs (265 babies, very low-certainty evidence) reported acetabular index at one year amongst stable or dislocatable hips. Both studies found there may be no evidence of a difference in splinting stable hips at first diagnosis compared to a strategy of active surveillance: one reported a mean difference (MD) of 0.10 (95% confidence interval (CI) -0.74 to 0.94), and the other an MD of 0.20 (95% CI -1.65 to 2.05). Two RCTs of stable hips (181 babies, very low-certainty evidence) reported there may be no evidence of a difference between groups for acetabular index at two years: one study reported an MD of -1.90 (95% CI -4.76 to 0.96), and another study reported an MD of -0.10 (95% CI -1.93 to 1.73), but did not take into account hips from the same child. No study reported data at five years. Four RCTs (434 babies, very low-certainty evidence) reported the need for surgical intervention. Three studies reported that no surgical interventions occurred. In the remaining study, two babies in the dynamic splinting group developed instability and were subsequently treated surgically. This study did not explicitly state if this treatment was to achieve concentric reduction or address residual dysplasia. Three RCTs (390 babies, very low-certainty evidence) reported no complications (avascular necrosis and femoral nerve palsy). Dynamic splinting versus static splinting One RCT and five non-RCTs compared dynamic versus static splinting. The RCT (118 hips) reported no occurrences of avascular necrosis (very low-certainty evidence) and did not report radiological outcomes or need for operative intervention. One quasi-RCT compared double nappies versus delayed or no splinting but reported no outcomes of interest. Other comparisons No RCTs compared static splinting versus delayed or no splinting or staged weaning versus immediate removal. AUTHORS' CONCLUSIONS There is a paucity of RCT evidence for splinting for the non-operative management of DDH: we included only six RCTs with 576 babies. Moreover, there was considerable heterogeneity between the studies, precluding meta-analysis. We judged the RCT evidence for all primary outcomes as being of very low certainty, meaning we are very uncertain about the true effects. Results from individual studies provide limited evidence of intervention effects on different severities of DDH. Amongst stable dysplastic hips, there was no evidence to suggest that treatment at any stage expedited the development of the acetabulum. For dislocatable hips, a delay in treatment onset to six weeks does not appear to result in any evidence of a difference in the development of the acetabulum at one year or increased risk of surgery. However, delayed splinting may reduce the number of babies requiring treatment with a harness. No RCTs compared static splinting with delayed or no splinting, staged weaning versus immediate removal or double nappies versus delayed or no splinting. There were few operative interventions or complications amongst the RCTs and the non-randomised studies. There's no apparent signal to indicate a higher frequency of either outcome in either intervention group. Given the frequency of this disease, and the fact that many countries undertake mandatory DDH screening, there is a clear need to develop an evidence-based pathway for treatment. Particular uncertainties requiring future research are the effectiveness of splinting amongst stable dysplastic hips, the optimal timing for the onset of splinting, the optimal type of splint to use and the need for 'weaning of splints'. Only once a robust pathway for treatment is established, can we properly assess the cost-effectiveness of screening interventions for DDH.
Collapse
Affiliation(s)
- Kerry Dwan
- Editorial & Methods Department, Cochrane Central Executive, London, UK
| | - Jamie Kirkham
- Centre for Biostatistics, The University of Manchester, Manchester Academic Health Science Centre, Manchester, UK
| | - Robin W Paton
- East Lancashire Hospitals NHS Trust, Burnley, UK
- School of Medicine, University of Central Lancashire, Preston, Lancashire, UK
| | | | | | - Daniel C Perry
- Faculty of Health and Life Sciences, University of Liverpool, Liverpool, UK
- Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS), University of Oxford, Oxford, UK
- Department of Orthopaedic Surgery, Alder Hey Hospital, Liverpool, UK
| |
Collapse
|