RESUMO
Background Missed fractures are a common cause of diagnostic discrepancy between initial radiographic interpretation and the final read by board-certified radiologists. Purpose To assess the effect of assistance by artificial intelligence (AI) on diagnostic performances of physicians for fractures on radiographs. Materials and Methods This retrospective diagnostic study used the multi-reader, multi-case methodology based on an external multicenter data set of 480 examinations with at least 60 examinations per body region (foot and ankle, knee and leg, hip and pelvis, hand and wrist, elbow and arm, shoulder and clavicle, rib cage, and thoracolumbar spine) between July 2020 and January 2021. Fracture prevalence was set at 50%. The ground truth was determined by two musculoskeletal radiologists, with discrepancies solved by a third. Twenty-four readers (radiologists, orthopedists, emergency physicians, physician assistants, rheumatologists, family physicians) were presented the whole validation data set (n = 480), with and without AI assistance, with a 1-month minimum washout period. The primary analysis had to demonstrate superiority of sensitivity per patient and the noninferiority of specificity per patient at -3% margin with AI aid. Stand-alone AI performance was also assessed using receiver operating characteristic curves. Results A total of 480 patients were included (mean age, 59 years ± 16 [standard deviation]; 327 women). The sensitivity per patient was 10.4% higher (95% CI: 6.9, 13.9; P < .001 for superiority) with AI aid (4331 of 5760 readings, 75.2%) than without AI (3732 of 5760 readings, 64.8%). The specificity per patient with AI aid (5504 of 5760 readings, 95.6%) was noninferior to that without AI aid (5217 of 5760 readings, 90.6%), with a difference of +5.0% (95% CI: +2.0, +8.0; P = .001 for noninferiority). AI shortened the average reading time by 6.3 seconds per examination (95% CI: -12.5, -0.1; P = .046). The sensitivity by patient gain was significant in all regions (+8.0% to +16.2%; P < .05) but shoulder and clavicle and spine (+4.2% and +2.6%; P = .12 and .52). Conclusion AI assistance improved the sensitivity and may even improve the specificity of fracture detection by radiologists and nonradiologists, without lengthening reading time. Published under a CC BY 4.0 license. Online supplemental material is available for this article. See also the editorial by Link and Pedoia in this issue.
Assuntos
Inteligência Artificial , Erros de Diagnóstico/prevenção & controle , Fraturas Ósseas/diagnóstico por imagem , Melhoria de Qualidade , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Conjuntos de Dados como Assunto , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Estudos Retrospectivos , Sensibilidade e EspecificidadeRESUMO
BACKGROUND: As the number of conventional radiographic examinations in pediatric emergency departments increases, so, too, does the number of reading errors by radiologists. OBJECTIVE: The aim of this study is to investigate the ability of artificial intelligence (AI) to improve the detection of fractures by radiologists in children and young adults. MATERIALS AND METHODS: A cohort of 300 anonymized radiographs performed for the detection of appendicular fractures in patients ages 2 to 21 years was collected retrospectively. The ground truth for each examination was established after an independent review by two radiologists with expertise in musculoskeletal imaging. Discrepancies were resolved by consensus with a third radiologist. Half of the 300 examinations showed at least 1 fracture. Radiographs were read by three senior pediatric radiologists and five radiology residents in the usual manner and then read again immediately after with the help of AI. RESULTS: The mean sensitivity for all groups was 73.3% (110/150) without AI; it increased significantly by almost 10% (P<0.001) to 82.8% (125/150) with AI. For junior radiologists, it increased by 10.3% (P<0.001) and for senior radiologists by 8.2% (P=0.08). On average, there was no significant change in specificity (from 89.6% to 90.3% [+0.7%, P=0.28]); for junior radiologists, specificity increased from 86.2% to 87.6% (+1.4%, P=0.42) and for senior radiologists, it decreased from 95.1% to 94.9% (-0.2%, P=0.23). The stand-alone sensitivity and specificity of the AI were, respectively, 91% and 90%. CONCLUSION: With the help of AI, sensitivity increased by an average of 10% without significantly decreasing specificity in fracture detection in a predominantly pediatric population.
Assuntos
Inteligência Artificial , Fraturas Ósseas , Adolescente , Adulto , Criança , Pré-Escolar , Fraturas Ósseas/diagnóstico por imagem , Humanos , Radiografia , Radiologistas , Estudos Retrospectivos , Adulto JovemRESUMO
OBJECTIVE: We aimed to perform an external validation of an existing commercial AI software program (BoneView™) for the detection of acute appendicular fractures in pediatric patients. MATERIALS AND METHODS: In our retrospective study, anonymized radiographic exams of extremities, with or without fractures, from pediatric patients (aged 2-21) were included. Three hundred exams (150 with fractures and 150 without fractures) were included, comprising 60 exams per body part (hand/wrist, elbow/upper arm, shoulder/clavicle, foot/ankle, leg/knee). The Ground Truth was defined by experienced radiologists. A deep learning algorithm interpreted the radiographs for fracture detection, and its diagnostic performance was compared against the Ground Truth, and receiver operating characteristic analysis was done. Statistical analyses included sensitivity per patient (the proportion of patients for whom all fractures were identified) and sensitivity per fracture (the proportion of fractures identified by the AI among all fractures), specificity per patient, and false-positive rate per patient. RESULTS: There were 167 boys and 133 girls with a mean age of 10.8 years. For all fractures, sensitivity per patient (average [95% confidence interval]) was 91.3% [85.6, 95.3], specificity per patient was 90.0% [84.0,94.3], sensitivity per fracture was 92.5% [87.0, 96.2], and false-positive rate per patient in patients who had no fracture was 0.11. The patient-wise area under the curve was 0.93 for all fractures. AI diagnostic performance was consistently high across all anatomical locations and different types of fractures except for avulsion fractures (sensitivity per fracture 72.7% [39.0, 94.0]). CONCLUSION: The BoneView™ deep learning algorithm provides high overall diagnostic performance for appendicular fracture detection in pediatric patients.
Assuntos
Aprendizado Profundo , Fraturas Ósseas , Algoritmos , Criança , Feminino , Fraturas Ósseas/diagnóstico por imagem , Humanos , Masculino , Curva ROC , Estudos Retrospectivos , Sensibilidade e EspecificidadeRESUMO
Osteoarthritis (OA) of the hip and knee is among the most common joint disorders. Intra-articular corticosteroid (IACS) injections are frequently performed to treat OA and other joint-related pain syndromes; however, there is conflicting evidence on their potential benefit. There is a lack of prospective and large retrospective studies evaluating potential joint findings, including increased risk for accelerated OA progression or adverse joint events, after treatment with IACS injection. Four main adverse joint findings have been structurally observed in patients after IACS injections: accelerated OA progression, subchondral insufficiency fracture, complications of osteonecrosis, and rapid joint destruction, including bone loss. Physicians, including radiologists, should be familiar with imaging findings and patient characteristics that may help them identify potential joints at risk for such events. The purpose of this report is to review the existing literature, describe observed adverse joint events after IACS injections, and provide an outlook on how this may affect clinical practice. Additional research endeavors are urgently needed to better understand and identify risk factors prior to intervention and to detect adverse joint events after injection as early as possible to prevent or minimize complications.
Assuntos
Corticosteroides/uso terapêutico , Osteoartrite do Quadril/tratamento farmacológico , Osteoartrite do Joelho/tratamento farmacológico , Corticosteroides/administração & dosagem , Humanos , Injeções Intra-Articulares , Osteoartrite do Quadril/diagnóstico por imagem , Osteoartrite do Joelho/diagnóstico por imagem , Manejo da Dor , Medição da DorRESUMO
OBJECTIVE: The purpose of this article is to describe knee abnormalities and the occurrence of MRI-detected sports-related knee abnormalities by evaluating MRI examinations performed during the Rio de Janeiro Olympic Games held in August 2016. CONCLUSION: There were 11,274 athletes at the Rio 2016 Olympic Games, and 113 of them underwent at least one knee MRI in the Olympic Village. Cartilage abnormalities, followed by meniscal tears and ligament sprains, were the most frequent abnormalities.
Assuntos
Traumatismos em Atletas/diagnóstico por imagem , Traumatismos do Joelho/diagnóstico por imagem , Imageamento por Ressonância Magnética/métodos , Esportes/estatística & dados numéricos , Adulto , Traumatismos em Atletas/epidemiologia , Brasil/epidemiologia , Feminino , Humanos , Traumatismos do Joelho/epidemiologia , MasculinoRESUMO
BACKGROUND: To use Magnetic Resonance Imaging (MRI) to characterize the severity, location, prevalence, and demographics of shoulder injuries in athletes at the Rio de Janeiro 2016 Summer Olympic Games. METHODS: This was a retrospective analysis of all routine shoulder MRIs obtained from the Olympic Village Polyclinic during the Rio 2016 Summer Olympics. Imaging was performed on 1.5 T and 3 T MRI, and interpretation was centrally performed by a board-certified musculoskeletal radiologist. Images were assessed for tendon, muscle, bone, bursal, joint capsule, labral, and chondral abnormality. RESULTS: A total of 11,274 athletes participated in the Games, of which 55 (5%) were referred for a routine shoulder MRI. Fifty-three (96%) had at least two abnormal findings. Seven (13%) had evidence of an acute or chronic anterior shoulder dislocation. Forty-nine (89%) had a rotator cuff partial tear and / or tendinosis. Subacromial / subdeltoid bursitis was present in 29 (40%). Thirty (55%) had a tear of the superior labrum anterior posterior (SLAP). CONCLUSION: Our study demonstrated a high prevalence of both acute and chronic shoulder injuries in the Olympic athletes receiving shoulder MRI. The high rates of bursal, rotator cuff, and labral pathology found in these patients implies that some degree of glenohumeral instability and impingement is occurring, likely due to fatigue and overuse of the dynamic stabilizers. Future studies are needed to better evaluate sport-specific trends of injury.
Assuntos
Atletas , Traumatismos em Atletas/diagnóstico por imagem , Traumatismos em Atletas/epidemiologia , Imageamento por Ressonância Magnética , Lesões do Ombro/diagnóstico por imagem , Lesões do Ombro/epidemiologia , Adolescente , Adulto , Traumatismos em Atletas/fisiopatologia , Brasil , Humanos , Valor Preditivo dos Testes , Prevalência , Estudos Retrospectivos , Índice de Gravidade de Doença , Lesões do Ombro/fisiopatologia , Fatores de Tempo , Adulto JovemRESUMO
Nerve growth factor (a-NGF) inhibitors have been developed for pain treatment including symptomatic osteoarthritis (OA) and have proven analgesic efficacy and improvement in functional outcomes in patients with OA. However, despite initial promising data, a-NGF clinical trials focusing on OA treatment had been suspended in 2010. Reasons were based on concerns regarding accelerated OA progression but were resumed in 2015 including detailed safety mitigation based on imaging. In 2021, an FDA advisory committee voted against approving tanezumab (one of the a-NGF compounds being evaluated) and declared that the risk evaluation and mitigation strategy was not sufficient to mitigate potential safety risks. Future clinical trials evaluating the efficacy of a-NGF or comparable molecules will need to define strict eligibility criteria and will have to include strategies to monitor safety closely. While disease-modifying effects are not the focus of a-NGF treatments, imaging plays an important role to evaluate eligibility of potential participants and to monitor safety during the course of these studies. Aim is to identify subjects with on-going safety findings at the time of inclusion, define those potential participants that are at increased risk for accelerated OA progression and to withdraw subjects from on-going studies in a timely fashion that exhibit imaging-confirmed structural safety events such as rapid progressive OA. OA efficacy- and a-NGF studies apply imaging for different purposes. In OA efficacy trials image acquisition and evaluation aims at maximizing sensitivity in order to capture structural effects between treated and non-treated participants in longitudinal fashion. In contrast, the aim of imaging in a-NGF trials is to enable detection of structural tissue alterations that either increase the risk of a negative outcome (eligibility) or may result in termination of treatment (safety).
RESUMO
Objective: Describe the radiograph-based screening program and frequencies of ineligibility in 3 large, international, randomized, double-blind, phase 3 studies of subcutaneous tanezumab in patients with osteoarthritis (OA). Design: Standardized bilateral shoulder, hip, and knee screening radiographs were obtained by trained imaging technologists and centrally read by 1 of 5 musculoskeletal radiology experts trained using a program-specific imaging atlas. Inter-reader consistency was tracked with test cases blindly inserted into the reader queue. Readers attended quarterly calibration meetings. Protocol-specified radiographic exclusion criteria included rapidly progressive OA (RPOA) or risk factors for RPOA (including severe malalignment of the knee, subchondral insufficiency fracture, atrophic OA, and osteonecrosis). Patients reporting disproportionate pain to radiographic evidence of OA in the hip or knee (without other pathology) were ineligible under a nonradiographic exclusion criterion. Results: At >480 international sites, 23,079 patients entered screening and 13,797 were radiographically assessed. Across 6 sets of quarterly testing, pairwise central reader agreement on radiographic eligibility was 72-87% (kappa: 0.41-0.71) and on radiographic OA grading 77-84% (kappa: 0.68-0.75). Among the 5,773/13,797 (41.8%) patients who met exclusionary criteria, 27% had disproportionate pain to radiographic findings (~10% of knee/hip radiographs). RPOA or risk factors for RPOA were each identified in <5% of patients (usually 1 joint) and <3% of knee/hip/shoulders. Conclusions: The phase 3 tanezumab screening program demonstrated the utility of radiographs to screen patients entering NGF inhibitor trials. A high degree of reader concordance was achieved. RPOA and risk factors for RPOA were not commonly observed. NCT02697773, NCT02709486, NCT02528188.
RESUMO
BACKGROUND: Patellofemoral joint (PFJ) disease is a common ailment in elite athletes. Our aim is to report the frequency of superolateral Hoffa's fat pad (SHFP) edema, and PFJ cartilage damage and bone marrow lesions (BML), among Olympian athletes, and to study the association between measurements of trochlear morphology and vertical patellar position and a) PFJ cartilage damage or BML, and b) SHFP edema. METHODS: All knee MRI, performed in the Olympic Village and polyclinics, of participating athletes in the 2016 Olympic Games of Rio de Janeiro were included. MRI were scored for PFJ cartilage damage and BML, and SHFP edema. Trochlear morphology measurements included sulcus angle, trochlear angle, lateral trochlear inclination, and medial trochlear inclination. Insall-Salvati ratio was also assessed. RESULTS: One hundred twenty-one knee MRIs were included (62 female, 51.2 %). The highest frequencies of PFJ cartilage damage, combination of PFJ cartilage damage and BML, and SHFP edema were found among Beach Volleyball and Volleyball athletes. SHFP edema was more common among female compared to male Olympian athletes. We found no statistically significant associations between different measurements of trochlear morphology/vertical patellar position, and 1. SHFP edema, and 2. PFJ cartilage damage/BML. CONCLUSION: SHFP edema and the combination of PFJ cartilage damage and BML are highly frequent among Olympic athletes especially those competing in Beach volleyball and Volleyball. SHFP edema is more common among female athletes. Further studies are needed to determine whether PFJ cartilage damage has a stronger association to sports disciplines rather than trochlear morphology.
RESUMO
PURPOSE: To report the MRI patterns of knee cartilage damage and concomitant internal derangement in athletes participating at the Rio de Janeiro 2016 Olympic Games. METHODS: Knee MRIs obtained at the core imaging facility of the International Olympic Committee were blindly, retrospectively reviewed by a board-certified musculoskeletal radiologist for meniscal, ligamentous, and tendon abnormalities. Cartilage assessment was based on the modified Outerbridge criteria. RESULTS: Of 122 athletes who received a knee MRI, 64 (52.4 %) had cartilage damage. Cartilage damage was more prevalent in the patellofemoral compartment (52 athletes, 42.6 %), followed by lateral (23 athletes, 18.9 %) and medial tibiofemoral compartments (12 athletes, 9.8 %). Patellofemoral cartilage damage was most prevalent in beach-volleyball (100 %), followed by volleyball (8 athletes, 66.7 %) and weightlifting (7 athletes, 70 %). Patellofemoral cartilage damage was most prevalent with quadriceps (8 athletes, 72.7 %) and patellar tendinosis (11 athletes, 61.1 %). Medial and lateral tibiofemoral cartilage damage was significantly associated with medial (8 athletes, 29.6 %) and lateral meniscal tears (16 athletes, 55.2 %), respectively. There was a trend for the percentage of athletes with cartilage damage to increase with age. CONCLUSION: The majority of athletes at the 2016 Rio Summer Olympics who had a knee MRI showed cartilage damage. Patellofemoral compartment cartilage damage was most common and frequently observed in certain sports including volleyball, beach volleyball, and weightlifting. Overuse in these sports can contribute to patellofemoral cartilage damage and subsequent development of anterior knee pain. Cartilage damage was also observed with concomitant meniscal tears and older age.
RESUMO
BACKGROUND: The cause of mucoid degeneration (MD) of the anterior cruciate ligament (ACL), which is commonly observed on magnetic resonance imaging (MRI) of patients with knee pain, has yet to be elucidated. Despite the limited evidence on the relationship between ACL lesions (injury and MD) and tibial morphologic features (ie, posterior tibial slope), the potential association between the presence of ACL MD and medial and lateral tibial slope (MTS and LTS) has not been well-established. PURPOSE: To investigate whether MTS and LTS measurements are associated with the presence of ACL MD. STUDY DESIGN: Cross-sectional study; Level of evidence, 3. METHODS: Consecutive knee MRI examinations of patients referred by an orthopaedic surgeon for potential internal joint derangements were identified within a 4-year period. The presence of ACL MD and the MTS/LTS values were assessed by independent expert observers in consensus in a blinded fashion. From 413 consecutive knee MRI scans, a sample of 80 knees, including 32 knees with ACL MD (cases) and 48 knees with normal ACL (controls), were selected using propensity score matching method for age, sex, body mass index, and presence of severe medial tibiofemoral compartment cartilage damage. The association between ACL MD and MTS/LTS was evaluated using conditional regression models. RESULTS: Knees with ACL MD had higher values of LTS (mean ± SD, 7.18° ± 3.58°) in comparison with control knees (5.32° ± 3.35°). Conditional regression analysis revealed a significant association between LTS measurements (not MTS) and ACL MD; every 1° increase in LTS was associated with a 17% (95% CI, 1%-35%) higher probability of having ACL MD. CONCLUSION: Excessive LTS was associated with the presence of ACL MD, independent of participants' age, sex, BMI, and cartilage damage severity.
RESUMO
PURPOSE OF REVIEW: Injuries to the labrum, joint capsule (in particular the inferior glenohumeral ligament), cartilage, and glenoid periosteum are associated with anterior shoulder instability. The goal of this review is to provide common radiographic images and findings in patients with anterior shoulder instability. Furthermore, we will demonstrate the best methods for measuring anterior glenoid bone loss. RECENT FINDINGS: Magnetic resonance (MR) imaging is highly relied upon for evaluating anterior shoulder instability and can diagnose soft tissue injuries with high sensitivity. While 3D computed tomography (CT) scan has been considered the optimal tool for evaluating osseous defects, certain MR imaging sequences have been shown to have similar diagnostic accuracy. Repair of Bankart lesions is critical to stabilizing the shoulder, and in the recent years, there has been an increasing focus on imaging to accurately characterize and measure glenoid bone loss to properly indicate patients for either arthroscopic repair or anterior bony reconstruction. Furthermore, Hill-Sachs lesions are commonly seen with shoulder instability, and importance must be placed on measuring the size and depth of these lesions along with possible engagement, as these factors will dictate management. The labral-ligamentous complex and rotator cuff are primary stabilizers of the shoulder. With anterior shoulder instability, the labrum is frequently injured. MRI with an arthrogram or provocative maneuvers is the gold standard for diagnosis. Various imaging modalities and methods can be performed to identify and measure Bankart and Hill-Sachs lesions, which can then be used for surgical planning and treating shoulder instability.
RESUMO
OBJECTIVE: Nerve growth factor antibodies (NGF-ab) have shown promising analgesic efficacy. Aim was to describe reader training efforts and present reliability data focusing on radiographic eligibility in the tanezumab program. METHODS: A multi-step process was used for reader calibration and reliability testing. First, a reference standard set of cases was created and diagnostic performance was evaluated. A second exercise focused on agreement of ordinal assessment (Kellgren-Lawrence grading) of radiographic osteoarthritis. Subsequently, 11 readers were trained and read a test set of 100 cases focused on eligibility assessments. Additional reliability testing and calibration of five core readers assessing eligibility of 30 cases was performed 3 and 6 months after study start. RESULTS: Sensitivity for the reference standard readings ranged from 0.50 to 0.90 and specificity from 0.40 to 0.83. Overall agreement for Kellgren-Lawrence grading ranged from 71.4% to 82.9%. For the 11 reader exercise, in 76% of cases at least 8 of 11 readers agreed on eligibility status. For the reliability testing 3 months after study start, in 80.0% of cases at least 4 of 5 readers agreed on eligibility with a κ = 0.43 (95% CI: 0.32-0.54). For the reliability testing after 6 months, in 83.3% of cases at least 4 of 5 readers agreed on eligibility with a κ = 0.52 (95% CI: 0.41-0.63). CONCLUSIONS: After intense efforts spent in the development of an imaging program for an NGF-ab clinical program, the achieved reliability for eligibility assessment is substantial but not perfect. Ongoing efforts of calibration prior to including additional readers to the program and during study conduct between current readers will be needed to ensure agreement on potential adverse events and radiographic disease severity.