RESUMO
BACKGROUND: Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model's predictions to clinical experts during interpretation. METHODS AND FINDINGS: Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson's chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts' specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts. CONCLUSIONS: Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.
Assuntos
Lesões do Ligamento Cruzado Anterior/diagnóstico por imagem , Aprendizado Profundo , Diagnóstico por Computador/métodos , Interpretação de Imagem Assistida por Computador/métodos , Joelho/diagnóstico por imagem , Imageamento por Ressonância Magnética/métodos , Lesões do Menisco Tibial/diagnóstico por imagem , Adulto , Automação , Bases de Dados Factuais , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Valor Preditivo dos Testes , Reprodutibilidade dos Testes , Estudos Retrospectivos , Adulto JovemRESUMO
The development of deep learning algorithms for complex tasks in digital medicine has relied on the availability of large labeled training datasets, usually containing hundreds of thousands of examples. The purpose of this study was to develop a 3D deep learning model, AppendiXNet, to detect appendicitis, one of the most common life-threatening abdominal emergencies, using a small training dataset of less than 500 training CT exams. We explored whether pretraining the model on a large collection of natural videos would improve the performance of the model over training the model from scratch. AppendiXNet was pretrained on a large collection of YouTube videos called Kinetics, consisting of approximately 500,000 video clips and annotated for one of 600 human action classes, and then fine-tuned on a small dataset of 438 CT scans annotated for appendicitis. We found that pretraining the 3D model on natural videos significantly improved the performance of the model from an AUC of 0.724 (95% CI 0.625, 0.823) to 0.810 (95% CI 0.725, 0.895). The application of deep learning to detect abnormalities on CT examinations using video pretraining could generalize effectively to other challenging cross-sectional medical imaging tasks when training data is limited.
Assuntos
Algoritmos , Apendicite/diagnóstico , Apendicite/metabolismo , Aprendizado Profundo , Adulto , Estudos Transversais , Feminino , Humanos , Masculino , Pessoa de Meia-IdadeRESUMO
Importance: Deep learning has the potential to augment clinician performance in medical imaging interpretation and reduce time to diagnosis through automated segmentation. Few studies to date have explored this topic. Objective: To develop and apply a neural network segmentation model (the HeadXNet model) capable of generating precise voxel-by-voxel predictions of intracranial aneurysms on head computed tomographic angiography (CTA) imaging to augment clinicians' intracranial aneurysm diagnostic performance. Design, Setting, and Participants: In this diagnostic study, a 3-dimensional convolutional neural network architecture was developed using a training set of 611 head CTA examinations to generate aneurysm segmentations. Segmentation outputs from this support model on a test set of 115 examinations were provided to clinicians. Between August 13, 2018, and October 4, 2018, 8 clinicians diagnosed the presence of aneurysm on the test set, both with and without model augmentation, in a crossover design using randomized order and a 14-day washout period. Head and neck examinations performed between January 3, 2003, and May 31, 2017, at a single academic medical center were used to train, validate, and test the model. Examinations positive for aneurysm had at least 1 clinically significant, nonruptured intracranial aneurysm. Examinations with hemorrhage, ruptured aneurysm, posttraumatic or infectious pseudoaneurysm, arteriovenous malformation, surgical clips, coils, catheters, or other surgical hardware were excluded. All other CTA examinations were considered controls. Main Outcomes and Measures: Sensitivity, specificity, accuracy, time, and interrater agreement were measured. Metrics for clinician performance with and without model augmentation were compared. Results: The data set contained 818 examinations from 662 unique patients with 328 CTA examinations (40.1%) containing at least 1 intracranial aneurysm and 490 examinations (59.9%) without intracranial aneurysms. The 8 clinicians reading the test set ranged in experience from 2 to 12 years. Augmenting clinicians with artificial intelligence-produced segmentation predictions resulted in clinicians achieving statistically significant improvements in sensitivity, accuracy, and interrater agreement when compared with no augmentation. The clinicians' mean sensitivity increased by 0.059 (95% CI, 0.028-0.091; adjusted P = .01), mean accuracy increased by 0.038 (95% CI, 0.014-0.062; adjusted P = .02), and mean interrater agreement (Fleiss κ) increased by 0.060, from 0.799 to 0.859 (adjusted P = .05). There was no statistically significant change in mean specificity (0.016; 95% CI, -0.010 to 0.041; adjusted P = .16) and time to diagnosis (5.71 seconds; 95% CI, 7.22-18.63 seconds; adjusted P = .19). Conclusions and Relevance: The deep learning model developed successfully detected clinically significant intracranial aneurysms on CTA. This suggests that integration of an artificial intelligence-assisted diagnostic model may augment clinician performance with dependable and accurate predictions and thereby optimize patient care.
Assuntos
Aprendizado Profundo , Aneurisma Intracraniano/diagnóstico , Competência Clínica/normas , Simulação por Computador , Estudos Cross-Over , Diagnóstico por Computador/métodos , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Exame Neurológico/métodos , Neurologistas/normas , Estudos RetrospectivosRESUMO
CONTEXT: Identifying patients in the out-of-hospital setting who have no realistic hope of surviving an out-of-hospital cardiac arrest could enhance utilization of scarce health care resources. OBJECTIVE: To validate 2 out-of-hospital termination-of-resuscitation rules developed by the Ontario Prehospital Life Support (OPALS) study group, one for use by responders providing basic life support (BLS) and the other for those providing advanced life support (ALS). DESIGN, SETTING, AND PATIENTS: Retrospective cohort study using surveillance data prospectively submitted by emergency medical systems and hospitals in 8 US cities to the Cardiac Arrest Registry to Enhance Survival (CARES) between October 1, 2005, and April 30, 2008. Case patients were 7235 adults with out-of-hospital cardiac arrest; of these, 5505 met inclusion criteria. MAIN OUTCOME MEASURES: Specificity and positive predictive value of each termination-of-resuscitation rule for identifying patients who likely will not survive to hospital discharge. RESULTS: The overall rate of survival to hospital discharge was 7.1% (n = 392). Of 2592 patients (47.1%) who met BLS criteria for termination of resuscitation efforts, only 5 (0.2%) patients survived to hospital discharge. Of 1192 patients (21.7%) who met ALS criteria, none survived to hospital discharge. The BLS rule had a specificity of 0.987 (95% confidence interval [CI], 0.970-0.996) and a positive predictive value of 0.998 (95% CI, 0.996-0.999) for predicting lack of survival. The ALS rule had a specificity of 1.000 (95% CI, 0.991-1.000) and positive predictive value of 1.000 (95% CI, 0.997-1.000) for predicting lack of survival. CONCLUSION: In this validation study, the BLS and ALS termination-of-resuscitation rules performed well in identifying patients with out-of-hospital cardiac arrest who have little or no chance of survival.
Assuntos
Reanimação Cardiopulmonar , Serviços Médicos de Emergência , Parada Cardíaca/mortalidade , Parada Cardíaca/terapia , Futilidade Médica , Suspensão de Tratamento , Adulto , Suporte Vital Cardíaco Avançado/normas , Idoso , Reanimação Cardiopulmonar/normas , Protocolos Clínicos , Serviços Médicos de Emergência/normas , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Sistema de Registros , Estudos Retrospectivos , Análise de Sobrevida , Estados Unidos , Suspensão de Tratamento/normasRESUMO
BACKGROUND: Increased recognition of ionizing radiation risks has placed an emphasis on the appropriate use of myocardial perfusion imaging (MPI). Hospitalists frequently order MPI in the evaluation of chest pain and are thus at the forefront of its inpatient utilization. METHODS: We collected baseline figures for a group MPI rate (March 2010-February 2011) as well as individual MPI rates for hospitalists caring for cardiac floor patients at a community teaching hospital. We performed a 2-part intervention; we presented the individual MPI rate data back to the hospitalist division and carried out longitudinal educational efforts on MPI appropriateness criteria. We then calculated the group MPI utilization rate for 3 postintervention periods (March 2011-February 2012, March 2012-February 2013, and March 2013-February 2014) and the MPI rate for the subgroup of cardiac floor patients. Finally, we calculated the percentage of inappropriately performed stress tests before and after our intervention. RESULTS: Group MPI rate declined from 6.1% to 5.0% in the first year after our intervention (P = 0.009); a decrease was maintained a year later-MPI rate 4.9% (P = 0.004)-and became even more pronounced 2 years later-MPI rate 3.9% (P < 0.0001). The MPI rate for the subgroup of patients on the cardiac floor similarly decreased from 8.0% to 6.7% (P = 0.039). Finally, we report a particularly encouraging and significant trend of a 46% postintervention decrease (from 16.5% to 9%, P = 0.034) in the proportion of inappropriate stress tests ordered. CONCLUSIONS: Analyzing individual ordering rates and combining them with educational efforts was an effective strategy for impacting MPI utilization in the hospitalist group studied.