Summary Background Deep learning offers considerable promise for medical diagnostics. We aimed to evaluate the diagnostic accuracy of deep learning algorithms versus health-care professionals in classifying diseases using medical imaging. Methods In this systematic review and meta-analysis, we searched Ovid-MEDLINE, Embase, Science Citation Index, and Conference Proceedings Citation Index for studies published from Jan 1, 2012, to June 6, 2019. Studies comparing the diagnostic performance of deep learning models and health-care professionals based on medical imaging, for any disease, were included. We excluded studies that used medical waveform data graphics material or investigated the accuracy of image segmentation rather than disease classification. We extracted binary diagnostic accuracy data and constructed contingency tables to derive the outcomes of interest: sensitivity and specificity. Studies undertaking an out-of-sample external validation were included in a meta-analysis, using a unified hierarchical model. This study is registered with PROSPERO, CRD42018091176. Findings Our search identified 31 587 studies, of which 82 (describing 147 patient cohorts) were included. 69 studies provided enough data to construct contingency tables, enabling calculation of test accuracy, with sensitivity ranging from 9·7% to 100·0% (mean 79·1%, SD 0·2) and specificity ranging from 38·9% to 100·0% (mean 88·3%, SD 0·1). An out-of-sample external validation was done in 25 studies, of which 14 made the comparison between deep learning models and health-care professionals in the same sample. Comparison of the performance between health-care professionals in these 14 studies, when restricting the analysis to the contingency table for each study reporting the highest accuracy, found a pooled sensitivity of 87·0% (95% CI 83·0–90·2) for deep learning models and 86·4% (79·9–91·0) for health-care professionals, and a pooled specificity of 92·5% (95% CI 85·1–96·4) for deep learning models and 90·5% (80·6–95·7) for health-care professionals. Interpretation Our review found the diagnostic performance of deep learning models to be equivalent to that of health-care professionals. However, a major finding of the review is that few studies presented externally validated results or compared the performance of deep learning models and health-care professionals using the same sample. Additionally, poor reporting is prevalent in deep learning studies, which limits reliable interpretation of the reported diagnostic accuracy. New reporting standards that address specific challenges of deep learning could improve future studies, enabling greater confidence in the results of future evaluations of this promising technology. Funding None.

A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis

Supervised artificial intelligence (AI) methods for evaluation of medical images require a curation process for data to optimally train, validate, and test algorithms. The chief obstacles to develo...

https://pubs.rsna.org/doi/pdf/10.1148/radiol.2020192224

Preparing Medical Imaging Data for Machine Learning.

The rapid increase in telemedicine coupled with recent advances in diagnostic artificial intelligence (AI) create the imperative to consider the opportunities and risks of inserting AI-based support into new paradigms of care Here we build on recent achievements in the accuracy of image-based AI for skin cancer diagnosis to address the effects of varied representations of AI-based support across different levels of clinical expertise and multiple clinical workflows We find that good quality AI-based support of clinical decision-making improves diagnostic accuracy over that of either AI or physicians alone, and that the least experienced clinicians gain the most from AI-based support We further find that AI-based multiclass probabilities outperformed content-based image retrieval (CBIR) representations of AI in the mobile technology environment, and AI-based support had utility in simulations of second opinions and of telemedicine triage In addition to demonstrating the potential benefits associated with good quality AI in the hands of non-expert clinicians, we find that faulty AI can mislead the entire spectrum of clinicians, including experts Lastly, we show that insights derived from AI class-activation maps can inform improvements in human diagnosis Together, our approach and findings offer a framework for future studies across the spectrum of image-based diagnostics to improve human-computer collaboration in clinical practice

/pdf/human-computer-collaboration-for-skin-cancer-recognition-2ibavkytcu.pdf

Human–computer collaboration for skin cancer recognition

Machine learning models for medical image analysis often suffer from poor performance on important subsets of a population that are not identified during training or testing. For example, overall performance of a cancer detection model may be high, but the model may still consistently miss a rare but aggressive cancer subtype. We refer to this problem as hidden stratification, and observe that it results from incompletely describing the meaningful variation in a dataset. While hidden stratification can substantially reduce the clinical efficacy of machine learning models, its effects remain difficult to measure. In this work, we assess the utility of several possible techniques for measuring hidden stratification effects, and characterize these effects both via synthetic experiments on the CIFAR-100 benchmark dataset and on multiple real-world medical imaging datasets. Using these measurement techniques, we find evidence that hidden stratification can occur in unidentified imaging subsets with low prevalence, low label quality, subtle distinguishing features, or spurious correlates, and that it can result in relative performance differences of over 20% on clinically important subsets. Finally, we discuss the clinical implications of our findings, and suggest that evaluation of hidden stratification should be a critical component of any machine learning deployment in medical imaging.

Hidden stratification causes clinically meaningful failures in machine learning for medical imaging

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; adjustedP = .01), mean accuracy increased by 0.038 (95% CI, 0.014-0.062; adjustedP = .02), and mean interrater agreement (Fleiss κ) increased by 0.060, from 0.799 to 0.859 (adjustedP = .05). There was no statistically significant change in mean specificity (0.016; 95% CI, −0.010 to 0.041; adjustedP = .16) and time to diagnosis (5.71 seconds; 95% CI, 7.22-18.63 seconds; adjustedP = .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.

/pdf/deep-learning-assisted-diagnosis-of-cerebral-aneurysms-using-4shumfklf0.pdf

Deep Learning-Assisted Diagnosis of Cerebral Aneurysms Using the HeadXNet Model

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.

/pdf/deep-learning-assisted-diagnosis-for-knee-magnetic-resonance-4r90kif854.pdf

Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet.

Pd60-05 automated detection of prostate cancer on multiparametric mri using deep neural networks trained on spatial coordinates and pathology of biopsy cores

Nicholas Bien

Papers

Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet.

Pd60-05 automated detection of prostate cancer on multiparametric mri using deep neural networks trained on spatial coordinates and pathology of biopsy cores