Deep Learning Detection of Active Pulmonary Tuberculosis at Chest Radiography Matched the Clinical Performance of Radiologists.

Background The World Health Organization (WHO) recommends chest radiography to facilitate tuberculosis (TB) screening. However, chest radiograph interpretation expertise remains limited in many regions. Purpose To develop a deep learning system (DLS) to detect active pulmonary TB on chest radiographs and compare its performance to that of radiologists. Materials and Methods A DLS was trained and tested using retrospective chest radiographs (acquired between 1996 and 2020) from 10 countries. To improve generalization, large-scale chest radiograph pretraining, attention pooling, and semisupervised learning ("noisy-student") were incorporated. The DLS was evaluated in a four-country test set (China, India, the United States, and Zambia) and in a mining population in South Africa, with positive TB confirmed with microbiological tests or nucleic acid amplification testing (NAAT). The performance of the DLS was compared with that of 14 radiologists. The authors studied the efficacy of the DLS compared with that of nine radiologists using the Obuchowski-Rockette-Hillis procedure. Given WHO targets of 90% sensitivity and 70% specificity, the operating point of the DLS (0.45) was prespecified to favor sensitivity. Results A total of 165 754 images in 22 284 subjects (mean age, 45 years; 21% female) were used for model development and testing. In the four-country test set (1236 subjects, 17% with active TB), the receiver operating characteristic (ROC) curve of the DLS was higher than those for all nine India-based radiologists, with an area under the ROC curve of 0.89 (95% CI: 0.87, 0.91). Compared with these radiologists, at the prespecified operating point, the DLS sensitivity was higher (88% vs 75%, P < .001) and specificity was noninferior (79% vs 84%, P = .004). Trends were similar within other patient subgroups, in the South Africa data set, and across various TB-specific chest radiograph findings. In simulations, the use of the DLS to identify likely TB-positive chest radiographs for NAAT confirmation reduced the cost by 40%-80% per TB-positive patient detected. Conclusion A deep learning method was found to be noninferior to radiologists for the determination of active tuberculosis on digital chest radiographs. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by van Ginneken in this issue.

Synthetic gene circuits for cell state detection and protein tuning in human pluripotent stem cells.

During development, cell state transitions are coordinated through changes in the identity of molecular regulators in a cell type- and dose-specific manner. The ability to rationally engineer such transitions in human pluripotent stem cells (hPSC) will enable numerous applications in regenerative medicine. Herein, we report the generation of synthetic gene circuits that can detect a desired cell state using AND-like logic integration of endogenous miRNAs (classifiers) and, upon detection, produce fine-tuned levels of output proteins using an miRNA-mediated output fine-tuning technology (miSFITs). Specifically, we created an "hPSC ON" circuit using a model-guided miRNA selection and circuit optimization approach. The circuit demonstrates robust PSC-specific detection and graded output protein production. Next, we used an empirical approach to create an "hPSC-Off" circuit. This circuit was applied to regulate the secretion of endogenous BMP4 in a state-specific and fine-tuned manner to control the composition of differentiating hPSCs. Our work provides a platform for customized cell state-specific control of desired physiological factors in hPSC, laying the foundation for programming cell compositions in hPSC-derived tissues and beyond.

Generalized pustular psoriasis: Quality and readability of online health information.

Generalized pustular psoriasis (GPP) is a multisystem disease with potentially life-threatening adverse effects. As patients increasingly seek health information online, and as the landscape for GPP changes, the quality of online health information (OHI) becomes progressively more important. This paper is the first of its kind to examine the quality, comprehensiveness and readability of online health information for GPP. Similar to pre-existing studies evaluating OHI, this paper examines 5 key search terms for GPP- 3 medical and 2 laymen. For each search term, the results were evaluated based on HONcode accreditation, an enhanced DISCERN analysis and a number of readability indices. Of the 500 websites evaluated, 84 (16.8%) were HONcode-accredited. Mean DISCERN scores of all websites were 74.9% and 38.6% for website reliability and treatment sections, respectively, demonstrating key gaps in comprehensiveness and reliability of GPP-specific OHI. Additionally, only 4/100 websites (4%) analysed for readability were written at the NIH-recommended sixth-grade level. Academic websites were significantly more difficult to read than governmental websites. This further exacerbates the patient information gap, particularly for patients with low health literacy, who may already be at higher risk of not receiving timely medical care.

Simplified Transfer Learning for Chest Radiography Models Using Less Data.

Background Developing deep learning models for radiology requires large data sets and substantial computational resources. Data set size limitations can be further exacerbated by distribution shifts, such as rapid changes in patient populations and standard of care during the COVID-19 pandemic. A common partial mitigation is transfer learning by pretraining a "generic network" on a large nonmedical data set and then fine-tuning on a task-specific radiology data set. Purpose To reduce data set size requirements for chest radiography deep learning models by using an advanced machine learning approach (supervised contrastive [SupCon] learning) to generate chest radiography networks. Materials and Methods SupCon helped generate chest radiography networks from 821 544 chest radiographs from India and the United States. The chest radiography networks were used as a starting point for further machine learning model development for 10 prediction tasks (eg, airspace opacity, fracture, tuberculosis, and COVID-19 outcomes) by using five data sets comprising 684 955 chest radiographs from India, the United States, and China. Three model development setups were tested (linear classifier, nonlinear classifier, and fine-tuning the full network) with different data set sizes from eight to 85. Results Across a majority of tasks, compared with transfer learning from a nonmedical data set, SupCon reduced label requirements up to 688-fold and improved the area under the receiver operating characteristic curve (AUC) at matching data set sizes. At the extreme low-data regimen, training small nonlinear models by using only 45 chest radiographs yielded an AUC of 0.95 (noninferior to radiologist performance) in classifying microbiology-confirmed tuberculosis in external validation. At a more moderate data regimen, training small nonlinear models by using only 528 chest radiographs yielded an AUC of 0.75 in predicting severe COVID-19 outcomes. Conclusion Supervised contrastive learning enabled performance comparable to state-of-the-art deep learning models in multiple clinical tasks by using as few as 45 images and is a promising method for predictive modeling with use of small data sets and for predicting outcomes in shifting patient populations. © RSNA, 2022 Online supplemental material is available for this article.

Recurrent aphthous stomatitis: A comprehensive review and recommendations on therapeutic options.

Recurrent aphthous stomatitis (RAS) is a common chronic disease in the oral mucosa that affects about 20% of the population. It is characterized by solitary or multiple, recurrent, small ulcers with erythematous haloes and yellow/gray floors. RAS can be managed through a wide variety of preventative measures and therapies, intending to reduce ulcer pain, stimulate ulcer healing, and/or prevent ulcer recurrence. First-line treatment options include topical medications in the form of corticosteroids (triamcinolone acetonide), anti-inflammatory drugs (amlexanox), antibiotics (doxycycline), and antiseptics (lidocaine). In more severe cases of RAS where local treatment is insufficient, systemic drugs in the form of corticosteroids (prednisone), immunomodulatory drugs (thalidomide), and antibiotics/antimicrobials (clofazimine) can prove effective. This review will summarize current treatment options for RAS with discussion of prevention, topical measures, natural treatments, systemic therapies, and new potential therapies. Furthermore, this review will provide recommendations on therapeutic options for RAS based on disease severity and patient circumstances.

Electromagnetic Navigational Bronchoscopy versus CT-guided Percutaneous Sampling of Peripheral Indeterminate Pulmonary Nodules: A Cohort Study.

Purpose To compare the diagnostic yield and complication rates of electromagnetic navigational bronchoscopic (ENB)-guided and computed tomography (CT)-guided percutaneous tissue sampling of lung nodules. Materials and Methods Retrospectively identified were 149 patients sampled percutaneously with CT guidance and 146 patients who underwent ENB with transbronchial biopsy of a lung lesion between 2013 and 2015. Clinical data, incidence of complications, and nodule pathologic analyses were assessed through electronic medical record review. Lung nodule characteristics were reviewed through direct image analysis. Molecular marker studies and pathologic analyses from surgical excision were reviewed when available. Multiple-variable logistic regression models were built to compare the diagnostic yield and complication rates for each method and for different patient and disease characteristics. Results CT-guided sampling was more likely to be diagnostic than ENB-guided biopsy (86.0% [129 of 150] vs 66.0% [99 of 150], respectively), and this difference remained significant even after adjustments were made for patient and nodule characteristics (P < .001). Age, American Society of Anesthesiologists class, emphysema grade, nodule size, and distance from pleura were not significant predictors of increased diagnostic yield. Intraprocedural time for physicians was significantly lower with CT-guided sampling (P < .001). Similar yield for molecular analyses was noted with the two approaches (ENB-guided sampling, 88.9% [32 of 36]; CT-guided sampling, 82.0% [41 of 50]). The two groups had similar rates of major complications (symptomatic hemorrhage, P > .999; pneumothorax requiring chest tube and/or admission, P = .417). Conclusion CT-guided transthoracic biopsy provided higher diagnostic yield in the assessment of peripheral pulmonary nodules than navigational bronchoscopy with a similar rate of clinically relevant complications. © RSNA, 2017 Online supplemental material is available for this article.

Historical Evolution of Imaging Techniques for the Evaluation of Pulmonary Embolism.

As we celebrate the 100th anniversary of the founding of the Radiological Society of North America (RSNA), it seems fitting to look back at the major accomplishments of the radiology community in the diagnosis of pulmonary embolism. Few diseases have so consistently captured the attention of the medical community. Since the first description of pulmonary embolism by Virchow in the 1850s, clinicians have struggled to reach a timely diagnosis of this common condition because of its nonspecific and often confusing clinical picture. As imaging tests started to gain importance in the 1900s, the approach to diagnosing pulmonary embolism also began to change. Rapid improvements in angiography, ventilation-perfusion imaging, and cross-sectional imaging modalities such as computed tomography (CT) and magnetic resonance imaging have constantly forced health care professionals to rethink how they diagnose pulmonary embolism. Needless to say, the way pulmonary embolism is diagnosed today is distinctly different from how it was diagnosed in Virchow's era; and imaging, particularly CT, now forms the cornerstone of diagnostic evaluation. Currently, radiology offers a variety of tests that are fast and accurate and can provide anatomic and functional information, thus allowing early diagnosis and triage of cases. This review provides a historical journey into the evolution of these imaging tests and highlights some of the major breakthroughs achieved by the radiology community and RSNA in this process. Also highlighted are areas of ongoing research and development in this field of imaging as radiologists seek to combat some of the newer challenges faced by modern medicine, such as rising health care costs and radiation dose hazards.

Building a bridge to save a failing ventricle: radiologic evaluation of short- and long-term cardiac assist devices.

Heart failure is recognized with increasing frequency worldwide and often progresses to an advanced refractory state. Although the reference standard for treatment of advanced heart failure remains cardiac transplantation, the increasing shortage of donor organs and the unsuitability of many patients for transplantation surgery has led to a search for alternative therapies. One such therapy is mechanical circulatory support, which helps relieve the load on the ventricle and thereby allows it to recover function. In addition, there is increasing evidence supporting the use of mechanical devices as a bridge to recovery in patients with acute refractory heart failure. In this article, the imaging evaluation of various commonly used short- and long-term cardiac assist devices is discussed, and their relevant mechanisms of action and physiology are described. Imaging, particularly computed tomography (CT), plays a crucial role in preoperative evaluation for assessment of candidacy for implantation of a left ventricular assist device (LVAD) or total artificial heart (TAH). Also, echocardiography and CT are indispensable in assessment of complications associated with cardiac devices. Complications commonly associated with short-term assist devices include bleeding and malpositioning, whereas long-term devices such as LVADs may be associated with infection, pump thrombosis, and cannula malfunction, as well as bleeding. CT is also commonly performed for preoperative planning before LVAD or TAH explantation, replacement of a device or one of its components, and cardiac transplantation. Online supplemental material is available for this article.

International association for the study of lung cancer (IASLC) lymph node map: radiologic review with CT illustration.

Accurate clinical or pretreatment stage classification of lung cancer leads to optimal treatment outcomes and improved prognostication. Such classification requires an accurate assessment of the clinical extent of regional lymph node metastasis. Consistent and reproducible regional lymph node designations facilitate reliable assessment of the clinical extent of regional lymph node metastasis. Regional lymph node maps, such as the Naruke lymph node map and the Mountain-Dresler modification of the American Thoracic Society lymph node map, were proposed for this purpose in the past. The most recent regional lymph node map to be published is the International Association for the Study of Lung Cancer (IASLC) lymph node map. The IASLC lymph node map supersedes all previous maps and should be used in tandem with the current seventh edition of the tumor, node, metastasis stage classification for lung cancer.

Treatment of calcinosis cutis associated with autoimmune connective tissue diseases.

Calcinosis cutis is a condition that is commonly associated with autoimmune connective tissue diseases. It is characterized by the deposition of insoluble calcium salts in the skin and subcutaneous tissue, which can cause pain, impair function, and have significant impacts on quality of life. Calcinosis cutis is difficult to manage because there is no generally accepted treatment: evidence supporting treatments is mostly comprised of case reports and case series, sometimes yielding mixed findings. Both pharmacologic and procedural interventions have been proposed to improve calcinosis cutis, and each may be suited to different clinical scenarios. This review summarizes current treatment options for calcinosis cutis, with discussion of recommendations based on patient-specific factors and disease severity.

Using generative AI to investigate medical imagery models and datasets.

BACKGROUND: AI models have shown promise in performing many medical imaging tasks. However, our ability to explain what signals these models have learned is severely lacking. Explanations are needed in order to increase the trust of doctors in AI-based models, especially in domains where AI prediction capabilities surpass those of humans. Moreover, such explanations could enable novel scientific discovery by uncovering signals in the data that aren't yet known to experts. METHODS: In this paper, we present a workflow for generating hypotheses to understand which visual signals in images are correlated with a classification model's predictions for a given task. This approach leverages an automatic visual explanation algorithm followed by interdisciplinary expert review. We propose the following 4 steps: (i) Train a classifier to perform a given task to assess whether the imagery indeed contains signals relevant to the task; (ii) Train a StyleGAN-based image generator with an architecture that enables guidance by the classifier ("StylEx"); (iii) Automatically detect, extract, and visualize the top visual attributes that the classifier is sensitive towards. For visualization, we independently modify each of these attributes to generate counterfactual visualizations for a set of images (i.e., what the image would look like with the attribute increased or decreased); (iv) Formulate hypotheses for the underlying mechanisms, to stimulate future research. Specifically, present the discovered attributes and corresponding counterfactual visualizations to an interdisciplinary panel of experts so that hypotheses can account for social and structural determinants of health (e.g., whether the attributes correspond to known patho-physiological or socio-cultural phenomena, or could be novel discoveries). FINDINGS: To demonstrate the broad applicability of our approach, we present results on eight prediction tasks across three medical imaging modalities-retinal fundus photographs, external eye photographs, and chest radiographs. We showcase examples where many of the automatically-learned attributes clearly capture clinically known features (e.g., types of cataract, enlarged heart), and demonstrate automatically-learned confounders that arise from factors beyond physiological mechanisms (e.g., chest X-ray underexposure is correlated with the classifier predicting abnormality, and eye makeup is correlated with the classifier predicting low hemoglobin levels). We further show that our method reveals a number of physiologically plausible, previously-unknown attributes based on the literature (e.g., differences in the fundus associated with self-reported sex, which were previously unknown). INTERPRETATION: Our approach enables hypotheses generation via attribute visualizations and has the potential to enable researchers to better understand, improve their assessment, and extract new knowledge from AI-based models, as well as debug and design better datasets. Though not designed to infer causality, importantly, we highlight that attributes generated by our framework can capture phenomena beyond physiology or pathophysiology, reflecting the real world nature of healthcare delivery and socio-cultural factors, and hence interdisciplinary perspectives are critical in these investigations. Finally, we will release code to help researchers train their own StylEx models and analyze their predictive tasks of interest, and use the methodology presented in this paper for responsible interpretation of the revealed attributes. FUNDING: Google.