RESUMEN

Chest X-rays (CXRs) are the most commonly performed diagnostic examination to detect cardiopulmonary abnormalities. However, the presence of bony structures such as ribs and clavicles can obscure subtle abnormalities, resulting in diagnostic errors. This study aims to build a deep learning (DL)-based bone suppression model that identifies and removes these occluding bony structures in frontal CXRs to assist in reducing errors in radiological interpretation, including DL workflows, related to detecting manifestations consistent with tuberculosis (TB). Several bone suppression models with various deep architectures are trained and optimized using the proposed combined loss function and their performances are evaluated in a cross-institutional test setting using several metrics such as mean absolute error (MAE), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and multiscale structural similarity measure (MS-SSIM). The best-performing model (ResNet-BS) (PSNR = 34.0678; MS-SSIM = 0.9828) is used to suppress bones in the publicly available Shenzhen and Montgomery TB CXR collections. A VGG-16 model is pretrained on a large collection of publicly available CXRs. The CXR-pretrained model is then fine-tuned individually on the non-bone-suppressed and bone-suppressed CXRs of Shenzhen and Montgomery TB CXR collections to classify them as showing normal lungs or TB manifestations. The performances of these models are compared using several performance metrics such as accuracy, the area under the curve (AUC), sensitivity, specificity, precision, F-score, and Matthews correlation coefficient (MCC), analyzed for statistical significance, and their predictions are qualitatively interpreted through class-selective relevance maps (CRMs). It is observed that the models trained on bone-suppressed CXRs (Shenzhen: AUC = 0.9535 ± 0.0186; Montgomery: AUC = 0.9635 ± 0.0106) significantly outperformed (p < 0.05) the models trained on the non-bone-suppressed CXRs (Shenzhen: AUC = 0.8991 ± 0.0268; Montgomery: AUC = 0.8567 ± 0.0870).. Models trained on bone-suppressed CXRs improved detection of TB-consistent findings and resulted in compact clustering of the data points in the feature space signifying that bone suppression improved the model sensitivity toward TB classification.

RESUMEN

Deep learning (DL) has drawn tremendous attention for object localization and recognition in both natural and medical images. U-Net segmentation models have demonstrated superior performance compared to conventional hand-crafted feature-based methods. Medical image modality-specific DL models are better at transferring domain knowledge to a relevant target task than those pretrained on stock photography images. This character helps improve model adaptation, generalization, and class-specific region of interest (ROI) localization. In this study, we train chest X-ray (CXR) modality-specific U-Nets and other state-of-the-art U-Net models for semantic segmentation of tuberculosis (TB)-consistent findings. Automated segmentation of such manifestations could help radiologists reduce errors and supplement decision-making while improving patient care and productivity. Our approach uses the publicly available TBX11K CXR dataset with weak TB annotations, typically provided as bounding boxes, to train a set of U-Net models. Next, we improve the results by augmenting the training data with weak localization, postprocessed into an ROI mask, from a DL classifier trained to classify CXRs as showing normal lungs or suspected TB manifestations. Test data are individually derived from the TBX11K CXR training distribution and other cross-institutional collections, including the Shenzhen TB and Montgomery TB CXR datasets. We observe that our augmented training strategy helped the CXR modality-specific U-Net models achieve superior performance with test data derived from the TBX11K CXR training distribution and cross-institutional collections (p < 0.05). We believe that this is the first study to i) use CXR modality-specific U-Nets for semantic segmentation of TB-consistent ROIs and ii) evaluate the segmentation performance while augmenting the training data with weak TB-consistent localizations.

RESUMEN

Pulmonary hypertension (PH) is defined by a mean pulmonary artery pressure greater than 20 mm Hg and classified into five different groups sharing similar pathophysiologic mechanisms, hemodynamic characteristics, and therapeutic management. Radiologists play a key role in the multidisciplinary assessment and management of PH. A working group was formed from within the Fleischner Society based on expertise in the imaging and/or management of patients with PH, as well as experience with methodologies of systematic reviews. The working group identified key questions focusing on the utility of CT, MRI, and nuclear medicine in the evaluation of PH: (a) Is noninvasive imaging capable of identifying PH? (b) What is the role of imaging in establishing the cause of PH? (c) How does imaging determine the severity and complications of PH? (d) How should imaging be used to assess chronic thromboembolic PH before treatment? (e) Should imaging be performed after treatment of PH? This systematic review and position paper highlights the key role of imaging in the recognition, work-up, treatment planning, and follow-up of PH. This article is a simultaneous joint publication in Radiology and European Respiratory Journal. The articles are identical except for stylistic changes in keeping with each journal's style. Either version may be used in citing this article. © 2021 RSNA and the European Respiratory Society. Online supplemental material is available for this article.

RESUMEN

Pulmonary hypertension (PH) is defined by a mean pulmonary artery pressure greater than 20 mmHg and classified into five different groups sharing similar pathophysiologic mechanisms, haemodynamic characteristics, and therapeutic management. Radiologists play a key role in the multidisciplinary assessment and management of PH. A working group was formed from within the Fleischner Society based on expertise in the imaging and/or management of patients with PH, as well as experience with methodologies of systematic reviews. The working group identified key questions focusing on the utility of CT, MRI, and nuclear medicine in the evaluation of PH: a) Is noninvasive imaging capable of identifying PH? b) What is the role of imaging in establishing the cause of PH? c) How does imaging determine the severity and complications of PH? d) How should imaging be used to assess chronic thromboembolic PH before treatment? e) Should imaging be performed after treatment of PH? This systematic review and position paper highlights the key role of imaging in the recognition, work-up, treatment planning, and follow-up of PH.

Asunto(s)

RESUMEN

Data-driven deep learning (DL) methods using convolutional neural networks (CNNs) demonstrate promising performance in natural image computer vision tasks. However, their use in medical computer vision tasks faces several limitations, viz., (i) adapting to visual characteristics that are unlike natural images; (ii) modeling random noise during training due to stochastic optimization and backpropagation-based learning strategy; (iii) challenges in explaining DL black-box behavior to support clinical decision-making; and (iv) inter-reader variability in the ground truth (GT) annotations affecting learning and evaluation. This study proposes a systematic approach to address these limitations through application to the pandemic-caused need for Coronavirus disease 2019 (COVID-19) detection using chest X-rays (CXRs). Specifically, our contribution highlights significant benefits obtained through (i) pretraining specific to CXRs in transferring and fine-tuning the learned knowledge toward improving COVID-19 detection performance; (ii) using ensembles of the fine-tuned models to further improve performance over individual constituent models; (iii) performing statistical analyses at various learning stages for validating results; (iv) interpreting learned individual and ensemble model behavior through class-selective relevance mapping (CRM)-based region of interest (ROI) localization; and, (v) analyzing inter-reader variability and ensemble localization performance using Simultaneous Truth and Performance Level Estimation (STAPLE) methods. We find that ensemble approaches markedly improved classification and localization performance, and that inter-reader variability and performance level assessment helps guide algorithm design and parameter optimization. To the best of our knowledge, this is the first study to construct ensembles, perform ensemble-based disease ROI localization, and analyze inter-reader variability and algorithm performance for COVID-19 detection in CXRs.

Asunto(s)

RESUMEN

We demonstrate use of iteratively pruned deep learning model ensembles for detecting pulmonary manifestation of COVID-19 with chest X-rays. This disease is caused by the novel Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) virus, also known as the novel Coronavirus (2019-nCoV). A custom convolutional neural network and a selection of ImageNet pretrained models are trained and evaluated at patient-level on publicly available CXR collections to learn modality-specific feature representations. The learned knowledge is transferred and fine-tuned to improve performance and generalization in the related task of classifying CXRs as normal, showing bacterial pneumonia, or COVID-19-viral abnormalities. The best performing models are iteratively pruned to reduce complexity and improve memory efficiency. The predictions of the best-performing pruned models are combined through different ensemble strategies to improve classification performance. Empirical evaluations demonstrate that the weighted average of the best-performing pruned models significantly improves performance resulting in an accuracy of 99.01% and area under the curve of 0.9972 in detecting COVID-19 findings on CXRs. The combined use of modality-specific knowledge transfer, iterative model pruning, and ensemble learning resulted in improved predictions. We expect that this model can be quickly adopted for COVID-19 screening using chest radiographs.

Asunto(s)

Asunto(s)

Asunto(s)

RESUMEN

BACKGROUND: Updating is important to ensure clinical guideline (CG) recommendations remain valid. However, little research has been undertaken in this field. We assessed CGs produced by the National Institute for Health and Care Excellence (NICE) to identify and describe updated recommendations and to investigate potential factors associated with updating. Also, we evaluated the reporting and presentation of recommendation changes. METHODS: We performed a descriptive analysis of original and updated CGs and recommendations, and an assessment of presentation formats and methods for recording information. We conducted a case-control study, defining cases as original recommendations that were updated ('new-replaced' recommendations), and controls as original recommendations that were considered to remain valid ('not changed' recommendations). We performed a comparison of main characteristics between cases and controls, and we planned a multiple regression analysis to identify potential predictive factors for updating. RESULTS: We included nine updated CGs (1,306 recommendations) and their corresponding original versions (1,106 recommendations). Updated CGs included 812 (62%) recommendations 'not reviewed', 368 (28.1%) 'new' recommendations, 104 (7.9%) 'amended' recommendations, and 25 (1.9%) recommendations reviewed but unchanged. The presentation formats used to indicate the changes in recommendations varied widely across CGs. Changes in 'amended', 'deleted', and 'new-replaced' recommendations (n = 296) were reported infrequently, mostly in appendices. These changes were recorded in 167 (56.4%) recommendations; and were explained in 81 (27.4%) recommendations. We retrieved a total of 7.1% (n = 78) case recommendations ('new-replaced') and 2.4% (n = 27) control recommendations ('not changed') in original CGs. The updates were mainly from 'Fertility CG', about 'gynaecology, pregnancy and birth' topic, and 'treatment' or 'prevention' purposes. We did not perform the multiple regression analysis as originally planned due to the small sample of recommendations retrieved. CONCLUSION: Our study is the first to describe and assess updated CGs and recommendations from a national guideline program. Our results highlight the pressing need to standardise the reporting and presentation of updated recommendations and the research gap about the optimal way to present updates to guideline users. Furthermore, there is a need to investigate updating predictive factors.

Asunto(s)

RESUMEN

OBJECTIVES: This study aims to quantify the unique useful yield from the Cumulative Index to Nursing and Allied Health Literature (CINAHL) database to National Institute for Health and Clinical Excellence (NICE) clinical guidelines. A secondary objective is to investigate the relationship between this yield and different clinical question types. It is hypothesized that the unique useful yield from CINAHL is low, and this database can therefore be relegated to selective rather than routine searching. STUDY DESIGN AND SETTING: A retrospective sample of 15 NICE guidelines published between 2005 and 2009 was taken. Information on clinical review question type, number of references, and reference source was extracted. RESULTS: Only 0.33% (95% confidence interval: 0.01-0.64%) of references per guideline were unique to CINAHL. Nursing- or allied health (AH)-related questions were nearly three times as likely to have references unique to CINAHL as non-nursing- or AH-related questions (14.89% vs. 5.11%), and this relationship was found to be significant (P<0.05). No significant relationship was found between question type and unique CINAHL yield for drug-related questions. CONCLUSIONS: The very low proportion of references unique to CINAHL strongly suggests that this database can be safely relegated to selective rather than routine searching. Nursing- and AH-related questions would benefit from selective searching of CINAHL.

Asunto(s)

RESUMEN

BACKGROUND: Healthcare decision makers face challenges when using guidelines, including understanding the quality of the evidence or the values and preferences upon which recommendations are made, which are often not clear. METHODS: GRADE is a systematic approach towards assessing the quality of evidence and the strength of recommendations in healthcare. GRADE also gives advice on how to go from evidence to decisions. It has been developed to address the weaknesses of other grading systems and is now widely used internationally. The Developing and Evaluating Communication Strategies to Support Informed Decisions and Practice Based on Evidence (DECIDE) consortium (http://www.decide-collaboration.eu/), which includes members of the GRADE Working Group and other partners, will explore methods to ensure effective communication of evidence-based recommendations targeted at key stakeholders: healthcare professionals, policymakers, and managers, as well as patients and the general public. Surveys and interviews with guideline producers and other stakeholders will explore how presentation of the evidence could be improved to better meet their information needs. We will collect further stakeholder input from advisory groups, via consultations and user testing; this will be done across a wide range of healthcare systems in Europe, North America, and other countries. Targeted communication strategies will be developed, evaluated in randomized trials, refined, and assessed during the development of real guidelines. DISCUSSION: Results of the DECIDE project will improve the communication of evidence-based healthcare recommendations. Building on the work of the GRADE Working Group, DECIDE will develop and evaluate methods that address communication needs of guideline users. The project will produce strategies for communicating recommendations that have been rigorously evaluated in diverse settings, and it will support the transfer of research into practice in healthcare systems globally.

Asunto(s)

RESUMEN

OBJECTIVES: Grading of Recommendations Assessment, Development and Evaluation (GRADE) is a system for rating the confidence in estimates of effect and grading guideline recommendations. It promotes evaluation of the quality of the evidence for each outcome and an assessment of balance between desirable and undesirable outcomes leading to a judgment about the strength of the recommendation. In 2007, the National Institute for Health and Clinical Excellence began introducing GRADE across its clinical guideline program to enable separation of judgments about the evidence quality from judgments about the strength of the recommendation. STUDY DESIGN AND SETTING: We describe the process of implementing GRADE across guidelines. RESULTS: Use of GRADE has been positively received by both technical staff and guideline development group members. CONCLUSION: A shift in thinking about confidence in the evidence was required leading to a more structured and transparent approach to decision making. Practical problems were also encountered; these have largely been resolved, but some areas require further work, including the application of imprecision and presenting results from analyses considering more than two alternative interventions. The use of GRADE for nonrandomized and diagnostic accuracy studies needs to be refined.

Asunto(s)

Asunto(s)

Asunto(s)

Asunto(s)

RESUMEN

GRADE requires a clear specification of the relevant setting, population, intervention, and comparator. It also requires specification of all important outcomes--whether evidence from research studies is, or is not, available. For a particular management question, the population, intervention, and outcome should be sufficiently similar across studies that a similar magnitude of effect is plausible. Guideline developers should specify the relative importance of the outcomes before gathering the evidence and again when evidence summaries are complete. In considering the importance of a surrogate outcome, authors should rate the importance of the patient-important outcome for which the surrogate is a substitute and subsequently rate down the quality of evidence for indirectness of outcome.

Asunto(s)

Asunto(s)

Asunto(s)