The 2021 SIIM-FISABIO-RSNA Machine Learning COVID-19 Challenge: Annotation and Standard Exam Classification of COVID-19 Chest Radiographs.

Lakhani, Paras; Mongan, John; Singhal, Chinmay; Zhou, Quan; Andriole, Katherine P.; Auffermann, William F.; Prasanna, Prasanth; Pham, Tessie; Peterson, Michael; Bergquist, Peter; Cook, Tessa S.; Ferraciolli, Suely Fazio; Corradi, Gustavo César de Antonio; Takahashi, Marcelo Straus; Workman, Spencer S; Parekh, Maansi; Kamel, Sarah; Galant, Joaquin Herrero; Mas-Sanchez, Alba; Benítez, Emi C.; Sánchez-Valverde, Mariola; Jaques, Lara; Panadero, María; Vidal, Marta; Culiáñez-Casas, María; Angulo-Gonzalez, Diego M.; Langer, Steve G.; Vayá, María de la Iglesia; Shih, George

doi:10.31219/osf.io/532ek

Cited by 17 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By including dark-field images, AUC values were significantly higher compared to attenuation images only (p = 3.9e-6 for dark-field alone, p = 3.5e-9 for combination). For comparison, we additionally applied the winning neural network of the SIIM-FISABIO-RSNA COVID-19 Detection Challenge 24 , trained on conventional attenuation images, to the attenuation images of both the COVID-19 patients and healthy controls. In this setting, an AUC value of 0.88 was achieved, which can also be found in literature 25 .…”

Section: Reader Study and Quantitative Analysismentioning

confidence: 99%

“…Values 1 to 3 were counted as negatives, while values 4 to 6 were counted as positives. Attenuation-based images were additionally evaluated by using the winning neural network of the SIIM-FISABIO-RSNA COVID-19 Detection Challenge 24 , which provides a probability for the presence of COVID-19-pneumonia for each patient. The quantitative dark-field coefficient was calculated according to Gassert et al 13 .…”

Section: Image Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Dark-field Chest X-ray Imaging for the Assessment of COVID‑19‑Pneumonia

Frank

Gassert

Urban

et al. 2022

Preprint

View full text Add to dashboard Cite

Medical imaging is performed in daily clinical routine for the assessment of pulmonary involvement in patients infected with COVID-19. As conventional (attenuation-based) chest radiography provides only a low sensitivity for COVID-19-pneumonia, CT is the gold standard for lung imaging in COVID-19 patients. However, CT imaging exposes the patient to a considerable amount of radiation, and is not as widely available as plain chest X-rays. Therefore, alternative low-dose imaging X-ray techniques are highly desirable. Here we present the first clinical results employing a novel dark-field chest X-ray imaging method for the assessment of COVID-19-pneumonia. The work is based on recent technological advancements in a human-scale X-ray dark-field chest imaging prototype that enable the acquisition of quantitative dark-field radiographs with diagnostic image quality at a radiation dose comparable to conventional X-rays. In a reader study, we found that dark-field imaging has a higher sensitivity for COVID-19-pneumonia than attenuation-based imaging, and that the combination of both is superior to one imaging modality alone. Furthermore, a quantitative image analysis showed a significant reduction of signal in X-ray dark-field chest radiographs of COVID-19 patients. While our results demonstrate that dark-field chest radiography presents an ultra-low-dose alternative to CT imaging for the assessment of COVID-19-pneumonia, we anticipate that the presented technique will also be useful for therapy follow-up of patients with long-COVID-syndrome and, more generally, for the imaging of other pulmonary pathologies.

show abstract

Section: Reader Study and Quantitative Analysismentioning

confidence: 99%

Section: Image Evaluationmentioning

confidence: 99%

Dark-field Chest X-ray Imaging for the Assessment of COVID‑19‑Pneumonia

Frank

Gassert

Urban

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The SIIM-FISABIO-RSNA COVID-19 Detection dataset was curated by an international group of 22 radiologists (Lakhani et al, 2021). It includes data from the Valencian Region Medical ImageBank (BIMCV) (de la Iglesia Vayá et al, 2020) and the Medical Imaging Data Resource Center (MIDRC) -RSNA International COVID-19 Open Radiology Database (RICORD) (Tsai et al, 2021).…”

Section: Datasetmentioning

confidence: 99%

Challenges in COVID-19 Chest X-Ray Classification: Problematic Data or Ineffective Approaches?

Muhammad¹,

Bawazir²,

Navarette³

et al. 2022

Preprint

View full text Add to dashboard Cite

The value of quick, accurate, and confident diagnoses cannot be undermined to mitigate the effects of COVID-19 infection, particularly for severe cases. Enormous effort has been put towards developing deep learning methods to classify and detect COVID-19 infections from chest radiography images. However, recently some questions have been raised surrounding the clinical viability and effectiveness of such methods. In this work, we carry out extensive experiments on a large COVID-19 chest X-ray dataset to investigate the challenges faced with creating reliable solutions from both the data and machine learning perspectives. Accordingly, we offer an in-depth discussion into the challenges faced by some widely-used deep learning architectures associated with chest X-Ray COVID-19 classification. Finally, we include some possible directions and considerations to improve the performance of the models and the data for use in clinical settings.

show abstract

“…To tackle the challenge, object localization by HINT can be used to see whether the identified responsible regions overlap with the lesion regions drawn by doctors. With the COVID19 dataset from SIIM-FISABIO-RSNA COVID-19 Detection competition [34], we trained models used by high-ranking teams and other baseline models for classification. The localization results of COVID19 cases with typical symptoms by Effi-cientNet [64] are shown in Figure S.1 (c).…”

Section: D3 Covid19 Classification Model Evaluationmentioning

confidence: 99%

HINT: Hierarchical Neuron Concept Explainer

Wang¹,

Lee²,

Qi³

2022

Preprint

View full text Add to dashboard Cite

To interpret deep networks, one main approach is to associate neurons with human-understandable concepts. However, existing methods often ignore the inherent relationships of different concepts (e.g., dog and cat both belong to animals), and thus lose the chance to explain neurons responsible for higher-level concepts (e.g., animal). In this paper, we study hierarchical concepts inspired by the hierarchical cognition process of human beings. To this end, we propose HIerarchical Neuron concepT explainer (HINT) to effectively build bidirectional associations between neurons and hierarchical concepts in a low-cost and scalable manner. HINT enables us to systematically and quantitatively study whether and how the implicit hierarchical relationships of concepts are embedded into neurons, such as identifying collaborative neurons responsible to one concept and multimodal neurons for different concepts, at different semantic levels from concrete concepts (e.g., dog) to more abstract ones (e.g., animal). Finally, we verify the faithfulness of the associations using Weakly Supervised Object Localization, and demonstrate its applicability in various tasks such as discovering saliency regions and explaining adversarial attacks. Code is available on https://github.com/AntonotnaWang/HINT.

show abstract

The 2021 SIIM-FISABIO-RSNA Machine Learning COVID-19 Challenge: Annotation and Standard Exam Classification of COVID-19 Chest Radiographs.

Cited by 17 publications

References 14 publications

Dark-field Chest X-ray Imaging for the Assessment of COVID‑19‑Pneumonia

Dark-field Chest X-ray Imaging for the Assessment of COVID‑19‑Pneumonia

Challenges in COVID-19 Chest X-Ray Classification: Problematic Data or Ineffective Approaches?

HINT: Hierarchical Neuron Concept Explainer

Contact Info

Product

Resources

About