dc.contributor.author | Lee, Hyunkwang | |
dc.contributor.author | Yune, Sehyo | |
dc.contributor.author | Mansouri, Mohammad | |
dc.contributor.author | Kim, Myeongchan | |
dc.contributor.author | Tajmir, Shahein H. | |
dc.contributor.author | Guerrier, Claude E. | |
dc.contributor.author | Ebert, Sarah A. | |
dc.contributor.author | Pomerantz, Stuart | |
dc.contributor.author | Romero, Javier | |
dc.contributor.author | Kamalian, Mohammad | |
dc.contributor.author | Gonzalez, Ramon | |
dc.contributor.author | Lev, Michael | |
dc.contributor.author | Do, Synho | |
dc.date.accessioned | 2022-03-21T16:48:57Z | |
dc.date.issued | 2018-12-17 | |
dc.identifier.citation | Lee, Hyunkwang, Sehyo Yune, Mohammad Mansouri, Myeongchan Kim, Shahein H. Tajmir, Claude E. Guerrier, Sarah A. Ebert et al. "An explainable deep-learning algorithm for the detection of acute intracranial haemorrhage from small datasets." Nature Biomedical Engineering 3, no. 3 (2018): 173-182. DOI: 10.1038/s41551-018-0324-9 | |
dc.identifier.issn | 2157-846X | en_US |
dc.identifier.uri | https://nrs.harvard.edu/URN-3:HUL.INSTREPOS:37371238 | * |
dc.description.abstract | Owing to improvements in image recognition via deep learning, machine-learning algorithms could eventually be applied to automated medical diagnoses that can guide clinical decision-making. However, these algorithms remain a 'black box' in terms of how they generate the predictions from the input data. Also, high-performance deep learning requires large, high-quality training datasets. Here, we report the development of an understandable deep-learning system that detects acute intracranial haemorrhage (ICH) and classifies five ICH subtypes from unenhanced head computed-tomography scans. By using a dataset of only 904 cases for algorithm training, the system achieved a performance similar to that of expert radiologists in two independent test datasets containing 200 cases (sensitivity of 98% and specificity of 95%) and 196 cases (sensitivity of 92% and specificity of 95%). The system includes an attention map and a prediction basis retrieved from training data to enhance explainability, and an iterative process that mimics the workflow of radiologists. Our approach to algorithm development can facilitate the development of deep-learning systems for a variety of clinical applications and accelerate their adoption into clinical practice. | en_US |
dc.language.iso | en_US | en_US |
dc.publisher | Springer Science and Business Media LLC | en_US |
dc.relation | Nature Biomedical Engineering | en_US |
dc.relation.isversionof | doi:10.1038/s41551-018-0324-9 | en_US |
dash.license | META_ONLY | |
dc.subject | Computer Science Applications | en_US |
dc.subject | Biomedical Engineering | en_US |
dc.subject | Medicine (miscellaneous) | en_US |
dc.subject | Bioengineering | en_US |
dc.subject | Biotechnology | en_US |
dc.title | An explainable deep-learning algorithm for the detection of acute intracranial haemorrhage from small datasets | en_US |
dc.type | Journal Article | en_US |
dc.description.version | Version of Record | en_US |
dc.relation.journal | Nature Biomedical Engineering | en_US |
dash.depositing.author | Gonzalez, Ramon | |
dc.date.available | 2022-03-21T16:48:57Z | |
dash.affiliation.other | Harvard John A. Paulson School of Engineering and Applied Sciences | en_US |
dc.identifier.doi | 10.1038/s41551-018-0324-9 | |
dc.source.journal | Nat Biomed Eng | |
dash.waiver.reason | Nature Biomedical Engineering editorial teams request an open-access waiver. | en_US |
dash.source.volume | 3 | en_US |
dash.source.page | 173-182 | en_US |
dash.source.issue | 3 | en_US |
dash.contributor.affiliated | Kamalian, Mohammad | |
dash.contributor.affiliated | Romero, Javier | |
dash.contributor.affiliated | Gonzalez, Ramon | |
dash.contributor.affiliated | Pomerantz, Stuart | |
dash.contributor.affiliated | Lev, Michael | |
dash.contributor.affiliated | Do, Synho | |