{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/longitudinal-detection-of-radiological","title":"Longitudinal detection of radiological abnormalities with time-modulated LSTM","arxiv_id":"1807.06144","date":"2018-07-16","proceeding":null,"authors":["Ruggiero Santeramo","Samuel Withey","Giovanni Montana"],"abstract":"Convolutional neural networks (CNNs) have been successfully employed in\nrecent years for the detection of radiological abnormalities in medical images\nsuch as plain x-rays. To date, most studies use CNNs on individual examinations\nin isolation and discard previously available clinical information. In this\nstudy we set out to explore whether Long-Short-Term-Memory networks (LSTMs) can\nbe used to improve classification performance when modelling the entire\nsequence of radiographs that may be available for a given patient, including\ntheir reports. A limitation of traditional LSTMs, though, is that they\nimplicitly assume equally-spaced observations, whereas the radiological exams\nare event-based, and therefore irregularly sampled. Using both a simulated\ndataset and a large-scale chest x-ray dataset, we demonstrate that a simple\nmodification of the LSTM architecture, which explicitly takes into account the\ntime lag between consecutive observations, can boost classification\nperformance. Our empirical results demonstrate improved detection of commonly\nreported abnormalities on chest x-rays such as cardiomegaly, consolidation,\npleural effusion and hiatus hernia.","url_abs":"http://arxiv.org/abs/1807.06144v1","url_pdf":"http://arxiv.org/pdf/1807.06144v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"longitudinal-detection-of-radiological","repo_url":"https://github.com/WMGDataScience/tLSTM","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.06144","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}