{"id":2754,"date":"2021-07-02T00:09:00","date_gmt":"2021-07-01T14:09:00","guid":{"rendered":"https:\/\/harrison.ai\/?p=2754"},"modified":"2025-10-05T13:04:16","modified_gmt":"2025-10-05T03:04:16","slug":"effect-of-a-comprehensive-deep-learning-model-on-the-accuracy-of-chest-x-ray-interpretation-by-radiologists-a-retrospective-multireader-multicase-study","status":"publish","type":"post","link":"https:\/\/harrison.ai\/effect-of-a-comprehensive-deep-learning-model-on-the-accuracy-of-chest-x-ray-interpretation-by-radiologists-a-retrospective-multireader-multicase-study\/","title":{"rendered":"Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists: a retrospective, multireader multicase study"},"content":{"rendered":"    <section id=\"evidence-block-block_47c8fed505ae18614f1dee66430ff0dc\" class=\"study-block   text-\" >\n        <div class=\"container  container--tab\">\n            <div class=\"container container--tab\">\n                <div class=\"connect__decor hide-md\">\n                    <span class=\"pixel-decor\" style=\"background-color: rgba(9, 114, 241, 0.75)\"><\/span>\n                    <span class=\"pixel-decor\" style=\"background-color: #0972f1\"><\/span>\n                    <span class=\"pixel-decor\" style=\"background-color: rgba(9, 114, 241, 0.5)\"><\/span>\n                    <span class=\"pixel-decor hide-sm\" style=\"background-color: rgba(9, 114, 241, 0.5)\"><\/span>\n                <\/div>\n                <div class=\"study__share hide-sm\" data-aos=\"fade-up\">\n                                    <\/div>\n                <div class=\"study__row\">\n                    <div class=\"study__left\" data-aos=\"fade-up\"><\/div>\n                    <div class=\"study__right\" data-aos=\"fade-up\">\n                        <div class=\"text b1\">\n                        <h5>Authors<\/h5>\n<p>Seah, J. C. Y., Tang, C. H. M., Buchlak, Q. D., Holt, X. G., Wardman, J. B., Aimoldin, A., Esmaili, N., Ahmad, H., Pham, H., Lambert, J. F., Hachey, B., Hogg, S. J. F., Johnston, B. P., Bennett, C., Oakden-Rayner, L., Brotchie, P., &amp; Jones, C. M.<\/p>\n<p>The Lancet Digital Health, August 2021<\/p>\n                        <\/div>\n                                            <\/div>\n                <\/div>\n                                                            <div class=\"study__row\">\n                            <div class=\"study__left\" data-aos=\"fade-up\">\n                                <h2 class=\"h5\">Summary<\/h2>\n                            <\/div>\n                            <div class=\"study__right\" data-aos=\"fade-up\">\n                                <div class=\"text b1\"><p>Chest x-rays are widely used in clinical practice; however, interpretation can be hindered by human error and a lack of experienced thoracic radiologists. Deep learning has the potential to improve the accuracy of chest x-ray interpretation. We therefore aimed to assess the accuracy of radiologists with and without the assistance of a deep-learning model.<\/p>\n<\/div>                            <\/div>\n                        <\/div>\n                                            <div class=\"study__row\">\n                            <div class=\"study__left\" data-aos=\"fade-up\">\n                                <h2 class=\"h5\">Methods<\/h2>\n                            <\/div>\n                            <div class=\"study__right\" data-aos=\"fade-up\">\n                                <div class=\"text b1\"><p>In this retrospective study, a deep-learning model was trained on 821 681 images (284 649 patients) from five data sets from Australia, Europe, and the USA. 2568 enriched chest x-ray cases from adult patients (\u226516 years) who had at least one frontal chest x-ray were included in the test dataset; cases were representative of inpatient, outpatient, and emergency settings. 20 radiologists reviewed cases with and without the assistance of the deep-learning model with a 3-month washout period. We assessed the change in accuracy of chest x-ray interpretation across 127 clinical findings when the deep-learning model was used as a decision support by calculating area under the receiver operating characteristic curve (AUC) for each radiologist with and without the deep-learning model. We also compared AUCs for the model alone with those of unassisted radiologists. If the lower bound of the adjusted 95% CI of the difference in AUC between the model and the unassisted radiologists was more than \u20130\u00b705, the model was considered to be non-inferior for that finding. If the lower bound exceeded 0, the model was considered to be superior.<\/p>\n<\/div>                            <\/div>\n                        <\/div>\n                                            <div class=\"study__row\">\n                            <div class=\"study__left\" data-aos=\"fade-up\">\n                                <h2 class=\"h5\">Findings<\/h2>\n                            <\/div>\n                            <div class=\"study__right\" data-aos=\"fade-up\">\n                                <div class=\"text b1\"><p>Unassisted radiologists had a macroaveraged AUC of 0\u00b7713 (95% CI 0\u00b7645\u20130\u00b7785) across the 127 clinical findings, compared with 0\u00b7808 (0\u00b7763\u20130\u00b7839) when assisted by the model. The deep-learning model statistically significantly improved the classification accuracy of radiologists for 102 (80%) of 127 clinical findings, was statistically non-inferior for 19 (15%) findings, and no findings showed a decrease in accuracy when radiologists used the deep-learning model. Unassisted radiologists had a macroaveraged mean AUC of 0\u00b7713 (0\u00b7645\u20130\u00b7785) across all findings, compared with 0\u00b7957 (0\u00b7954\u20130\u00b7959) for the model alone. Model classification alone was significantly more accurate than unassisted radiologists for 117 (94%) of 124 clinical findings predicted by the model and was non-inferior to unassisted radiologists for all other clinical findings.<\/p>\n<\/div>                            <\/div>\n                        <\/div>\n                                            <div class=\"study__row\">\n                            <div class=\"study__left\" data-aos=\"fade-up\">\n                                <h2 class=\"h5\">Disclaimer<\/h2>\n                            <\/div>\n                            <div class=\"study__right\" data-aos=\"fade-up\">\n                                <div class=\"text b1\"><p>Harrison.ai Radiology Solutions were previously marketed as Annalise.ai solutions.<\/p>\n<\/div>                            <\/div>\n                        <\/div>\n                                                                        <div class=\"study__row\">\n                        <div class=\"study__left\" data-aos=\"fade-up\"><\/div>\n                        <div class=\"study__right\" data-aos=\"fade-up\">\n                            <div class=\"study__btn\">\n                                <a href=\"https:\/\/www.thelancet.com\/journals\/landig\/article\/PIIS2589-7500(21)00106-0\/fulltext\" target=\"_blank\" class=\"btn btn--xl font-rules\" rel=\"noopener\">Full study <span class=\"icon-arrow-right-up\"><\/span><\/a>\n                            <\/div>\n                        <\/div>\n                    <\/div>\n                            <\/div>\n        <\/div>\n    <\/section>\n\n\n\n    <section id=\"cta-block-block_3c1bef7edef7257a176377f6111bc954\" class=\"cta-block   text-\" >\n        <div class=\"container  container-\">\n            <div class=\"cta__content\">\n                <div class=\"cta__decor cta__decor--alt\">\n                    <div class=\"pixel-md-holder\">\n                        <span class=\"pixel-md-decor pixel-blur\" style=\"background-color: rgba(9, 114, 241, 0.15)\"><\/span>\n                        <span class=\"pixel-md-decor pixel-blur hide-md\" style=\"background-color: rgba(9, 114, 241, 0.45)\"><\/span>\n                        <span class=\"pixel-md-decor pixel-blur\" style=\"background-color: rgba(9, 114, 241, 0.45)\"><\/span>\n                    <\/div>\n                    <div class=\"pixel-lg-holder\">\n                        <span class=\"pixel-lg-decor pixel-blur\" style=\"background-color: rgba(9, 114, 241, 0.15)\"><\/span>\n                        <span class=\"pixel-lg-decor pixel-blur\" style=\"background-color: rgba(9, 114, 241, 0.15)\"><\/span>\n                        <span class=\"pixel-lg-decor pixel-blur\" style=\"background-color: rgba(9, 114, 241, 0.15)\"><\/span>\n                    <\/div>\n                <\/div>\n                <div class=\"cta__bg\"><img decoding=\"async\" src=\"https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-scaled.jpg\" class=\"attachment-full size-full\" alt=\"\" srcset=\"https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-scaled.jpg 2560w, https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-460x193.jpg 460w, https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-1024x429.jpg 1024w, https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-768x322.jpg 768w, https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-1536x644.jpg 1536w, https:\/\/harrison.ai\/wp-content\/uploads\/2025\/08\/cta-2-2048x859.jpg 2048w\" sizes=\"(max-width: 2560px) 100vw, 2560px\" \/><\/div>\n                <div class=\"cta__inner\">\n                    <h2 class=\"cta__title display-2 font-pixel\"  data-aos=\"fade-up\">Ready to unlock infinite medical possibility?<\/h2>                                            <div class=\"cta__btn\" data-aos=\"fade-up\">\n                            <a class=\"btn btn--secondary\" href=\"#\" target=\"_self\">Explore the product <span class=\"icon-arrow-right-up\"><\/span><\/a>\n                        <\/div>\n                                    <\/div>\n            <\/div>\n        <\/div>\n    <\/section>\n\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":2,"featured_media":2733,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[29],"tags":[],"region":[34,35,32,36,33],"product":[40],"class_list":["post-2754","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-evidence","region-apac","region-emea","region-oceania","region-uk","region-us","product-cxr"],"acf":[],"_links":{"self":[{"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/posts\/2754","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/comments?post=2754"}],"version-history":[{"count":3,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/posts\/2754\/revisions"}],"predecessor-version":[{"id":4962,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/posts\/2754\/revisions\/4962"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/media\/2733"}],"wp:attachment":[{"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/media?parent=2754"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/categories?post=2754"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/tags?post=2754"},{"taxonomy":"region","embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/region?post=2754"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/harrison.ai\/wp-json\/wp\/v2\/product?post=2754"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}