Show simple item record

dc.contributor.authorLee, Changhee
dc.contributor.authorLight, Alexander
dc.contributor.authorSaveliev, Evgeny S
dc.contributor.authorvan der Schaar, Mihaela
dc.contributor.authorGnanapragasam, Vincent J
dc.date.accessioned2022-08-06T15:01:09Z
dc.date.available2022-08-06T15:01:09Z
dc.date.issued2022-08-06
dc.date.submitted2022-03-01
dc.identifier.citationnpj Digital Medicine, volume 5, issue 1, article-number 110
dc.identifier.others41746-022-00659-w
dc.identifier.other659
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/339908
dc.description.abstract<jats:title>Abstract</jats:title><jats:p>Active Surveillance (AS) for prostate cancer is a management option that continually monitors early disease and considers intervention if progression occurs. A robust method to incorporate “live” updates of progression risk during follow-up has hitherto been lacking. To address this, we developed a deep learning-based individualised longitudinal survival model using Dynamic-DeepHit-Lite (DDHL) that learns data-driven distribution of time-to-event outcomes. Further refining outputs, we used a reinforcement learning approach (Actor-Critic) for temporal predictive clustering (AC-TPC) to discover groups with similar time-to-event outcomes to support clinical utility. We applied these methods to data from 585 men on AS with longitudinal and comprehensive follow-up (median 4.4 years). Time-dependent C-indices and Brier scores were calculated and compared to Cox regression and landmarking methods. Both Cox and DDHL models including only baseline variables showed comparable C-indices but the DDHL model performance improved with additional follow-up data. With 3 years of data collection and 3 years follow-up the DDHL model had a C-index of 0.79 (±0.11) compared to 0.70 (±0.15) for landmarking Cox and 0.67 (±0.09) for baseline Cox only. Model calibration was good across all models tested. The AC-TPC method further discovered 4 distinct outcome-related temporal clusters with distinct progression trajectories. Those in the lowest risk cluster had negligible progression risk while those in the highest cluster had a 50% risk of progression by 5 years. In summary, we report a novel machine learning approach to inform personalised follow-up during active surveillance which improves predictive power with increasing data input over time.</jats:p>
dc.languageen
dc.publisherSpringer Science and Business Media LLC
dc.subjectArticle
dc.subject/692/699/67/589/466
dc.subject/692/499
dc.subjectarticle
dc.titleDeveloping machine learning algorithms for dynamic estimation of progression during active surveillance for prostate cancer
dc.typeArticle
dc.date.updated2022-08-06T15:01:09Z
prism.publicationNamenpj Digital Medicine
dc.identifier.doi10.17863/CAM.87330
dcterms.dateAccepted2022-07-14
rioxxterms.versionofrecord10.1038/s41746-022-00659-w
rioxxterms.versionVoR
rioxxterms.licenseref.urihttp://creativecommons.org/licenses/by/4.0/
dc.contributor.orcidLee, Changhee [0000-0002-8681-4739]
dc.contributor.orcidSaveliev, Evgeny S [0000-0003-2887-0342]
dc.contributor.orcidGnanapragasam, Vincent J [0000-0003-4722-4207]
dc.identifier.eissn2398-6352
cam.issuedOnline2022-08-06


Files in this item

Thumbnail
Thumbnail
Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record