{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,1]],"date-time":"2026-05-01T17:41:47Z","timestamp":1777657307051,"version":"3.51.4"},"reference-count":26,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2022,8,6]],"date-time":"2022-08-06T00:00:00Z","timestamp":1659744000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2022,8,6]],"date-time":"2022-08-06T00:00:00Z","timestamp":1659744000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["npj Digit. Med."],"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 \u201clive\u201d 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 (\u00b10.11) compared to 0.70 (\u00b10.15) for landmarking Cox and 0.67 (\u00b10.09) 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>","DOI":"10.1038\/s41746-022-00659-w","type":"journal-article","created":{"date-parts":[[2022,8,6]],"date-time":"2022-08-06T11:03:54Z","timestamp":1659783834000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":24,"title":["Developing machine learning algorithms for dynamic estimation of progression during active surveillance for prostate cancer"],"prefix":"10.1038","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8681-4739","authenticated-orcid":false,"given":"Changhee","family":"Lee","sequence":"first","affiliation":[]},{"given":"Alexander","family":"Light","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2887-0342","authenticated-orcid":false,"given":"Evgeny S.","family":"Saveliev","sequence":"additional","affiliation":[]},{"given":"Mihaela","family":"van der Schaar","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4722-4207","authenticated-orcid":false,"given":"Vincent J.","family":"Gnanapragasam","sequence":"additional","affiliation":[]}],"member":"297","published-online":{"date-parts":[[2022,8,6]]},"reference":[{"key":"659_CR1","doi-asserted-by":"publisher","first-page":"196","DOI":"10.21037\/tau.2020.02.21","volume":"9","author":"Q Cai","year":"2020","unstructured":"Cai, Q. et al. Estimates of over-time trends in incidence and mortality of prostate cancer from 1990 to 2030. Transl. Androl. Urol. Apr 9, 196\u2013209 (2020).","journal-title":"Transl. Androl. Urol. Apr"},{"key":"659_CR2","doi-asserted-by":"publisher","first-page":"432.e1","DOI":"10.1016\/j.urolonc.2020.11.024","volume":"39","author":"Y Liu","year":"2021","unstructured":"Liu, Y., Hall, I. J., Filson, C. & Howard, D. H. Trends in the use of active surveillance and treatments in Medicare beneficiaries diagnosed with localized prostate cancer. Urol. Oncol. 39, 432.e1\u2013432.e10 (2021).","journal-title":"Urol. Oncol."},{"key":"659_CR3","doi-asserted-by":"publisher","first-page":"337","DOI":"10.1016\/j.eururo.2021.12.007","volume":"81","author":"PM Willemse","year":"2022","unstructured":"Willemse, P. M. et al. Systematic review of active surveillance for clinically localised prostate cancer to develop recommendations regarding inclusion of intermediate-risk disease, biopsy characteristics at inclusion and monitoring, and surveillance repeat biopsy strategy. Eur. Urol. 81, 337\u2013346 (2022).","journal-title":"Eur. Urol."},{"key":"659_CR4","doi-asserted-by":"publisher","first-page":"758","DOI":"10.1111\/bju.14800","volume":"124","author":"VJ Gnanapragasam","year":"2019","unstructured":"Gnanapragasam, V. J. et al. Using prognosis to guide inclusion criteria, define standardised endpoints and stratify follow-up in active surveillance for prostate cancer. BJU Int. 124, 758\u2013767 (2019).","journal-title":"BJU Int."},{"key":"659_CR5","doi-asserted-by":"publisher","first-page":"693","DOI":"10.1016\/j.eururo.2019.07.041","volume":"76","author":"FH Drost","year":"2019","unstructured":"Drost, F. H., Nieboer, D., Morgan, T. M., Carroll, P. R. & Roobol, M. J. Movember Foundation\u2019s Global Action Plan Prostate Cancer Active Surveillance (GAP) consortium. Predicting biopsy outcomes during active surveillance for prostate cancer: external validation of the canary prostate active surveillance study risk calculators in five large active surveillance cohorts. Eur. Urol. 76, 693\u2013702 (2019).","journal-title":"Eur. Urol."},{"key":"659_CR6","doi-asserted-by":"publisher","first-page":"96","DOI":"10.1111\/bju.15136","volume":"127","author":"A Tomer","year":"2021","unstructured":"Tomer, A. et al. Movember Foundation\u2019s Global Action Plan Prostate Cancer Active Surveillance (GAP3) consortium. Personalised biopsy schedules based on risk of Gleason upgrading for patients with low-risk prostate cancer on active surveillance. BJU Int. 127, 96\u2013107 (2021).","journal-title":"BJU Int."},{"key":"659_CR7","doi-asserted-by":"publisher","first-page":"e203187","DOI":"10.1001\/jamaoncol.2020.3187","volume":"6","author":"MR Cooperberg","year":"2020","unstructured":"Cooperberg, M. R. et al. Tailoring intensity of active surveillance for low-risk prostate cancer based on individualized prediction of risk stability. JAMA Oncol. 6, e203187 (2020).","journal-title":"JAMA Oncol."},{"key":"659_CR8","doi-asserted-by":"publisher","first-page":"715","DOI":"10.1038\/s41416-019-0569-4","volume":"121","author":"DR Thurtle","year":"2019","unstructured":"Thurtle, D. R., Jenkins, V., Pharoah, P. D. & Gnanapragasam, V. J. Understanding of prognosis in non-metastatic prostate cancer: a randomised comparative study of clinician estimates measured against the PREDICT prostate prognostic model. Br. J. Cancer 121, 715\u2013718 (2019).","journal-title":"Br. J. Cancer"},{"key":"659_CR9","doi-asserted-by":"publisher","first-page":"1548","DOI":"10.1016\/j.jvir.2021.07.025","volume":"32","author":"QF Chen","year":"2021","unstructured":"Chen, Q. F. et al. Surveillance strategy after complete ablation of initial recurrent hepatocellular carcinoma: a risk-based machine learning study. J. Vasc. Inter. Radio. 32, 1548\u20131557.e2 (2021).","journal-title":"J. Vasc. Inter. Radio."},{"key":"659_CR10","doi-asserted-by":"publisher","first-page":"20","DOI":"10.1016\/j.lungcan.2021.04.012","volume":"156","author":"HA Miller","year":"2021","unstructured":"Miller, H. A. et al. Evaluation of disease staging and chemotherapeutic response in non-small cell lung cancer from patient tumor-derived metabolomic data. Lung Cancer 156, 20\u201330 (2021).","journal-title":"Lung Cancer"},{"key":"659_CR11","doi-asserted-by":"publisher","first-page":"e196972","DOI":"10.1001\/jamanetworkopen.2019.6972","volume":"2","author":"L Wang","year":"2019","unstructured":"Wang, L. et al. Development and validation of a Deep Learning Algorithm for Mortality Prediction in Selecting Patients With Dementia for Earlier Palliative Care Interventions. JAMA Netw. Open 2, e196972 (2019).","journal-title":"JAMA Netw. Open"},{"key":"659_CR12","doi-asserted-by":"publisher","first-page":"e23483","DOI":"10.2196\/23483","volume":"23","author":"OT Jones","year":"2021","unstructured":"Jones, O. T. et al. Artificial intelligence techniques that may be applied to primary care data to facilitate earlier diagnosis of cancer: systematic review. J. Med Internet Res 23, e23483 (2021).","journal-title":"J. Med Internet Res"},{"key":"659_CR13","doi-asserted-by":"publisher","DOI":"10.1186\/s13073-021-00968-x","volume":"27","author":"KA Tran","year":"2021","unstructured":"Tran, K. A. et al. Deep learning in cancer diagnosis, prognosis and treatment selection. Genome Med 27, 152 (2021).","journal-title":"Genome Med"},{"key":"659_CR14","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1111\/bju.14166","volume":"122","author":"D Thurtle","year":"2018","unstructured":"Thurtle, D. et al. Progression and treatment rates using an active surveillance protocol incorporating image-guided baseline biopsies and multiparametric magnetic resonance imaging monitoring for men with favourable-risk prostate cancer. BJU Int 122, 59\u201365 (2018).","journal-title":"BJU Int"},{"key":"659_CR15","doi-asserted-by":"publisher","first-page":"2696","DOI":"10.1007\/s00330-020-07336-0","volume":"31","author":"I Caglic","year":"2021","unstructured":"Caglic, I. et al. MRI-derived PRECISE scores for predicting pathologically-confirmed radiological progression in prostate cancer patients on active surveillance. Eur. Radio. 31, 2696\u20132705 (2021).","journal-title":"Eur. Radio."},{"key":"659_CR16","doi-asserted-by":"publisher","first-page":"648","DOI":"10.1016\/j.eururo.2016.06.011","volume":"71","author":"CM Moore","year":"2017","unstructured":"Moore, C. M. et al. Reporting magnetic resonance imaging in men on active surveillance for prostate cancer: the PRECISE recommendations-A report of a European school of oncology task force. Eur. Urol. 71, 648\u2013655 (2017).","journal-title":"Eur. Urol."},{"key":"659_CR17","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1111\/bju.14707","volume":"1","author":"SWD Merriel","year":"2019","unstructured":"Merriel, S. W. D. et al. Prostate cancer UK expert reference group on active surveillance. Best practice in active surveillance for men with prostate cancer: a Prostate Cancer UK consensus statement. BJU Int 1, 47\u201354 (2019). 2019.","journal-title":"BJU Int"},{"key":"659_CR18","doi-asserted-by":"publisher","first-page":"161.e1","DOI":"10.1016\/j.urolonc.2021.08.007","volume":"40","author":"M Nayan","year":"2022","unstructured":"Nayan, M. et al. A machine learning approach to predict progression on active surveillance for prostate cancer. Urol. Oncol. 40, 161.e1\u2013161.e7 (2022).","journal-title":"Urol. Oncol."},{"key":"659_CR19","first-page":"00367","volume":"S1078-1439","author":"LD Venderbos","year":"2021","unstructured":"Venderbos, L. D., Luiting, H., Hogenhout, R. & Roobol, M. J. Interaction of MRI and active surveillance in prostate cancer: Time to re-evaluate the active surveillance inclusion criteria. Urol. Oncol. S1078-1439, 00367\u20137 (2021).","journal-title":"Urol. Oncol."},{"key":"659_CR20","first-page":"353:i3235","volume":"8","author":"M Pavlou","year":"2016","unstructured":"Pavlou, M. et al. How to develop a more accurate risk prediction model when there are few events. BMJ 8, 353:i3235 (2016).","journal-title":"BMJ"},{"key":"659_CR21","doi-asserted-by":"publisher","first-page":"523","DOI":"10.1016\/j.eururo.2018.10.025","volume":"75","author":"M Van Hemelrijck","year":"2019","unstructured":"Van Hemelrijck, M. et al. Members of the movember foundation\u2019s global action plan prostate cancer active surveillance GAP3 consortium; members of the movember foundation\u2019s global action plan prostate cancer active surveillance GAP3 consortium. Reasons for discontinuing active surveillance: assessment of 21 centres in 12 countries in the movember GAP3 consortium. Eur. Urol. 75, 523\u2013531 (2019).","journal-title":"Eur. Urol."},{"key":"659_CR22","volume":"13","author":"Y Philippou","year":"2015","unstructured":"Philippou, Y., Raja, H. & Gnanapragasam, V. J. Active surveillance of prostate cancer: a questionnaire survey of urologists, clinical oncologists and urology nurse specialists across three cancer networks in the United Kingdom. BMC Urol. 13, 15:52 (2015).","journal-title":"BMC Urol."},{"key":"659_CR23","doi-asserted-by":"publisher","first-page":"122","DOI":"10.1109\/TBME.2019.2909027","volume":"67","author":"C Lee","year":"2020","unstructured":"Lee, C., Yoon, J. & Schaar, M. V. Dynamic-deephit: A deep learning approach for dynamic survival analysis with competing risks based on longitudinal data. IEEE Trans. Biomed. Eng. 67, 122\u2013133 (2020).","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"659_CR24","doi-asserted-by":"crossref","unstructured":"Hosmer, D. W., Lemeshow, S. & May, S. Applied Survival Analysis: Regression Modelling of Time-to-Event Data. (John Wiley & Sons, New York 2008).","DOI":"10.1002\/9780470258019"},{"key":"659_CR25","doi-asserted-by":"publisher","first-page":"1105","DOI":"10.1002\/sim.4154","volume":"10","author":"H Uno","year":"2011","unstructured":"Uno, H., Cai, T., Pencina, M. J., D\u2019Agostino, R. B. & Wei, L. J. On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data. Stat. Med. 10, 1105\u201317 (2011).","journal-title":"Stat. Med."},{"key":"659_CR26","first-page":"5767","volume":"535","author":"C Lee","year":"2020","unstructured":"Lee, C. & van der Schaar, M. Temporal phenotyping using deep predictive clustering of disease progression. Proc. 37th Int. Conf. Mach. Learn. 535, 5767\u20135777 (2020).","journal-title":"Proc. 37th Int. Conf. Mach. Learn."}],"container-title":["npj Digital Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41746-022-00659-w.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-022-00659-w","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-022-00659-w.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,11,25]],"date-time":"2022-11-25T08:12:27Z","timestamp":1669363947000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41746-022-00659-w"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,8,6]]},"references-count":26,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2022,12]]}},"alternative-id":["659"],"URL":"https:\/\/doi.org\/10.1038\/s41746-022-00659-w","relation":{},"ISSN":["2398-6352"],"issn-type":[{"value":"2398-6352","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,8,6]]},"assertion":[{"value":"1 March 2022","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"14 July 2022","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 August 2022","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare no competing interests.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"110"}}