{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,18]],"date-time":"2025-11-18T12:28:22Z","timestamp":1763468902836},"reference-count":43,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2023,12,19]],"date-time":"2023-12-19T00:00:00Z","timestamp":1702944000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"}],"content-domain":{"domain":["aapm.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Medical Physics"],"published-print":{"date-parts":[[2024,2]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:sec><jats:title>Background<\/jats:title><jats:p>Dose distributions calculated with electronic portal imaging device (EPID)\u2010based in vivo dosimetry (EIVD) differ from planned dose distributions due to generic and plan\u2010specific deviations. Generic deviations are characteristic to a class of plans. Examples include limitations in EIVD dose reconstruction, inaccuracies in treatment planning system (TPS) calculations and systematic machine deviations. Plan\u2010specific deviations have an unpredictable character. Examples include discrepancies between the patient model used for dose calculation and the patient position or anatomy during delivery, random machine deviations, and data transfer, human or software errors. During the inspection work performed with traditional \u03b3\u2010evaluation statistical methods: (i) generic deviations raise alerts that need to be inspected but that rarely lead to action as their root cause is usually understood and (ii) the detection of relevant plan\u2010specific deviations may be hindered by the presence of generic deviations.<\/jats:p><\/jats:sec><jats:sec><jats:title>Purpose<\/jats:title><jats:p>To investigate whether deep learning\u2010based tools can help in identifying \u03b3\u2010alerts raised by generic deviations and in improving the detectability of plan\u2010specific deviations.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>A 3D U\u2010Net was trained as an autoencoder to reconstruct underlying patterns of generic deviations in \u03b3\u2010distributions. The network was trained for four treatment disease sites differently affected by generic deviations: volumetric modulated arc therapy (VMAT) lung (no known deviations), VMAT prostate (TPS inaccuracies), VMAT head\u2010and\u2010neck (EIVD limitations) and intensity modulated radiation therapy (IMRT) breast (large EIVD limitations). The network was trained with virtual non\u2010transit \u03b3\u2010distributions: 60 train\/10 validation for the VMAT sites and 30 train\/10 validation for IMRT breast. It was hypothesized that in vivo \u03b3\u2010distributions obtained in the presence of plan\u2010specific deviations would differ from those seen during training. For each disease site, the sensitivity of \u03b3\u2010analysis and the network to detect (synthetically introduced) patient\u2010related deviations was compared by receiver operator characteristic analysis. The investigated deviations were patient positioning errors, weight gain or loss, and tumor volume changes. The clinical relevance was illustrated qualitatively with 793 in vivo clinical cases (141 lung, 136 head\u2010and\u2010neck, 209 prostate and 307 breast).<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>Error detectability of patient\u2010related deviations was better with the network than with <jats:italic>\u03b3<\/jats:italic>\u2010analysis. The average area under the curve values over all sites were 0.86\u00a0\u00b1\u00a00.12(1SD) and 0.69\u00a0\u00b1\u00a00.25(1SD), respectively. Regarding in vivo clinical results, the percentage of cases differently classified by \u03b3\u2010analysis and the network was 1%, 19%, 18% and 64% for lung, head\u2010and\u2010neck, prostate, and breast, respectively. In head\u2010and\u2010neck and breast cases, 45 \u03b3\u2010only alerts were examined, of which 43 were attributed to EPID dose reconstruction limitations. For prostate, all 15 investigated \u03b3\u2010only alerts were due to known TPS inaccuracies. All 59 investigated network alerts were explained by either patient\u2010related deviations or EPID acquisition incidents. Some patient\u2010related deviations detected by the network were not detected by \u03b3\u2010analysis.<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusions<\/jats:title><jats:p>Deep learning\u2010based tools trained to reconstruct underlying patterns of generic deviations in \u03b3\u2010distributions can be used to (i) automatically identify false positives within the set of \u03b3\u2010alerts and (ii) improve the detection of plan\u2010specific deviations, hence minimizing the likelihood of false negatives. The presented method provides clear additional value to the \u03b3\u2010alert management process for large scale EIVD systems.<\/jats:p><\/jats:sec>","DOI":"10.1002\/mp.16895","type":"journal-article","created":{"date-parts":[[2023,12,19]],"date-time":"2023-12-19T12:45:22Z","timestamp":1702989922000},"page":"854-869","update-policy":"http:\/\/dx.doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Deep learning\u2010based tools to distinguish plan\u2010specific from generic deviations in EPID\u2010based in vivo dosimetry"],"prefix":"10.1002","volume":"51","author":[{"given":"Igor","family":"Olaciregui\u2010Ruiz","sequence":"first","affiliation":[{"name":"Department of Radiation Oncology Netherlands Cancer Institute  Amsterdam The Netherlands"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Rita","family":"Sim\u00f5es","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Netherlands Cancer Institute  Amsterdam The Netherlands"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sonke","family":"Jan\u2010Jakob","sequence":"additional","affiliation":[{"name":"Department of Radiation Oncology Netherlands Cancer Institute  Amsterdam The Netherlands"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2023,12,19]]},"reference":[{"key":"e_1_2_8_2_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2008.07.007"},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2017.03.049"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2021.10.021"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2017.01.035"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2020.09.011"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2020.08.003"},{"key":"e_1_2_8_8_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4935199"},{"key":"e_1_2_8_9_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3148482"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2009.12.020"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.12441"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.2128498"},{"key":"e_1_2_8_13_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.1637969"},{"key":"e_1_2_8_14_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.1984335"},{"key":"e_1_2_8_15_1","doi-asserted-by":"publisher","DOI":"10.1120\/jacmp.v15i6.4895"},{"key":"e_1_2_8_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prro.2015.07.001"},{"key":"e_1_2_8_17_1","doi-asserted-by":"publisher","DOI":"10.1007\/978\u20103\u2010319\u201046723\u20108_49"},{"key":"e_1_2_8_18_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.3665709"},{"key":"e_1_2_8_19_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2017.04.016"},{"key":"e_1_2_8_20_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.12080"},{"key":"e_1_2_8_21_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13893"},{"key":"e_1_2_8_22_1","doi-asserted-by":"publisher","DOI":"10.1002\/acm2.12610"},{"key":"e_1_2_8_23_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.2721657"},{"key":"e_1_2_8_24_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.594853"},{"key":"e_1_2_8_25_1","doi-asserted-by":"publisher","DOI":"10.1007\/978\u20103\u2010319\u201024574\u20104_28"},{"key":"e_1_2_8_26_1","doi-asserted-by":"publisher","DOI":"10.1088\/1361\u20106560\/aac71a"},{"key":"e_1_2_8_27_1","doi-asserted-by":"publisher","DOI":"10.1088\/1742\u20106596\/1627\/1\/012014"},{"key":"e_1_2_8_28_1","unstructured":"KingmaDP AdamBaJA method for stochastic optimization.http:\/\/arxiv.org\/abs\/1412.6980"},{"key":"e_1_2_8_29_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13265"},{"key":"e_1_2_8_30_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4792203"},{"key":"e_1_2_8_31_1","doi-asserted-by":"publisher","DOI":"10.1118\/1.4862515"},{"key":"e_1_2_8_32_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijrobp.2018.05.033"},{"key":"e_1_2_8_33_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejmp.2021.08.014"},{"key":"e_1_2_8_34_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.13338"},{"key":"e_1_2_8_35_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.14416"},{"key":"e_1_2_8_36_1","doi-asserted-by":"publisher","DOI":"10.1002\/mp.15031"},{"key":"e_1_2_8_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.phro.2022.03.004"},{"key":"e_1_2_8_38_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.radonc.2020.09.048"},{"key":"e_1_2_8_39_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2021.3107163"},{"key":"e_1_2_8_40_1","unstructured":"FernandoT GammulleH DenmanS SridharanS FookesC Deep learning for medical anomaly detection a survey.http:\/\/arxiv.org\/abs\/2012.02364"},{"key":"e_1_2_8_41_1","doi-asserted-by":"publisher","DOI":"10.1109\/IC3.2014.6897155"},{"key":"e_1_2_8_42_1","unstructured":"KingmaDP WellingM Auto\u2010encoding variational bayes.http:\/\/arxiv.org\/abs\/1312.6114"},{"key":"e_1_2_8_43_1","unstructured":"Di MattiaF GaleoneP De SimoniM GhelfiE A survey on GANs for anomaly detection.http:\/\/arxiv.org\/abs\/1906.11632"},{"key":"e_1_2_8_44_1","unstructured":"ZhaoH GalloO FrosioI KautzJ Loss functions for neural networks for image processing.http:\/\/arxiv.org\/abs\/1511.08861"}],"container-title":["Medical Physics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/mp.16895","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,2,8]],"date-time":"2024-02-08T20:15:16Z","timestamp":1707423316000},"score":1,"resource":{"primary":{"URL":"https:\/\/aapm.onlinelibrary.wiley.com\/doi\/10.1002\/mp.16895"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,12,19]]},"references-count":43,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2024,2]]}},"alternative-id":["10.1002\/mp.16895"],"URL":"https:\/\/doi.org\/10.1002\/mp.16895","archive":["Portico"],"relation":{},"ISSN":["0094-2405","2473-4209"],"issn-type":[{"value":"0094-2405","type":"print"},{"value":"2473-4209","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,12,19]]},"assertion":[{"value":"2023-07-17","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-12-01","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-12-19","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}