Lina M. Åström , Patrik Sibolt , Hannah Chamberlin , Eva Serup-Hansen , Claus E. Andersen , Marcel van Herk , Lene S. Mouritsen , Marianne C. Aznar , Claus P. Behrens
{"title":"人工智能生成的目标和膀胱癌在线自适应放疗中观察者之间的差异","authors":"Lina M. Åström , Patrik Sibolt , Hannah Chamberlin , Eva Serup-Hansen , Claus E. Andersen , Marcel van Herk , Lene S. Mouritsen , Marianne C. Aznar , Claus P. Behrens","doi":"10.1016/j.phro.2024.100640","DOIUrl":null,"url":null,"abstract":"<div><h3>Background and purpose</h3><p>Daily target re-delineation in online adaptive radiotherapy (oART) introduces uncertainty. The aim of this study was to evaluate artificial intelligence (AI) generated contours and inter-observer target variation among radiotherapy technicians in cone-beam CT (CBCT) guided oART of bladder cancer.</p></div><div><h3>Materials and methods</h3><p>For each of 10 consecutive patients treated with oART for bladder cancer, one CBCT was randomly selected and retrospectively included. The bladder (CTV-T) was AI-segmented (CTV-T<sub>AI</sub>). Seven radiotherapy technicians independently reviewed and edited CTV-T<sub>AI</sub>, generating CTV-T<sub>ADP</sub>. Contours were benchmarked against a ground truth contour (CTV-T<sub>GT</sub>) delineated blindly from scratch. CTV-T<sub>ADP</sub> and CTV-T<sub>AI</sub> were compared to CTV-T<sub>GT</sub> using volume, dice similarity coefficient, and bidirectional local distance. Dose coverage (D<sub>99%</sub>>95 %) of CTV-T<sub>GT</sub> was evaluated for treatment plans optimized for CTV-T<sub>AI</sub> and CTV-T<sub>ADP</sub> with clinical margins. Inter-observer variation among CTV-T<sub>ADP</sub> was assessed using coefficient of variation and generalized conformity index.</p></div><div><h3>Results</h3><p>CTV-T<sub>GT</sub> ranged from 48.7 cm<sup>3</sup> to 211.6 cm<sup>3</sup>. The median [range] volume difference was 4.5 [−17.8, 42.4] cm<sup>3</sup> for CTV-T<sub>ADP</sub> and −15.5 [−54.2, 4.3] cm<sup>3</sup> for CTV-T<sub>AI</sub>, compared to CTV-T<sub>GT</sub>. Corresponding dice similarity coefficients were 0.87 [0.71, 0.95] and 0.84 [0.64, 0.95]. CTV-T<sub>GT</sub> was adequately covered in 68/70 plans optimized on CTV-T<sub>ADP</sub> and in 6/10 plans optimized on CTV-T<sub>AI</sub> with clinical margins. The median [range] coefficient of variation was 0.08 [0.05, 0.11] and generalized conformity index was 0.78 [0.71, 0.88] among CTV-T<sub>ADP</sub>.</p></div><div><h3>Conclusions</h3><p>Target re-delineation in CBCT-guided oART of bladder cancer demonstrated non-isotropic inter-observer variation. Manual adjustment of AI-generated contours was necessary to cover ground truth targets.</p></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":null,"pages":null},"PeriodicalIF":3.4000,"publicationDate":"2024-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2405631624001106/pdfft?md5=65ea09c6823d780c0e3cf276f0b90698&pid=1-s2.0-S2405631624001106-main.pdf","citationCount":"0","resultStr":"{\"title\":\"Artificial intelligence-generated targets and inter-observer variation in online adaptive radiotherapy of bladder cancer\",\"authors\":\"Lina M. Åström , Patrik Sibolt , Hannah Chamberlin , Eva Serup-Hansen , Claus E. Andersen , Marcel van Herk , Lene S. Mouritsen , Marianne C. Aznar , Claus P. Behrens\",\"doi\":\"10.1016/j.phro.2024.100640\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><h3>Background and purpose</h3><p>Daily target re-delineation in online adaptive radiotherapy (oART) introduces uncertainty. The aim of this study was to evaluate artificial intelligence (AI) generated contours and inter-observer target variation among radiotherapy technicians in cone-beam CT (CBCT) guided oART of bladder cancer.</p></div><div><h3>Materials and methods</h3><p>For each of 10 consecutive patients treated with oART for bladder cancer, one CBCT was randomly selected and retrospectively included. The bladder (CTV-T) was AI-segmented (CTV-T<sub>AI</sub>). Seven radiotherapy technicians independently reviewed and edited CTV-T<sub>AI</sub>, generating CTV-T<sub>ADP</sub>. Contours were benchmarked against a ground truth contour (CTV-T<sub>GT</sub>) delineated blindly from scratch. CTV-T<sub>ADP</sub> and CTV-T<sub>AI</sub> were compared to CTV-T<sub>GT</sub> using volume, dice similarity coefficient, and bidirectional local distance. Dose coverage (D<sub>99%</sub>>95 %) of CTV-T<sub>GT</sub> was evaluated for treatment plans optimized for CTV-T<sub>AI</sub> and CTV-T<sub>ADP</sub> with clinical margins. Inter-observer variation among CTV-T<sub>ADP</sub> was assessed using coefficient of variation and generalized conformity index.</p></div><div><h3>Results</h3><p>CTV-T<sub>GT</sub> ranged from 48.7 cm<sup>3</sup> to 211.6 cm<sup>3</sup>. The median [range] volume difference was 4.5 [−17.8, 42.4] cm<sup>3</sup> for CTV-T<sub>ADP</sub> and −15.5 [−54.2, 4.3] cm<sup>3</sup> for CTV-T<sub>AI</sub>, compared to CTV-T<sub>GT</sub>. Corresponding dice similarity coefficients were 0.87 [0.71, 0.95] and 0.84 [0.64, 0.95]. CTV-T<sub>GT</sub> was adequately covered in 68/70 plans optimized on CTV-T<sub>ADP</sub> and in 6/10 plans optimized on CTV-T<sub>AI</sub> with clinical margins. The median [range] coefficient of variation was 0.08 [0.05, 0.11] and generalized conformity index was 0.78 [0.71, 0.88] among CTV-T<sub>ADP</sub>.</p></div><div><h3>Conclusions</h3><p>Target re-delineation in CBCT-guided oART of bladder cancer demonstrated non-isotropic inter-observer variation. Manual adjustment of AI-generated contours was necessary to cover ground truth targets.</p></div>\",\"PeriodicalId\":36850,\"journal\":{\"name\":\"Physics and Imaging in Radiation Oncology\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":3.4000,\"publicationDate\":\"2024-07-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.sciencedirect.com/science/article/pii/S2405631624001106/pdfft?md5=65ea09c6823d780c0e3cf276f0b90698&pid=1-s2.0-S2405631624001106-main.pdf\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Physics and Imaging in Radiation Oncology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2405631624001106\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"ONCOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Physics and Imaging in Radiation Oncology","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2405631624001106","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ONCOLOGY","Score":null,"Total":0}
Artificial intelligence-generated targets and inter-observer variation in online adaptive radiotherapy of bladder cancer
Background and purpose
Daily target re-delineation in online adaptive radiotherapy (oART) introduces uncertainty. The aim of this study was to evaluate artificial intelligence (AI) generated contours and inter-observer target variation among radiotherapy technicians in cone-beam CT (CBCT) guided oART of bladder cancer.
Materials and methods
For each of 10 consecutive patients treated with oART for bladder cancer, one CBCT was randomly selected and retrospectively included. The bladder (CTV-T) was AI-segmented (CTV-TAI). Seven radiotherapy technicians independently reviewed and edited CTV-TAI, generating CTV-TADP. Contours were benchmarked against a ground truth contour (CTV-TGT) delineated blindly from scratch. CTV-TADP and CTV-TAI were compared to CTV-TGT using volume, dice similarity coefficient, and bidirectional local distance. Dose coverage (D99%>95 %) of CTV-TGT was evaluated for treatment plans optimized for CTV-TAI and CTV-TADP with clinical margins. Inter-observer variation among CTV-TADP was assessed using coefficient of variation and generalized conformity index.
Results
CTV-TGT ranged from 48.7 cm3 to 211.6 cm3. The median [range] volume difference was 4.5 [−17.8, 42.4] cm3 for CTV-TADP and −15.5 [−54.2, 4.3] cm3 for CTV-TAI, compared to CTV-TGT. Corresponding dice similarity coefficients were 0.87 [0.71, 0.95] and 0.84 [0.64, 0.95]. CTV-TGT was adequately covered in 68/70 plans optimized on CTV-TADP and in 6/10 plans optimized on CTV-TAI with clinical margins. The median [range] coefficient of variation was 0.08 [0.05, 0.11] and generalized conformity index was 0.78 [0.71, 0.88] among CTV-TADP.
Conclusions
Target re-delineation in CBCT-guided oART of bladder cancer demonstrated non-isotropic inter-observer variation. Manual adjustment of AI-generated contours was necessary to cover ground truth targets.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
