FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4229376477> ?p ?o ?g. }

• W4229376477 abstract "Abstract Background: Radiotherapy has been crucial in the treatment of prostate cancer. However, manual segmentation is labor intensive and highly variable among radiation oncologists, especially for lymph node volumes. Hence, in this study, a deep learning based automated contouring model is constructed with style adaptation algorithm for clinical target volumes (CTVs) of intact and postoperative prostate cancer. Methods: Computed tomography (CT) data sets for 197 prostate cancer patients, treated by one senior physician with more than 15 years of radiation work experience, during the period of November 2012 to January 2020 were selected, retrospectively. Of these patients, 167 cases were used for the training set and 30 for the test set. Two auto-delineation models were built for radical radiotherapy and postoperative radiotherapy of prostate cancer, including CTV1 for pelvic lymph nodes and CTV2 for prostate tumors or prostate tumor beds. In addition, delineation quality and processing time of the artificial intelligence (AI) algorithm were compared with those of a junior oncologist with 3 years of working experience. To further explore the number of patients required for the style adaptation model, to achieve another senior oncologist’s counting style, cases of the style transfer mode were gradually increased, adding ten patients to the training model each time. Results: No significant differences were found for the volumetric dice (VD) coefficients of the two models, CTV1 and CTV2, among DeeplabV3+, Unet++, and 3D U-net (p > 0.05). In the radical radiotherapy model, the VD coefficient of CTV1 calculated by AI, was higher than that of the one delineated by the junior physician (0.85 vs. 0.82, p = 0.018), while the VD coefficients of CTV2 counted by the junior physician and AI were the same (0.84 vs. 0.84, p = 0.958); In the postoperative radiotherapy model, the quantitative parameter of CTV1 and CTV2, counted by AI, was better than that of the junior physician (VD, 0.86 vs 0.83 for CTV1, p = 0.041; 0.79 vs 0.73 for CTV2, p = 0.003, respectively). The median delineation time for AI was 0.23 min in the postoperative model and 0.26 min in the radical model, which were significantly shorter than those of the physician (50.40 min and 45.43 min, respectively, p < 0.001). The correction time of the senior physician for AI was much shorter compared with that of the junior physician in both models (p < 0.001). For style adaptation from one physician to another, approximately 20 cases were required to construct a personal style based on the basic delineation model. Conclusion: Using deep active learning and attention mechanism, a highly consistent and time-saving contouring model was built for CTVs of prostate cancer. It was possible to train the style adaptation mode from one physician’s style to that of another with limited patient samples." @default.
• W4229376477 created "2022-05-10" @default.
• W4229376477 creator A5006822602 @default.
• W4229376477 creator A5017921159 @default.
• W4229376477 creator A5018091164 @default.
• W4229376477 creator A5042385997 @default.
• W4229376477 creator A5074803923 @default.
• W4229376477 creator A5075657121 @default.
• W4229376477 creator A5076001800 @default.
• W4229376477 creator A5084313763 @default.
• W4229376477 date "2022-05-09" @default.
• W4229376477 modified "2023-10-09" @default.
• W4229376477 title "Deep learning based clinical target volumes contouring and model adaptation for prostate cancer: easy and efficient application" @default.
• W4229376477 doi "https://doi.org/10.21203/rs.3.rs-1586096/v1" @default.
• W4229376477 hasPublicationYear "2022" @default.
• W4229376477 type Work @default.
• W4229376477 citedByCount "0" @default.
• W4229376477 crossrefType "posted-content" @default.
• W4229376477 hasAuthorship W4229376477A5006822602 @default.
• W4229376477 hasAuthorship W4229376477A5017921159 @default.
• W4229376477 hasAuthorship W4229376477A5018091164 @default.
• W4229376477 hasAuthorship W4229376477A5042385997 @default.
• W4229376477 hasAuthorship W4229376477A5074803923 @default.
• W4229376477 hasAuthorship W4229376477A5075657121 @default.
• W4229376477 hasAuthorship W4229376477A5076001800 @default.
• W4229376477 hasAuthorship W4229376477A5084313763 @default.
• W4229376477 hasBestOaLocation W42293764771 @default.
• W4229376477 hasConcept C121608353 @default.
• W4229376477 hasConcept C121684516 @default.
• W4229376477 hasConcept C126322002 @default.
• W4229376477 hasConcept C126838900 @default.
• W4229376477 hasConcept C19527891 @default.
• W4229376477 hasConcept C2776235491 @default.
• W4229376477 hasConcept C2779104521 @default.
• W4229376477 hasConcept C2779466945 @default.
• W4229376477 hasConcept C2780192828 @default.
• W4229376477 hasConcept C2780849966 @default.
• W4229376477 hasConcept C2780920918 @default.
• W4229376477 hasConcept C41008148 @default.
• W4229376477 hasConcept C509974204 @default.
• W4229376477 hasConcept C71924100 @default.
• W4229376477 hasConceptScore W4229376477C121608353 @default.
• W4229376477 hasConceptScore W4229376477C121684516 @default.
• W4229376477 hasConceptScore W4229376477C126322002 @default.
• W4229376477 hasConceptScore W4229376477C126838900 @default.
• W4229376477 hasConceptScore W4229376477C19527891 @default.
• W4229376477 hasConceptScore W4229376477C2776235491 @default.
• W4229376477 hasConceptScore W4229376477C2779104521 @default.
• W4229376477 hasConceptScore W4229376477C2779466945 @default.
• W4229376477 hasConceptScore W4229376477C2780192828 @default.
• W4229376477 hasConceptScore W4229376477C2780849966 @default.
• W4229376477 hasConceptScore W4229376477C2780920918 @default.
• W4229376477 hasConceptScore W4229376477C41008148 @default.
• W4229376477 hasConceptScore W4229376477C509974204 @default.
• W4229376477 hasConceptScore W4229376477C71924100 @default.
• W4229376477 hasLocation W42293764771 @default.
• W4229376477 hasOpenAccess W4229376477 @default.
• W4229376477 hasPrimaryLocation W42293764771 @default.
• W4229376477 hasRelatedWork W108895284 @default.
• W4229376477 hasRelatedWork W1983636976 @default.
• W4229376477 hasRelatedWork W2088520467 @default.
• W4229376477 hasRelatedWork W2125193799 @default.
• W4229376477 hasRelatedWork W2175808153 @default.
• W4229376477 hasRelatedWork W2271751063 @default.
• W4229376477 hasRelatedWork W2305959020 @default.
• W4229376477 hasRelatedWork W3093990103 @default.
• W4229376477 hasRelatedWork W3187823854 @default.
• W4229376477 hasRelatedWork W4246083279 @default.
• W4229376477 isParatext "false" @default.
• W4229376477 isRetracted "false" @default.
• W4229376477 workType "article" @default.