Robert Poel1,2, Elias Rüfenacht3, Ekin Ermis4, Michael Müller3, Michael K Fix5, Daniel M Aebersold4, Peter Manser5, Mauricio Reyes3. 1. Department of Radiation Oncology, Inselspital, Bern University Hospital, University of Bern, Freiburgstrasse 18, 3010, Bern, Switzerland. Robert.poel@insel.ch. 2. ARTORG Center for Biomedical Research, University of Bern, Bern, Switzerland. Robert.poel@insel.ch. 3. ARTORG Center for Biomedical Research, University of Bern, Bern, Switzerland. 4. Department of Radiation Oncology, Inselspital, Bern University Hospital, University of Bern, Freiburgstrasse 18, 3010, Bern, Switzerland. 5. Division of Medical Radiation Physics and Department of Radiation Oncology, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland.
Abstract
AIMS: To save time and have more consistent contours, fully automatic segmentation of targets and organs at risk (OAR) is a valuable asset in radiotherapy. Though current deep learning (DL) based models are on par with manual contouring, they are not perfect and typical errors, as false positives, occur frequently and unpredictably. While it is possible to solve this for OARs, it is far from straightforward for target structures. In order to tackle this problem, in this study, we analyzed the occurrence and the possible dose effects of automated delineation outliers. METHODS: First, a set of controlled experiments on synthetically generated outliers on the CT of a glioblastoma (GBM) patient was performed. We analyzed the dosimetric impact on outliers with different location, shape, absolute size and relative size to the main target, resulting in 61 simulated scenarios. Second, multiple segmentation models where trained on a U-Net network based on 80 training sets consisting of GBM cases with annotated gross tumor volume (GTV) and edema structures. On 20 test cases, 5 different trained models and a majority voting method were used to predict the GTV and edema. The amount of outliers on the predictions were determined, as well as their size and distance from the actual target. RESULTS: We found that plans containing outliers result in an increased dose to healthy brain tissue. The extent of the dose effect is dependent on the relative size, location and the distance to the main targets and involved OARs. Generally, the larger the absolute outlier volume and the distance to the target the higher the potential dose effect. For 120 predicted GTV and edema structures, we found 1887 outliers. After construction of the planning treatment volume (PTV), 137 outliers remained with a mean distance to the target of 38.5 ± 5.0 mm and a mean size of 1010.8 ± 95.6 mm3. We also found that majority voting of DL results is capable to reduce outliers. CONCLUSIONS: This study shows that there is a severe risk of false positive outliers in current DL predictions of target structures. Additionally, these errors will have an evident detrimental impact on the dose and therefore could affect treatment outcome.
AIMS: To save time and have more consistent contours, fully automatic segmentation of targets and organs at risk (OAR) is a valuable asset in radiotherapy. Though current deep learning (DL) based models are on par with manual contouring, they are not perfect and typical errors, as false positives, occur frequently and unpredictably. While it is possible to solve this for OARs, it is far from straightforward for target structures. In order to tackle this problem, in this study, we analyzed the occurrence and the possible dose effects of automated delineation outliers. METHODS: First, a set of controlled experiments on synthetically generated outliers on the CT of a glioblastoma (GBM) patient was performed. We analyzed the dosimetric impact on outliers with different location, shape, absolute size and relative size to the main target, resulting in 61 simulated scenarios. Second, multiple segmentation models where trained on a U-Net network based on 80 training sets consisting of GBM cases with annotated gross tumor volume (GTV) and edema structures. On 20 test cases, 5 different trained models and a majority voting method were used to predict the GTV and edema. The amount of outliers on the predictions were determined, as well as their size and distance from the actual target. RESULTS: We found that plans containing outliers result in an increased dose to healthy brain tissue. The extent of the dose effect is dependent on the relative size, location and the distance to the main targets and involved OARs. Generally, the larger the absolute outlier volume and the distance to the target the higher the potential dose effect. For 120 predicted GTV and edema structures, we found 1887 outliers. After construction of the planning treatment volume (PTV), 137 outliers remained with a mean distance to the target of 38.5 ± 5.0 mm and a mean size of 1010.8 ± 95.6 mm3. We also found that majority voting of DL results is capable to reduce outliers. CONCLUSIONS: This study shows that there is a severe risk of false positive outliers in current DL predictions of target structures. Additionally, these errors will have an evident detrimental impact on the dose and therefore could affect treatment outcome.
Authors: David N Teguh; Peter C Levendag; Peter W J Voet; Abrahim Al-Mamgani; Xiao Han; Theresa K Wolf; Lyndon S Hibbard; Peter Nowak; Hafid Akhiat; Maarten L P Dirkx; Ben J M Heijmen; Mischa S Hoogeman Journal: Int J Radiat Oncol Biol Phys Date: 2010-10-06 Impact factor: 7.038
Authors: Amy V Young; Angela Wortham; Iddo Wernick; Andrew Evans; Ronald D Ennis Journal: Int J Radiat Oncol Biol Phys Date: 2011-03-01 Impact factor: 7.038
Authors: Tim Lustberg; Johan van Soest; Mark Gooding; Devis Peressutti; Paul Aljabar; Judith van der Stoep; Wouter van Elmpt; Andre Dekker Journal: Radiother Oncol Date: 2017-12-05 Impact factor: 6.280
Authors: Mingyao Zhu; Karl Bzdusek; Carsten Brink; Jesper Grau Eriksen; Olfred Hansen; Helle Anita Jensen; Hiram A Gay; Wade Thorstad; Joachim Widder; Charlotte L Brouwer; Roel J H M Steenbakkers; Hubertus A M Vanhauten; Jeffrey Q Cao; Gail McBrayne; Salil H Patel; Donald M Cannon; Nicholas Hardcastle; Wolfgang A Tomé; Matthias Guckenberg; Parag J Parikh Journal: Int J Radiat Oncol Biol Phys Date: 2013-11-15 Impact factor: 7.038
Authors: William J Beasley; Alan McWilliam; Adam Aitkenhead; Ranald I Mackay; Carl G Rowbottom Journal: J Appl Clin Med Phys Date: 2016-03-08 Impact factor: 2.102