Multi-task learning-based three-dimensional dose distribution prediction for multiple organs in a single model
Guo Futong1, Li Yongbao2, Jia Qiyuan1, Qi Mengke1, Wu Aiqian1, Kong Fantu1, Mai Yanhua, Song Ting1, Zhou Linghong1
1Department of Biomedical Engineering, Southern Medical University, Guangzhou 510515,China; 2Department of Radiation Oncology, Sun Yat-sen University Cancer Center, Guangzhou 510060,China
Abstract：Objective To establish a three-dimensional (3D) dose prediction model, which can predict multiple organs simultaneously in a single model and automatically learn the effect of the geometric anatomical structure on dose distribution.Methods Clinical radiotherapy plans of patients diagnosed with the same type of tumors were collected and retrospectively analyzed. For every plan, each organs at risk (OAR) voxel was regarded as the study sample and its deposited dose was considered as the dosimetric feature. A regularized multi-task learning method than could learn the relationship among different tasks was employed to establish the relationship matrix among tasks and the correlation between geometric structure and dose distribution among organs. In this experiment, the spinal cord, brainstem and bilateral parotids involved in the intensity-modulated radiotherapy (IMRT) plan of 15 nasopharyngeal cancer patients were utilized to establish the multi-organ prediction model. The relative percentage error between the predicted dose of voxel and the clinical planning dose was calculated to assess the feasibility of the model.Results Ten cases receiving IMRT plans were utilized as the training data, and the remaining five cases were used as the test data. The test results demonstrated a higher prediction accuracy and less data demand. And the average voxel dose errors among the spinal cord, brainstem and the left and right parotids were (2.01±0.02)%,(2.65±0.02)%,(2.45±0.02)% and (2.55±0.02)%,respectively. Conclusion The proposed model can accurately predict the dose of multiple organs in a single model and avoid the establishment of multiple single-organ prediction models, laying a solid foundation for patient-specific plan quality control and knowledge-based treatment planning.
Guo Futong,Li Yongbao,Jia Qiyuan et al. Multi-task learning-based three-dimensional dose distribution prediction for multiple organs in a single model[J]. Chinese Journal of Radiation Oncology, 2019, 28(6): 432-437.
[1] Wong FC,Ng AW,Lee VH,et al. Whole-field simultaneous integrated-boost intensity-modulated radiotherapy for patients with nasopharyngeal carcinoma[J].Int J Radiat Oncol Biol Phys,2010,76(1):138-145. DOI:10.1016/j. ijrobp.2009.01.084. [2] Kam MK,Teo PM,Chau RM,et al. Treatment of nasopharyngeal carcinoma with intensity-modulated radiotherapy:the Hong Kong experience[J].Int J Radiat Oncol Biol Phys,2004,60(5):1440-1450.DOI:10.1016/j.ijrobp.2004.05.022. [3] Fredriksson A.Automated improvement of radiation therapy treatment plans by optimization under reference dose constraints[J].Phys Med Biol,2012,57(23):7799-7881.DOI:10.1088/0031-9155/57/23/7799. [4] Moore KL,Brame RS,Low DA,et al. Experience-based quality control of clinical intensity-modulated radiotherapy planning[J].Int J Radiat Oncol Biol Phys,2011,81(2):545-551.DOI:10.1016/j.ijrobp.2010.11.030. [5] Wang H,Dong P,Liu H,et al. Development of an autonomous treatment planning strategy for radiation therapy with effective use of population-based prior data[J].Med Phys,2017,44(2):389-396.DOI:10.1002/mp.12058. [6] Mcintosh C,Purdie TG.Voxel-based dose prediction with multi-patient atlas selection for automated radiotherapy treatment planning[J].Phys Med Biol,2016,62(2):415-431.DOI:10.1088/1361-6560/62/2/415. [7] Skarpman MJ,Sjölund J.Dose-volume histogram prediction using density estimation[J].Phys Med Biol,2015,60(17):6923-6936.DOI:10.1088/0031-9155/60/17/6923. [8] Appenzoller LM,Michalski JM,Thorstad WL,et al. Predicting dose-volume histograms for organs-at-risk in IMRT planning[J].Med Phys,2012,39(12):7446-7461. DOI:10. 1118/ 1.4761864. [9] Shiraishi S,Moore KL.Knowledge-based prediction of three-dimensional dose distributions for external beam radiotherapy[J].Med Phys,2016,43(1):378-387.DOI:10. 1118/1.4938 583. [10] Zhang Y,Yeung DY.A convex Formulation for learning task relationships in multi-task learning[C].Proc the 26th conference on uncertainty in artificial intelligence,2010:733-742. [11] Caruana R.Multitask learning[J].Mach Learn,1997,28(1):41-75.DOI:10.1023/a:1007379 606734. [12] Pong TK,Tseng P,Ji S,et al. Trace norm regularization:reformulations,algorithms,and multi-task learning[J]. Siam J Optimiz,2009,20(6):3465-3489.DOI:10.1137/ 090763184. [13] He D,Kuhn D,Parida L.Novel applications of multitask learning and multiple output regression to multiple genetic trait prediction[J].Bioinformatics,2016,32(12):i37-i43.DOI:10.1093/bioinformatics/btw249. [14] Yan Y,Ricci E,Subramanian R,et al. A Multi-task learning framework for head pose estimation under target motion[J].IEEE Trans Pattern Anal Mach Intell,2016, 38(6):1070-1083.DOI:10.1109/TPAMI.2015.2477843. [15] Zhang Y,Yang Q.An overview of multi-task learning[J]. Natl Sci Rev,2018,5(1):30-43.DOI:10. 1093/nsr/nwx105. [16] Hunt MA,Jackson A,Narayana A,et al. Geometric factors influencing dosimetric sparing of the parotid glands using IMRT[J].Int J Radiat Oncol Biol Phys,2006,66(1):296-304.DOI:10.1016/j.ijrobp.2006.05.028. [17] Lian J,Yuan L,Ge Y,et al. Modeling the dosimetry of organ at risk in head and neck IMRT planning:an intertechnique and interinstitutional study[J].Med Phys,2013,40(12):121704.DOI:10.1118/1.4828788. [18] Saeys Y,Inza I,LarrañrañagaP.A review of feature selection techniques in bioinformatics[J].Bioinformatics,2007, 23(19):2507-2517.DOI:10.1093/bioinformatics/btm344. [19] Li J,Liu H.Challenges of feature selection for big data analytics[J].IEEE Intell Syst,2017,32(2):9-15. DOI:10.1109/MIS.2017.38. [20] Guyon I,Weston J,Barnhill S,et al. Gene selection for cancer classification using support vector machines[J].Mach Learn,2002,46(1-3):389-422. [21] Lee AW,Ng WT,Pan JJ,et al. International guideline for the delineation of the clinical target volumes (CTV) for nasopharyngeal carcinoma[J].Radiother Oncol,2017,126(1):25-36.DOI:10.1016/j.radonc.2017.10.032.
2019, Vol. 28 
