A deep learning approach to radiation dose estimation

PURPOSE Currently methods for predicting absorbed dose after administering a radiopharmaceutical are rather crude in daily clinical practice. Most importantly individual tissue density distributions as well as local variations of the concentration of the radiopharmaceutical are commonly neglected. The current study proposes machine learning techniques like Green's function-based empirical mode decomposition and deep learning methods on U-net architectures in conjunction with soft tissue kernel Monte Carlo (MC) simulations to overcome current limitations in precision and reliability of dose estimations for clinical dosimetric applications. METHODS We present a hybrid method (DNN-EMD) based on deep neural networks (DNN) in combination with empirical mode decomposition (EMD) techniques. The algorithm receives Computed Tomography (CT) tissue density maps and dose maps, estimated according to the MIRD protocol, i.e., employing whole organ S-values and related time-integrated activities (TIAs), and from measured SPECT distributions of 177-Lu radionuclei, and learns to predict individual absorbed dose distributions. Density maps are replaced by their intrinsic modes as deduced from an EMD analysis. The system is trained using individual full MC simulations as reference. RESULTS Data from a patient cohort of 26 subjects are reported in this study. The proposed methods were validated employing a leave-one-out cross-validation technique. Deviations of estimated dose from corresponding MC results corroborate a superior performance of the newly proposed hybrid DNN-EMD method compared to its related MIRD DVK dose calculation. Not only are the mean deviations much smaller with the new method, but also the related variances are much reduced. If intrinsic modes of the tissue density maps are input to the algorithm, variances become even further reduced though the mean deviations are less affected. CONCLUSION The newly proposed hybrid DNN-EMD method for individualized radiation dose prediction outperforms the MIRD DVK dose calculation method. It is fast enough to be of use in daily clinical practice.

[1]  Ke Li,et al.  Low-dose CT denoising with convolutional neural network , 2016, 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017).

[2]  Elmar Wolfgang Lang,et al.  A green's function-based Bi-dimensional empirical mode decomposition , 2016, Inf. Sci..

[3]  B. Wessels,et al.  Patient-specific scaling of reference S-values for cross-organ radionuclide S-values: what is appropriate? , 2007, Radiation protection dosimetry.

[4]  A. Maier,et al.  A comparison of methods for adapting 177Lu dose-voxel-kernels to tissue inhomogeneities , 2019, Physics in medicine and biology.

[5]  Otto Muzik,et al.  Corrigendum: Hybrid Imaging: Instrumentation and Data Processing , 2018, Front. Phys..

[6]  Mauro Valente,et al.  Dosimetry for Beta-Emitter Radionuclides by Means of Monte Carlo Simulations , 2011 .

[7]  N. Huang,et al.  The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis , 1998, Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[8]  Jae Sung Lee,et al.  Computed tomography super-resolution using deep convolutional neural network , 2018, Physics in medicine and biology.

[9]  Georgios Kalantzis,et al.  Toward IMRT 2D dose modeling using artificial neural networks: a feasibility study. , 2011, Medical physics.

[10]  J. Hornegger,et al.  Quantitative SPECT/CT Imaging of 177Lu with In Vivo Validation in Patients Undergoing Peptide Receptor Radionuclide Therapy , 2015, Molecular Imaging and Biology.

[11]  Vasant Kearney,et al.  DoseNet: a volumetric dose prediction algorithm using 3D fully-convolutional neural networks , 2018, Physics in medicine and biology.

[12]  Tomas Kron,et al.  Deep Learning Renal Segmentation for Fully Automated Radiation Dose Estimation in Unsealed Source Therapy , 2018, Front. Oncol..

[13]  T. Bortfeld,et al.  Correlation between CT numbers and tissue parameters needed for Monte Carlo simulations of clinical dose distributions. , 2000, Physics in medicine and biology.

[14]  M. Kolstein,et al.  Gamos: A framework to do Geant4 simulations in different physics fields with an user-friendly interface , 2014 .

[15]  Min Sun Lee,et al.  Deep-dose: a voxel dose estimation method using deep convolutional neural network for personalized internal dosimetry , 2019, Scientific Reports.

[16]  Ryan Whalen,et al.  Priority Queuing on the Docket: Universality of Judicial Dispute Resolution Timing , 2018, Front. Phys..

[17]  A Mairani,et al.  Calculation of electron and isotopes dose point kernels with FLUKA Monte Carlo code for dosimetry in nuclear medicine therapy. , 2011, Medical physics.

[18]  Elmar Wolfgang Lang,et al.  Sliding Empirical Mode Decomposition for On-line Analysis of Biomedical Time Series , 2011, IWANN.

[19]  George Loudos,et al.  A dose point kernel database using GATE Monte Carlo simulation toolkit for nuclear medicine applications: comparison with other Monte Carlo codes. , 2012, Medical physics.

[20]  Steve B. Jiang,et al.  Dose Prediction with U-net: A Feasibility Study for Predicting Dose Distributions from Contours using Deep Learning on Prostate IMRT Patients , 2017, ArXiv.

[21]  Gilmer Valdes,et al.  An unsupervised convolutional neural network-based algorithm for deformable image registration , 2018, Physics in medicine and biology.

[22]  N. Lanconelli,et al.  Differences in 3D dose distributions due to calculation method of voxel S-values and the influence of image blurring in SPECT , 2015, Physics in medicine and biology.

[23]  Theresia I. Götz,et al.  Factors affecting accuracy of S values and determination of time-integrated activity in clinical Lu-177 dosimetry , 2019, Annals of Nuclear Medicine.

[24]  T. Solberg,et al.  A Multi-Task CNN Model for Autosegmentation of Prostate Patients , 2018, International Journal of Radiation Oncology*Biology*Physics.

[25]  W E Bolch,et al.  MIRD pamphlet No. 17: the dosimetry of nonuniform activity distributions--radionuclide S values at the voxel level. Medical Internal Radiation Dose Committee. , 1999, Journal of nuclear medicine : official publication, Society of Nuclear Medicine.

[26]  T. Zhuang,et al.  An image-based deep learning framework for individualising radiotherapy dose: a retrospective analysis of outcome prediction , 2019, The Lancet Digital Health.

[27]  Norden E. Huang,et al.  The Multi-Dimensional Ensemble Empirical Mode Decomposition Method , 2009, Adv. Data Sci. Adapt. Anal..