Objective:To utilize radiomics features to enhance the prediction of personalized organ doses from CT scans,in order to explore methods for improving neural network-based models.Methods:Patient CT DICOM files were pro...Objective:To utilize radiomics features to enhance the prediction of personalized organ doses from CT scans,in order to explore methods for improving neural network-based models.Methods:Patient CT DICOM files were processed using DeepViewer to define regions of interest(ROIs)in their organs.Radiomics features were extracted from the CT images and ROIs,and benchmark organ doses were calculated using Monte Carlo simulations.Fully-connected neural networks(FCNN)were trained with radiomics features to predict organ doses.The FCNN model was optimized by adjusting the number of input radiomics features and FCNN layers.Performance was evaluated using relative root mean squared error(RRMSE)and Rsquared(R^(2)).Results:Higher RRMSE and lower R^(2) values are observed when fewer than 30 input radiomics features are used for head CTs and fewer than 10 for chesst CTs.Increasing input features didn't significantly improve FCNN's performance.For head CTs,FCNN's layer quantities affected predictive stability,with better robustness observed with 4-and 5-layer FCNN.Specifically,the median RRMSE was reduced to 8.14%for the brain,10.27%for the left eye,and 10.16%for the right eye when using 30 or more radiomics features.For chest CTs,the model's predictive stability was less sensitive to the number of layers,with median RRMSE values of 9.58%for the left lung and 9.44%for the right lung,and R^(2) values of 0.76 for both lungs.Conclusions:Optimizing feature quantities and neural network layers enhances performance in predicting organ doses from CT scans.Specifically,head CTs show optimal results with 4–5 layers,while chest CTs do not significantly benefit from increased layers.展开更多
During the last few years,active personal dosimeters have been developed and have replaced passive personal dosimeters in some external monitoring systems,frequently using silicon diode detectors.Incident photons inte...During the last few years,active personal dosimeters have been developed and have replaced passive personal dosimeters in some external monitoring systems,frequently using silicon diode detectors.Incident photons interact with the constituents of the diode detector and produce electrons.These photon-induced electrons deposit energy in the detector＇s sensitive region and contribute to the response of diode detectors.To achieve an appropriate photon dosimetry response,the detectors are usually covered by a metallic layer with an optimum thickness.The metallic cover acts as an energy compensating shield.In this paper,a software process is performed for energy compensation.Selective data sampling based on pulse height is used to determine the photon dose equivalent.This method is applied to improve the energy response in photon dosimetry.The detector design is optimized for the response function and determination of the photon dose equivalent.Photon personal dose equivalent is determined in the energy range of 0.3-6 MeV.The error values of the calculated data for this wide energy range and measured data for ^133Ba,^137Cs,^60Co and ^241Am-Be sources respectively are up to 20%and 15%.Fairly good agreement is seen between simulation and dose values obtained from our process and specifications from several photon sources.展开更多
基金supported by the National Natural Science Foundation of China(12075064)the National Key R&D Program of China(2019YFC0117304).
文摘Objective:To utilize radiomics features to enhance the prediction of personalized organ doses from CT scans,in order to explore methods for improving neural network-based models.Methods:Patient CT DICOM files were processed using DeepViewer to define regions of interest(ROIs)in their organs.Radiomics features were extracted from the CT images and ROIs,and benchmark organ doses were calculated using Monte Carlo simulations.Fully-connected neural networks(FCNN)were trained with radiomics features to predict organ doses.The FCNN model was optimized by adjusting the number of input radiomics features and FCNN layers.Performance was evaluated using relative root mean squared error(RRMSE)and Rsquared(R^(2)).Results:Higher RRMSE and lower R^(2) values are observed when fewer than 30 input radiomics features are used for head CTs and fewer than 10 for chesst CTs.Increasing input features didn't significantly improve FCNN's performance.For head CTs,FCNN's layer quantities affected predictive stability,with better robustness observed with 4-and 5-layer FCNN.Specifically,the median RRMSE was reduced to 8.14%for the brain,10.27%for the left eye,and 10.16%for the right eye when using 30 or more radiomics features.For chest CTs,the model's predictive stability was less sensitive to the number of layers,with median RRMSE values of 9.58%for the left lung and 9.44%for the right lung,and R^(2) values of 0.76 for both lungs.Conclusions:Optimizing feature quantities and neural network layers enhances performance in predicting organ doses from CT scans.Specifically,head CTs show optimal results with 4–5 layers,while chest CTs do not significantly benefit from increased layers.
文摘During the last few years,active personal dosimeters have been developed and have replaced passive personal dosimeters in some external monitoring systems,frequently using silicon diode detectors.Incident photons interact with the constituents of the diode detector and produce electrons.These photon-induced electrons deposit energy in the detector＇s sensitive region and contribute to the response of diode detectors.To achieve an appropriate photon dosimetry response,the detectors are usually covered by a metallic layer with an optimum thickness.The metallic cover acts as an energy compensating shield.In this paper,a software process is performed for energy compensation.Selective data sampling based on pulse height is used to determine the photon dose equivalent.This method is applied to improve the energy response in photon dosimetry.The detector design is optimized for the response function and determination of the photon dose equivalent.Photon personal dose equivalent is determined in the energy range of 0.3-6 MeV.The error values of the calculated data for this wide energy range and measured data for ^133Ba,^137Cs,^60Co and ^241Am-Be sources respectively are up to 20%and 15%.Fairly good agreement is seen between simulation and dose values obtained from our process and specifications from several photon sources.
