Objective:To achieve the conversion from dual-energy cone-beam CT (DECBCT) at the kilovolt (KV) level to projections at the megavolt (MV) level using an improved CycleGAN network, in order to provide a potential reference benchmark and real-time monitoring of in vivo doses delivered by exit beams for the safe implementation of advanced techniques such as online adaptive radiotherapy. Methods:Simulated patient data were generated using a 4D extended cardiac torso (XCAT) model, and projections were generated based on the geometric parameters of Varian′s onboard cone-beam CT. Furthermore, relative electron density (RED) images were derived from DECBCT images using an iterative dual-energy decomposition algorithm. The SE-CycleGAN and CycleGAN networks were trained to generate MV projection images using DECBCT projections and RED images, respectively. The performance of both methods was evaluated using metrics including structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and root mean square error (RMSE).Results:SE-CycleGAN significantly outperformed CycleGAN in all evaluation metrics ( Z = -23.92, -26.17, -25.54, -26.80, -11.54, -11.21, P<0.05), particularly in learning global information. Besides, although both methods generated satisfactory MV projections, training using DECBCT projections as input yielded better effects than training using RED images. For all the 3 636 sets of projections in the test set, the SE-CycleGAN and CycleGAN networks using DECBCT projections as input respectively yielded SSIMs of 0.997 7±0.000 7 and 0.997 1±0.001 6, PSNRs of 39.625 0±4.684 4 and 36.272 2±5.566 3, and RMSEs of 0.004 1±0.002 7 and 0.006 3±0.0043, respectively. In contrast, the SE-CycleGAN and CycleGAN networks using RED projections as input respectively yielded SSIMs of 0.996 8±0.001 0 and 0.996 2±0.001 5, PSNRs of 38.548 7±3.637 4 and 36.007 3±4.437 8, and RMSEs of 0.004 3±0.002 2 and 0.006 1±0.0037, respectively. Conclusions:This study proposed a new method to establish reference benchmarks for in vivo dose monitoring based on DECBCT and deep learning technologies. This method is accurate and effective according to the preliminary validation using virtual simulation experiments.

Objective:To achieve the conversion from dual-energy cone-beam CT (DECBCT) at the kilovolt (KV) level to projections at the megavolt (MV) level using an improved CycleGAN network, in order to provide a potential reference benchmark and real-time monitoring of in vivo doses delivered by exit beams for the safe implementation of advanced techniques such as online adaptive radiotherapy. Methods:Simulated patient data were generated using a 4D extended cardiac torso (XCAT) model, and projections were generated based on the geometric parameters of Varian′s onboard cone-beam CT. Furthermore, relative electron density (RED) images were derived from DECBCT images using an iterative dual-energy decomposition algorithm. The SE-CycleGAN and CycleGAN networks were trained to generate MV projection images using DECBCT projections and RED images, respectively. The performance of both methods was evaluated using metrics including structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and root mean square error (RMSE).Results:SE-CycleGAN significantly outperformed CycleGAN in all evaluation metrics ( Z = -23.92, -26.17, -25.54, -26.80, -11.54, -11.21, P<0.05), particularly in learning global information. Besides, although both methods generated satisfactory MV projections, training using DECBCT projections as input yielded better effects than training using RED images. For all the 3 636 sets of projections in the test set, the SE-CycleGAN and CycleGAN networks using DECBCT projections as input respectively yielded SSIMs of 0.997 7±0.000 7 and 0.997 1±0.001 6, PSNRs of 39.625 0±4.684 4 and 36.272 2±5.566 3, and RMSEs of 0.004 1±0.002 7 and 0.006 3±0.0043, respectively. In contrast, the SE-CycleGAN and CycleGAN networks using RED projections as input respectively yielded SSIMs of 0.996 8±0.001 0 and 0.996 2±0.001 5, PSNRs of 38.548 7±3.637 4 and 36.007 3±4.437 8, and RMSEs of 0.004 3±0.002 2 and 0.006 1±0.0037, respectively. Conclusions:This study proposed a new method to establish reference benchmarks for in vivo dose monitoring based on DECBCT and deep learning technologies. This method is accurate and effective according to the preliminary validation using virtual simulation experiments.

Objective:To analyze the effects of two decomposition algorithms of dual-energy cone beam CT (DECBCT) (direct decomposition and iterative decomposition) on the image quality and material decomposition accuracy of different sizes of phantoms.Methods:Different sizes of imaging parts of patients were simulated using the combination of CatPhan604 phantoms and customized annuluses. CBCT with high energy of 140 kVp and low energy of 100 kVp were acquired using the Varian Edge CBCT system. Then the material decomposition of DECBCT images was performed using the two algorithms. The electron density (ED) and contrast-to-noise ratio (CNR) of each material in the CTP682 module were calculated. They were used to assess the decomposition accuracy and image quality of the two algorithms.Results:Based on the values in the Catphan604 manual, both algorithms have high ED accuracy. Only the ED accuracy of four materials of the smallest sized phantom showed statistical difference ( z = -4.21, 4.30, 2.87, 5.45, P < 0.05), but the average relative error was less than 1%. The CNR of the iterative decomposition algorithm was significantly higher than that of the direct decomposition, increasing by 51.8%-703.47%. The increase in the phantom size significantly reduced the accuracy of ED, and the increased amplitude of the relative error was up to a maximum of 2.52%. The large phantom size also reduced the image quality of iterative decomposition, and the decreased amplitude of CNR was up to a maximum of 39.71. Conclusions:Compared with the direct decomposition, the iterative decomposition algorithm can significantly reduce the image noise and improve the contrast without losing the accuracy of electron density in the DECBCT construction of different sizes of phantoms.

Objective: To establish comprehensive laboratory reference intervals for Chinese children. Methods: This was a cross-sectional multicenter study. From June 2013 to December 2014, eligible healthy children aged from 6-month to 17-year were enrolled from 20 medical centers with informed consent. They were assessed by physical examination, questionnaire survey and abdominal ultrasound for eligibility. Fasting blood samples were collected and delivered to central laboratory. Measurements of 15 clinical laboratory parameters were performed, including estradiol (E2), testosterone(T), luteinizing hormone(LH), follicle-stimulating hormone(FSH), alanine transaminase(ALT), serum creatinine(Scr), cystatin C, immunoglobulin A(IgA), immunoglobulin G(IgG), immunoglobulin M(IgM), complement (C3, C4), alkaline phosphatase(ALP), uric acid(UA) and creatine kinase(CK). Reference intervals were established according to central 95% confidence intervals for reference population, stratified by age and sex. Results: In total, 2 259 children were enrolled. Finally, 1 648 children were eligible for this study, including 830 boys and 818 girls, at a mean age of 7.4 years. Age- and sex- specific reference intervals have been established for the parameters. Reference intervals of sex hormones increased gradually with age. Concentrations of ALT, cystatin C, ALP and CK were higher in children under 2 years old. Serum levels of sex hormones, creatinine, immunoglobin, CK, ALP and urea increased rapidly in adolescence, with significant sex difference. In addition, reference intervals were variable depending on assay methods. Concentrations of ALT detected by reagents with pyridoxal 5'-phosphate(PLP) were higher than those detected by reagents without PLP. Compared with enzymatic method, Jaffe assay always got higher results of serum creatinine, especially in children younger than 9 years old. Conclusion This study established age- and sex- specific reference intervals, for 15 clinical laboratory parameters based on defined healthy children.