Estimation of the radiation dose in pregnancy: an automated patient-specific model using convolutional neural networks

Abstract

Objectives The conceptus dose during diagnostic imaging procedures for pregnant patients raises health concerns owing to the high radiosensitivity of the developing embryo/fetus. The aim of this work is to develop a methodology for automated construction of patient-specific computational phantoms based on actual patient CT images to enable accurate estimation of conceptus dose. Methods We developed a 3D deep convolutional network algorithm for automated segmentation of CT images to build realistic computational phantoms. The neural network architecture consists of analysis and synthesis paths with four resolution levels each, trained on manually labeled CT scans of six identified anatomical structures. Thirty-two CT exams were augmented to 128 datasets and randomly split into 80%/20% for training/testing. The absorbed doses for six segmented organs/tissues from abdominal CT scans were estimated using Monte Carlo calculations. The resulting radiation doses were then compared between the computational models generated using automated segmentation and manual segmentation, serving as reference. Results The Dice similarity coefficient for identified internal organs between manual segmentation and automated segmentation results varies from 0.92 to 0.98 while the mean Hausdorff distance for the uterus is 16.1 mm. The mean absorbed dose for the uterus is 2.9 mGy whereas the mean organ dose differences between manual and automated segmentation techniques are 0.07%, − 0.45%, − 1.55%, − 0.48%, − 0.12%, and 0.28% for the kidney, liver, lung, skeleton, uterus, and total body, respectively. Conclusion The proposed methodology allows automated construction of realistic computational models that can be exploited to estimate patient-specific organ radiation doses from radiological imaging procedures. Key Points • The conceptus dose during diagnostic radiology and nuclear medicine imaging procedures for pregnant patients raises health concerns owing to the high radiosensitivity of the developing embryo/fetus. • The proposed methodology allows automated construction of realistic computational models that can be exploited to estimate patient-specific organ radiation doses from radiological imaging procedures. • The dosimetric results can be used for the risk-benefit analysis of radiation hazards to conceptus from diagnostic imaging procedures, thus guiding the decision-making process.

Standard SPECT myocardial perfusion estimation from half-time acquisitions using deep convolutional residual neural networks

“…Recent developments in machine/deep learning have successfully introduced a paradigm shift in medical image analysis techniques. A number of studies have assessed the relevance of machine/deep learning in various areas of medical image analysis including disease classification, 9,10 image denoising, 11 resolution recovery, 12 image reconstruction, 13 segmentation, 14,15 and PET attenuation correction. 16,17 Ma et al 18 used a convolutional neural network (CNN) in the diagnosis of different thyroid diseases (Graves' disease, Hashimoto disease, and subacute thyroiditis) using SPECT images.…”

The promise of artificial intelligence and deep learning in PET and SPECT imaging

“…In the context of individualized dose profiling, the construction of patient-specific computational models is the first step toward this goal [155]. Numerous works focused on the development of pipelines for the construction of patient-specific computational models applicable in personalized dosimetry in either therapy or diagnostic procedures [156][157][158]. Fu et al developed a framework for automated generation of computational phantoms from CT images [159].…”

Quantitative Molecular Positron Emission Tomography Imaging Using Advanced Deep Learning Techniques

“…Image segmentation tasks that can be addressed using DL techniques include but are not limited to the development of computational anthropomorphic anatomical models for imaging physics and radiation dosimetry research (84) and segmentation of malignant lesions from PET images (85). A recent study reported an automated DL-guided algorithm for segmentation of CT images of pregnant female patients that was developed specifically for calculation of the radiation dose to the fetus from CT imaging procedures (86). Figure 6 shows 3D coronal and sagittal views of this computational model, segmented using manually and DL-based segmentation techniques.…”