Attenuation correction is a notable challenge associated with simultaneous PET/MRI, particularly in neuroimaging, where sharp boundaries between air and bone volumes exist. This leads to concerns regarding the visual and, more specifically, quantitative accuracy of PET reconstructions for data obtained with PET/MRI. Recently developed techniques can synthesize attenuation maps using only MRI data and are likely adequate for clinical use, however, little work has been conducted to assess their suitability for the dynamic PET studies frequently employed in research to derive physiological information such as binding potential of neuroreceptors in a region. At the same time, existing PET/MRI attenuation correction methods are predicated upon synthesizing CT data, which is not ideal, as CT data are acquired with much lower energy photons than PET data and thus do not optimally reflect the PET attenuation map. We trained a convolutional neural network to generate patient-specific transmission data from T1-weighted MRI. Using the trained network, we generated transmission data for a testing set comprising 11 subjects scanned withC-WAY-100635 and 10 subjects scanned with C-DASB. We assessed both static and dynamic reconstructions. For dynamic PET data, we report differences in BPND and BPF forC-WAY-100635 and VT for C-DASB. The mean bias for generated transmission data was -1.06±0.81%. Global biases in static PET uptake were -0.49±1.7%, and -1.52±0.73% for C-WAY-100635 andC-DASB, respectively. Our neural network approach is capable of synthesizing patient-specific transmission data with sufficient accuracy for both static and dynamic PET studies.
Background: Bone marrow fat increases when bone mass decreases, which could be attributed to the fact that adipogenesis competes with osteogenesis. Bone marrow fat has the potential to predict abnormal bone density and osteoporosis. Purpose: To investigate the predictive value of using vertebral bone marrow fat fraction(BMFF) obtained from modified Dixon(mDixon) Quant in the determination of abnormal bone density and osteoporosis. Study type: Prospective. Population: 257 subjects (age: 20-79 years old; BMI: 16.6-32.9 kg/m 2 ;181 females,76 males) without known spinal tumor, history of trauma, dysplasia, spinal surgery or hormone therapy. Field strength/sequence: 3.0T/mDixon. Assessment: BMFF was measured at the L1, L2 and L3 vertebral body on fat fraction maps of the lumbar spine. Bone mineral density (BMD) was obtained using quantitative computed tomography, which served as the reference standard. Statistical tests: The BMFF between the three groups (normal bone density, osteopenia and osteoporosis) was tested using one-way analysis of variance in SPSS. The correlation and partial correlation of BMFF and BMD were analyzed before and after controlling for age, sex and BMI. Logistic regression analysis using independent training and validation data was conducted to evaluate the performance of predicting abnormal BMD or osteoporosis using BMFF. Results: There was a significant difference in vertebral BMFF between the three groups (P < 0.001). Moderate inverse correlation was found between vertebral BMFF and BMD after controlling age, sex and BMI (r = -0.529; P < 0.001). The mean area under the curve, sensitivity, specificity and negative predictive value (NPV) for predicting abnormal bone density were 0.940, 0.877, 0.896, and 0.890, respectively. The corresponding results for predicting subjects with osteoporosis were 0.896, 0.848, 0.853, and 0.969, respectively. Data conclusion: mDixon Quant is a fast, simple, noninvasive and nonionizing method to access vertebral BMFF and has a high predictive power for identifying abnormal bone density and osteoporosis. Level of Evidence: 1 Technical Efficacy: Stage 2
Purpose
Radiomics allows for powerful data‐mining and feature extraction techniques to guide clinical decision making. Image segmentation is a necessary step in such pipelines and different techniques can significantly affect results. We demonstrate that a convolutional neural network (CNN) segmentation method performs comparably to expert manual segmentations in an established radiomics pipeline.
Methods
Using the manual regions of interest (ROIs) of an expert radiologist (R1), a CNN was trained to segment breast lesions from dynamic contrast‐enhanced MRI (DCE‐MRI). Following network training, we segmented lesions for the testing set of a previously established radiomics pipeline for predicting lymph node metastases using DCE‐MRI of breast cancer. Prediction accuracy of CNN segmentations relative to manual segmentations by R1 from the original study, a resident (R2), and another expert radiologist (R3) were determined. We then retrained the CNN and radiomics model using R3’s manual segmentations to determine the effects of different expert observers on end‐to‐end prediction.
Results
Using R1’s ROIs, the CNN achieved a mean Dice coefficient of 0.71 ± 0.16 in the testing set. When input to our previously published radiomics pipeline, these CNN segmentations achieved comparable prediction performance to R1’s manual ROIs, and superior performance to those of the other radiologists. Similar results were seen when training the CNN and radiomics model using R3’s ROIs.
Conclusion
A CNN architecture is able to provide DCE‐MRI breast lesion segmentations which are suitable for input to our radiomics model. Moreover, the previously established radiomics model and CNN can be accurately trained end‐to‐end using ground truth data provided by distinct experts.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.