Evaluation of Deep Learning–Based Approaches to Segment Bowel Air Pockets and Generate Pelvic Attenuation Maps from CAIPIRINHA-Accelerated Dixon MR Images

• Published in: Journal of Nuclear Medicine Vol. 63; no. 3; pp. 468 - 475
• Main Authors: Sari, Hasan, Reaungamornrat, Ja, Catalano, Onofrio A., Vera-Olmos, Javier, Izquierdo-Garcia, David, Morales, Manuel A., Torrado-Carvajal, Angel, Ng, Thomas S.C., Malpica, Norberto, Kamen, Ali, Catana, Ciprian
• Format: Journal Article
• Language: English
• Published: United States Society of Nuclear Medicine 01.03.2022
• Subjects: Air pockets; Artificial neural networks; Attenuation; Clinical Investigation; Computed tomography; Deep Learning; Gastrointestinal tract; Humans; Image processing; Image Processing, Computer-Assisted - methods; Image reconstruction; Image segmentation; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Medical imaging; Neural networks; Pelvis; Pelvis - diagnostic imaging; Positron emission; Positron emission tomography; Positron-Emission Tomography - methods; Synthesis; Tomography; Tomography, X-Ray Computed; pseudo-CT; deep learning; PET/MRI; attenuation correction; PET quantification
• DOI: 10.2967/jnumed.120.261032
• ISSN: 0161-5505; 1535-5667

Attenuation correction remains a challenge in pelvic PET/MRI. In addition to the segmentation/model-based approaches, deep learning methods have shown promise in synthesizing accurate pelvic attenuation maps (μ-maps). However, these methods often misclassify air pockets in the digestive tract, potentially introducing bias in the reconstructed PET images. The aims of this work were to develop deep learning-based methods to automatically segment air pockets and generate pseudo-CT images from CAIPIRINHA-accelerated MR Dixon images. A convolutional neural network (CNN) was trained to segment air pockets using 3-dimensional CAIPIRINHA-accelerated MR Dixon datasets from 35 subjects and was evaluated against semiautomated segmentations. A separate CNN was trained to synthesize pseudo-CT μ-maps from the Dixon images. Its accuracy was evaluated by comparing the deep learning-, model-, and CT-based μ-maps using data from 30 of the subjects. Finally, the impact of different μ-maps and air pocket segmentation methods on the PET quantification was investigated. Air pockets segmented using the CNN agreed well with semiautomated segmentations, with a mean Dice similarity coefficient of 0.75. The volumetric similarity score between 2 segmentations was 0.85 ± 0.14. The mean absolute relative changes with respect to the CT-based μ-maps were 2.6% and 5.1% in the whole pelvis for the deep learning-based and model-based μ-maps, respectively. The average relative change between PET images reconstructed with deep learning-based and CT-based μ-maps was 2.6%. We developed a deep learning-based method to automatically segment air pockets from CAIPIRINHA-accelerated Dixon images, with accuracy comparable to that of semiautomatic segmentations. The μ-maps synthesized using a deep learning-based method from CAIPIRINHA-accelerated Dixon images were more accurate than those generated with the model-based approach available on integrated PET/MRI scanners.