The impact of MR-based attenuation correction on PET quantitation accuracy is an ongoing cause of concern for advanced brain research with PET/MR. The purpose of this study was to evaluate a new, template-enhanced zero-echo-time attenuation correction method for PET/MR scanners

Multiple studies have shown in the past that PET/MR is particularly well suited for the investigation of brain function, degenerative diseases, and oncology.

In this region, the detailed anatomical information provided by MR complements the high sensitivity and functional targeting capabilities of PET, as well as enabling such enhancements as motion and partial volume effect correction

30 subjects underwent a clinically-indicated 18F-FDG-PET/CT, followed by PET/MR on a GE SIGNA PET/MR. For each patient, a 42-s zero echo time (ZTE) sequence was used to generate two attenuation maps: one with the standard ZTE segmentation-based method; and another with a modification of the method, wherein pre-registered anatomical templates and CT data were used to enhance the segmentation.

CT data were used as gold standard. Reconstructed PET images were qualified visually and quantified in 68 volumes-of-interest using a standardized brain atlas.


Attenuation maps were successfully generated in all cases, without manual intervention or parameter tuning. One patient was excluded from the quantitative analysis due to the presence of multiple brain metastases.

The PET bias with template-enhanced ZTE attenuation correction was measured to be -0.9%?±?0.9%, compared with -1.4%?±?1.1% with regular ZTE attenuation correction. In terms of absolute bias, the new method yielded 1.1%?±?0.7%, compared with 1.6%?±?0.9% with regular ZTE.

Statistically significant bias reduction was obtained in the frontal region (from -2.0% to -1.0%), temporal (from -1.2% to -0.2%), parietal (from -1.9% to -1.1%), occipital (from -2.0% to -1.1%) and insula (from -1.4% to -1.1%).

These results indicate that the co-registration of pre-recorded anatomical templates to ZTE data is feasible in clinical practice and can be effectively used to improve the performance of segmentation-based attenuation correction.