Comparative study of total-body PET and PET/MR in the diagnosis of liver metastases.

OBJECTIVE

To compare the diagnostic differences between total-body PET/CT (positron emission tomography/computed tomography) and PET/MR (positron emission tomography/magnetic resonance) in detecting liver metastases.

METHODS

The study analyzed data from patients with malignancies who underwent both conventional total-body PET/CT and liver PET/MR imaging between June 2020 and December 2020. A total of 20 patients with confirmed liver metastases were included, 9 of whom also underwent 2-hour delayed imaging of the liver. Paired t-tests were used to compare the signal-to-noise ratio (SNR) and tissue-to-background ratio (T/B) between PET/MR and conventional total-body PET/CT. Wilcoxon non-parametric tests were used to compare the standardized uptake value maximum (SUVmax) between the two imaging modalities. The McNemar test was employed to assess diagnostic performance differences between PET/MR and conventional total-body PET/CT, as well as between PET/MR and 2-hour delayed total-body PET/CT.

RESULTS

A total of 20 patients with confirmed liver metastases were included, with 39 suspicious lesions identified, and 27 lesions confirmed as liver metastases through biopsy or follow-up. The sensitivity of total-body PET/CT was 66.7% (18/27), while PET/MR had a sensitivity of 96.3% (26/27). The specificity of total-body PET/CT was 83.3% (10/12), and PET/MR had a specificity of 91.7% (11/12). The McNemar test revealed a significant difference in diagnostic performance between the two modalities, with PET/MR outperforming conventional total-body PET/CT (p=0.016). In 9 patients who underwent 2-hour delayed total-body PET/CT, 10 suspicious lesions were identified, 8 of which were confirmed as liver metastases. The sensitivity of delayed total-body PET/CT was 75% (6/8), and PET/MR had a sensitivity of 87.5% (7/8). Both modalities had a specificity of 50% (1/2). The McNemar test for delayed imaging showed no statistically significant difference (p=1). Wilcoxon non-parametric testing showed that the SUVmax of total-body PET/CT was significantly higher than that of PET/MR (Z=-2.355, p=0.019). Paired t-tests indicated no significant differences in SNR (t=-1.565, p=0.156) and T/B ratio (t=-1.689, p=0.115) between the two modalities.

CONCLUSION

Total-body PET/CT demonstrated higher detector sensitivity compared to PET/MR. However, PET/MR showed superior diagnostic performance for detecting liver metastases. The delayed 2-hour PET/CT imaging could partially compensate for the lower diagnostic efficiency of conventional PET/CT compared to PET/MR.