Comparison of integrated PET/MRI with PET/CT in evaluation of endometrial cancer: a retrospective analysis of 81 cases

Patients

This retrospective study was approved by the institutional review board, and the need for patient informed consent was waived (Ethical Application Ref: GH301-18052). Our primary patient selection criteria was pathologically proven diagnoses of endometrial cancer patients who underwent pretreatment ¹⁸F-FDG PET/CT or integrated¹⁸F-FDG PET/MRI for initial staging between April 2013 and May 2018. According to the primary criteria, 81 consecutive patients were selected. Of these, 37 patients underwent the ¹⁸F-FDG PET/CT scanning, and 44 patients underwent the integrated ¹⁸F-FDG PET/MRI scanning.

PET/CT

Whole body PET/CT images were obtained using a PET/CT scanner Biograph 64 PET/CT (Siemens Healthcare Sector, Erlangen, Germany) as previously reported (Stecco et al., 2016). Patients fasted for at least 6 h before tracer injection. After injection of 2.22∼4.44 MBq/kg of ¹⁸F-FDG, PET images were obtained after an approximately 60-min uptake period with the patient’s arms raised to cover the orbitomeatal line to the proximal third of the femurs. After obtaining a scout view (120-140 kVp, 30 mAs), the PET protocol comprised five to six bed positions (3 min each) depending on the patients height. Three-dimensional image reconstructions were acquired using the iterative reconstruction algorithms. The duration of PET/CT acquisition was approximately 20 min.

Integrated PET/MRI

Simultaneous PET/MRI images were obtained using an integrated PET/MRI scanner BiographmMR (Siemens Healthcare Sector, Erlangen, Germany) as previously reported (Schwartz et al., 2018). Patients fasted for at least 6 h before tracer injection. After injection of 2.22∼4.44 MBq/kg of ¹⁸F-FDG, simultaneous PET /MRI scan was conducted. A coronal 3D Dixon volumetric interpolated breath- hold examination (VIBE) T1-weighted imaging sequence in- and out-of -phase was acquired as a template for attenuation correction (AC). The Dixon MR acquisition, used for fat-water separation, was then segmented into four distinct tissue classifications for the MR- based AC map. The following additional MR sequences were used: half-Fourier acquire single-shot turbo spin echo (HASTE) T2 weighted imaging (axial and coronal); axial, coronal and sagittal T2; axial T2 Turbo spin echo (TSE) fat suppressed; diffusion weighted imaging (DWI) using various b values (50/800 s/mm³); 3D axial in and out of phase; axial VIBE pre- and post- contrast and sagittal VIBE post- contrast. PET acquisition was simultaneous with MR acquisition. PET protocol comprised five to six bed positions (5 min each). No gadolinium contrast was used for the MRI portion of the study. The PET/MRI examination lasted approximately 60 min.

Image analysis and standard of reference

Two radiologists/nuclear medicine physicians (five years of experience working in PET/CT and PET/MRI) who were especially experienced in gynecological imaging, consensually and retrospectively evaluated PET/CT and Integrated PET/MRI images for the following findings: (a) presence of the primary tumor, (b) tumor extension into the myometrium, cervical stroma, uterine serosa or adnexa, vagina or parametrium, and urinary bladder or rectum mucosa as well as (c) pelvic lymph nodes (Kitajima et al., 2013). Neither reader was aware of the results of other imaging studies, histopathologic findings or clinical data.

Histopathological correlation was available in all 81 patients and was used as the reference standard.

Statistical analysis

All statistical analyses were performed using IBM SPSS version 23.0 (SPSS Inc, Armonk, NY, USA). The McNemar test was used to determine the statistical significance of differences in the accuracy of staging as determined by PET/CT and Integrated PET/MRI. Compare and analyze the data by the means of Frequencies, Crosstabs and Chi square test. Differences at P < 0.05 were considered to be statistically significant.

Primary tumor detection

Both PET/CT and Integrated PET/MRI detected 100% of the primary tumors (Table 3).

Regional lymph node metastasis and abdominal metastasis detection

The overall accuracy of regional lymph node metastasis detection for PET/CT and Integrated PET/MRI was 86.5% and 95.5%, respectively (Table 3). Of three positive pelvic lymph nodes, PET/CT correctly identified one as positive, with two false negative lesions, resulting in a sensitivity per lesion of 33.3%. Of four positive pelvic lymph nodes, Integrated PET/MRI correctly identified two as positive, with two false negative lesions, resulting in a sensitivity per lesion of 50.0%. Integrated PET/MRI proved more accurate than PET/CT, although the difference was not significant ( P = 0.113). Integrated PET/MRI proved significantly more sensitivity and specificity than PET/CT ( P = 0.015 and P < 0.001, respectively).

The overall accuracy of abdominal metastasis detection for PET/CT and Integrated PET/MRI was 97.3% and 100%, respectively (Table 3). When correlated with histopathology, the only one false positive of PET/CT was liver metastasis. Integrated PET/MRI proved more accurate than PET/CT, although the difference was not significant ( P = 0.081). Integrated PET/MRI proved more specificity than PET/CT.

Deep myometrial invasion detection

PET/CT over-staged the myometrial invasion in six patients (16.2%) and under-staged it in 11 patients (29.7%). Integrated PET/MRI over-staged the myometrial invasion in three patients (6.8%) and under-staged it in5 patients (11.4%). The overall accuracy of myometrial invasion detection for PET/CT and Integrated PET/MRI was 45.9% and 81.8%, respectively (Table 4). Integrated PET/MRI proved significantly more accurate than PET/CT ( P < 0.001).

Two representative cases are shown (Fig. 1).

Limitations

This study has several limitations. First, we did not allow patients to simultaneously test the PET/CT and integrated PET/MRI, but only randomly asked them to assess one method, due to patient compliance and economic reasons. This may increase the test’s interference factor. Furthermore, this study is a retrospective analysis, and the number of patients according to the inclusion criteria is relatively small. Thus, our conclusions can only be considered preliminary. It is imperative to study the differences in the diagnosis of endometrial cancer between PET/CT and integrated PET/MRI using large-scale prospective clinical trials.