Diagnostic efficiency of whole-body 18F-FDG PET/MRI, MRI alone, and SUV and ADC values in staging of primary uterine cervical cancer

Abstract

Background: The use of PET/MRI for gynecological cancers is emerging. The purpose of this study was to assess the additional diagnostic value of PET over MRI alone in local and whole-body staging of cervical cancer, and to evaluate the benefit of standardized uptake value (SUV) and apparent diffusion coefficient (ADC) in staging.

Methods: Patients with histopathologically-proven cervical cancer and whole-body ¹⁸F-FDG PET/MRI obtained before definitive treatment were retrospectively registered. Local tumor spread, nodal involvement, and distant metastases were evaluated using PET/MRI or MRI dataset alone. Histopathology or clinical consensus with follow-up imaging were used as reference standard. Tumor SUVmax and ADC were measured and SUVmax/ADC ratio calculated. Area under the curve (AUC) was determined to predict diagnostic performance and Mann-Whitney U test was applied for group comparisons.

Results: In total, 33 patients who underwent surgery (n = 23) or first-line chemoradiation (n = 10) were included. PET/MRI resulted in higher AUC compared with MRI alone in detecting parametrial (0.89 versus 0.73), vaginal (0.85 versus 0.74), and deep cervical stromal invasion (0.96 versus 0.74), respectively. PET/MRI had higher diagnostic confidence than MRI in identifying patients with radical cone biopsy and no residual at hysterectomy (sensitivity 89% versus 44%). PET/MRI and MRI showed equal AUC for pelvic nodal staging (both 0.73), whereas AUC for distant metastases was higher using PET/MRI (0.80 versus 0.67). Tumor SUVmax/ADC ratio, but not SUVmax or ADC alone, was significantly higher in the presence of metastatic pelvic lymph nodes (P < 0.05).

Conclusions: PET/MRI shows higher accuracy than MRI alone for determining local tumor spread and distant metastasis emphasizing the added value of PET over MRI alone in staging of cervical cancer. Tumor SUVmax/ADC ratio may predict pelvic nodal involvement.

Keywords: 3 T MRI; ADC; Cervical cancer; PET/MRI; SUV; Staging.

Fig. 1

A 38-year old woman with stage T0 cervical cancer at hysterectomy that was misinterpreted as stage T1b1 on MRI. Coronal contrast enhanced T1-weighted image shows a hypoenhancing area in the cervix interpreted as a small residual following loop electrosurgical excision procedure (a, white arrow). PET/MRI shows no FDG uptake confirming post-procedural changes without evidence for a residual (b) (patient no. 9 on Table 2)

Fig. 2

A 48-year old woman with stage T1b1 cervical cancer that was misinterpreted as stage T2b on MRI. Axial T2-weighted image shows the cervical cancer with some irregularity of the stromal rim (a, white arrows) interpreted as parametrial invasion by MRI. PET/MRI shows the tumor confined to the cervix with no FDG uptake in areas of stromal irregularities (b). The patient underwent surgical staging including parametrial biopsies and sentinel node dissection with ultrastaging, followed by curative chemoradiation (patient no. 15 on Table 2)

Fig. 3

A 60-year old woman with a locally advanced disease and a small FDG avid liver metastasis in segment IV (white arrows). Whole-body PET/MRI images (with T1-weighted 3D water Dixon MRI) in axial (a) and coronal (b) plane

Fig. 4

SUVmax/ADC ratio was significantly higher (P < 0.05) among patients with metastatic pelvic lymph nodes (stage N1) compared with patients without regional nodal involvement (stage N0). The difference was not significant when using SUVmax or ADC alone