Diagnostic Value of 18F-FDG PET/MRI for Revised 2018 FIGO Staging in Patients with Cervical Cancer

Purpose: To evaluate the diagnostic potential of PET/MRI with 18F-fluorodeoxyglucose (18F-FDG) in cervical cancer based on the revised 2018 International Federation of Gynecology and Obstetrics (FIGO) staging system. Materials and Methods: Seventy-two patients with biopsy-proven primary cervical cancer underwent pretreatment 18F-FDG PET/MRI, CT, and pelvic MRI. The diagnostic performance of 18F-FDG PET/MRI and MRI for assessing extent of the primary tumor and 18F-FDG PET/MRI and CT for assessing nodal and distant metastases was evaluated by two experienced readers. Histopathological and follow-up imaging results were used as the gold standard. McNemar test was employed for statistical analysis. Results: Accuracy for the invasion of vagina, parametrium, side wall, and adjacent organs was 97.2%, 93.1%, 97.2%, and 100% for 18F-FDG PET/MRI; and 97.2%, 91.7%, 97.2%, and 100% for pelvic MRI, respectively (p > 0.05). Patient-based accuracy for metastasis to pelvic and paraaortic lymph nodes and distant organs was 95.8%, 98.6%, and 100% for 18F-FDG PET/MRI; and 83.3%, 95.8%, and 97.2% for CT, respectively; metastasis to pelvic lymph nodes was statistically significant (p < 0.01). Lesion-based sensitivity, specificity, and accuracy for lymph nodes were 83.3%, 95.9%, and 94.8% for 18F-FDG PET/MRI; and 29.2%, 98.9% and 93.1% for CT, respectively; sensitivity was statistically significant (p < 0.001). After excluding patients diagnosed by conization, accuracy for revised FIGO staging 2018 was significantly better for 18F-FDG PET/MRI (82.1%) than for CT and MRI (60.7%) (p < 0.01). Conclusions: 18F-FDG PET/MRI offers higher diagnostic value for revised 2018 FIGO staging, suggesting that 18F-FDG PET/MRI might provide an optimal diagnostic strategy for preoperative staging.


Introduction
Cervical cancer is the fourth most common cancer in women worldwide, with more than half a million cases diagnosed annually, even though prophylactic vaccines against human papillomavirus have reduced the occurrence of this disease [1]. Previously, cervical cancer had been clinically staged based on the 2009 International Federation of Gynecology and Obstetrics (FIGO) classifications [2,3]. Although preoperative assessment of disease is essential, the 2009 FIGO system did not yet include descriptions of imaging findings other than hydronephrosis and distant metastasis. Moreover, no information was included regarding lymph node involvement, which is associated with poor prognosis. The 2018 FIGO system highlighted the utility of imaging and permitted its use, when available, as part of clinical staging [4,5]. Compared with the 2009 classifications, the revised system better recognizes metastatic or recurrence risk by including a greater number of tumor size subdivisions (IB1, IB2, and IB3) and by taking into account the status of regional lymph nodes detected radiographically or pathologically (IIIC1 and IIIC2). Therefore, the revised system has afforded a greater importance to imaging findings in the planning of optimal treatment. Cervical cytology and biopsy are the standard methods internationally for detecting and diagnosing disease. Imaging is less useful in screening and for evaluation of microinvasion, particularly in early-stage cancers, such as FIGO IA and part of IB1, which is detectable only by microscopy. When we assess local disease by including tumor size (IB1, IB2, and IB3), vaginal and parametrial invasion (IIA, IIB, IIIA, and IIIB), and the exclusion of bladder and rectal invasion (IVA), MRI provides higher sensitivity and comparable specificity compared with clinical assessment and could be one of the most reliable imaging modalities for assessing local extension in cervical cancer [6][7][8][9].
Lymph node metastasis was not included in the former FIGO classifications, although it is strongly associated with poor prognosis and 10-30% of patients have lymph node metastases on histology even in early cervical cancer [10]. In the revised 2018 FIGO staging system, the presence of regional lymph node metastases is designated as stage IIIC, suggesting the importance of accurate detection by the most suitable imaging modality. CT has been widely used for this purpose because of its cost-and time-effectiveness, whereas 18 F-fluorodeoxyglucose ( 18 F-FDG) PET/CT and MRI each have greater sensitivity and specificity than CT for detecting node metastasis. Moreover, the sensitivity and specificity of 18 F-FDG PET/CT for aortic node metastasis are greater than those for MRI and CT, suggesting that 18 F-FDG PET/CT is a valid alternative for detection of lymph node metastasis in cervical cancer [11,12].
The new PET modality of 18 F-FDG PET/MRI provides high soft-tissue contrast along with the functional imaging of FDG uptake, and has shown potentially better diagnostic performance compared with conventional imaging. Integrated PET/MRI provides comparable diagnostic value in evaluating local extent and better sensitivity and specificity in the detection of nodal metastasis compared with MRI in cervical cancer [13][14][15], which suggests that integrated PET/MRI may have a critical role to play in preoperative diagnosis based on the revised 2018 FIGO staging system for cervical cancer.
The aim of our study is, thus, to evaluate the diagnostic utility of integrated 18 F-FDG PET/MRI for the revised 2018 FIGO staging in cervical cancer, and to compare the diagnostic accuracy of integrated 18 F-FDG PET/MRI with those of CT and MRI.

Patients
We retrospectively reviewed the medical records of 72 patients (mean age, 53.0 years; age range, 29-91 years) with biopsy-proven primary cervical cancer. They had undergone 18 F-FDG PET/MRI, CT, and pelvic MRI with obtained informed consent for the initial staging based on the Japanese Imaging Guidelines of the Japan Radiological Society between November 2015 and June 2020. Patients had completed 18 F-FDG PET/MRI, CT, and MRI within 3 months prior to treatment. The maximum interval among 18 F-FDG PET/MRI, CT, and MRI was 109 days (mean, 7.6 days; range, 0-109 days). Two patients underwent conization only, five patients underwent simple hysterectomy, and 39 underwent radical hysterectomy; 39 underwent pelvic and three underwent pelvic and para-aortic lymphadenectomy. Six patients underwent neoadjuvant chemotherapy, 18 underwent definitive radiotherapy with or without chemotherapy, and two refused definitive treatments. This was a multi-center study, as 27 patients with data from CT and/or MRI were referred from other institutions. All patients underwent 18  Fifty minutes after injection, patients were transferred to a whole-body 3.0-T PET/MR scanner (Signa PET/MR; GE Healthcare, Waukesha, WI, USA). Anatomical coverage was from the vertex to the mid-thigh. PET acquisition was performed in three-dimensional (3D) mode with 5.5 min/bed position (89 slices/bed) in 5-6 beds with a 24-slice overlap. A two-point Dixon 3D volumetric interpolated T1-weighted fast spoiled gradient echo sequence was acquired at each table position and was used to generate MR attenuation correction (MR-AC) maps. Dixon-based MR-AC classifies body tissues into soft tissue, fat, and air. PET data were reconstructed by ordered subset expectation maximization (OSEM), selecting 14 subsets and three iterations, and post-smoothing with a 3-mm Gaussian filter. Reconstructed images were then converted to semiquantitative images corrected by the injected dose and the bodyweight of the subject as the standardized uptake value (SUV).

Pelvic PET/MRI
After whole-body scanning and a brief break for urination, the patient was repositioned in the PET/MR scanner. The pelvic PET scan was performed as a 3D acquisition in list mode with 15 min/bed position (89 slices/bed) in 1-2 beds with a 24-slice overlap. Regional PET data were reconstructed with OSEM selecting 16 subsets and four iterations, and post-smoothing with a 4-mm Gaussian filter. The reconstructed images were then converted to SUV images. For pelvic MRI, T2-weighted images were acquired in the sagittal, transaxial, and coronal planes, using the following T2-weighted image parameters: TR, 4000-7000 ms; TE, 146 ms; section thickness, 4 mm; section overlap, 0 mm; flip angle, 100 • ; FOV, 240 × 240 mm; matrix, 384 × 384; two excitations; and bandwidth, 83.3 kHz.

CT
CT examinations covering the chest, abdomen, and pelvis were performed using a 64-slice multidetector CT scanner (Discovery CT 750HD; GE Medical Systems, Milwaukee, WI, USA).

Image Interpretation
Images were analyzed on a dedicated workstation (Advantage Workstation 4.6; GE Healthcare). Two board-certificated radiologists/nuclear medicine physicians, each with double certifications and specializing in gynecological imaging, evaluated the 18 F-FDG PET/MRI, CT, and MRI images retrospectively and reached consensus decisions. Images were evaluated for the following: (a) presence of the primary tumor; (b) tumor invasion into the vagina (IIA or IIIA); (c) tumor invasion into the parametrium (IIIA); (d) tumor extension to the pelvic wall (IIIB); (e) pelvic lymph node metastasis (IIIC1); (f) para-aortic lymph node metastasis (IIIC2); (g) tumor extension to adjacent organs such as the bladder or rectum (IVA); and (h) distant metastasis (IVB). The diagnostic performance of 18 F-FDG PET/MRI and MRI for assessing the extent of the primary tumor and that of 18 F-FDG PET/MRI and CT for assessing nodal and distant metastases was evaluated. Both readers were blinded to the results of other imaging studies, the histopathological findings, and clinical data. Each dataset was reviewed as the consensus decision of the two readers after a minimum interval of three weeks to avoid any decision threshold bias due to readingorder effects. For CT and MRI interpretation, several previous standard criteria related to primary tumor and nodal or distant metastatic staging of cervical cancer were used as the reference criteria [6,11]. Swollen lymph nodes larger than 1 cm in short-axis diameter were graded as malignant. For 18 F-FDG PET/MRI interpretations, the classification of lymph nodes as cancer-positive was based on the presence of focally appreciable metabolic activity above that of normal muscle; or asymmetric metabolic activity greater than that of normal-appearing lymph nodes at the same level in the contralateral pelvis, in a location corresponding to the lymph node chains on CT or MRI images, with reference to previous reports [13,14]. Tumor invasion of neighboring structures was decided primarily on the basis of the CT or MRI findings, with reference to the 18 F-FDG PET findings.

Reference Standard
Histopathological results were used as the standard of reference. Because clinical and ethical standards of patient management do not require surgery or sampling of all detected lesions, a modified reference standard was used for lesions without histopathological sampling to take into account all prior and follow-up imaging. A decrease in size and/or SUVmax under therapy or an increase in size and/or SUVmax without therapy was regarded as a sign of malignancy. PET-negative and inconspicuous lesions with constant size were rated as benign.

Statistical Analysis
The McNemar test was used to determine the statistical significance of differences in the accuracy of staging as determined by PET/MRI, CT, and MRI. Statistical analysis was performed using PRISM version 6.0 software (GraphPad, San Diego, CA, USA). Differences at the level of p < 0.05 were considered statistically significant.

Discussion
To the best of our knowledge, this is the first study to investigate the diagnostic value of 18 F-FDG PET/MRI for cervical cancer based on the revised 2018 FIGO staging system in comparison with the conventional imaging modalities of MRI and CT. 18 F-PET/MRI offered significantly superior accuracy to MRI and CT for the revised 2018 FIGO staging of cervical cancer. The accuracy of 18 F-PET/MRI for metastasis to pelvic lymph nodes was significantly superior to that of CT, whereas the accuracy of 18 F-PET/MRI was equivalent to those of MRI and CT for the detection of local extent and metastasis to distant organs. These findings suggest that 18 F-FDG PET/MRI might provide a useful alternative to conventional imaging modalities in staging cervical cancer.
Cervical cancer had previously been clinically staged based on the 2009 FIGO staging system, which did not include imaging in the staging. Unlike in other cancers such as those of the uterine endometrium or ovary, in cervical cancer treatment options (e.g., radical hysterectomy, trachelectomy, or radiation with or without chemotherapy) are applied depending on the extent of disease. Therefore, there has been an increasing necessity for new techniques that can indicate the optimal treatment. The revised 2018 FIGO staging system recommended the use of imaging techniques as well as clinical evaluation or pathological measurement to assess tumor size and extent and detect lymph node metastases. Although MRI is recommended for assessment of the size and extent of the primary tumor, the choice of imaging modality for nodal evaluation has not yet been fixed by FIGO [4]. Therefore, it is necessary to establish the optimal method for accurate preoperative staging.
The spread of cervical cancer into adjacent organs such as the vagina, parametrium, side wall, bladder, and rectum can be better appreciated on MRI than on CT. Reported pooled sensitivities and specificities of MRI for assessing all aspects of local extent have ranged between 0.71-0.88 and 0.86-0.95, respectively [11]. In particular, MRI has shown significantly better pooled sensitivity (84%) for the evaluation of parametrial invasion compared with that of clinical examination (40%), although pooled specificities were comparable between clinical examination and MRI [9]. Despite of the paucity of data, in the hands of experienced operators 18 F-FDG PET/CT and ultrasound also have shown similar pooled sensitivity and specificity for the evaluation of parametrial invasion [11]. In terms of 18 F-FDG PET/MRI, Grueneisen et al. showed that PET/MRI could detect all 27 primary tumor lesions in the uterine cervix and correctly determined the T-stage in 23/27 (85%) patients [14]. Moreover, Sarabhai et al. showed that PET/MRI correctly determined the T stage in 45/53 (85%) patients, whereas MRI alone correctly identified the tumor stage in 46/53 (87%) patients [13]. In the present study, accuracy for the invasion of the vagina, parametrium, side wall, and adjacent organs was 97.2%, 93.1%, 97.2%, and 100% for 18 F-FDG PET/MRI; and 97.2%, 91.7%, 97.2%, and 100% for pelvic MRI, respectively. However, there were no significant differences between these modalities (p > 0.05), suggesting that PET/MRI has similar diagnostic accuracy to that of conventional imaging modalities such as MRI, ultrasound, and 18 F-FDG PET/CT.
In terms of the detection of lymph node metastasis, PET and PET/CT each showed the highest pooled sensitivity (82%) and specificity (95%), whereas CT showed 50% and 92%, and MRI showed 56% and 91%, respectively [12]. A recent meta-analysis has reported high specificity but poor sensitivity for the detection of lymph node metastases for CT, MRI, and PET [11]. The pooled sensitivities and specificities were 0.51 and 0.87 for CT; 0.57 and 0.93 for MRI; and 0.57 and 0.95, for PET, respectively. However, this review purposefully avoided incorporating cutting-edge, non-standard methods that include acquisition techniques such as integrated PET/MRI because they need validation and are unlikely to be available in regions where cervical cancer is most prevalent [11]. Grueneisen et al. have reported sensitivity, specificity, and diagnostic accuracy for the detection of node-positive patients of 91%, 94%, and 93%, respectively [14]. Moreover, Sarabhai et al. have shown sensitivity, specificity, and accuracy of PET/MRI of 83%, 90%, and 87%; whereas the values for MRI alone were 71%, 83%, and 77%, respectively [13]. In other cancers, including nasopharyngeal cancer, 18 F-FDG PET/MRI has been reported as superior compared with 18 F-FDG PET/CT for detecting lymph node metastasis [16]. In the present study, accuracy for detecting metastases in the pelvic and paraaortic lymph nodes was 95.8% and 98.6% for 18 F-FDG PET/MRI, and 83.3% and 95.8% for CT, respectively. Metastasis to pelvic lymph nodes showed significant difference (p < 0.01), suggesting that PET/MRI might have superior diagnostic accuracy compared with the conventional modalities for detecting lymph node metastasis.
In terms of distant metastasis, 18 F-FDG PET/CT is reported to have high diagnostic accuracy (94%) for identifying metastatic lesions, superior to that MRI and CT [17,18]. For detecting bone metastases, 18 F-FDG PET/CT has better sensitivity than CT and better specificity than MRI [19]. In terms of 18 F-FDG PET/MRI, Grueneisen et al. showed that PET/MRI correctly identified non-regional lymph node metastases, including inguinal lymph nodes, in 100% of patients [14]. Moreover, Sarabhai et al. reported that PET/MRI showed higher values for the detection of distant metastases (i.e., metastatic spread beyond regional lymph nodes) compared with MRI alone (sensitivity, 87% vs. 67%; specificity, 92% vs. 90%; diagnostic accuracy, 91% vs. 83%) [13]. In other cancers including breast and colorectal cancers, 18 F-FDG PET/MRI has been reported as superior compared with 18 F-FDG PET/CT for detecting distant metastasis to such as the liver [16]. In the present study, accuracy for detecting distant metastasis including to the lung, thoracic spine, and non-regional lymph nodes (except the paraaortic lymph nodes) was 100% for 18 F-FDG PET/MRI and 97.2% for CT, although there was no significant difference. A possible reason for the lack of significant difference could be the small number of events in our study. Although further studies with larger sample sizes are warranted, we consider that PET/MRI possibly has superior diagnostic accuracy compared with conventional modalities for detecting distant metastasis.
In the revised 2018 FIGO staging system that permits the use of imaging data, MRI has been reported as the best imaging modality for assessing primary tumor of diameter >10 mm [6]. In the present study, the accuracy for local staging was >90% for 18 F-FDG PET/MRI and also for MRI alone. In the assessment of lymph node metastasis, however, 18 F-FDG PET/MRI had greater accuracy than CT, which can evaluate only the size of lymph nodes. Therefore, the accuracy for 2018 FIGO staging was significantly better for 18 F-FDG PET/MRI (82.1%) than for CT and MRI (60.7%), which suggests that 18 F-FDG PET/MRI might have additional diagnostic value compared with conventional imaging modalities such as MRI and CT.
This study had several limitations. First, this investigation used a retrospective design, and not all MRI examinations were performed at our institution. However, our readers reevaluated the images from other hospitals and were blinded to the initial imaging findings. Second, the sample size was relatively small, and further prospective studies are needed. Third, we could not evaluate histopathological correlations with imaging in patients who had not yet undergone lymphadenectomy. We thus performed node-specific comparisons between imaging and histopathology in all other patients.

Conclusions
18 F-PET/MRI offered significantly superior accuracy compared with MRI and CT for the revised 2018 FIGO staging of cervical cancer. The accuracy of 18 F-PET/MRI for detecting metastasis to pelvic lymph nodes was significantly superior to CT, and the accuracy of 18 F-PET/MRI was equivalent to that of MRI and CT for the detection of local extent and of metastasis to distant organs. These findings suggest that 18 F-FDG PET/MRI might provide a useful alternative to conventional imaging modalities in cervical cancer.

Data Availability Statement:
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.