Multiparametric dual-time-point [18F]FDG-PET/MRI for lymph node staging in patients with untreated FIGO I/II cervical carcinoma

Abstract Purpose:[18F]FDG PET/MRI comes with limited sensitivity for initial detection of lymph node metastases (LNM) in early stage cervical carcinoma. Therefore, the aim of this prospective study was to investigate the additional value of multiparametric dual-time-point PET/MRI and to asses potential influencing factors for LNM detection.Methods: 63 patients underwent whole-body dual-time-point [18F]FDG-PET/MRI ≈60+90min p.i. (Biograph mMR®, Siemens Healthineers). 251 LN were evaluated visually by nuclear medicine and radiology expert, quantified multiparametrically and were correlated with histology. Multiparametric analysis was performed using a logistic regression-based malignancy score (MS).Results:Overall prevalence of LNM was 17.1% and dependent on tumor grade (Gx 1/6, G1:0/19, G2:18/106, G3:24/120). Grading of the primary tumor (G2/G3) had a significant impact on visual detection (sens: 8.3%/31%).Best single parameter for LNM detection was SUVavg, however with a significant loss of discriminatory power in G2 vs. G3 tumors (AUC: 0.673/ 0.901). For multiparametric analysis, the independent predictors “SUVavg, ∆SUVpeak (60 to 90 min p.i.), LN sphericity, ADC and histologic grade” were included in the MS. Application of the MS enhanced AUCs especially in G2 tumors (AUC: G2:0.769; G3:0.877) and improved accuracy for single LNM from 34.5% to 55.5% compared to the best univariate parameter SUVavg.Equated to the expert consensus, use of the malignancy score increased overall sensitivity from 31.0% to 79.3% (Youden optimum) with a moderate decrease in specificity from 98.3% to 75.6%.Conclusion:Low histologic grad (G2 vs. G3) hampers visual detectability of LNM in [18F]FDG-PET/MRI. Multiparametric evaluation of dual-time-point PET/MRI has the potential to improve accuracy compared to visual interpretation and enables sufficient N-staging also in G2 tumors.Trial registration:DRKS00014346. Registered April 4, 2018 - retrospectively registered,https://www.drks.de/drks_web/navigate.do?navigationId=trial.HTML&TRIAL_ID=DRKS00014346