318 consecutive women with EC were included when referred to three Danish tertiary gynecological centers for surgical treatment. Preoperatively they were PET/CT-, MRI-, and 2DUS scanned. The imaging results were compared to the final pathological findings. This study was approved by the National Committee on Health Research Ethics.
For predicting myometrial invasion, we found sensitivity, specificity, PPV, NPV, and accuracy for PET/CT to be 93 % , 49 % , 41 % , 95 % and 61 % , for MRI to be 87 % , 57 % , 44 % , 92 % , and 66 % and for 2DUS to be 71 % , 72 % , 51 % , 86 % and 72 % . For predicting cervical invasion, the values were 43 % , 94 % , 69 % , 85 % and 83 % , respectively, for PET/CT, 33 % , 95 % , 60 % , 85 % , and 82 % , respectively, for MRI, and 29 % , 92 % , 48 % , 82 % and 78 % for 2DUS. Finally, for lymph node metastases, the values were 74 % , 93 % , 59 % , 96 % , and 91 % for PET/CT and 59 % , 93 % , 40 % , 97 % and 90 % for MRI. When comparing the diagnostic performance we found PET/CT, MRI and 2DUS to be comparable in predicting myometrial invasion. For cervical invasion and lymph node metastases, however, PET/CT was the best.
None of the modalities can yet replace surgical staging. However, they all contributed to important knowledge and were, furthermore, able to upstage low-risk patients who would not have been recommended lymph node resection based on histology and grade alone.