Diagnosis and dosimetry in differentiated thyroid carcinoma using 124I PET: comparison of PET/MRI vs PET/CT of the neck.

PURPOSE

This study compares intrinsically coregistered 124I positron emission tomography (PET) and CT (PET/CT) and software coregistered 124I PET and MRI (PET/MRI) images for the diagnosis and dosimetry of thyroid remnant tissues and lymph node metastases in patients with differentiated thyroid carcinoma (DTC).

METHODS

After thyroidectomy, 33 high-risk DTC patients (stage III or higher) received 124I PET/CT dosimetry prior to radioiodine therapy to estimate the absorbed dose to lesions and subsequently underwent a contrast-enhanced MRI examination of the neck. Images were evaluated by two experienced nuclear medicine physicians and two radiologists to identify the lesions and to categorize their presumable provenience, i.e. thyroid remnant tissue (TT), lymph node metastasis (LN) and inconclusive tissue. The categorization and dosimetry of lesions was initially performed with PET images alone (PET only). Subsequently lesions were reassessed including the CT and MRI data.

RESULTS

The analyses were performed on a patient and on a lesion basis. Patient-based analyses showed that 26 of 33 (79%) patients had at least one lesion categorized as TT on PET only. Of these patients, 11 (42%) and 16 (62%) had a morphological correlate on CT and MRI, respectively, in at least one TT PET lesion. Twelve patients (36%) had at least one lesion classified as LN on PET only. Nine (75%) of these patients had a morphological correlate on both CT and MRI in at least one LN PET lesion. Ten patients (30%) showed at least one lesion on PET only classified as inconclusive. The classification was changed to a clear classification in two patients (two LN) by CT and in four (two TT, two LN) patients by MRI. Lesion-based analyses (n=105 PET positive lesions) resulted in categorization as TT in 61 cases (58%), 16 (26%) of which had a morphological correlate on CT and 33 (54%) on MRI. A total of 29 lesions (27%) were classified as LN on PET, 18 (62%) of which had a morphological correlate on CT and 24 (83%) on MRI. In 16 lesions (15%) PET alone allowed no definite categorization. Categorization was achieved with the aid of CT and MRI, respectively, in five (one TT, four LN) and in six (two TT, four LN) lesions. In direct comparison, 23 lesions were not discernible on CT but clearly visible on MRI, 15 of which were smaller than 10 mm and about two thirds were classified as TT. Redoing dosimetry based on the volume information from MRI for these small lesions would have changed the initial therapy regime in five patients. These patients would have received (131)I therapy with standardized activities of 3.7 GBq or 7.4 GBq instead of activities higher than 10 GBq and would have benefited from reduced radiation exposure.

CONCLUSION

PET/MRI is superior to PET/CT in terms of tracing back a PET focus to a morphological correlate. For this reason PET/MRI enhances diagnostic certainty for lesions<10 mm and improves pretherapeutic lesion dosimetry in DTC.