Comparison of sequential planar 177Lu-DOTA-TATE dosimetry scans with 68Ga-DOTA-TATE PET/CT images in patients with metastasized neuroendocrine tumours undergoing peptide receptor radionuclide therapy.

Purpose: The aim of the study was to compare sequential (177)Lu-DOTA-TATE planar scans ((177)Lu-DOTA-TATE) in patients with metastasized neuroendocrine tumours (NET) acquired during peptide receptor radionuclide therapy (PRRT) for dosimetry purposes with the pre-therapeutic (68)Ga-DOTA-TATE positron emission tomography (PET)/CT ((68)Ga-DOTA-TATE) maximum intensity projection (MIP) images obtained in the same patients concerning the sensitivity of the different methods.
Methods: A total of 44 patients (59 ± 11 years old) with biopsy-proven NET underwent (68)Ga-DOTA-TATE and (177)Lu-DOTA-TATE imaging within 7.9 ± 7.5 days between the two examinations. (177)Lu-DOTA-TATE planar images were acquired at 0.5, 2, 24, 48 and 72 h post-injection; lesions were given a score from 0 to 4 depending on the uptake of the radiopharmaceutical (0 being lowest and 4 highest). The number of tumour lesions which were identified on (177)Lu-DOTA-TATE scans (in relation to the acquisition time after injection of the therapeutic dose as well as with regard to the body region) was compared to those detected on (68)Ga-DOTA-TATE studies obtained before PRRT.
Results: A total of 318 lesions were detected; 280 (88%) lesions were concordant. Among the discordant lesions, 29 were (68)Ga-DOTA-TATE positive and (177)Lu-DOTA-TATE negative, whereas 9 were (68)Ga-DOTA-TATE negative and (177)Lu-DOTA-TATE positive. The sensitivity, positive predictive value and accuracy for (177)Lu-DOTA-TATE as compared to (68)Ga-DOTA-TATE were 91, 97 and 88%, respectively. Significantly more lesions were seen on the delayed (72 h) (177)Lu-DOTA-TATE images (91%) as compared to the immediate (30 min) images (68%). The highest concordance was observed for bone metastases (97%) and the lowest for head/neck lesions (75%). Concordant lesions (n = 77; mean size 3.8 cm) were significantly larger than discordant lesions (n = 38; mean size 1.6 cm) (p < 0.05). No such significance was found for differences in maximum standardized uptake value (SUV(max)). However, concordant liver lesions with a score from 1 to 3 in the 72-h (177)Lu-DOTA-TATE scan had a lower SUV(max) (n = 23; mean 10.9) than those metastases with a score of 4 (n = 97; mean SUV(max) 18) (p < 0.05).
Conclusion: Although (177)Lu-DOTA-TATE planar dosimetry scans exhibited a very good sensitivity for the detection of metastases, they failed to pick up 9% of lesions seen on the (68)Ga-DOTA-TATE PET/CT. Three-dimensional dosimetry using single photon emission computed tomography/CT could be applied to investigate this issue further. Delayed (72 h) images are most suitable for drawing regions of interest for dosimetric calculations.