. Author manuscript; available in PMC: 2017 Jan 12.

Published in final edited form as: J Nucl Med. 2016 Jan 21;57(5):715–720. doi: 10.2967/jnumed.115.167445

TABLE 3.

Comparison of GLP-1R PET/CT and Conventional Imaging in Patients with Insulinoma

	Patient-based analysis					Lesion-based analysis
Parameter	GLP-1R PET/CT (n = 43)	CT (n = 43)	MR (n = 25)	EUS (n = 25)	SRS (n = 41)	GLP-1R PET/CT (n = 45)	CT (n = 45)	MR (n = 25)	EUS (n = 26)	SRS (n = 43)
Sensitivity	97.7%	74.4%	56.0%	84.0%	19.5%	97.8%	73.3%	56.0%	84.6%	16.3%
95% CI	87.7–99.9	58.8–86.5	34.9–75.6	63.9–95.5	8.8–34.9	88.2–99.9	58.1–85.4	34.9–75.6	65.1–95.6	6.8–30.7
P	NA	0.006	0.006	0.125	<0.001	NA	0.003	0.006	0.125	<0.001
PPV	—	—	—	—	—	100%	94.3%	93.3%	95.7%	100%
95% CI	—	—	—	—	—	92.0–100	80.8–99.3	68.1–99.8	78.1–99.9	59.0–100
P	NA	NA	NA	NA	NA	NA	0.147	0.306	0.308	NA

CI = confidence interval; NA = not applicable.

Patient 15, with multiple distant metastases, was excluded from lesion-based analysis. Positive predictive value (PPV) was not assessed in patient-based analysis because all patients were diagnosed with insulinoma in this cohort. P value was analyzed between GLP-1R PET/CT and each type of conventional imaging.