Given the high frequency of liver metastases in neuroendocrine tumor patients, we aimed to determine whether hepatic intraarterial administration of 90Y-DOTATOC peptide receptor radionuclide therapy (PRRT) would increase treatment efficacy while reducing systemic toxicity compared with systemic toxicity from intravenous administration as previously reported in the literature. Methods: PRRT-naïve adult neuroendocrine tumor patients with liver-dominant metastases were enrolled in a prospective single-center, open-label pilot study. The patients underwent baseline PET/CT using intravenous 68Ga-DOTATOC. Then, 3.5 ± 0.2 GBq (94.7 ± 5.4 mCi) of 90Y-DOTATOC were administered into the proper hepatic artery over 30 min. The first 5 patients also received intraarterial 68Ga-DOTATOC and underwent PET/CT. All patients were followed for response (RECIST, version 1.1) (primary aim 2, safety) and toxicity (Common Terminology Criteria for Adverse Events, version 4.0) (primary aim 1, efficacy) for at least 6 mo, with optional follow-up for up to 1 y. In the subset of 5 patients who underwent both intravenous and intraarterial 68Ga-DOTATOC PET/CT, tumor SUVmax was compared between intravenous and intraarterial administration for hepatic tumors, intrahepatic tumors, and uninvolved background organs (secondary aim, intravenous vs. intraarterial uptake). Results: The study was terminated after a planned analysis of the first 10 patients because of lack of efficacy. The best response was stable disease in 90% (9/10 patients) and progressive disease in 10% (1/10 patients) at 3 mo, and stable disease in 8 of 10 patients and progressive disease in 2 of 10 patients at 6 mo. One additional patient developed progressive disease after the 6-mo follow-up period but within the optional 1-y follow-up period. No partial response or complete response was observed. The 2 patients with the highest liver tumor burden died within 6 mo of treatment, with treatment considered a possible contributor. Patients who received intraarterial administration failed to demonstrate higher uptake by hepatic metastases than patients who received intravenous administration, with a median intraarterial-to-intravenous SUVmax ratio of 0.81 (range, 0.36-2.09) on a lesion level. Conclusion: Our study found that administration of PRRT via the proper hepatic artery did not reproduce the increase in hepatic tumor uptake that was previously reported. In addition, the single treatment using 90Y-DOTATOC did not induce tumor shrinkage, indicating that more treatment cycles may be required. Possible safety concerns in patients with a high liver tumor burden should inform patient selection for future studies.