Purpose:To determine if baseline patient, tumor, and pretreatment evaluation characteristics could help identify patients who require technetium-99m (Tc-99m) macroaggregated albumin Tc-(99m MAA) imaging before selective internal radiation therapy (SIRT). Materials and Methods: In this retrospective analysis, 532 consecutive patients with primary (n = 248) or metastatic (n = 284) liver tumors were evaluated between 2006 and 2015. Variables were compared between patients in whom Tc-99m MAA imaging results contraindicated/modified SIRT administration with yttrium-90 (Y-90) resin microspheres and those who were treated as initially planned. The Tc-99m MAA findings that contraindicated/modified SIRT were a lung shunt fraction (LSF) > 20%, gastrointestinal Tc-99m MAA uptake, or a mismatch between Tc-99m MAA uptake and intrahepatic tumor distribution. Results: LSF > 20% and gastrointestinal MAA uptake were observed in 7.5% and 3.9% of patients, respectively, and 11% presented a mismatch. Presence of a single lesion (odds ratio [OR] = 2.4) and vascular invasion (OR = 5.5) predicted LSF > 20%, and GI MAA uptake was predicted by the presence of liver metastases (OR = 3.7) and Tc-99m MAA injection through the common/proper hepatic artery (OR = 4.7). Vascular invasion (OR = 4.1) was the only predictor of LSF > 20% and/or GI MAA uptake (sensitivity = 49.2%, specificity = 80.3%, negative predictive value = 92.4%). Previous antiangiogenic treatment (OR = 2.4) and presence of a single lesion (OR = 2.6) predicted mismatch. Conclusions: Imaging with Tc-99m MAA is essential in SIRT workup because baseline characteristics may not adequately predict Tc-99m MAA results. Nevertheless, the absence of vascular invasion potentially identifies a group of patients at low risk of SIRT contraindication/modification in whom performing SIRT in a single session (ie, pretreatment evaluation and SIRT on the same day) should be explored.