Background and Objective Lu-177 DOTATATE (Lutathera ®) is a radiolabeled analog of somatostatin administered intravenously in patients with somatostatin receptor-positive gastroenteropancreatic neuroendocrine tumors. Biodistribution of Lu-177 DOTATATE in tumor and healthy tissues can be monitored by serial post-injection scintigraphy imaging. Patient exposure to the drug is variable with the recommended fixed dosage, and hence there is a variable response to treatment. The aim of this work was to study the pharmacokinetics of Lu-177 DOTATATE by a population modeling approach, based on single-photon emission computed tomography (SPECT)/computed tomography (CT) images used as surrogate of plasma concentrations to study the interindividual variability and finally optimize an individual dosage. Methods From a retrospective study, SPECT/CT images were acquired at 4 h, 24 h, 72 h, and 192 h postadministration. From these images, volumic activities were calculated in blood and bone marrow. An individual non-compartmental pharmacokinetic analysis was performed, and the mean pharmacokinetic parameters of each tissue were compared together and with reference data. Blood volumic activities were then used to perform a population pharmacokinetic analysis (NONMEM). Results The pharmacokinetic parameters (non-compartmental analysis) obtained from blood (clearance [CL] = 2.65 L/h, volume of distribution at steady state [Vss] = 309 L, elimination half-life [t 1/2 ] = 86.3 h) and bone marrow (CL =1.68 L/h, Vss = 233 L, t 1/2 = 98.8 h) were statistically different from each other and from reference values (CL = 4.50 L/h, Vss = 460 L, t 1/2 = 71.0 h) published in the literature. SPECT/CT blood images were used as a surrogate of plasma concentrations to develop a population pharmacokinetic model. Weight was identified as covariate on volume of the central compartment, reducing the interindividual variability of all population pharmacokinetic parameters. Conclusion This study is a proof of concept that obtaining pharmacokinetic parameters with image-based blood concentration is possible. Obtaining observed concentrations from SPECT/CT images, without the need for blood sampling, is a real advantage for the patient and the drug monitoring. Pharmacokinetic modeling could be combined with a deep learning model for automatic contouring and allow precise patient-specific dose adjustment in a non-invasive manner.