Background: Cancer cell line-derived feeder cells enhance natural killer (NK) cell expansion; however, concerns regarding viable residual feeder cells in the final product limit their use. Evidence supporting the safety of NK-sensitive K562-based feeders, even when irradiated, is scarce. We optimized an NK cell expansion protocol using genetically engineered K562-mbIL-18/-21 (GE-K562) feeder cells and clinical-grade media and confirmed the absence of residual feeder cells. Methods: NK cell expansion efficiency was compared between feeder-free and feederbased systems using CTS NK-Xpander Medium. To achieve optimal NK expansion, various peripheral blood mononuclear cell (PBMC)-to-feeder ratios and re-stimulation frequencies were tested over 21 days. Flow cytometry and BCR::ABL1 quantitative reverse transcription PCR (RT-qPCR) were used to confirm the absence of feeder cells in the final NK cell product. Results: Feeder-based systems showed superior NK cell fold expansion compared with that of feeder-free systems. Among feeder-based conditions, NK cells expanded 5,224-fold at a 2:1 PBMC-to-feeder ratio after 3 weeks, relative to 1,450-fold at a 6:1 ratio (P < 0.05).Re-stimulation on days 7 and 14 further increased expansion up to 261,457-fold. Irradiated feeder cells showed no proliferation and were eliminated within 3–6 days. On day 21, flow cytometry and BCR::ABL1 RT-qPCR results confirmed the absence of residual feeder cells. Conclusions Our optimized NK cell expansion protocol using irradiated GE-K562 feeder cells and clinical-grade media offers a safe and scalable approach to generating large numbers of NK cells, supporting its potential use in clinical immunotherapy applications.