PURPOSE: To demonstrate the utility of a QA program for seed localization, and to design a 3D-printed template to improve the accuracy of seed placement on custom-built METHODS AND MATERIALS: A seed localization analysis tool (SLAT) was developed in MATLAB to detect variations in seed placement relative to a treatment plan. A flexible seed placement template (3D-FSPT) was designed in CAD and printed using a Formlabs Form-3 3D printer. The 3D-FSPT and SLAT were tested using 3D-printed model eye plaques with nonradioactive seeds arranged following clinically-relevant treatment plans. Five clinical plaques were also evaluated. RESULTS: SLAT detected submillimeter scale variations in seed position with 2.3% error relative to the plan's seed coordinates, and with an uncertainty of ± 0.01 mm. The average seed displacement on the model plaques with free-handed seed placement was 1.31 mm (SD = 0.61), and the average seed orientation difference was 5.27 degrees (SD = 4.77). The average seed displacement on the clinical plaques was 0.77 mm (SD = 0.42), and the average seed orientation difference was 4.41 degrees (SD = 3.00). For the clinical plaques, changes in dosimetry to the tumor apex and critical eye structures were within acceptable tolerances. Seed displacement (mm) (p <
0.001) and seed orientation differences (degrees) (p = 0.008) were significantly lower using the template to guide seed placement on the model plaques compared to free-handed seed placement. CONCLUSIONS: The feasibility of a 3D-FSPT and SLAT is demonstrated for improving seed placement accuracy relative to a treatment plan.