Renal function reconstruction and modeling in dynamic scintigraphy

Dynamic renal scintigraphy is a key imaging technique for assessing renal function using time-activity curves (TACs), which represent radiotracer uptake and clearance. TAC accuracy depends on the region of interest (ROI) selection and the modeling approach used. This study aims to: (i) Reconstruct TACs manually using gray-level values in scintigraphic images and compare them to machine-generated TACs using key kinetic parameters (Tmax, T1/2, and the 30-min min/max ratio); and (ii) evaluate the effectiveness of a one-compartment empirical mathematical model for TAC fitting and its physiological relevance. Twelve clinical cases were analyzed, with TACs reconstructed manually using a rectangular ROI selection method and compared to those automatically generated by the scintigraphy machine. An empirical mathematical fitting function was developed to improve TAC fitting. Manually reconstructed TACs showed better dynamic behavior and physiological accuracy over machine-generated TACs, particularly due to differences in ROI selection and signal processing. Using gray-level values instead of raw radioactive counts enhanced the depiction of kidney dynamics. The proposed mathematical model demonstrated a strong correlation (R2 close to 1) and low error metrics, confirming its suitability for renal function assessment. While a free-hand ROI selection may improve accuracy, the rectangular method gives valuable results for the considered cases. This study highlights the importance of ROI selection in TAC reconstruction and demonstrates how manual methods and mathematical modeling can enhance renal functional assessment in clinical practice. Future work should validate these findings in larger datasets and assess the reproducibility of the proposed approach across different patient populations and imaging systems.