Improved computer analysis of solid phase gastric emptying scans.

Abstract

OBJECTIVES: The solid phase gastric emptying scan (GES) is used to confirm the clinical impression of abnormal gastric emptying. There is variability in the interpretation of GES. Determination of initial lag phase of the GES and the emptying half-time (t1/2) is generally performed by curve inspection and thus may suffer from lack of objectivity. The purpose of this study was to develop a physiological model for interpretation of the GES using nonlinear curve fitting. This model resulted in computer-generated best fits for lag time and t1/2, which were analyzed in a group of patients with suspected gastroparesis. METHODS: All gastric emptying scans performed at our institution over a 3.5-yr period were studied. Raw data from these studies were analyzed by nonlinear curve fitting. Using the equation: If (x < xo, plateau, plateau * exp( - K * (x - xo))) data were best fit to a function describing a lag followed by a log linear decay. This model generated four parameters; lag, K, t1/2, and T50%. Forty patients with less than 50% emptying at 1 h (group II) were compared with 31 patients with normal emptying (group I). RESULTS: The nonlinear model resulted in better curve fitting (higher r2) in 59 of 71 studies (81%) when compared with a monoexponential decay after a lag of 0 min. Mean lag for patients in group I was 8.5 +/- 1.2 min and was 25.9 +/- 3.1 min in group II (p < 0.0005). Mean t1/2 was 31.7 +/- 1.8 min in group I and 69.7 +/- 5.0 min in group II (p < 0.007). By adding 2 SD to lag and t1/2 in group I, normal values for these parameters were 21.9 and 52.2 min, respectively. Eleven patients in group II had a prolonged lag alone, 13 had a prolonged t1/2, and 13 had prolongation of both parameters. CONCLUSIONS: A new physiological model for the interpretation of GES is presented. Individual patients with delayed gastric emptying may have increased lag times, a decreased rate of antral emptying, or both abnormalities.