Time-dependent MR diffusion analysis of functioning and nonfunctioning pituitary adenomas/pituitary neuroendocrine tumors.

BACKGROUND AND PURPOSE

Differentiation between functioning and nonfunctioning pituitary adenomas/pituitary neuroendocrine tumors (PAs) is clinically relevant. The goal of this study was to determine the feasibility of using time-dependent diffusion MRI (dMRI) for microstructural characterization of PAs.

METHODS

The study included 54 participants, 24 with functioning PA and 30 with nonfunctioning PA. Time-dependent dMRI of the pituitary gland was performed using an inner field-of-view echo-planar imaging based on 2-dimensional-selective radiofrequency excitations with oscillating gradient and pulsed gradient preparation (effective diffusion time: 7.1 and 36.3 ms) at b-values of 0 and 1000 seconds/mm. Each tumor had its apparent diffusion coefficients (ADCs) measured at two diffusion times (ADC and ADC), its ADC change (cADC), and relative ADC change. The mean values of diffusion parameters were compared between functioning and nonfunctioning PAs. We compared the diffusion parameters of nonfunctioning PAs with those of each type of hormone-producing PAs. The diagnostic performances of the diffusion parameters were assessed.

RESULTS

The cADC was significantly higher in functioning PAs than nonfunctioning PAs (p = .0124). The receiver operating characteristic (ROC) curve analysis revealed that cADC (area under the ROC curve [AUC] = .677, p = .017) is effective in distinguishing between functioning and nonfunctioning PAs. The cADC was significantly higher in growth hormone (GH)-producing PAs compared to nonfunctioning PAs (p = .006). The ROC curve analysis indicated that cADC (AUC = .771, p < .001) effectively distinguishes between GH-producing and nonfunctioning PAs.

CONCLUSIONS

The cADC derived from time-dependent dMRI could distinguish between functioning and nonfunctioning PAs, particularly those producing GH.