Abstract
The aim of this study was a comprehensive analysis of a number of MR patterns in the diagnosis of trigeminal neuralgia (TN) by the method of logistic regression. Study population consisted of 133 patients: with trigeminal neuralgia (TN) (86, surgically (n = 56) and clinically (n = 30) verified; 55 ± 11 years) and reference group (47, 51 ± 16 years). The data were obtained using the FIESTA sequence (slice thickness 0,5 mm) on 3 T and 1,5 T (Discovery 750w and Optima 450w, General Electric). To construct the statistical models, 7 separate MR patterns were used: lengths of intercisternal portions of trigeminal nerves, cross-sectional area and ratio of nerve diameters in REZ regions, diameters of adjacent arterials, length of neurovascular contact, angles of these loops and angles of transverse intersection of trigeminal nerves with arteries in the posterior cranial fossa. Three logistic models were built separately, differing in the inclusion of control groups. The total model included the maximum number of nerves for analysis (209 nerves), the second model with the inclusion of the parameters of the contralateral nerves (128 nerves) and the third model with nerves of the healthy individuals (147 nerves). In all three models, the greatest odds ratio was observed for the diameters of the adjacent arteries (6.13; 5.51; 9.56). The second most important predictor of TN for the general model and contralateral comparisons was the ratio of nerve diameters (1,97 and 3,3). For the third model (comparison with the healthy individuals), the second most important predictor was the length of neurovascular contact (1,88). Cross-sectional areas also had significance in all models (0,64; 0,71; 0,63). Values of angles and lengths of nerves had lower statistical significance for all models. Thus, the most important MR patterns for the diagnosis of TN were the diameters of the adjacent arterial loops and the ratio of nerve diameters in REZ. This fact may reflect the importance of nerve deformation for first two models. It may confirm the mechanical pressure as the main mechanism for classical TN.