INTRODUCTION:In our institutional experience, determination of the alpha (alpha) angle at MR arthrography as an indicator of the likelihood of cam-type femoroacetabular impingement (FAI) is fraught with inconsistency. The aims of this study were to quantify the degree of variability in and calculate the diagnostic accuracy of the alpha angle in suggesting a diagnosis of cam impingement, to determine the accuracy of a positive clinical impingement test, and to suggest alternative MR arthrographic measures of femoral head-neck overgrowth and determine their diagnostic utilities. MATERIALS AND METHODS:We carried out a retrospective analysis of MR arthrographic studies performed during a 4-year period, combined with chart analysis, which allowed identification of 78 patients in whom surgical correlation was also available. The status of a preoperative clinical impingement test was also noted. Patients were designated as having cam-type FAI (Group A, n = 39) if intra-operative femoral head-neck junction bony osteochondroplasty/arthoscopic femoral debridement was performed. Group B (n = 39) acted as controls. Three radiologists independently and blindly performed a series of measurements (alpha angle and two newly proposed measurements) in each patient on two separate occasions. An alpha angle of greater than 55 degrees was considered indicative of the presence of cam-type FAI. RESULTS:Performance values for alpha angle measurement were poor for each observer. There was considerable (up to 30% of the mean value) intra-observer variability between the first and second alpha angle measurements for each subject. Binary logistic regression analysis confirmed that the alpha angle is of no value in predicting the presence or absence of cam-FAI. A statistically significant difference existed between Groups A and B with regard to the newly proposed anterior femoral distance (AFD; p = 0.004). Using an AFD value of 3.60 mm or greater as being indicative of the presence of cam-FAI yields a 0.67 performance measure (95% confidence interval 0.55-0.79). The second proposed parameter (femoral neck ratio) was of no value in suggesting the presence or absence of this condition. The sensitivity, specificity, and positive and negative predictive values of the clinical impingement test were 76.9%, 87.2%, 85.7% and 79.1% respectively. CONCLUSIONS:Femoral alpha angle measurement is associated with considerable variability. This index performed poorly in our patient population and was statistically of no value in suggesting the presence or absence of cam-FAI. One of our proposed measures, the AFD, outperformed the alpha angle, though to an insufficient degree to suggest its routine incorporation into clinical practice. Our experience suggests that the clinical impingement test remains the most reliable predictor of the presence of this condition.