The aim of this study was to compare the performance of conebeam computed tomography (CBCT), clinical and surgical probing in assessing maxillary molar furcation involvement (FI). Furcation defects (n= 120) were assessed through CBCT, clinical and intra-surgical evaluation (ISE). Furcation Involvement, vertical and horizontal bone loss were assessed through clinical probing, CBCT and probing during ISE. Three trained radiologists evaluated CBCT images and intra- and interobserver agreement were calculated by Kappa test and Intraclass Correlation Coefficient (ICC). McNemar and Wilcoxon tests were used to compare clinical probing, ISE and CBCT. Accuracy, sensitivity, specificity, positive and negative predictive values were calculated to detect FI. Clinical findings showed 28 sites with Degree I, 25 sites with Degree II, and 8 sites with Degree III. Good intra- (k=1.00) and interobserver agreement (k=0.773) were observed. Intraobserver and interobserver agreement for horizontal bone loss were moderate, k=0.485 and k=0.549, respectively. Intra-surgical findings showed Degree I at 21 sites, and Degree II and Degree III FI at fifteen sites each. Clinical evaluation showed 75% agreement with ISE and 78% with CBCT. Accuracy for clinical detection of FI was 75%, while for CBCT evaluation ranged from 72.5% to 77.5%, considering the 3 observers. Significant differences were found at distal sites using CBCT (p<0.05). Clinical evaluation and CBCT showed similar results for the presence or absence of FI. Concerning horizontal and vertical bone loss, CBCT was not considered a precise examination method for incipient bone defects.