Purpose: The purpose of this study was to measure the angular differences between the horizontal reference planes on the CBCT generated PA cephalogram and the modified interpupillary plane, which was usually used in the clinical examination, and to evaluate the validity of the horizontal reference planes. Methods: The CBCT generated PA cephalogram was used to measure the angles between the FH, Lo and IP planes. The subjects consisted of 42 patients with facial asymmetry (males: 21, females: 21, mean-age: 21.6 years). The control groups were also assessed (males: 10, females: 10, mean-age: 23.8 years). The distance of the interpupil was measured on the soft-tissue volume rendered image. The angular differences were statistically analyzed using the $Mann-Whitney$ $U$ $test$ for inter-group comparisons and the $Friedman$ $test$ for intra-group comparisions. Results: The angle between the FH plane and IP plane (the angle of the FH-IP line) showed a statistically significant difference between the two groups ($p$ <0.05). There was no statistical differences between each angle (angle of the FH-IP line, angle of the FH-Lo line, angle of the Lo-IP line) on the intra-group comparision ($p$ >0.05). Conclusion: The angle between the Lo line and IP line (angle of the Lo-IP line) showed no statistically significant difference in both the control and asymmetry groups. Therefore, the Lo line could be used as a horizontal reference plane in CBCT generated PA cephalograms.