Background and purpose Irradiation of the Pharyngeal Superior Constrictor Muscle (PSCM) seems to play a crucial role in radiation-related swallowing dysfunctions. Purpose of our study was to quantify operator-related variability in the contouring of PSCM on Computed Tomography (CT) scans and adherence with contours derived from MR images. Materials and methods Three sets of treatment planning CT and their corresponding MR images were selected. Contouring of the PSCM was performed using both a literature-based method, derived from literature review, and an optimized method, derived from Magnetic Resonance (MR) images thus obtaining "literature-based" and "optimized" contours. Each operator contoured the PSCM on CT scans according to both methods for three times in three different days. Inter- and intra-operator variability and adherence to a contour obtained from MR images (named "MR-derived" contour) were analyzed. Results Thirty-four operators participated and 612 contours were obtained. Both intra- and inter-operator variability and adherence to the "MR-derived" contour were significantly different between the two methods (p ≤ 0.05). The "optimized" method showed a lower intra- and inter-operator variability and a higher adherence to the "MR-derived" contour. Conclusions The "optimized" method ameliorates both operator-related variability and adherence with MR images.