Session Item

Our current prone rectum radiotherapy patients are aligned using traditional three-points skin markings plane. However, the setup errors differ hugely from day to day. We would like to study if the new technology surface-guided positioning approach could mitigate this challenge.

Two groups of patients were included for comparison. In Group 1, 30 prone rectum patients were selected randomly. All patients were aligned using surface-guided system (Varian OSMS, Palo Alto, CA). All 6 degree of freedoms were corrected to nearly zero using ‘Move Couch’ function on the OSMS. 230 of CBCT images were taken, the 6D shift errors were recorded. Retrospectively, 30 prone rectum patients who were treated using skin-marking positioning (Group 2) between July 2021 to September 2021 were extracted randomly. Similarly, 230 of CBCT images were retrieved. The shifts errors were recorded. The data were analysed using SPSS19. Both shift errors were compared in terms of mean, standard deviations, maximum displacement, and residual errors. Residual error was defined as the shift error that was more than 3° on based on CBCT image registration.

The result showed that the mean and standard deviations for Group 1 and 2 were (-0.06±0.18)cm, (0.04± 0.24)cm, (-0.01±0.18)cm, (-0.14±0.81)°, (0.28±1.06)°, (-0.01±0.87)° and (0.02±0.42)cm, (-0.06± 0.47)cm, (0.05±0.33)cm, (0.04±1.07)°, (1.56±1.43)°, (-0.43±1.06)° in vertical, lateral, longitudinal, roll, pitch and rotation directions. Data analysis (Table 1) has shown statistically significant (p<0.05) in longitudinal, roll and pitch directions. The maximum displacements for Group 1 and 2 were (- 0.58/0.58)cm, (-0.66/0.57)cm, (-0.56/0.69)cm, (-2.8/2.7)°, (-2.6/2.8)°, (-2.7/2.6)° and (-0.98/0.92)cm, (-1.02/0.87)cm, (-0.86/0.91)cm, (-3.8/2.7)°, (-2.3/5.7)°, (-3.5/2.6)° in vertical, lateral, longitudinal, roll, pitch and rotation directions. In terms of residual error rates, Group 1 had no patient with the discrepancies that were more than 3° in all three rotational directions; while there were 9.6% and 3.9% of CBCT images were larger than 3° in pitch and roll direction in Group 2. The average positioning time using surface-guided was (102.5±12.9)s.

Table 1: The mean and standard deviations of SGRT-based positioning (Group 1) method and skin marking-based positioning (Group 2) method.

Surface-guided positioning approach could improve the positioning accuracy and reduce large shift errors in both translational and rotational directions. The positioning time using surface-guided method is acceptable in our institution as well.