Room: ePoster Forums
Purpose: Six positional parameters are tracked during 3D optical surface imaging (OSI) for patient motion monitoring. Because of translation-rotation coupling, a drift in one can cause chain variations in the others. This intrinsic interdependency greatly affects the patient shift correction sequence and effectiveness. We investigated the coupling correlations among these parameters in order to define an effective shift correction strategy.
Methods: An anthropomorphic phantom was placed on the couch and imaged with OSI. Couch rotations of 0Â°, 1Â°, 2Â°, and 3Â° were used to investigate rotational coupling and translations of 0,1, 2, and 3 cm were used to investigate translational coupling. Because of mathematical symmetry, only positive shifts were applied. About 100 tracking data points were acquired per rotation or translation. Pearson's correlation coefficients were calculated for each pair of positional parameters
Results: For rotational coupling, we found that VRT-LNG, VRT-LAT, VRT-ROLL, LNG-LAT, LNG-ROLL, and LAT-ROLL had a strong positive correlation (r>0.70). The correlation coefficients were 0.89, 0.82, 0.85, 0.79, 0.88, and 0.94, respectively. VRT-PITCH, LNG-PITCH, LAT-PITCH, and ROLL-PITCH had a strong negative correlation. The correlation coefficients were -0.84, -0.91, -0.86, and -0.94. For translational coupling, we found that RTN-PITCH and ROLL-PITCH had a very strong negative correlation. Both were -0.83. Using the number of strong correlations that one parameter had with others, called the coupling strength, we ranked these six parameters as follows: PITCH, ROLL, VRT, LNG, LAT, and RTN.
Conclusion: All six tracking parameters are interdependent. A shift in one can induce shifts in other parameters. Rotational shifts have a greater impact on translational shifts. Contrary to common perception, a translational shift could also cause rotational shifts. Among the six parameters, PITCH, ROLL, and VRT are the most critical ones so the most effective order of shift corrections would be PITCH, ROLL, and VRT, followed by other three.
Funding Support, Disclosures, and Conflict of Interest: This work was partially supported by the MSK Cancer Center Support Grant/Core Grant (P30 CA008748)