Room: Exhibit Hall | Forum 2
Purpose: Magnetic-resonance image guided on-line adaptive radiotherapy (On-ART) has recently been introduced clinically in several institutions. On-ART can lead to compromised plans because insufficient time is available for the iterative process of plan optimization used in offline planning. The purpose of this work was to evaluate to what extent On-ART plans can be improved by comparison to plans created offline without time constraints.
Methods: In our clinical On-ART process, adaptive re-planning is performed by re-contouring structures and repeating fluence and segment optimization without adjusting structure weights and thresholds. Fifty (50) consecutive adapted fractions from pancreatic treatments were analyzed. Of these, the 10 with greatest deviations from physician constraints were chosen for off-line re-planning, which included adjustment of structure weights and thresholds only. Per clinical protocol, primary consideration was given to 4 evaluation metrics: target coverage, maximum dose, and the volume of small bowel and stomach receiving 35 Gy. The primary goal of offline re-planning was to produce an objectively improved plan by improving one or more of the 4 evaluation metrics without degrading the others. The secondary goal of offline re-planning was to subjectively improve the plan by improving one or more of the evaluation metrics without exceeding the physicianâ€™s constraints for the others.
Results: Of the 50 treated adaptive plans analyzed, the 10 with the greatest deviation from constraints had on average 26.0% reduction in target coverage, and 1 exceeded the small bowel constraint by 39%. 4 plans were objectively improved by offline re-planning by an average of 11.5% (range: 4.2% - 27.1%) in target coverage and one plan was subjectively improved.
Conclusion: Off-line re-planning was able to improve dosimetric quality of On-ART plans in 10% of evaluated fractions. Work to improve the temporal efficiency of On-ART planning and develop techniques to better anticipate ideal re-optimization parameters is continuing.