TH‐C‐137‐04: Separating the Dosimetric Consequences of Changing Tumor Anatomy Vs. Positional Uncertainty for Conventionally Fractionated Lung Cancer Patients

Purpose: To differentiate between the dosimetric consequences of positional uncertainty and changing tumor anatomy for conventionally fractionated lung cancer treatment plans. Methods: Modified IMRT photon treatment plans were retrospectively generated for 12 lung cancer patients (prescription range = 45–66 Gy). 10 of 12 patients had centrally located tumors, and the average internal target volume (ITV) was 119.5 cc (standard deviation = 133.3 cc). To simulate a large and immediate tumor response, all voxels corresponding to tumor within the ITV were reassigned an electron density equivalent to lung. The original, unchanged treatment plan was then calculated on this modified image set. To simulate positional uncertainty, isocenters in each plan were shifted by 2 mm and 5 mm in each of 6 directions: superior/inferior, left/right, anterior/posterior. Shifted plans were calculated for both the original and modified ITV. Paired t‐tests were used to compare the original and modified plans, both with and without simulated positional uncertainty. For the purposes of this abstract, changes to the volume of the PTV receiving 95% of prescription dose (PTV V95%), mean lung dose, the percent volume of lung receiving 20 Gy (Lung V20%), the volume in ccs of esophagus receiving 55 Gy (V55Gy(cc)), and the volume in ccs of spinal cord receiving 45 Gy (V45Gy(cc)) are reported. Results: Reduced tumor volumes alone resulted in minimal dosimetric impact. Shifts of 2 and 5 mm reduced PTV coverage by 0.54% and 3.3% respectively when compared to the original (unshifted) plan. When comparing modified and original plans which are both shifted in the same direction, small increases in dose were seen. Conclusion: The dosimetric consequences of positional uncertainty are larger than volumetric tumor reduction. Adaptive replanning techniques based solely on the perceived dosimetric consequences of shrinking tumor anatomy are unwarranted for conventionally fractionated lung cancer patients.