Motion-Inclusive Treatment Planning to Assess Normal Tissue Dose for Central Lung Stereotactic Body Radiation Therapy

Abstract

Purpose

For lung stereotactic body radiation therapy, 4-dimensional computed tomography is often used to delineate target volumes, whereas organs at risk (OARs) are typically outlined on either average intensity projection (AIP) or midventilation (MidV = 30% phase) images. AIP has been widely adopted as it represents a true average, but image blurring often precludes accurate contouring of critical structures such as central airways. Here, we compare AIP versus MidV planning for centrally located tumors via respiratory motion-inclusive (RMI) plans to better evaluate dose delivered throughout the breathing cycle.

Methods and Materials

Independently contoured and optimized AIP and MidV plans were created for 16 treatments and rigidly copied to each of the 10 breathing phase-specific computed tomography image sets. Resulting dose distributions were deformably registered back to the MidV image set (used as reference because of clearer depiction of anatomy compared with motion-blurred AIP) and averaged to create RMI plans. Doses to central OARs were compared between plans.

Results

Mean absolute dose differences were low for all comparisons (range, 0.01-2.87 Gy); however, individual plans exhibited differences >20 Gy. Dose differences >5 Gy were observed most often for plan comparisons involving AIP-based plans (MidV vs AIP 23, AIP RMI vs AIP 12, MidV RMI vs AIP RMI 7, and MidV RMI vs MidV 8 times). Inclusion of respiratory motion reduced large dose differences. Standard OAR thresholds were exceeded up to 5 times for each plan comparison scenario and always involved proximal bronchial tree D4 cc tolerance dose. AIP-based contours were larger by, on average, 3% to 15%.

Conclusions

Large dose differences were observed when plans with AIP-based contours were compared with MidV-based contours, indicating that observed dose differences were likely due to contoured volume differences rather than the effect of motion. Because of blurring with AIP images, MidV RMI-based planning may offer a more accurate method to determine dose to critical OARs in the presence of respiratory motion.