Minimization of Thoracic Vertebral Bone Marrow Irradiation during Intensity-Modulated Radiation Therapy of Esophageal and Thoracic Malignancies

Purpose: Thoracic vertebral bodies represent about one-quarter of an adult human’s bone marrow. As previous work has demonstrated, treatment of the lumbar spine and pelvis with radiation can induce a fatty replacement or fibrosis of the bone marrow in these regions; this may result in long-term immunosuppressive effects. In this study, thoracic vertebral body bone marrow was contoured as an OAR (Organ at Risk) to evaluate practical feasibility of implementing bone marrow sparing without compromising treatment dose

Methodology: Patients with previously treated thoracic cancer were included in this study. The thoracic bone marrow was contoured using PET scan to verify avid areas and a new plan was created. Guideline constraints for the thoracic bone marrow were V10 ( < 300cc), V20 ( < 200 cc), and V30 ( < 150cc). If bone marrow optimization caused a violation of standard OAR guidelines then the bone marrow constraints were not met. The median changes in thoracic bone marrow and OARs (lung, heart, spinal cord) were compared with the Wilcoxon sign rank test for non-parametric paired samples.

Results: We included 20 patients, 10 with esophageal cancer and 10 with NSCLC (Non-Small Cell Lung Cancer). Only three cases were able to meet the prespecified bone marrow constraints after optimization (15%). Bone marrow mean V10 was reduced by 20.29cc (-7.64%). More dramatic decreases were seen V20 and V30 dose-volume reductions, with mean decreases of 38.33 and 25.69cc (-17.03% for both), respectively. These chances were found to be significant for V30 (Paired T < 0.00001), V20 (Paired T =0.001), and V10 (p=0.0005). Though most experienced a decrease in mean heart dose, there was an overall mean increase of 20.52cGy, attributable to a single outlier(P=1). Mean max cord dose was decreased by 253.47cGy (-6.27%). There was a minimal non-significant decrease in mean Lung V20 of 0.57% as well as minimal mean decrease in the mean lung dose (0.07%).

TABLE 1: Mean Dosimetric Changes Pre- and Post-Vertebral Sparing
Before Sparing After Sparing Absolute Change Percent Change Significance
V10 (cc) 265.48 245.19 -20.29 -7.64% p=0.0005
V20 (cc) 225.13 186.80 -38.33 -17.03% Paired T= 0.001
V30 (cc) 150.89 125.20 -25.69 -17.03% Paired T < 0.00001
Mean Heart (cGy) 1497.83 1518.36 20.52 1.37% P=0.24
Cord Max (cGy) 4041.47 3788.00 -253.47 -6.27% p=0.8
Total Lung Mean (cGy) 1422.57 1421.59 -0.98 -0.07% p=0.83
V20 lung (cc) 25.75 25.61 -0.15 -0.57% p=0.6

Conclusions: This retrospective dosimetric analysis shows the feasibility of bone marrow optimization. Though we did not entirely eliminate thoracic vertebral body exposure, we did note significant decreases in mean dosages without compromise of OAR or treatment dosage. Importantly, bone marrow optimization was able to be performed without leading to significant increase in dose to the heart, lungs, nor spinal cord.

References: