On-Treatment Changes in Gross Tumor Volume of Pediatric Parameningeal Rhabdomyosarcoma Treated With Upfront Proton Therapy

Abstract

<h3>Purpose/Objective(s)</h3> To characterize the time trends of on-treatment morphologic change in pediatric parameningeal rhabdomyosarcoma treated with upfront proton therapy and examine the feasibility of identifying patients who may benefit from on-treatment CT/MRI to address significant changes. <h3>Materials/Methods</h3> We retrospectively analyzed 69 pretreatment and weekly on-treatment MRI as well as simulation and repeat CT acquired from 10 patients (aged 1 – 21 years) with parameningeal rhabdomyosarcoma, who were treated with chemotherapy and upfront proton therapy in a prospective trial. A radiation oncologist delineated GTV on each MRI by modifying the GTV contour from the initial simulation MRI (reference). Longitudinal changes in the volume and surface of GTV from the reference GTV were quantified to examine the feasibility of predicting the on-treatment behavior. The surface displacement was measured by the 90% Hausdorff distance. We investigated the association between the pattern grouping and the clinical need for repeat CT and/or replanning as well as potential image features of simulation CT in differentiating the groups. <h3>Results</h3> The median time was 6.5 days from the start of chemotherapy to CT/MR simulation (ranged -6 to 12 days) and 14 days from simulation to the start of proton therapy (ranged 6 to 21 days). Most (8/10) patients showed continuously decreasing trends in GTV from the simulation (slope, 0.13 – 0.88 %/day), except for two whose GTV reduced only after the start of proton therapy. The GTV reduction was > 20% in 6 patients, which occurred in 13 - 31 days from the simulation, while it was < 10% in the other 4 patients up to 38 days. Excluding a case with orbital involvement, all patients with early significant GTV reduction also showed a distinctively greater surface displacement (> 3 mm) than those in the other group (< 2 mm) at as early as 13 days from the simulation. Image features that most differentiated the group of significant reduction from the other group, was a lower mean HU intensity (70 ± 19 vs. 115 ± 40) and a higher inverse-elongation parameter (0.81 ± 0.05 vs. 0.63 ± 0.12), which suggested a potential relationship between the reduction with a lower occupancy of bony structures and a rounded shape. Among the 6 patients exhibiting significant GTV reduction, 5 received a repeat CT in 10 days from the start of proton therapy for replanning (n = 4) or plan verification (n = 1), whereas any repeat CT prescribed for the other group, on 21 - 23 days from the start of proton therapy, was solely for the scheduled boost phase simulation. <h3>Conclusion</h3> The early significant GTV change is common in pediatric parameningeal rhabdomyosarcoma receiving upfront proton therapy, which often needs repeat CT for dose recalculation and even adaptive replanning. Image features reflecting bone content and shape of the initial GTV appear promising for predicting the extent of GTV reduction.