Pre- and Post-treatment PET-CT Image Analysis of Pulsed-low-dose-rate (PLDR) Compared to Conventional RT Delivery for Initial Treatment of Esophageal Tumors

Purpose/Objective(s): The purpose of this study is to compare the response of PET-defined tumor to PLDR radiotherapy to patients treated with conventional delivery through evaluation of pre (scan-1) and post-treatment (scan-2) PET-CT scans of patients not previously irradiated. Materials/Methods: Data for 9 patients treated for esophageal cancer on an IRB-approved Phase I study of neoadjuvant PLDR RT with concurrent chemotherapy for treatment of NSCLC or esophageal cancer were evaluated. The patients were treated to 50.4 Gy in 28 fractions via IMRT or VMAT, prior to surgery. All scan-1 and scan-2 data sets were fused based on boney anatomy with RT planning software. PET volumes were delineated by an experienced Nuclear Medicine physician and a standardized uptake value (SUV) threshold of 3.0 was used to designate tumor. Both scan-1 and 2 PET-defined volumes were delineated within the PTV assuring they received the prescribed dose and all volumes superimposed onto each CT data set through the fusion process and corrected for low density inclusions (air). The average volumes and mean HU values on each CT scan were obtained. The post-treatment mean HU values were corrected for difference in scanner performance. A CT-normalized mean SUV was determined for the volumes on both scan sets according to aveSUVCT-norm = aveSUV/aveHU. This value reflects active tumor regardless of density or thickness differences that may not be observed from CT alone. Data were compared with those from 18 sequentially selected, esophageal cancer patients conventionally treated to the same dose with concurrent chemotherapy using Fisher’s exact and Wilcoxon rank sum tests. Volume reduction and complete response (CR) (no voxels with SUV >3.0) were evaluated using Tweedie generalized linear models and logistic regression, respectively. Results: Under direct comparison, no significant difference was found between demographic and tumor characteristics and the two groups are well balanced. The mean pre to post-treatment percent volume reduction was 90.3% for PLDR and 74.7% for conventional delivery (p=0.022). Volume reduction was also greater in node negative (p=0.037) and younger patients (p=0.065). The CR rate of PLDR was 33.3% vs. 16.7% for conventional delivery (p=0.203). The mean HU value decrease was not significant. The average normalized mean SUV decrease from scan-1 to 2 for pre-treatment volumes was greater for PLDR delivery by approximately 22.8% (ns). This parameter for the post-treatment volume was greater for PLDR delivery by 6.8% (ns). Conclusion: The decreases in PET-defined volumes and the trends in improved normalized mean SUV and CR rate for PLDR are indications suggestive of treatment effectiveness. Given these results and the small sample size, PLDR delivery may even prove superior to conventional delivery for these patients.