P2.02-046 Prognostic Value of Early Tumor Regression during Chemo-Radiotherapy in Locally Advanced Non-Small Cell Lung Cancer

Abstract

Volumetric changes are observed on serial cone-beam computed tomography (CBCT) images obtained for image-guidance throughout the course of radical radiotherapy for non-small cell lung cancer (NSCLC). This study aims to a) examine whether the magnitude of tumor regression is correlated with disease control and survival; b) explore the potential difference between adenocarcinoma and non-adenocarcinoma NSCLC subtypes. In a previous study from our institution, primary tumor volumes were assessed on weekly CBCT images of 60 NSCLC patients treated with radical radiotherapy from January 2006 to June 2007. We performed a retrospective review of these patients, documenting patient-, tumor-, and treatment-related details. Outcome measures included loco-regional failure free survival (LRFFS), distant failure free survival (DFFS), disease free survival (DFS) and overall survival (OS), which were calculated using Kaplan-Meier method. Uni-variable analysis (UVA) and multivariable analysis (MVA) were performed using Cox regression model. Further analysis was performed for the adenocarcinoma and non-adenocarcinoma subgroups. Forty-five patients with locally advanced NSCLC were included in this study. Median follow-up was 22.1 months for all patients, and 90 months for alive patients (range: 0.9-108). The distribution of 7th ed. AJCC stage was as follows: stage II 8.9%; IIIA 66.7%; IIIB 24.4%. Twenty patients (44.4%) had adenocarcinoma, while 25 patients (55.6%) had non-adenocarcinoma. Twenty-eight patients (62.3%) received total radiation dose ≥ 60Gy, 15 patients (33.3%) received 45Gy as neoadjuvant therapy, and 2 patients (4.4%) received 58-59Gy due to missed fractions. 23 patients (51.1%) had more than 30% regression by fraction 15 and 32 patients (71.1%) by treatment completion. In UVA, adenocarcinoma (p=0.03) was associated with better LRFFS; young age was associated with better LRFFS (p=0.02), DFFS (p=0.048) and OS (p=0.04). In MVA, large regression by fraction 15 was associated with better DFS (p=0.047). For patients with adenocarcinoma, MVA showed that large regression by fraction 15 was associated with better DFFS (p=0.01), DFS (p=0.01) and OS (p=0.02). For patients with non-adenocarcinoma, larger regression by treatment completion and trimodality therapy (radiation dose of 45Gy) were associated with better LRFFS (p=0.02, 0.04). Evaluation of tumor regression on CBCT images during radiotherapy may be predictive of treatment response. Early tumor regression, as indicated by regression ≥ 30% by fraction 15, was associated with better DFS for all patients; and this was associated with better DFFS, DFS and OS for the adenocarcinoma cohort. This observation may provide insight into when and how to best utilize adaptive radiotherapy.