Dissimilarities of Regression Between Primary Tumor and Regionally-Involved Lymph Nodes in Locally Advanced Non–Small Cell Lung Cancer Treated By Concurrent Chemoradiotherapy

Abstract

Image-guided individual treatment based on regional tumor control rate has been a research hotspot of locally-advanced NSCLC. Concurrent chemoradiotherapy has become standard treatment for locally advanced NSCLC. To investigate whether there was dissimilarity of volumetric shrinkage rate (VSR) between primary tumor and regionally-metastasized lymph nodes of locally advanced non-small cell lung cancer (NSCLC) treated by concurrent chemoradiotherapy and to explore the influential factors. A retrospective review identified patients diagnosed with locally advanced NSCLC of stage IIIA or IIIB in Shandong Cancer Hospital between August 1, 2014 and December 31, 2016. Patients who both underwent baseline 18F-FDG PET-CT scan within 4 weeks before treatment and treated with concurrent chemoradiotherapy were enrolled. The target lesions were delineated on each patient’s PET and CT images (enhanced planning CT and follow-up CT) respectively. Parameters such as metabolism parameters were automatically calculated and output as a table by a commercially available deformable registration algorithm. VSR and overall response rate (ORR) was calculated from gross tumor volume (GTV) delineated on CT scans. Relationship between VSR or ORR and metabolism parameters were analyzed. (1) The average VSR of primary tumor was significantly higher than the average VSR of regionally-metastasized lymph nodes (59.8% vs. 34.14%, Z= -3.235, P =0.001). SUVmax of lymph node (SUVmax-nd) was the only significant influential factor to VSR in multivariate analysis. (2) ORR in primary tumor was 56.3% and in regionally-metastasized lymph nodes was 31.0%. ORR was much more significant in primary tumor (χ2 = 6.326, P = 0.012). Similarly, SUVmax was the only significant influential factor to primary tumor and regionally-involved lymph nodes multivariate analysis. (3) The receiver operating characteristic curve (ROC) analysis showed SUVmax-nd=6.91 was an optimally predictive cut-off value of short-term efficacy in regionally-metastasized lymph nodes, but one was not found in primary tumor. The average VSR of regionally-involved lymph nodes with SUVmax-nd≥6.91 was 57.23% (95%CI: 45.47%-68.98%) versus those with SUVmax-nd＜6.91 was 20.04% (95%CI: 3.33%-36.74%). Two groups had statistical significance (Z = -3.438, P =0.001). ORR of regionally-involved lymph nodes with SUVmax-nd≥6.91 was 51.5% versus those with SUVmax-nd＜6.91 was 20.4% and they were statistically significant (χ2 =9.103, P =0.003). In locally advanced NSCLC with concurrent chemoradiotherapy, VSR in primary tumor was more pronounced than regionally-involved lymph nodes and pretreatment SUVmax-nd may be a predictive variable. Similarly, SUVmax-nd was an effective prognostic factor for short-term-efficacy of regionally-involved lymph nodes treated with concurrent chemoradiotherapy. It was proved that SUVmax-nd=6.91 was an optimally predictive cut-off value.