Dose Response Relationship and Prognostic Factors of Nodal Control Rate of Metastatic Lymph Nodes in Cervical Cancer

Abstract

In contrast to excellent local control in definitive aim radiation therapy (RT) of cervical cancer, control rate of metastatic lymph nodes remains insufficient. Therefore, radiation boost to metastatic lymph nodes has been attempted, but little is known about the optimal dose and prognostic factors of regional nodal control. In this study, we retrospectively investigated the dose response relationship of metastatic lymph nodes as well as prognostic factors of nodal control, and attempted to propose optimal dose for individual metastatic lymph nodes. We identified 90 patients (295 lymph nodes) who received definitive aim RT for node positive cervical cancer between 2013 and 2017. Total prescribed dose to each metastatic lymph node was converted to 2 Gy equivalent dose (EQD2, α/β = 10 Gy). Each node was tracked individually from initial diagnosis to any recurrence in the same specific node. Any tumor or treatment factors predictive of nodal control were searched. In addition, we attempted to stratify risk groups in order to find optimal dose for nodal control in a certain risk group. The median number of metastatic lymph nodes per patient was 2 (range, 1-14). Regarding total lymph nodes, the most common pathology was squamous cell carcinoma (85.6%) followed by adenocarcinoma (8.9%) and others (5.6%). The mean nodal volume at initial diagnosis was 5.0 ± 9.2 mL, and the mean total prescribed dose in EQD2 for each node was 55.0 ± 6.8 Gy. With median follow up of 25 months (range, 3-57 months), 32 nodal failures occurred, and 2-year nodal control rate was 89%. In multivariate analysis, neutrophil-to-lymphocyte ratio (NLR) higher than 3.0 (p<0.001), para-aortic nodes (vs. pelvic nodes, p=0.013), larger initial nodal volume (p<0.001) and lower total dose (HR 0.90 for increase of 1 Gy in EQD2; 95% CI 0.85-0.95; p<0.001) were identified as risk factors for poor nodal control. Each node was stratified into two risk groups; low risk group with none or one risk factor, and high risk group with two or three risk factors. For risk group analysis, initial nodal volume was dichotomized (≤ 5.3 mL or > 5.3 mL). ROC curve analysis showed that 53.1 Gy and 62.0 Gy was the optimal cut-off dose for predicting nodal control rate in low and high risk group, respectively. In low risk group, doses higher than 53.1 Gy showed significantly superior nodal control rate (p=0.050) with 2-year nodal control rate of 99% (vs. 93% for ≤ 53.1 Gy). In high risk group, doses equal to or higher than 62.0 Gy showed superior nodal control rate with marginal significance (p=0.053) with 2-year nodal control rate of 80% (vs. 62% for < 62.0 Gy). Our data demonstrated that nodal control depends on the prescribed dose to lymph node, NLR, location of lymph node, and initial nodal volume. Therefore, these factors should be considered to enhance nodal control at the stage of treatment plan.