Impact of Para-Aortic Lymph Node Count on Survival in Patients with Cervical Cancer and Para-Aortic Lymph Node Involvement Treated with Definitive Radiation Therapy

Abstract

<h3>Purpose/Objective(s)</h3> Definitive chemoradiation is the standard treatment for loco-regionally advanced cervical cancer. One of the major changes in the 2018 FIGO staging update was the inclusion of para-aortic lymph node (PAN) involvement in a new IIIC2 category. However, the impact of the extent of para-aortic involvement on treatment outcomes remains unclear. This study aims to correlate features of PAN metastases with treatment outcomes in order to inform future treatment strategies for these patients. <h3>Materials/Methods</h3> Electronic medical databases were queried to identify all cervical cancer patients with FDG-avid PAN metastases treated with curative intent using extended field radiotherapy +/- concurrent chemotherapy at our institution between 2004 to 2018. Patients with neuroendocrine histology or distant metastases were excluded. We determined the number of FDG-avid PANs; distance between the highest PAN and the renal vessels or aortic bifurcation; and the size of the largest PAN. Analysis of variance, Kaplan-Meier method, log-rank tests, and Cox proportional hazard tests were used for analysis. <h3>Results</h3> The 5-year OS of the 102 patients in this study was 42% (95% CI, 32-52%), and the 5-year DMFS was 49% (95% CI, 39-59%). The 5-year OS for patients with non-squamous cancers metastatic to the PANs (N=18) vs. patients with squamous cancers metastatic to the PANs (N=84) were 28% and 45%, respectively (<i>P</i> = 0.075). None of the PAN characteristics studied were significantly associated with OS. However, DMFS was strongly associated with the number of FDG-avid PANs; DMFS of patients with <5 or ≥5 PANs were 56% and 17%, respectively (<i>P</i> = 0.0002). PAN size and the distance between the highest PAN and the renal vessels were not significantly correlated with DMFS. On univariate analysis, age (<i>P</i> = 0.12), BMI (<i>P =</i> 0.91), FIGO 1994 stage (<i>P</i> = 0.35), primary lesion size (<i>P</i> = 0.31), or histologic type (<i>P</i> = 0.99) were not significant predictors of PAN count. None of the 3 patients with more than 5 PAN mets from non-squamous cervical cancers survived. For patients with < 5 positive PANs, those with non-squamous cancers (15 patients) had a 5-year OS of 33%, vs. 44% for 77 patients with squamous histology (<i>P</i> = N.S.). On multivariate analysis, PAN count was significantly associated with 5-year DMFS but not OS. <h3>Conclusion</h3> Although the number of FDG+ PANs was correlated with poor DMFS, proximity to the renal vessels and nodal size appear to have less influence on outcome. The presence of both PAN count 5 or higher and adenocarcinoma histology conferred very poor outcomes. Although even patients with extensive PAN metastases are frequently curable with RT, these data suggest that patients with many PAN metastases from non-squamous cervical cancers may benefit from strategies that emphasize systemic treatments.