Affiliation:
1. Sun Yat-sen University Cancer Center
2. The Fifth Affiliated Hospital of Guangzhou Medical University
Abstract
Abstract
Objectives
To investigate the value of negative lymph node (LNneg) in prognostic prediction and therapeutic implications in the N0 subgroup (T3–4N0M0) of locoregionally advanced nasopharyngeal carcinoma (NPC).
Materials and Methods
All patients in this retrospective cohort study were consecutively extracted from an NPC-specific database (N = 15,126) and treated using concurrent chemoradiotherapy (CCRT) with/without induction chemotherapy (IC) during 2009–2017. Cervical LNneg distribution was dichotomized per MRI-based features as the concentrated and dispersed types. The association of overall survival (OS) with LNneg size, distribution, and regression was investigated using Cox analysis. LNneg regression was explored from three aspects: speed, extent, and overall pattern.
Results
In 724 included patients (mean age: 47 +/- 11 [standard deviation], 533 men), the dispersed type of cervical LNneg had a significant higher 5-year OS than the concentrated type (95.0% vs. 89.0%; P-value = .005), which was mainly due to its smaller nodal metastasis rate (3.2% vs. 13.0%) and validated in low-infection status (albumin > 40g/L, C-reactive protein ≤ 3mg/L, lactate dehydrogenase ≤ 250U/L). IC + CCRT and CCRT with ≥ 200mg cisplatin benefited the concentrated (HR = 0.47 [95% CI = 0.22–0.98]; P-value = .004) and dispersed types (HR = 0.18 [0.06–0.54]; P-value = .002), respectively. IC + CCRT induced a generally greater and faster LNneg regression than CCRT. The concentrated type preferred to show an overall regression pattern than the dispersed type. Reduction in short/long axial diameter of the largest cervical LNneg of ≥ 3.0mm/4.0mm was an OS improvement indicator.
Conclusion
MRI-based cervical LNneg distribution and regression predicted prognosis and identified high-risk cases of the N0 patients with locoregionally advanced NPC (i.e., concentrated type) to receive IC + CCRT.
Publisher
Research Square Platform LLC