Substaging Of Newly Diagnosed Metastatic Nasopharyngeal Carcinoma

Objective:Nearly 10% of new nasopharyngeal carcinoma(NPC)patients are accompanied by distant metastasis,and these patients often have a poor prognosis.There is obvious heterogeneity in de novo metastatic nasopharyngeal carcinoma.Some patients can survive for a long time,while some patients have a short survival time.In the current staging of nasopharyngeal carcinoma,all newly diagnosed metastatic nasopharyngeal carcinomas are classified as M1 stage,which brings limitations to prognostic evaluation and guidance of treatment.Therefore,this study aimed to further sub-stratify patients with M1 nasopharyngeal carcinoma,establish a substaging model,and explore its value in assessing prognosis and treatment guidance.Methods:This study retrospectively collected patients with newly diagnosed metastatic nasopharyngeal carcinoma in Jiangxi Cancer Hospital and Sun Yat-sen University Cancer Center from 2010 to 2019,and divided them into training cohort and validation cohort.Inclusion criteria included:(1)Histopathology confirmed nasopharyngeal carcinoma patients;(2)All patients had complete pre-treatment whole-body imaging evaluation and confirmed distant metastasis;(3)received anti-tumor therapy,including at least 2 cycles Intravenous chemotherapy +/-radiotherapy;Exclusion criteria:(1)conventional radiotherapy in the nasopharyngeal and cervical lymph node regions;(2)with a history of other malignant tumors;(3)coexisting serious illness.Kaplan–Meier univariate analysis was used to evaluate the overall survival of newly diagnosed metastatic nasopharyngeal carcinoma,and the Log-rank test was used.Multivariate adjusted analyses were performed using Cox proportional hazards models to explore independent prognostic factors for metastatic nasopharyngeal carcinoma.Concordance index(C-index)and time-dependent ROC curve was used to evaluate the accuracy of different oligo metastatic models for prognostic assessment.Statistical analysis was performed using SPSS(IBM 26.0)and R language(version 3.1.6).All statistical tests were two-sided and p<0.05 was considered statistically significant.Results:The training cohort included 197 patients with a median follow-up of 46 months(range 3-123 months)and a 3-year overall survival(OS)of 47.1%;the validation cohort included 307 patients with a median follow-up of 53 months(range 4-138months),with a 3-year OS of 65.0%.The comparison results of different oligo metastasis models showed that: in the training cohort,the C-index of oligo metastasis model 5(with ≤2 metastatic organs and ≤5 metastatic lesions)was 0.6230,which was higher than the other five oligo metastasis models;The consistency index of model 5(C-index=0.572)in the validation cohort was only slightly inferior to model 3(C-index=0.575).Therefore,in this study,oligo metastasis were defined as no more than 2 metastatic organs and no more than 5 metastatic lesions.Multivariate analysis of non-coexistence of oligo metastasis status and liver metastasis status showed that in the training and validation cohorts,liver metastases(HR=1.740,95% CI1.128-2.685,p=0.012;HR=1.615,95% CI 1.117-2.336,P=0.011)and oligo metastasis(HR=2.104,95% CI 1.360-3.257,p=0.001;HR=1.555,95%CI=1.102-2.194,p=0.012)were independent prognostic factors for OS.Stratified analysis found that in the oligo metastasis group,3-year OS was inferior in both the training cohort(46.7% vs.66.0%,p=0.027)and the validation cohort(44.3% vs.77.1%,p<0.001)for patients with liver metastases and without liver metastases.In the case of multiple metastases,there was no significant difference in OS between patients with liver metastases and those without liver metastases in the training and validation cohorts.Based on the status of liver metastases and oligo metastasis,a new substratification model was constructed,defining M1 a as oligo metastasis without liver metastases,M1 b as oligo metastasis with liver metastases,and M1 c as multiple metastases.Kaplan–Meier univariate analysis found that M1 a,M1b,and M1 c had significantly different survival outcomes(3-year OS: 66.0% vs.46.7% vs.44.1%,p <0.001),and consistent conclusions were obtained in the validation cohort(3-year OS was 77.1% vs.44.3% vs.56.0%,p<0.001).After combining the training cohort and validation cohort data,patients with M1 a who received locoregional radiotherapy had better outcomes(3-year OS: 79.1% vs.58.7%,p<0.001),while locoregional radiotherapy did not improve OS in patients with M1b(3-year OS: 51.7% vs.38.0%,p=0.616)and M1c(46.6% vs.40.2%,p=0.064).Considering that the survival and prognosis of M1 b and M1 c groups are similar,and locoregional radiotherapy cannot improve the prognosis of patients,the study combined the two,and revised the substaging model to M1a(oligo metastasis and no liver metastasis)and M1b(liver metastasis or multiple metastasis).The proposed M1 substaging model has a C-index of 0.6258 and 0.6011 for prognostic assessment in the training and validation cohorts,respectively.conclusionWe established a sub-staging model of de novo metastatic nasopharyngeal carcinoma based on anatomical features: M1a(oligo metastasis and no liver metastasis)and M1b(liver metastasis or multiple metastasis).The new staging can not only effectively evaluate the overall survival of patients,but it also has potential value in guiding locoregional radiotherapy therapy.