| Pheochromocytoma and paraganglioma(termed as PPGL)are kinds of rare neuroendocrine tumors,originating from adrenal medullary chromaffin cells(called PHEO)or extra-adrenal ganglia(called PGL),respectively.Its annual incidence rate is about 6 per million,and the prevalence rate is about 0.05%.PPGL is characterized by the synthesis and secretion of excessive catecholamine and hypertension.All PPGL has metastatic potential,and the occurrence of metastasis in PPGL would lead to a significant decrease in survival rate.Thus,an early evaluation of metastatic potential is of vital importance for a better prognosis for patients with PPGL.However,it remains a challenge to obtain a definitive diagnosis of metastatic PPGL in clinical practice as malignancy can only be determined after the presence of distant metastasis in non-chromaffin tissues.Many factors had been reported to be related to metastatic PPGL,including clinical,pathological,genetic mutations,molecular markers,etc.Among them,genetic mutations play an important role in the development and progression of PPGL,while comprehensive prediction models including comprehensive genetic mutations characteristics are still lacking.Purpose: The purpose of this study was to identify the clinical,pathological,and genetic predictors as well as establish a comprehensive prediction model for metastatic PPGL,and further more explore the role of molecular markers in predicting PPGL metastasis and prognosis.Methods:(1)The complete clinical data,pathological reports and paraffin samples(or peripheral blood samples)of 242 patients with surgery removed and pathological diagnosed PPGL were retrospectively collected.Next-generation sequencing(NGS)was performed in all PPGL patients to comprehensively explore the genetic background.The gene mutation map and genotype-phenotype of PPGL patients was analyzed,the risk genes associated with metastatic PPGL were screened,and the predictive ability of these gene characteristics on PPGL metastasis was further analyzed.The Kaplan Meier analysis was used for the impact of these gene characteristics on the prognosis of PPGL patients.(2)Clinical data and pathological data of PPGL patients were analyzed by univariate and multivariable analysis to identify the independent predictors for metastatic PPGL.The predictive ability of these characteristics was analyzed,Kaplan Meier survival analysis and Cox proportional risk model were performed.The clinical pathological prediction model was established with metastatic PPGL independent predictors.The receiver operating characteristic curve(ROC)and the area under ROC(AUC)were used to evaluate the model’s differentiation ability,and the calibration curve was used to evaluate the model’s calibration ability.Risk genes were combing with the clinical pathological model to build the comprehensive prediction model,ROC curve and AUC were used to evaluate the differentiation ability of the model,calibration curve was performed to evaluate the calibration ability of the model.The decision curve analysis(DCA)was used to analyze the net benefit of patients.The prediction model with the highest benefit of patients and the best model performance was used to construct the nomograph.(3)Sixty paraffin samples from PPGL patients were selected according to the ratio of metastatic PPGL: non-metastatic PPGL=1:4 by using the propensity score matching(PSM)method.The expression of six molecular markers including Apelin,succinyl-Co A synthetase subunit beta(SUCLG2),human epidermal growth factor receptor-2(HER-2),contact 4(CNTN4),chromogranin B(CHGB)and succinate dehydrogenase B(SDHB)in metastatic and non-metastatic PPGL were analyzed by immunohistochemical(IHC).Kaplan Meier analysis and Cox proportional risk models were used to analyze the prognostic value of differentially expressed molecular markers in PPGL patients.Results:(1)Harvey rat sarcoma viral oncogene(HRAS)、von Hippel-Lindau(VHL)、rearranged during transfection(RET)、fibroblast growth factor receptor 1(FGFR1)and neurofibromatosis type 1(NF1)mutation were the top-five mutations in Chinese PPGL patients.The frequency of VHL mutation(38.5% vs 11.4%;p=0.016),succinate dehydrogenase subunits(SDHx)mutation(23.1% vs 4.4%,p=0.025),and Cluster 1 mutation(61.5%vs 20.5%,p=0.002)in patients with metastatic PPGL were significantly higher than those in patients with non-metastatic PPGL;Kaplan Meier survival analysis showed that patients with SDHx mutation,VHL mutation,SDHx mutation or VHL mutation and Cluster 1 mutation had significantly shorter metastasis free survival.(2)Three clinical or pathological characteristics including age ≤ 35years(odds ratio(OR)6.430,p=0.007),tumor size ≥ 7cm(OR 8.179,p=0.003),and vascular invasion(OR 7.687,p=0.002)were independently associated with metastatic PPGL.Cox proportional hazard model showed that age ≤ 35 years(hazard ratios(HR)6.080,p=0.002),tumor size ≥ 7cm(HR 7.331,p=0.001)and vascular invasion(HR 5.124,p=0.004)were independent predictors of poor prognosis of PPGL.The three independent predictors were used to construct the clinical and pathological prediction model of PPGL metastasis risk.The AUC of the model was 0.844(95% CI0.722,0.967),the specificity was 91.3% and the sensitivity was 69.2%.The AUC after calibration is 0.827,and the Hosmer-Lemeshow test p=0.976,indicating that the model fits well.The SDHx mutation,VHL mutation,SDHx mutation or VHL mutation and Cluster 1 mutation were included in the clinical pathological model to build 4 comprehensive prediction models of PPGL metastasis risk.The results showed that the SDHx mutation or VHL mutation combined with clinical and pathological characteristics had the highest comprehensive prediction performance,with an AUC of 0.884(95% CI 0.772,0.997),specificity of 83.8%,and sensitivity of 84.6%.DCA showed that compared with the clinical pathological model,the comprehensive prediction model produced higher net benefits,indicating the reliability of the comprehensive prediction model in predicting the metastatic risk of PPGL(3)Compared with patients with non-metastatic PPGL,the expression of Apelin,SDHB and CHGB in patients with metastatic PPGL was significantly lower(p<0.001),while the expression of HER-2 was significantly higher(p=0.042);there was no significant difference in the expression of CNTN4 and SUCLG2 between metastatic and nonmetastatic PPGL.Kaplan Meier analysis showed that the metastasis free survival rate of patients with negative expression of Apelin,SDHB and CHGB was significantly lower than that of patients with positive expression;Cox analysis showed that the low expression of SDHB and CHGB were independent factors regarding the poor prognosis of PPGL patients.Conclusion:(1)Age ≤ 35 years old,tumor size ≥ 7cm and vascular invasion are independent risk factors of metastatic PPGL and can be used as independent predictors of poor prognosis of PPGL;SDHx,VHL and Cluster 1 are risk genes for metastatic PPGL,and patients with SDHx,VHL or Cluster 1 mutations had poor prognosis.(2)Combining gene mutation features can provide additional value for prediction models to help predict the metastatic risk of PPGL;the comprehensive prediction model nomogram established by clinical,pathological and gene mutation features can be used as an independent tool to evaluate the metastasis risk of individual patients in clinical practice,and help clinicians make decisions at the same time.(3)The expression levels of Apelin,CHGB,SDHB and HER-2 may be molecular markers for the diagnosis of metastatic PPGL;patients with negative expression of Apelin,CHGB and SDHB have worse prognosis,and should be followed up more closely after operation.19 figures,16 tables,123 references... |
PDF Full Text Request