| Research BackgroundColon cancer is the fourth most common malignant tumor.In 2020,there were 1.76 million new cases of digestive tract tumors in China,of which about 560,000 new cases were colon cancer.Approximately 50%-60%of colon cancer patients have metastatic disease at the time of diagnosis,and 75-85%of them have unresectable metastases.Different from the strategy of early-stage cancer,which is mainly based on surgery,the treatment of advanced cancer is primarily multidisciplinary management.The treatment goals of advanced cancer patients are to improve their quality of life and prolong their survival period.Primary tumor resection(PTR)is advocated based on concerns about complications associated with an intact primary tumor.In patients with asymptomatic primary tumor and unresectable metastases,the impact of PTR on survival and quality of life remains controversial.This paper aims to study the effect of PTR on the survival of unresectable advanced colon cancer,how to select suitable patients for PTR and the potential mechanism of PTR.Objective(1)This paper analyzes the effect of PTR on the survival of patients with unresectable advanced colon cancer by conducting a retrospective study on the public database,and establishes a postoperative model to predict survival of patients undergoing PTR.At the same time,we conduct a pre-treatment prediction model and pre-surgery score to screen patients who are suitable for PTR.(2)We conduct a single-center prospective non-randomized controlled trial to validate the therapeutic value of PTR in patients with asymptomatic unresectable advanced colon cancer in our hospital.We also validate the predictive models and the PTR screening score in our hospital.(3)We evaluate the plasma exosome levels before and after PTR in patients with unresectable advanced colon cancer.We analyze the effect of plasma exosomes on CD4+T cells,in order to explore the mechanism by which PTR improves survival.Methods(1)We extract data from 18 registration sites in the Surveillance,Epidemiology,and End Results Program(SEER)database to identify patients with advanced metastatic colon cancer.Inclusion criteria:pathologically diagnosed colon cancer,stage M1,unresectable metastases,clear clinicopathological features and survival information,and no other primary tumors.The exclusion criteria:patients receiving combined resection of primary tumor and metastases,primary tumor destruction,unclear relevant information,and survival period less than 1 month.(2)We organize and recode the information of advanced colon cancer patients.Patients are divided into PTR group and non-PTR group,and the chi-square(χ2)test is used to analyze the statistical differences of baseline characteristics(such as age at diagnosis,gender,race,primary tumor site and tumor grade,etc.)between the two groups.To simulate randomization in randomized controlled trials as much as possible,we match the two groups of patients using propensity score matching(PSM)analysis to reduce confounding bias between the two groups.The difference of overall survival(OS)between the matched two groups is compared using the log-rank test,and a K-M survival curve is drawn.(3)To build the post-surgery model which could predict the survival of patients undergoing PTR,we first assign these patients 7:3 randomly to construct the derivation set and validation set.Multivariate Cox regression analysis is used to evaluate all potential risk factors for postoperative survival,and variables with P<0.05 are included in the construction of nomograms as independent prognostic factors.We use the bootstrap selfsampling method to internally validate the nomogram and calculate the C-index as a performance indicator for the nomogram predictive model.At the same time,calibration curves of the nomogram are created to show the agreement between the observed frequency of the true outcome and the predicted probability.The survival score obtained by the post-surgery model is called "post-surgery survival score".We use X-tile analysis to obtain its best cut-off value as the decision point for clinical judgment of prognosis.We divide patients undergoing PTR into different mortality risk subsets by this cut-off value.(4)We use the same method to build a pre-treatment predictive model based on the prognostic factors before treatment.We name the score obtained from the pre-treatment predictive model as the "pre-treatment survival score",which is used to predict patient survival at the time of diagnosis.(5)In addition,we use the pre-surgery factors in the pre-treatment predictive model to score the patients,named "pre-surgery score".Then,the best cut-off value of the "presurgery score" is obtained by X-tile analysis,and the patients are divided into preoperative high-risk and low-risk subsets.We then compare survival in high-risk and low-risk subsets of patients who undergo PTR with those who do not.We speculate that patients with a low "pre-surgery score" risk would benefit most from PTR,who are thought to be the PTR-benefit population.(6)We conduct a single-center,prospective,non-randomized controlled trial to compare the efficacy of PTR combined with postoperative chemotherapy versus chemotherapy alone in patients with asymptomatic unresectable advanced colon cancer.The performance status(ECOG)score of the included patients is<2,and these patients are divided into PTR group and non-PTR group,and the chemotherapy regimen is mFOLFOX6 or Bevacizumab+mFOLFOX6.The primary outcome is OS.Secondary outcomes were as follows:progression free survival(PFS),objective response rate(ORR),incidence of chemotherapy-induced adverse events,incidence of postoperative complications,and serious bowel complications symptoms(requiring emergency surgical intervention).In addition,we validate our predictive models by plotting the Receiver Operating Characteristic(ROC)curves.We also score the patients using Ppre-surgery,and analyze whether Ppre-surgery combined with its cut-off value could screen out the PTR-benefit population.(8)We collect blood samples before and after surgery from 6 patients in the PTR group.Plasma exosomes are obtained by differential ultracentrifugation,and identified by transmission electron microscopy(TEM),nanoparticle tracking analysis(NTA)and Western Blot.We compare plasma exosome levels in advanced colon cancer patients and healthy individuals.We also compare the plasma exosome levels in patients before surgery,7 days after surgery,and 15 days after surgery.We download the plasma exosome mRNA expression profile of colon cancer from the public database exoRbase.Then we use limma differential analysis to obtain the up-regulated mRNAs in the plasma exosome of colon cancer patients.We also analyze the differential expression of CD274(PD-L1)in plasma exosome between cancer patients and healthy individuals.We adopt flow cytometry to detect IL-2 and INF-γ production from activated CD4+T cells,which are co-cultured with plasma exosomes.In addition,we use Western Blot to examine the expression of PD-1 in naive CD4+T cells,activated CD4+T cells and activated CD4+T cells co-cultured with plasma exosomes.Results(1)We found that the patients who chose PTR for the initial treatment had longer survival than those without PTR(median OS:20 vs.6 months,P<0.001).This trend persisted in subgroup analyses:different metastatic sites,different diagnosed years,different age groups,and left or right colon cancer.(2)We constructed a postoperative nomogram predictive model to predict survival in patients undergoing PTR,and this predictive model performed well(the C-index was 0.715,and the calibration curve showed that the predicted and observed survival rates were basically the same).Postoperative survival score:Ppost-surgery=Page+Psite+Pgrade+PTstage+PLNexamined+PLNpositive+Pchemo Ppost-surgery with 22.3 points and 27.7 points can be used to stratify the postoperative risk of patients with PTR.(3)The diagnostic year of patients in which we constructed the pre-treatment model was 2010-2017.The nomogram prediction model based on pre-treatment risk factors could simulate the actual survival rate of patients well and C-index was 0.724.Ppre-treatment=P’age+P’site+P’grade+P’Tsize+P’cLN+P’liver+P’lung+P’bone+P’brain+P’PTR+P’chemo(4)We establish a preoperative score:Ppre-surgery=P’age+P’site+P’grade+P’Tsize+P’cLN+P’liver+P’lung+P’bone+P’brain.The best cut-off value of Ppre-surgery determined by X-tile was 13.2.If the patient’s Ppre-surgery≤13.2.the median OS of patients who received PTR was significantly longer than that of patients who did not receive surgery(HR=0.43,95%Cl:0.39-0.474,P<0.001).These patients belonged to the "surgery benefit population".(5)The results of the prospective study in our hospital showed that the median OS of the PTR group was longer than that of the non-PTR group(21.8 months vs.17.4 months,HR=0.59,95%CI,0.37-0.93,P=0.02).In the PTR group,the median survival time of Bevacizumab+mFOLFOX6 was 25.4 months,which was longer than that of mFOLFOX6.Although no complete remission was observed in both groups,ORR appeared to be higher in the PTR group than in the non-PTR group(21.25%vs.10.00%,P=0.05).We observed a longer median PFS in the PTR group than in the non-PTR group(10.9 months vs.8 months,HR=0.45,95%CI,0.28-0.73,P<0.001).There was no significant difference in the incidence of serious chemotherapy adverse events and the incidence of emergency surgery between the PTR group and the non-PTR group.(6)The baseline comparison between the unresectable colon cancer patients in our hospital and those in the SEER database showed that the age of the surgical patients in our hospital was younger than that of the surgical patients in the SEER database,62.3±10.279 years old vs.64.79±13.512 years old,P=0.035.Regional lymph node metastases and liver metastases were statistically different between the two populations,however,gender,tumor location,tumor grade,tumor size,lung metastases,bone metastases,and brain metastases were not statistically different between the two populations.(7)We scored the patients with pre-treatment score and plotted the ROC curve,1-year AUC=0.81,95%CI:0.67-0.95,1.5-year AUC=0.74,95%CI:0.63-0.85,2-year AUC=0.79,95%CI:0.68-0.89.The results show that the pre-treatment prediction model performs well in practical applications.(8)We further verified whether Ppre-surgery could screen out the PTR-benefit population.We scored the patients in our hospital with pre-surgery score,and divided them into PTR-benefit group(50 cases)and PTR non-benefit group(30 cases)with the cut-off score of 13.2.In the PTR-benefit group,survival of patients with PTR was significantly longer than those without PTR(median OS:27.8 months vs.19.1 months,HR=2.34.95%Cl:1.27-4.31.P<0.001).indicating that patients with Ppre-surgery less than 13.2 were eligible for PTR.For patients who did not benefit from surgery,there was no significant difference in survival between PTR and non-PTR(HR=1.58,95%CI:0.74-3.35,P=0.23).(9)We analyzed the prognostic factors of patients undergoing PTR in our hospital,which were basically consistent with those in the SEER database.In addition,we scored the patients with post-surgery score and plotted the ROC curve,1-year AUC=0.85,95%CI:0.63-1.0,1.5-year AUC=0.74,95%CI:0.55-0.93,2-year AUC=0.79,95%CI:0.640.93.The results showed that the postoperative prediction model performed well in practical applications.(10)We found that patients with unresectable colon cancer exhibited higher levels of plasma exosomes compared with healthy individuals(305.62±19.47 μg/mL vs.135.78±12.76 μg/mL,P<0.001).The level of plasma exosomes decreased significantly after PTR(7 days after PTR:228.98±30.18 μg/mL,P<0.001,15 days after PTR:204.72±39.12 μg/mL,P<0.001).(11)We performed the limma differential analysis on the plasma exosomal mRNAs of colon cancer and healthy controls,and 67 up-regulated differentially expressed mRNAs were identified(fold change of 2,FDR<0.05).CD274(PD-L1)was up-regulated in cancer compared to that in healthy controls.We further analyzed whether patients’ plasma exosomes containing PD-L1 could affect CD4+T cell activity through PD-1/PD-L1 interaction.When we added plasma exosomes to activated CD4+T cells,INF-γproduction of CD4+T cell was reduced in a dose-dependent manner.Furthermore,PD-1 expression was minimal in naive CD4+T cells,increased in activated CD4+T cells,and most pronounced in exosome-co-cultured activated CD4+T cells.This suggests that plasma exosomes may regulate IFN-y secretion in CD4+T cells via PD-1/PD-L1 interaction.Conclusions(1)For patients with asymptomatic unresectable advanced colon cancer,PTR as the initial treatment is beneficial to patients’ survival and has important therapeutic value.(2)The prediction model and Ppre-surgery≤13.2 could help select suitable patients for PTR.Patients with Ppre-surgery≤13.2 were PTR-benefit population.It provided a basis for treatment strategies.(3)Colon cancer patients’ plasma exosomes may inhibit the IFN-γ secretion of CD4+T cells through PD-1/PD-L1,leading to immunosuppression.PTR reduces the level of plasma exosomes in advanced colon cancer,which may be beneficial to reduce tumorinduced immunosuppression.This may be a potential mechanism for PTR to improve survival. |
PDF Full Text Request