PART ?:staging Optimization For Rectal Cancer After Neoadjuvant Therapy PART ?:prediction For Lymph Node Metastasis In Rectal Cancer Based On Transcriptomics

PART ?Staging optimization for rectal cancer after neoadjuvant therapyBackground and Purpose:Neoadjuvant chemoradiotherapy(NCRT)leads to a decreased number of retrieved lymph nodes(LNs)and possibly affects LN staging after total mesorectal excision in rectal cancer patients.This study compared the prognostic value of logarithmic odds of positive lymph nodes(LODDS)with the prognostic values of ypN staging and the lymph node ratio(LNR)in rectal cancer patients treated with NCRT.Materials and Methods:Four hundred forty-five patients with pathologically confirmed rectal cancer treated with NCRT followed by radical surgery from 2004 to 2015 in National Cancer Center/Cancer Hospital,Chinese Academy of Medical Sciences were included(NCC cohort).The data from 4881 patients from the Surveillance,Epidemiology and End Results(SEER cohort)database were used to verify our results.Firstly,we compare ypN staging,LNR(lymph node ratio)with LODDS(logarithmic odds of positive lymph node ratio)to evaluate the prognosis of DFS.We then validated our results through the SEER public database.Results:The median total number of lymph nodes(TLN)retrieved was 14(range 0 to 61),and 41.8%of patients(n=186)had TLN<12.The univariate analysis showed no significant association between ypN1 and ypN2 staging for DFS(P=0.252).However,LNR and LODDS were significantly associated with DFS(P<0.05 across all groups).Additionally,ypNO was divided into different subgroups by LODDS(P<0.05),and the LODDS system was significantly associated with DFS when TLN<12(P<0.05 across all groups).These results were verified using the SEER database.Multivariate analysis indicated that LODDS?TRG and TD were independent prognostic factors for DFS.Conclusion:LODDS was a better prognostic factor for DFS than both the current AJCC ypN staging system and the LNR-based approach used in locally advanced rectal cancer patients treated with NCRT,especially for patients with TLN<12 or with no LN metastasis.PART ?Prediction for lymph node metastasis in rectal cancer based on transcriptomicsBackground and Purpose:Accurate assessment of regional lymph node status is a huge challenge for preoperative staging in rectal cancer.The purpose of this study was to predict lymph node metastasis in rectal cancer by transcriptome.Materials and Methods:From October 2015 to December 2017,twenty-seven patients with primary rectal cancer who underwent surgical resection in National Cancer Center/Cancer Hospital,Chinese Academy of Medical Sciences were enrolled.There were 27 patients in the development cohort,mong the 27 patients,ten(37.0%)had pT3N0 stage disease and 17(63.0%)had pT3N+stage disease.Two libraries of microRNAs,coding RNA(RNA)and long-chain non-coding RNA(IncRNA)were constructed and sequenced with high throughput.The differentially expressed genes were screened by DESeq2 package and Wilcoxon test by R software.The risk prediction model of lymph node metastasis was constructed by difFerentially expressed genes.The data of transcription of 125 rectal cancer patients in TCGA data were used to construct and validate the risk prediction model.Results:In the development group,thirty-nine candidate microRNAs were screened,and finally 6 microRNAs were included in the risk model for prediction lymph node metastasis.The AUC of 6 microRNAs predicting lymph node metastasis in the training and validation group were 0.777 and 0.768?respectively.A total of 76 differential expression mRNAs were included in the development cohort.Six of them were used to construct the model after random forest and logistic stepwise regression.The AUC of predicting lymph node metastasis was 0.849 and 0.789 in the training and validation groups,respectively.A total of 66 differentially expressed IncRNAs were screened out in the development cohort.Seven IncRNAs were incorporated into the model after reduction by random forests.The AUC for predicting lymph node metastasis was 0.815 and 0.830 in the training and validation groups.After the integration of three RNAs,11 RNAs panel were included to construct the risk model for lymph node metastasis.The under area of ROC curve for predicting lymph node metastasis were 0.912 and 0.875 in the training and validation groups,respectively.The sensitivity,specificity,positive predictive value,negative predictive value and predictive accuracy were 0.952,0.778,0.769,0.955 and 0.854 for predicting lymph node metastasis in validation group.Conclusion:It is a feasible method to predict lymph node metastasis of rectal cancer based on transcriptome method.The 11 gene panel is a potential and accurate marker for predicting lymph node metastasis.