PT - JOURNAL ARTICLE AU - Yi, Mengshi AU - Zhao, Rui AU - Wan, Qianyi AU - Wu, Xiaoting AU - Zhuang, Wen AU - Yang, Hanshuo AU - Wu, Chuncheng AU - Xia, Lin AU - Chen, Yi AU - Zhou, Yong TI - Tumor-infiltrating immune cells signature predicts recurrence free survival after complete resection of localized primary gastrointestinal stromal tumors AID - 10.1101/2020.02.28.20025908 DP - 2020 Jan 01 TA - medRxiv PG - 2020.02.28.20025908 4099 - http://medrxiv.org/content/early/2020/02/29/2020.02.28.20025908.short 4100 - http://medrxiv.org/content/early/2020/02/29/2020.02.28.20025908.full AB - Background Growing evidence has proposed prognostic value of immune infiltration in gastrointestinal stromal tumors (GISTs). Therefore, we aimed to develop a novel immune-based prognostic classifier (IPC) to help better stratify and predict prognosis of GISTs.Methods The gene expression profiles of 22 immune features of GISTs were detected from GEO dataset. The IPC was constructed using the LASSO Cox regression model and validated in a cohort including 54 patients with complete resection of localized primary GISTs via immunohistochemistry process. The performance assessment of the IPC was estimated, then compared with conventional risk prognostic criteria.Results The IPC was established based on 4 features: CD8, CD8/CD3, CD68, CD163/CD68 and validated to be an independent predictor of RFS for GISTs (HR 5.2, 95%CI 1.99-13.65). Significant differences were found between low- and high-IPC group in 5-year RFS (92.6% vs 48.1%, p < 0.001). Using the IPC, the high-risk group of the Modified NIH classification was split into two groups in 5-year RFS (low-IPC vs high-IPC, 85.7% vs 30.0%, p < 0.001). The IPC showed a higher net benefit than both “treat all” or “treat none” methods for the threshold probability within a range of 0-0.62 and exhibited a performance (AUC 0.842) superior to modified NIH classification (AUC 0.763).Conclusion The IPC was effective to predict RFS after complete resection of localized primary GISTs, adding prognostic value to the routine clinical prognostic criteria. Prospective studies are needed to further validate the analytical accuracy and practicability of the IPC in estimating prognosis of GISTs.Competing Interest StatementThe authors have declared no competing interest.Funding StatementThis work was supported by Chinese Medical Board Grant on Evidence-Based Medicine, New York, USA (No. 98-680), National Natural Science Foundation of China (No. 30901427)Author DeclarationsAll relevant ethical guidelines have been followed; any necessary IRB and/or ethics committee approvals have been obtained and details of the IRB/oversight body are included in the manuscript.YesAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesThe data referred to in the manuscript was available from the GEO dataset.GISTgastrointestinal stromal tumorIPCimmune-based prognostic classifierRFSrecurrence free survivalTAMstumor-associated macrophagesTemseffector memory T cellsTcmscentral memory T cellsTgd cellsT Gamma/Delta cellsTILstumor infiltrating lymphocytesTregsregulatory cellsNIHNational Institutes of HealthUICCthe Union for International Cancer ControlAUCarea under the curvePD-L1programmed death-ligand1NKnatural killerLASSOleast absolute shrinkage and selection operator methodHRhazard ratioCIconfidence intervalROCreceiver operating characteristicDCAdecision curve analysis