The article provides a comprehensive review of the intention-to-treat (ITT) concept in randomized controlled trials (RCTs). ITT analysis includes all subjects randomized according to their treatment assignment, regardless of compliance or protocol deviations, ensuring that the prognostic balance generated from the original randomization is maintained. This approach aims to avoid overoptimistic estimates of treatment efficacy by accounting for noncompliance and protocol deviations, which are common in clinical practice. The article highlights the advantages of ITT analysis, such as unbiased estimates of treatment effects, preservation of sample size, and greater generalizability. However, it also discusses the limitations, including conservative estimates due to dilution from noncompliance and potential heterogeneity in the final analysis. The article emphasizes the importance of complete outcome data for optimal ITT application and the need to minimize missing responses and follow up withdrawn participants. It also reviews the per-protocol (PP) analysis, which excludes protocol violators, and the modified ITT (mITT) concept, which allows for justified exclusions. The article concludes by discussing the regulatory guidelines and consensus on the use of ITT and PP analyses, particularly in noninferiority trials.The article provides a comprehensive review of the intention-to-treat (ITT) concept in randomized controlled trials (RCTs). ITT analysis includes all subjects randomized according to their treatment assignment, regardless of compliance or protocol deviations, ensuring that the prognostic balance generated from the original randomization is maintained. This approach aims to avoid overoptimistic estimates of treatment efficacy by accounting for noncompliance and protocol deviations, which are common in clinical practice. The article highlights the advantages of ITT analysis, such as unbiased estimates of treatment effects, preservation of sample size, and greater generalizability. However, it also discusses the limitations, including conservative estimates due to dilution from noncompliance and potential heterogeneity in the final analysis. The article emphasizes the importance of complete outcome data for optimal ITT application and the need to minimize missing responses and follow up withdrawn participants. It also reviews the per-protocol (PP) analysis, which excludes protocol violators, and the modified ITT (mITT) concept, which allows for justified exclusions. The article concludes by discussing the regulatory guidelines and consensus on the use of ITT and PP analyses, particularly in noninferiority trials.