Background: Immunotherapy (IO)-related fatigue is a common yet underrecognized adverse event (AE). The objective of this study was to determine the frequency of IO-related fatigue and non-IO-related fatigue, contributing factors, and the association between fatigue and treatment outcome in patients with advanced tumors, who received different types of IO-based treatments. Methods: We performed a retrospective chart review of patients treated on IO-based early phase clinical trials between September 2016 and May 2021. We collected data on baseline patient characteristics, AEs, fatigue at baseline and on-treatment, response to treatment, progression free survival (PFS), and overall survival (OS). AEs attribution to treatment were determined by the investigator. We determined IO-related based on the timing of the highest-grade fatigue with respect to treatment start date. Cox proportional hazards were used to assess association between fatigue and survival endpoints. Results: A total of 511 patients with 45 different cancer types were included in this analysis. The most common tumor types were colon (n=50), pancreatic, ovarian, sarcoma (n=35 each), cervical (n=29), and melanoma (n=26). The median age was 62 years, and 98% of patients were ECOG 1. Of the 136 patients (26.6 %) who reported fatigue, 21 (4.1%) had baseline fatigue and 115 (22.5%) developed fatigue after treatment initiation. A total of 69 (13.5%) patients had IO-related and 67 (13.1%) had non-IO related fatigue. The majority of the patients (87.8%) were treated on immune checkpoint-inhibitor-based regimens. Among 69 patients with IO-related fatigue, 58% of the patients reported fatigue only after treatment initiation and 29% had worsening grades of fatigue. In univariate analysis, patients with IO-related fatigue had a higher frequency of baseline depression than those with non-IO-related (14.5% vs. 3%, p=0.0391). Among the 81 patients with fatigue who were evaluated for response by irRECIST, objective response rate was 11.1% (5/45) in patients with IO-related versus 8.3% (3/36) in those with non-IO related fatigue. The median PFS was 4.08 mo in patients with IO related versus 3.68 in non-IO-related fatigue (p=0.27). The median OS was 15 mo in patients with IO related vs 10.6 mo (p=0.8) in non-IO related fatigue. The median OS for all the patients on the study was 14.7 mo in patients with IO-related vs 9.4 mo in non-IO related fatigue (p=0.32). Conclusions: In half of the patients reporting fatigue on IO-based clinical trials, fatigue was related to treatment. Patients with IO-related fatigue had higher response rates, improved PFS and OS, although not significant. Ongoing analysis with patient-reported outcome may provide more insight into prevalence of fatigue in this patient population. Further understanding of underlying biology may help to develop treatment strategies.