Background: HPD is an unexpected and rapid acceleration of tumor growth in patients following treatment with ICI. The definition of HPD is not fully established and risk factors are uncertain. Our aim in this study was to find the incidence of HPD as well as its significant clinicopathological and genetic predictive factors. Methods: A total of 203 patients with NSCLC treated with ICI at Northwestern Memorial Hospital during 2015-2020 were included. Clinicopathological data were retrospectively gathered and analyzed. Tumor growth kinetics (TGK) of pre-immunotherapy and post-immunotherapy were collected and TGK ratio (TGKr) was calculated. HPD was defined as TGKr≥2. Results: HPD was observed in 16% (33/203) of patients. There was a statistically significant difference in overall survival(OS) between the two groups (non-HPD vs. HPD) [OR = 2.39, P < 0.01]. The progression-free survival(PFS) also showed a statistically significant difference in patients (non-HPD vs HPD) [OR = 3.4, P < 0.01]. The median OS and PFS of patients with HPD were 13.57 months and 2.2 months respectively. The median OS and PFS of patients without HPD were 30.67 months and 7.3 months respectively. HPD was significantly associated with performance status (3-4 vs. 1-2) [OR = 4.52, P = 0.01], presence of bone metastasis [OR = 2.51, P = 0.01], neutrophils/lymphocyte ratio(NLR) ≥5 [OR = 2.68, P = 0.01], thrombocytosis (platelet count > 450k) [OR = 5.31, P < 0.01], and treatment with PD-L1 inhibitor vs. PD-1 inhibitor [OR = 2.14, P = 0.06]. HPD was inversely associated with positive PD-L1 expression [OR = 0.34, P = 0.01], ICI combined with chemotherapy vs. ICI monotherapy [OR = 0.24, P = 0.04], and positive TTF-1 expression [OR = 0.46, P = 0.06]. HPD was not associated with histological subtypes of NSCLC, age(< 70, ≥70), sex, smoking history, number of metastatic lesions (< 3, ≥3), brain metastasis, liver metastasis, brain metastasis before treatment initiation, EGFR mutation, TP53 mutation, BRAF mutation, TNM staging, tumor mutational burden (TMB) (< 10, ≥10), microsatellite stability index (MSI), human leukocyte antigen-1 (HLA-1) heterozygosity, hemoglobin(≤12, > 12), albumin(≤3.5, > 3.5), and LDH(≤240, > 240). Conclusions: The incidence of HPD, 16%, was consistent with current literature. NLR ≥5, platelet count > 450k, poor performance status, presence of bone metastasis, treatment with PD-L1 inhibitor, negative PD-L1 expression, negative TTF-1 expression, and ICI monotherapy were predictors of HPD. To our knowledge, this is the first study to report an inverse correlation of TTF-1 expression, and non correlation of HLA-1 heterozygosity with HPD. The strength of our study is that we analyzed recently emerging next-generation sequencing (NGS) data to find novel predictors of HPD. Further studies on validating risk factors of HPD and exploring their associated mechanisms are warranted.