Risk Factors for Keratinocyte Carcinoma in Recipients of Allogeneic Hematopoietic Cell Transplants.

Importance: Allogeneic hematopoietic cell transplant (alloHCT) is known to increase the risk for keratinocyte carcinoma. The extent to which host characteristics, including pigmentary phenotype and UV radiation exposure, contribute is unknown. Objective: To identify and validate independent risk factors for keratinocyte carcinoma after alloHCT, including those associated with the transplant and the host. Design, Setting, and Participants: This retrospective cohort study analyzed a consecutive sample of alloHCT recipients from January 1, 2000, to December 31, 2014, at the Mayo Clinic, Rochester, Minnesota (n = 872) and University Hospitals Cleveland Medical Center, Cleveland, Ohio (n = 147). Participants from the Mayo Clinic were randomly allocated (2:1) into discovery (n = 581) and validation (n = 291) cohorts. Time to first keratinocyte carcinoma and information about transplant- and host-associated risk factors were extracted. A multivariate keratinocyte carcinoma risk model was created using a stepwise Cox proportional hazards regression model with P ≤ .05 for entry that incorporated all covariates that were individually statistically significant at α = 0.05 in the discovery cohort. The risk model was first internally validated using the Mayo Clinic validation cohort and then externally validated using the independent cohort of alloHCT recipients at University Hospitals Cleveland Medical Center. Data were analyzed from March 13, 2018, to June 12, 2019. Exposures: Allogeneic hematopoietic cell transplant. Main Outcomes and Measures: The primary outcome was time to development of the first cutaneous keratinocyte carcinoma after alloHCT; secondary outcome, time to development of the first individual basal and/or squamous cell carcinoma after alloHCT. Results: Of the 872 alloHCT recipients identified in the Mayo Clinic cohort (520 men [59.6%]; mean [SD] age, 48.3 [12.6] years), 95 (10.9%) developed keratinocyte carcinoma after alloHCT during 5349 person-years of follow-up. Of the 147 alloHCT recipients in the exernal validation cohort (86 men [58.5%]; mean [SD] age, 47.9 [17.5] years), 18 (12.2%) developed keratinocyte carcinoma after alloHCT in 880 person-years of follow up. Risk factors independently associated with keratinocyte carcinoma after alloHCT included age (hazard ratio [HR] per 10 years, 1.72; 95% CI, 1.21-2.42), chronic lymphocytic leukemia (HR, 2.47; 95% CI, 1.20-5.09), clinically photodamaged skin (HR, 3.47; 95% CI, 1.87-6.41), and history of cutaneous squamous cell carcinoma (HR, 2.60; 95% CI, 1.41-5.91). Harrell concordance statistics were 0.81 (95% CI, 0.72-0.90) and 0.86 (95% CI, 0.74-0.98) for internal and external validation of the keratinocyte carcinoma risk model, respectively. Conclusions and Relevance: This study found validated independent risk factors for keratinocyte carcinoma after alloHCT that are enriched with host- compared with transplant-associated risk factors. These findings highlight the importance of assessing host-associated risk factors for keratinocyte carcinoma in patients eligible for alloHCT. Future studies should examine whether keratinocyte carcinoma risk stratification before alloHCT may inform long-term surveillance strategies.