Background: Contemporary therapies produce high cure rates in HL. However, complications of treatment produce substantial morbidity and mortality. Identification of individuals at high risk for toxicity is a goal for precision HL therapy. Methods: We conducted a nested case-control study of 953 intermediate risk HL patients enrolled on the Childrens Oncology Group AHOD0031 protocol, assessing associations between severe lung toxicity and clinical, demographic, & genetic predictors. Chemotherapy included doxorubicin, bleomycin, vincristine, etoposide, cyclophosphamide, and prednisone. The frequency of 1936 variant alleles (1931 SNPs & 5 CNVs) in 225 drug metabolism enzyme and transporter genes were compared between patients with no/mild (n=929) & severe (n=24) lung toxicities. Significant polymorphisms (p<0.005) were combined into per-individual risk scores by summing risk alleles, and included in a predictive regression model with significant covariates. Results: We found a significant association between severe lung toxicity and polymorphisms in 3 genes; NAT2, SLCO1B1&ABCC1. Risk alleles were summed across the 3 polymorphisms, generating an allelic risk score ranging from 0-5. A regression model for prediction of lung toxicity included allelic risk score and erythrocyte sedimentation rate (ESR). The model identified a high-risk group with severe lung toxicity prevalence of >9%, and a low-risk group with prevalence of 1.5%. Within the high-risk subset, a second-stage test of polymorphisms in GSTA4 and SLCO1B1 identified a subset with a prevalence of >28%. Within the low-risk subset, polymorphisms in CYP11B1, SLC7A8, and SLCO1B1 identified individuals with a prevalence of >8%. To validate the results, we retested the model in 1000 bootstrapped data sets generated from the original data. Conclusions: A two-step algorithm combining genetic risk alleles in 8 polymorphisms, integrated with ESR, identifies HL patients at low (<1.5%) and high (>28%) risk for severe pulmonary toxicities. These results suggest potential for genomic-guided therapy regimen assignment to avoid drugs (eg bleomycin) associated with lung toxicity in high risk patients. This work serves as a model for approaches to other HL toxicities.