Exogenous arginine is required for growth in some argininosuccinate synthetase (ASS)-deficient cancers. Arginine deiminase (ADI) inhibits growth in various ASS-deficient cancers by depleting arginine. The efficacy of pegylated ADI (ADI-PEG20) in relapsed/refractory/poor-risk acute myeloid leukemia (AML) was evaluated in 43 patients in a prospective, phase II trial (NCT01910012 (10/07/2013), https://clinicaltrials.gov/ct2/show/NCT01910012?term = ADI-PEG20&rank = 12 ). Despite almost all pre-treatment tumor samples showing ASS deficiency, the best response among 21 evaluable patients was complete response (CR) in 2 (9.5%) and stable disease in 7 (33.3%), yielding a disease control rate (DCR) of 42.9%. The response durations of the two patients with CR were 7.5 and 8.8 months. DCR was correlated with a median of 8 weeks of arginine depletion to ≤10 μM. Using whole transcriptome sequencing, we compared gene expression profiling of pre- and post-treatment bone marrow samples of the two responders and three non-responders. The expression levels of some markers for AML subtypes and c-MYC regulated genes were considered potential predictors of response to ADI-PEG20. These results suggest that ASS deficiency is a prerequisite but not a sufficient condition for response to ADI-PEG20 monotherapy in AML. Predictive biomarkers and mechanistic explorations will be critical for identifying appropriate patients for future AML trials of ADI-PEG20.