Background Darolutamide and enzalutamide are second-generation androgen receptor inhibitors with activity in men with castrate-resistant prostate cancer (CRPC) and different toxicity profiles. Objective ODENZA is a prospective, randomized, multicenter, cross-over, phase 2 trial designed to assess preference between darolutamide and enzalutamide in men with asymptomatic or mildly symptomatic metastatic CRPC (mCRPC). Design, setting, and participants Patients were randomized 1:1 to receive either darolutamide 1200 mg/d for 12 wk followed by enzalutamide 160 mg/d for 12 wk or enzalutamide followed by darolutamide. In both arms, the second treatment was given in absence of cancer progression. Outcome measurements and statistical analysis The primary endpoint was patient preference between the two drugs, as assessed by a preference questionnaire (p value calculated with the Prescott test). After week 24, patients entered an extension period during which they received their preferred treatment until progression or toxicity. The main secondary objectives included reasons for patient preference, response at week 12, tolerance of each drug, and measurement compared with baseline of cognitive outcomes assessed using tablet questionnaires. Results and limitations Overall, 249 patients, with a median age of 72 yr, were randomized. Among the 200 patients who fulfilled the preplanned criteria for the evaluation of the primary endpoint of preference, 97 (49% [41; 56]), 80 (40% [33; 47]), and 23 (12% [7; 16]) chose darolutamide, chose enzalutamide, and had no preference, respectively (p = 0.92). Reduced fatigue, easier administration, and better quality of life were the main criteria that influenced patient choice. A moderate benefit in episodic memory from darolutamide was observed for the acquisition of new information (least square [LS] means difference = 2.2, effect size = 0.5) and for the recall of that information after a brief delay (LS means difference = 0.7, effect size = 0.3). Using the Brief Fatigue Inventory questionnaire, patients reported greater fatigue with enzalutamide (3.3 [3.0; 3.6]) than with darolutamide (2.7 [2.4; 3.0]). There was no difference in terms of depression, seizures, and falls. Conclusions The study did not show a difference in preference between the two treatments. In men with mCRPC, darolutamide was associated with a clinically meaningful benefit in episodic memory and less fatigue compared with enzalutamide. Patient summary Preference between darolutamide and enzalutamide was well balanced in men with castrate-resistant prostate cancer. Darolutamide was associated with a significant benefit in verbal learning and less fatigue compared with enzalutamide.