Disease	Hepatocellular carcinoma
Stage of Disease/Treatment	Metastatic/advanced
Prior Therapy	No designated number of regimens
Type of Study	Phase I, 3 + 3
Primary Endpoints	Safety, tolerability, recommended phase II dose
Secondary Endpoints	Efficacy, pharmacokinetics, correlative endpoint
Additional Details of Endpoints or Study Design Dose Escalation Design: For Part 1: Three patients will be enrolled at enzalutamide 160 mg daily. If no DLTs are observed, an additional three patients will be enrolled for confirmation of safety. If zero out of six or one out of six DLTs are observed, then 160 mg will be defined as the dose to move forward as the recommended phase II dose (RP2D). If two out of three or two out of six DLTs are observed, we will de‐escalate to 120 mg and again follow the 3 + 3 design. Upon defining the RP2D, 10 additional patients receive the enzalutamide at the RP2D to further assess safety, obtain additional PK analysis, and explore efficacy in the second‐line setting. In total, 16 patients were enrolled into part 1. Dose Escalation Design: For Part 2: A dose escalation scheme will be used whereby patients will be treated in sequential cohorts of three. If no patients experience a DLT at dose level 1 in an initial group of three patients, cohort 2 will open and re‐enroll three patients. If one of three patients experiences a DLT, the cohort will be expanded to six. If no further DLTs occur, this dose will be considered the RP2D. If two of six patients experience a DLT, the maxium tolerated dose (MTD) has been exceeded. The MTD will be defined as the highest dose for which no more than one of six patients develops a DLT. After the establishment of the RP2D and schedule, an expansion cohort was planned for a total of 39 additional patients who were treatment naive. Using a Simon minimax design with 39 patients, we can show an improvement in 4‐month PFS from 50% to 70% using a type I and type II error rate of 10% each. In the first stage, we would need 23 patients, out of whom we need 12 to be alive and progression free at 4 months, in which case we would accrue an additional 16 patients. If at the end of the study 24 or more are alive and progression free, we would call this promising. This cohort was not explored based on interim analysis of the study showing limited antitumor activity and a drug–drug interaction with the combination. Immunohistochemistry for AR: Deparaffinized tissue sections from HCC tumors were treated with antigen retrieval solution followed by incubation with standard blocking reagents. Primary antibody for AR (DAKO) was then applied with dilution 1:70 (clone AR441, Dako, catalog number M3562) and incubated overnight at 4°C. Appropriate secondary antibodies labeled with polymer −30’ were applied at room temperature (Envision Kit, Dako catalog number K4006) followed by detection using DAB as substrate‐chromogen. Positive and negative controls were performed in parallel (prostate, positive control for AR). The number of cells that are AR positive, the intensity of staining, and the percentage of nuclear staining was assessed. Nuclear staining >5% will be considered positive. AR testing was performed and analyzed without knowledge of the patients' clinical status. PK Determination: Whole blood samples were obtained at Cycle 1: Day −1, before dose and 1 hour, 4 hours, 8 hours, and 24 hours after dose (Cycle 1 Day 1) for sorafenib PKs and Cycle 2 Day 1, before dose and 1 hour, 4 hours, 8 hours, and 24 hours (Cycle 2 Day 2) for both sorafenib and enzalutamide PKs on the first 20 patients in the Part 2 dose escalation. A 7‐day run‐in of sorafenib alone allowed intrapatient comparison of sorafenib steady state PK on and off enzalutamide. Sorafenib and sorafenib N‐oxide in plasma was determined using a validated liquid chromatography tandem mass spectrometry method on an AB Sciex 5500 triple quadrupole. The software program used was Phoenix 64 WinNonlin (Pharsight Corp., St. Louis, MO, version 7.0).
Investigator's Analysis	Lack of efficacy and drug–drug interaction