Background: Tesetaxel, a novel oral taxane, yielded a 20% major objective response rate in pretreated patients (pts) with advanced gastric cancer when administered at a dose of 27 mg/m2 Q3 wks. Relative to docetaxel, tesetaxel is not a substrate for P-gp, is associated with lower neurotoxicity, and does not cause hypersensitivity or require premedications. We have conducted a dose-ranging study of tesetaxel to determine whether “flat” (rather than weight-based) dosing of this new agent would be feasible for phase III testing. Methods: Pts with advanced gastric cancer who had failed 1 prior fluoropyrimidine/cisplatin regimen were accrued to 1 of 3 cohorts. Tesetaxel was administered to pts in Cohorts 1 and 2 once every 3 weeks at a flat starting dose of 40-45 mg and 50-60 mg, respectively. In Cohort 3, the starting dose was 27 mg/m2. For analysis, flat doses were converted to weight-based (body-surface area [BSA]) equivalents. Standard parameters of response and safety were compared. Results: Tesetaxel was well tolerated. No episodes of ≥ grade 3 neutropenia occurred in either flat-dose cohort. We found substantial BSA variability in both flat-dose cohorts (median BSA = 1.9 and 2.0 mg/m2, respectively), which resulted in substantial underdosing compared with the weight-based regimen. Weight-based doses as low as 19 and 20 mg/m2, respectively, were administered in the 2 flat-dose cohorts. Among 11 evaluable pts in Cohort 1, the lower flat-dose regimen yielded 1 unconfirmed partial response (uPR) and 1 stable disease (SD). Among 13 evaluable pts in Cohort 2, the higher flat-dose regimen yielded 1 uPR and 4 SD; it is too early to evaluate 3 pts. Pts are currently being accrued to the weight-based regimen. Conclusions: On the Q3 week schedule, flat dosing of tesetaxel results in substantial underdosing of pts due to high BSA variability in Western subjects. Safety and response to tesetaxel in pts with advanced gastric cancer appear to be dose-related. A randomized Phase 3 study will proceed using the Q3 week schedule supported in this trial. Our data suggest a starting dose of 27 mg/m2, with escalation to 35 mg/m2 depending on tolerance, appears optimal.