Ketogenic diets (KDs) are emerging as effective therapies for several chronic diseases, including cancer. However, concerns regarding safety and adherence may prevent clinicians from prescribing KDs. We hypothesized that a KD does not negatively affect blood lipid profile compared to a lower-fat diet in ovarian and endometrial cancer patients, and that KD subjects would demonstrate acceptable adherence. Subjects were randomized to either a KD (70% fat, 25% protein, 5% carbohydrate), or the American Cancer Society diet (ACS; high-fiber and lower-fat). Blood lipids and ketones were measured at baseline and after 12 weeks of the assigned intervention. Adherence measures included urinary ketones in the KD and 4 days’ diet records. Diet records were also examined to identify general patterns of consumption. Differences between the diets on blood lipids and dietary intake were assessed with Analysis of covariance and independent t-tests. Correlation analyses were used to estimate associations between dietary intake and serum analytes. At 12 weeks, there were no significant differences between diet groups in blood lipids, after adjusting for baseline values and weight loss. Adherence among KD subjects ranged from 57% to 80%. These findings suggest that KDs may be a safe and achievable component of treatment for some cancer patients.