Background: Human cytomegalovirus (CMV) infection induces a subset of long-lived CD57⁺NKG2C⁺ adaptive NK cells that exhibit enhanced antibody-dependent cellular cytotoxicity and resistance to tumor-suppressive mechanisms. We developed a 7-day culture process using a GSK3 inhibitor and IL-15 to manufacture modulated adaptive NK cells (FATE-NK100) from CMV⁺ haploidentical donors for adoptive transfer. The phase I Apollo trial tests the maximum tolerated dose/maximum feasible dose (MTD/MFD) of FATE-NK100 administered intraperitoneally (IP) to treat platinum-sensitive or -resistant recurrent ovarian, fallopian tube, and primary peritoneal cancer. Methods: FATE-NK100 via IP port was tested using 3 dose cohorts ([DC]; 1 × 10⁷ cells/kg; >1 × 10⁷ cells/kg to ≤3 × 10⁷ cells/kg; or >3 × 10⁷ to ≤10 × 10⁷ cells/kg) after lympho-conditioning with fludarabine 25 mg/m² IV and cyclophosphamide 300 mg/m² IV on days −6 and −5. After FATE-NK100 infusion on day 0, rhIL-2 at 6 million IU was given IP 3 times a week for 6 doses for in vivo NK activation. IP fluid and peripheral blood were collected regularly until response assessment (day 28). Patients with stable disease or better were eligible for retreatment. Pre- and post-treatment tumor biopsies were collected. Results: Nine patients were treated with no dose-limiting toxicities (DLTs) to date. Retreatment based on clinical benefit was performed on 3 patients (33%), 2 following stable disease (DC 2) and 1 with partial remission (48% tumor reduction, DC 3). IP samples were collected for PK and functional analysis. FATE-NK100 product was detected by flow cytometry in 5 of 6 patients with evaluable samples (range 4.8%–91.2% donor NK cells at day +5-7). Retreatment samples were available in 1 patient, where FATE-NK100 persisted to day +21, demonstrating that repeated IP dosing did not accelerate clearance of the donor NK cells. In that same patient, measurement of NK cell CD107a degranulation or IFNg production in response to K562 targets demonstrated sustained enhanced in vivo function of FATE-NK100 compared to endogenous patient NK cells (e.g. at Day +12 CD107a⁺ NK were 39.0% vs. 22.5% cycle 1, and 40.3% vs. 18.2% retreatment cycle 2, and IFNg⁺ NK were 12.3% vs. 5.9% cycle 1, and 2.4% vs. 0.2% retreatment cycle 2). Conclusions: IP delivery of FATE-NK100 is safe, with clinical benefit in 3/9 patients treated. The allogeneic product cells persist and have enhanced function compared to patient NK cells for up to 21 days, even after retreatment. This phase I study in recurrent/refractory ovarian cancer shows promise for IP NK cell delivery. Clinical trial information: NCT00652899