Inclusion requires GCC+ (H-score &ge;10) advanced GI tumors (CRC, gastric/esophageal/pancreatic), &ge;18 years, adequate organ function (ANC &ge;1.5&times;10⁹/L, platelets &ge;100&times;10⁹/L, CrCl &ge;60 mL/min), ECOG 0-1, and &ge;12-week life expectancy. Exclusions: recent anticancer therapy (<28 days), CYP3A/P-gp modulators, bleeding disorders/biopsy contraindications, concurrent alcohol abuse, or other malignancies (exceptions: nonmelanoma skin/carcinoma in situ). For Part C (imaging substudy), &ge;2 cm extrahepatic metastasis is required.

The study evaluates dose-limiting toxicities (DLTs) based on NCI CTCAE v5, including Grade 4 neutropenia (ANC <500/mm ³), thrombocytopenia (<25,000/mm ³), febrile neutropenia, Grade ≥3 bleeding thrombocytopenia, and Grade ≥3 nausea/vomiting/diarrhea despite optimal care. Adverse event rates (≥Grade 3, drug-related, serious, and leading to discontinuation) were monitored over 22 months, alongside determination of the recommended Phase 2 dose (RP2D).

TAK-164 as a single intravenous infusion with a duration of up to 2 h, on day 1 of each 21-day cycle or every 3 weeks, until disease progression, unacceptable toxicity, or withdrawal from the study. Doses of 0.004 mg/kg, 0.008 mg/kg, 0.016 mg/kg, 0.032 mg/kg, 0.064 mg/kg, 0.12 mg/kg, 0.16 mg/kg, 0.19 mg/kg, 0.25 mg/kg, and 0.32 mg/kg were planned. To guide dose escalation, a method based on a Bayesian model of modified toxicity probability interval (mTPI) was used.