This phase I study is the first to evaluate the safety, tolerability, and preliminary efficacy of liposomal irinotecan combined with TAS-102 in patients with advanced solid tumors. With prophylactic use of G-CSF, both Level 4A and Level 4B were tolerable, with DLT being observed in one of six and none of three patients, respectively. Level 4A (70 mg/m2 of liposomal irinotecan plus 30 mg/m2 of TAS-102, twice daily, Days 1–5, in a 14-day treatment cycle) was selected as RP2D based on the decision of DSMC. TRAEs were observed in 95.3% of participants, with the majority (86.7%) being grade 1–2 per CTCAE criteria. The most reported TRAEs were gastrointestinal events, including diarrhea, nausea, vomiting, and abdominal discomfort, aligning with the known safety profiles of both agents [5, 11, 16, 17].

In the current study, although the RP2D of the combination was selected to be 70 mg/m2 liposomal irinotecan combined with 30 mg/m2 TAS-102 (Level 4A), the absence of DLT in the three patients at Level 4B supports that 60 mg/m2 liposomal irinotecan plus 35 mg/m2 TAS-102 could be served as an alternative dosing regimen, consist with that reported in the OniLon trial [18].

In this study, prophylactic short-acting G-CSF was incorporated based on the anticipated hematologic toxicity of both TAS-102 and liposomal irinotecan, particularly in Asian populations known to exhibit higher susceptibility to myelosuppression. For instance, in the pivotal trial of TAS-102 for metastatic gastric cancer, grade 3–4 neutropenia was reported in 52.2% of Japanese patients [19]. Similarly, in the NAPOLI-1 trial, liposomal irinotecan at 70 mg/m2 combined with 5-FU/LV was associated with a markedly higher incidence of Grade 3–4 neutropenia in the Asian subgroup (54.5%) compared to the overall study population (27.4%) [20]. In a Japanese Phase I study evaluating TAS-102 plus conventional irinotecan, the recommended dose was 25 mg/m2 twice daily for TAS-102 and 150 mg/m2 for irinotecan, with grade 3–4 neutropenia and febrile neutropenia occurring in 100% and 30% of patients, respectively [15]. Given the overlapping myelotoxicity of TAS-102 and liposomal irinotecan, prophylactic G-CSF was used to ensure safe dose escalation and maintain dose intensity. This likely contributed to the lower incidence of grade 3–4 neutropenia compared to other Asian studies without routine G-CSF. While this adds clinical and economic burden, its use was justified in a phase I setting prioritizing safety. Further studies are warranted to assess tolerability without G-CSF in real-world settings.

Previous studies have established key benchmarks for the use of liposomal irinotecan in various treatment settings. The PEP0201 study identified the MTD of liposomal irinotecan monotherapy as 100 mg/m2 (equivalent to 120 mg/m2 as the salt-containing form) every three weeks, demonstrating a manageable toxicity profile and potential for combination therapy [17]. Additionally, the PEP0203 phase I study, which determined the MTD of liposomal irinotecan as 70 mg/m2 at day 1 in combination with 24-h infusion of 5-FU/LV at Days 1 and 8 on a three-week cycle in solid tumors. Of the 6 patients at MTD dose level, only 10.6% experienced grade ≥ 3 AEs [21]. The lower incidence of severe AEs observed in PEP0203 compared to the present study may be attributed to the longer treatment intervals. In contrast, the NIFTY study, which evaluated liposomal irinotecan (70 mg/m2) plus 5-FU/LV every two weeks in metastatic cholangiocarcinoma, reported Grade ≥ 3 adverse events in 42% of patients, with neutropenia being the most common [22]. In our study, the overall AE rates were comparable. However, grade 3 diarrhea occurred more frequently (14%) with TAS-102 in our study compared to 5% in NIFTY and 3% of Asian subpopulation in NAPOLI-1 suggesting that TAS-102 may contribute to an increased gastrointestinal toxicity burden compared to 5-FU/LV [20, 22]. This highlights the importance of proactive gastrointestinal management in future studies evaluating this combination.

Pharmacokinetic analysis revealed that when TAS-102 was administered with 70 mg/m2 liposomal irinotecan, the Cmax of FTD on Day 1 was 1900 ± 828 ng/mL, which is approximately half of the reported Cmax (3677 ± 1459 ng/mL) observed in prior TAS-102 monotherapy studies at the same dose [23]. Similar reductions were observed for TPI, with a Cmax of 54.5 ± 26.0 ng/mL in the combination therapy, compared to 136 ± 77 ng/mL with TAS-102 monotherapy. These findings suggest a potential pharmacokinetic interaction between liposomal irinotecan and TAS-102, possibly affecting the bioavailability of FTD and TPI. Further supporting this, a Phase I study evaluating TAS-102 (25 mg/m2 BID) in combination with irinotecan and bevacizumab reported a Day 1 FTD Cmax of 3,362 ± 1,702 ng/mL, significantly higher than the levels observed in our study [24]. The variations in pharmacokinetic parameters across different studies suggest that individual patient factors, drug combinations, and dosing schedules may significantly impact FTD and TPI exposure.

The PEP0203 study reported CPT-11 Cmax of 29.2 ± 5.2 µg/mL and SN-38 Cmax of 7.98 ± 4.39 ng/mL at 70 mg/m2 liposomal irinotecan with 5-FU/LV [21]. In our study, at the same dose level, CPT-11 Cmax (31 ± 7.16 µg/mL) was comparable, while SN-38 levels (4.29 ± 1.76 ng/mL) were somewhat lower but still within a similar range when considering standard deviations. These variations could be attributed to differences in patient populations, sampling time points, or subtle protocol discrepancies between studies. Despite these differences, both data sets confirm that a 70 mg/m2 dose of liposomal irinotecan provides SN-38 exposure in a comparable range, supporting the feasibility of this dosing strategy in combination regimens. Furthermore, our PK analysis demonstrated that SN-38 exhibited a prolonged half-life (~ 48.2 h), consistent with the expected sustained circulation of liposomal irinotecan, compared to conventional irinotecan formulations (~ 12–22 h) [25]. However, considerable inter-patient variability in SN-38 clearance was observed, which may contribute to differences in treatment-related toxicities across patients.

Efficacy analysis of our study demonstrated a partial response (PR) rate of 18.4% and a DCR of 84.2%, with a median DoR of 18 weeks and median PFS of 30 weeks. These results compare favorably with previous studies evaluating irinotecan-based regimens. In the PEP0203 study, which examined liposomal irinotecan plus 5-FU/LV every three weeks, 13.3% of patients achieved PR, and the DCR was 73.3% [21]. Similarly, the TABASCO phase II trial, investigating TAS-102 with irinotecan and bevacizumab, reported an ORR of 18.4% and a median PFS of 7.9 months, albeit with a higher incidence of grade ≥ 3 toxicities (67%), primarily neutropenia and gastrointestinal events [26]. Another phase I study in Japanese patients identified a recommended dose of 50 mg/m2/day TAS-102 plus 150 mg/m2 irinotecan, reporting a partial response rate of 22% and neutropenia as the most frequent grade 3–4 toxicity [15]. The OniLon phase II trial, which evaluated liposomal irinotecan plus TAS-102 in metastatic colorectal cancer, reported an ORR of 15%, a DCR of 75%, and a median PFS of 9.7 months, further reinforcing the clinical potential of TAS-102 combined with irinotecan-based regimens [27].

The favorable antitumor activity observed in our study may be attributed to shorter treatment intervals and the addition of TAS-102, which has shown efficacy even in 5-FU-pretreated patients [11, 28]. Notably, two of the seven PR cases were in gastric cancer patients, suggesting potential clinical relevance in this setting. In current clinical practice, taxane-based therapy is the standard second-line treatment after platinum-based chemotherapy failure, with irinotecan as an alternative for patients with taxane intolerance due to neuropathy [29]. The TAGS trial has already validated TAS-102 as a third-line option for gastric cancer, demonstrating improvements in OS, PFS, and quality of life [28]. These findings support further evaluation of liposomal irinotecan plus TAS-102 as a second-line or later-line therapy for gastric cancer, forming the basis for an ongoing phase II study (NCT05927857) investigating this combination with ramucirumab.