The trial evaluated S-1, an oral fluorinated pyrimidine prodrug of 5-fluorouracil (5-FU), given together with intravenous cisplatin. S-1 is not a simple chemotherapy pill; it is a three-component formulation with an unusual pharmacological logic. Its first ingredient, tegafur, is an oral prodrug that is converted to the active cytotoxic metabolite 5-FU preferentially in tumor tissues. The second component, gimeracil, inhibits dihydropyrimidine dehydrogenase (DPD), the major catabolic enzyme of 5-FU that is found in high concentrations in the liver, thereby maintaining elevated systemic 5-FU concentrations. The third component, potassium oxonate, blocks orotate phosphoribosyltransferase in the gastrointestinal tract, reducing gut toxicity without compromising anti-tumor activity. Because gimeracil’s target enzyme is concentrated in the liver, and because tegafur itself is metabolized mainly by the cytochrome P450 isoenzyme CYP2A6, S-1’s behavior in patients with severe hepatic dysfunction has been a genuine pharmacological unknown.

Between September 2016 and November 2023, the investigators enrolled 61 patients with advanced solid tumors that were refractory to standard therapy or had no available treatment options. All patients had measurable disease by RECIST version 1.1 criteria, an Eastern Cooperative Oncology Group performance status of 0 or 1, adequate bone marrow reserve, and a calculated creatinine clearance above 60 milliliters per minute. Patients were stratified into four cohorts according to the National Cancer Institute Organ Dysfunction Working Group (NCI-ODWG) criteria: normal hepatic function (Cohort 1, 11 patients), mild dysfunction (Cohort 2, 10 patients), moderate dysfunction (Cohort 3, 13 patients), and severe dysfunction (Cohort 4, 27 patients). To avoid misclassifying potentially reversible causes of liver injury, patients with active hepatitis B or C infection or cholangitis were excluded, and those with malignant biliary obstruction were eligible only after biliary drainage, with cohort assignment based on post-drainage liver function tests.

The treatment regimen was deliberately conventional: intravenous cisplatin at 60 milligrams per square meter on day 1, combined with oral S-1 at 30 milligrams per square meter twice daily on days 1 through 14 of each 21-day cycle. After six cycles, patients without progressive disease could discontinue cisplatin and continue S-1 maintenance. The primary endpoints were safety and 5-FU pharmacokinetics, while secondary endpoints included the objective response rate (ORR), the 3-month clinical benefit rate (3M-CBR), and progression-free survival (PFS). A 3-plus-3 safety design was used, with larger planned enrollment in the moderate and severe dysfunction cohorts to compensate for anticipated higher attrition in these fragile patients.

The pharmacokinetic findings are the technical heart of the study. Blood samples were collected before treatment and after 7 and 14 days of S-1 dosing during cycle 1, at 0, 30 minutes, 60 minutes, 2 hours, 4 hours, and 6 hours after dosing. Plasma 5-FU concentrations were measured by liquid chromatography with tandem mass spectrometry, and individual parameter estimates were derived using compartmental and non-compartmental methods in WINNONLIN software. On cycle 1 day 15, patients with severe hepatic dysfunction had a significantly lower peak plasma concentration of 5-FU (Cmax of 135.8 versus 199.9 nanograms per milliliter, p = 0.018) and a significantly longer time to reach peak concentration (Tmax of 3.1 versus 1.9 hours, p = 0.045) than patients with normal hepatic function. These shifts suggest altered conversion of tegafur to 5-FU in the failing liver, although the authors note that this mechanism was not directly evaluated and that other factors affecting oral absorption, hepatic metabolism, or clearance may also contribute.

Critically, however, the area under the concentration-time curve (AUC), which reflects overall drug exposure, remained comparable across all four cohorts. On day 15, the AUC in the severe dysfunction cohort was 545.4 versus 738.4 hour-nanograms per milliliter in the normal-function cohort, a non-significant trend (p = 0.086). This distinction matters clinically: overall exposure to the active metabolite was preserved even as its absorption profile changed, and no significant association was found between the altered pharmacokinetic parameters and either severe toxicity or clinical benefit in multivariate analyses. Exploratory univariate analyses suggested that shorter Tmax and higher Cmax on day 8 were associated with clinical benefit, and that lower baseline bilirubin, higher ALT, and higher day-8 Cmax and AUC were associated with response in the severe dysfunction cohort, but none of these associations survived multivariate adjustment.

The safety results were equally reassuring. Treatment-related adverse events occurred in 83.6 percent of all patients, with 47.5 percent experiencing grade 3 or 4 events, and the profile in the severe dysfunction cohort was similar to that in the milder cohorts. Common toxicities included hematologic effects such as neutropenia, anemia, and thrombocytopenia, along with fatigue, nausea, anorexia, and electrolyte abnormalities. Dose-limiting toxicities occurred in only two patients across the entire study—one with grade 3 fatigue in the mild cohort and one with grade 3 vomiting attributed to cisplatin in the moderate cohort. Most importantly, no grade 5 treatment-related deaths occurred anywhere in the study. The authors also performed a complementary assessment using the ALBI grade, based on serum albumin and bilirubin, and found broad concordance with the NCI-ODWG classification: no patients in the normal or mild cohorts had ALBI grade 3, whereas 48 percent of the severe dysfunction cohort did.

The efficacy findings, while preliminary, were striking for a population usually written off as untreatable. In the severe hepatic dysfunction cohort, 22.2 percent of patients achieved a partial response, 40.7 percent derived clinical benefit at three months, and 22.2 percent maintained progression-free survival of six months or longer. Among the six patients with partial responses, five had breast cancer and one had nasopharyngeal carcinoma; notably, one responder had a baseline total bilirubin 4.13 times the upper limit of normal, while the other five had baseline AST or ALT above five times normal. Within the severe dysfunction cohort, breast cancer patients showed an ORR of 33.3 percent and a 3M-CBR of 46.7 percent, compared with 8.3 percent and 33.3 percent in non-breast cancers, a difference that was not statistically significant. Descriptive analyses also showed that 40 percent of patients with baseline AST above five times normal achieved clinical benefit, and 75 percent of the four patients with ALT above five times normal achieved partial responses.

The study also explored whether CYP2A6 genetic polymorphisms could explain variability in treatment outcomes. Genomic DNA from 58 patients was genotyped using the Infinium Global Screening Array, focusing on five SNPs in the CYP2A6 gene selected for known or suspected functional relevance. The most prevalent variant, rs28399433, was observed in 36.2 percent of patients, followed by rs8192720 in 19.0 percent and rs5031016 in 12.1 percent. No statistically significant association was found between any individual SNP and objective response, clinical benefit, or severe toxicity. The authors caution that these analyses were exploratory and underpowered, and that the absence of significant findings should not be interpreted as definitive evidence of no association, particularly since cisplatin’s cytotoxic mechanism does not depend on CYP2A6-mediated metabolism.

The investigators are candid about the study’s limitations. It was a single-arm, single-center trial in Singapore with a predominantly Asian population, and breast cancer accounted for 57.4 percent of all patients, making tumor-specific efficacy claims impossible. More than one-third of patients with severe hepatic dysfunction experienced rapid disease progression and could not complete even one treatment cycle, reflecting the harsh clinical reality of this population. The trial was closed after seven years because of slow accrual, and the number of patients with complete pharmacokinetic data fell short of the original target. Nevertheless, the prospective safety and pharmacokinetic data represent a meaningful advance in a population for whom evidence to guide treatment decisions remains scarce. The authors conclude that cisplatin combined with S-1 may be a feasible treatment approach for selected patients with severe hepatic dysfunction who have limited therapeutic options, and that larger prospective studies are needed to confirm these findings and further define the regimen’s safety, pharmacokinetics, and clinical activity.

Subject of Research: Oncology

Article Title: Prospective evaluation of combination oral S-1 and intravenous cisplatin in cancer patients with severe hepatic dysfunction – a feasibility, safety, and pharmacokinetic study

Article References: Gwee, Y. X., Wong, A. L. A., Goh, B. C., Soo, R. A., Yong, W. P., Tan, D. S. P., Chee, C. E., Lim, S. E., Ang, Y., Huang, Y., Chong, W. Q., Lim, J., Chan, G., Ngoi, N., Ow, S., Wong, R., Sooi, K., Choo, J., & Lee, S. C. (2026). Prospective evaluation of combination oral S-1 and intravenous cisplatin in cancer patients with severe hepatic dysfunction – a feasibility, safety, and pharmacokinetic study. Frontiers in Oncology, 16, Article 1948986. https://doi.org/10.3389/fonc.2026.1948986

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1948986

Keywords: S-1, cisplatin, hepatic dysfunction, pharmacokinetics, 5-fluorouracil, clinical trial, oncology, Prospective, evaluation, combination, oral, intravenous