*3.3. Toxicity*

All 39 patients were included in the safety assessment. As shown in Table 2, myelosuppression was the principal toxic effect observed. However, the degree of myelosuppression was generally mild. Grade 3 or 4 leukocytopenia, neutropenia, thrombocytopenia, and anemia occurred in 17 (44%), 26 (67%), 0 (0%), and 0 (0%) patients, respectively. Nonhematologic toxicity was also generally mild and less frequent. Grade 3 or 4 toxicity primarily included loss of appetite (7.7%), fever (5.1%), and interstitial pneumonia (5.1%).


**Table 2.** Toxicities in the phase II study.

#### *3.4. E*ffi*cacy*

Tumor response was evaluated in 39 patients. The median follow-up period was 8 months (range, 1–39 months). The following results were found for treatment response: complete response (CR), 0; partial response (PR), 3; stable disease (SD), 24; PD, 5; and not evaluable (NE), 7. The overall response rate was 3/39 or 7.7% (95% confidence interval (CI), 1.6 –20.9%, *p* = 0.31), a value that was not significantly higher than the threshold response rate statistically estimated from previous studies. The rate of disease control, CR + PR + SD, was 27/39 or 69%. Median PFS and OS were 18.0 weeks (95% CI, 11.3–22.9 weeks) and 53.0 weeks (95% CI, 40.9–134.6 weeks), respectively. The survival curves are shown in Figures 1 and 2.

**Figure 1.** Progression-free survival (PFS) analyzed by the Kaplan-Meyer method is presented as a solid line. Median PFS was 18.0 weeks (95% confidence interval (CI), 11.3–22.9 weeks). The 95% CI is presented as two dashed lines.

**Figure 2.** Overall survival (OS) analyzed by the Kaplan-Meyer method is presented as a solid line. Median OS was 53.0 weeks (95% confidence interval (CI), 40.9–134.6 weeks). The 95% CI is presented as two dashed lines.

#### **4. Discussion**

In the present study, response rate did not meet the 18% criterion needed to establish a significant improvement relative to the previously reported rate (i.e., the primary endpoint for the efficacy in the phase II study). However, combination chemotherapy was well tolerated and resulted in encouraging survival data.

Based on large-scale randomized controlled trials that compared 75 mg/m2 docetaxel to optimal supportive care or other anticancer drugs, docetaxel monotherapy is considered to be a standard treatment for advanced non-small cell lung cancer in second-line settings [1,3]. Furthermore, S-1 was confirmed to be effective for previously treated advanced non-small cell lung cancer. Evidence for the efficacy of combination chemotherapy with docetaxel and S-1 has mainly been derived from studies on advanced gastric cancer. S-1 is one of the preferred agents for the treatment of gastric cancer. In fact, docetaxel displayed synergism with S-1 in vitro, improving the response rates in several phase II trials for advanced gastric cancer [7,9–11]. Wada et al. reported the decrease in expression of thymidylate synthase (TS) and dihydropyrimidine dehydrogenase (DPD) and an increase in expression of orotate phosphoribosyl transferase (OPRT) after co-treatment with docetaxel plus FU compared to 5-FU alone in gastric cancer cell lines [9]. DPD catalyzes the metabolic inactivation of 5-FU, while OPRT directly converts 5-FU to 5-fluorouridine-5 -monophosphate (FUMP), an active metabolite that displays anticancer effects. These changes in enzymes that metabolize 5-FU might clarify the enhanced effect of S-1 combined with docetaxel. As reported by Hasegawa et al., a similar synergistic effect is observed in castration-resistant prostate cancer [12]. Interestingly, in the xenograft model, S-1 with low-dose docetaxel could enhance the antitumor effect of S-1.

Several clinical trials have been conducted using docetaxel and S-1 for chemotherapy-naive or previously treated non-small cell lung cancer [9,13–17]. In fact, a triweekly schedule was employed with 40 mg/m<sup>2</sup> as the starting docetaxel dose. Previously, Oki et al. reported the usefulness of biweekly administration of docetaxel [16]. Table 3 contains a summary of results from these clinical studies. Other studies used a fixed docetaxel dose of 40 mg/m2 + S-1 80 mg/m2. The relatively low efficacy in these studies was perhaps due to the low fixed dose of chemotherapeutic agents. Therefore, we employed docetaxel dose escalation in the phase I part of the study, and confirmed that 50 mg/m<sup>2</sup> of docetaxel was effective. Moreover, S-1 dosage was determined depending on body surface area (BSA) in each patient in this study, because the fixed dose of S-1 may be too toxic for patients with smaller BSA or ineffective for the patients with larger BSA. Although the response rate in the current study was relatively low compared to that in other studies, the disease control rate was better than those of other studies (69% in this study, 68.9% in Atagi et al., 61% in Segawa et al., 84% in Yanagihara et al., and 49% in Oki et al.) and favorable PFS and OS were obtained. In addition, the PFS was longer than that reported in a previous phase III study [18] where treatment efficacy was compared between docetaxel and gefitinib for previously treated patients with non-small cell lung cancer in Japan; the PFS in the arm treated with docetaxel was 2.0 months. For toxicity, we found that ≥grade 3 myelosuppression was more frequent in our study and may be due to the higher docetaxel dose employed relative to other studies. Nonetheless, nonhematologic toxicity was demonstrated to be mild, and 50 mg/m2 docetaxel plus S-1 was generally well tolerated.


ORR: overall response rate; OS: overall survival; PFS: progression-free survival.

Takeda et al. reported that in advanced non-small cell lung cancer, the expression of TS and DPD are associated with the response to S-1 and carboplatin [19]. At a low level, their expression was found to be associated with a better response and a longer survival in patients treated with S-1 and carboplatin. Altogether, the expression levels of TS and DPD may be predictive markers for the response to regimens containing S-1. Histologically, squamous cell carcinoma and high-grade carcinoma display higher expression levels for the TS protein and mRNA in non-small cell lung cancer [20,21]. Therefore, a selected patient population with lower expression levels of TS may be most suitable for treatment regimens containing TS-inhibiting agents such as the combination of docetaxel and S-1. We have a plan to conduct a study considering this in the future.

**Author Contributions:** Conceptualization, K.T. and Y.N.; methodology, K.T.; software, S.T.; validation, S.T., K.T., and J.U.; formal analysis, S.T.; investigation, M.F., T.I., R.M., N.E., S.S., K.M., H.W., and T.Y.; data curation, S.T.; writing—original draft preparation, K.T.; writing—review and editing, J.U.; supervision, Y.N.; funding acquisition, Y.N.

**Funding:** Financial support was obtained from the Clinical Research Support Center Kyushu.

**Acknowledgments:** We thank the staff at the Clinical Research Support Center Kyushu for their secretarial assistance and for preparing the manuscript.

**Conflicts of Interest:** K. Takayama received grants from Chugai-Roche Co. and personal fees from AstraZeneca Co. and Chugai-Roche Co. outside of the submitted work. J. Uchino received grants from Eli Lilly Japan K.K. that are outside of the submitted work. The other authors have no conflicts of interest.
