*3.2. E*ffi*cacy*

The ORR from our previous report was 58.3% (95% CI: 42.2–72.9), which included a complete response rate of 2.8% and a partial response rate of 55.6%. The stable disease rate was 38.9%, and the DCR was 97.2%. The median response duration was 54.9 weeks (95% CI: 26.9–69.1), and a waterfall plot revealed that 33 patients (91.6%) experienced tumor shrinkage, which indicated favorable antitumor activity. Sixteen patients (44.4%) continued treatment beyond progression.

The median PFS was 11.9 months (95% CI: 7.9–17.5), with 1-year PFS rate of 50.0% and 2-year PFS rate of 18.3% (Figure 1). The median OS was 22.0 months (95% CI: 16.0–not reached), with 1-year OS rate of 77.8% and 2-year OS rate of 49.5% (Figure 2).

**Figure 1.** Progression-free survival.

**Figure 2.** Overall survival.

#### *3.3. Safety*

Adverse events occurred in 31 cases (86.1%), with Grade 3 or higher adverse events observed in 10 cases (27.8%). Seven patients (19.4%) required dose reductions, 10 patients (27.8%) discontinued treatment because of adverse events, and 1 patient died (2.8%). The adverse event leading to death was a pulmonary infection, although this was judged unlikely to have been caused by the osimertinib treatment. There were no death events caused by drug-induced lung injury. The most frequent adverse event was anemia/hypoalbuminemia (27 patients, 75.0%), which was followed by thrombocytopenia (21 patients, 58.3%), paronychia/anorexia/diarrhea/neutropenia (15 patients, 41.7%), leukopenia/aspartate aminotransferase increase (14 patients, 38.9%), fatigue/acneiform eruption (13 patients, 36.1%), and alanine aminotransferase increase/alkaline phosphatase increase/creatinine increase (11 patients, 30.6%). The Grade 3–4 adverse events included fatigue, anorexia, diarrhea, cardiac ejection fraction decreased, prolonged QT, leukopenia, neutropenia, and aspartate aminotransferase increase. The cases of cardiac ejection fraction were decreased and the cases of prolonged QT were different cases, and delirium and hallucinations were observed in the same patient. Pneumonitis was observed in four patients (11.1%), including two patients (5.6%) with Grade 3–4 pneumonitis (Table 2).




#### **4. Discussion**

Treatment of NSCLC has advanced dramatically after the introduction of molecularly targeted drugs, such as EGFR-TKIs for EGFR-mutated cases. The first- and second-generation of EGFR-TKIs proved to be highly effective in several studies, although the effects tended to only last for approximately 1 year [1–11]. Approximately one-half of the resistant cases involved a gatekeeper mutation in exon 20 (T790M), and osimertinib was developed and approved for the treatment of these cases [12,13]. The results of the FLAURA trials positioned osimertinib as a standard treatment option, and even as an initial treatment option [14,15]. However, many cases still involve treatment in the second line or later, as the T790M mutation was identified via re-biopsy in patients who received first-generation or second-generation EGFR-TKIs as their initial treatment. When the T790M mutation was identified in these cases, patients typically received osimertinib.

Aging populations are becoming increasingly common worldwide, and many lung cancer cases involve older patients [26,27]. There are concerns that older patients have a higher risk of developing adverse events, which may necessitate dose reduction or treatment discontinuation, and subsequently result in decreased efficacy. Thus, this phase II study aimed to evaluate the safety and efficacy of osimertinib in elderly patients with EGFR-mutated lung cancer involving the T790M mutation. The primary endpoint was the ORR, and our previous report found that the ORR was 58.3% (95% CI: 42.2–72.9), which fulfilled the efficacy criterion (the lower limit of the CI exceeded the threshold response rate of 35%) [19]. This report describes the secondary endpoints, which include the DCR (97.2%), median PFS (11.9 months), and median OS (22.0 months). In terms of efficacy, the pooled results from the AURA expansion and AURA2 studies revealed an ORR of 66%, a DCR of 91%, a median PFS of 9.9 months, and a median OS of 26.8 months [23]. In addition, phase 3 AURA3 studies revealed an ORR of 70.6%, a DCR of 93.2%, a median PFS of 10.1 months, and a median OS of 26.8 months [13,28]. Thus, while our ORR was lower than that shown in the previous studies, it agrees with the slightly lower ORR (61.1%) that was retrospectively observed in another sample of elderly Japanese patients [29]. Furthermore, our findings regarding PFS and OS do not appear inferior to the results from previous studies, thereby suggesting that osimertinib was effective in elderly Japanese patients. Regarding the effects based on the PS, the ORR of PS0 and PS1 was 75% and 53.6%, respectively, and the PFS was 13.7 months and 11.9 months, respectively. Since there were few cases, it was impossible to discuss the significant differences, but the PS0 group tended to be superior.

It is also important to compare the results from osimertinib treatment to those from cytotoxic anticancer drugs, which are the alternative options if osimertinib is not used for T790M-positive cases. For example, the control group for the AURA3 study received platinum plus pemetrexed, which provided an ORR of 31%, a DCR of 74%, a median PFS of 4.4 months, and a median OS of 22.5 months [13,28]. A subgroup analysis of ≥70-year-old Japanese patients from the JACAL study evaluated carboplatin plus pemetrexed and revealed an ORR of 24%, a DCR of 68%, a median PFS of 5.2 months, and a median OS of 16.8 months [30,31]. Thus, our OS findings may be comparable to the results from the entire AURA3 population, although our ORR, DCR, and PFS outcomes are comparable or even slightly better. Interestingly, 71% of the patients in the group that received platinum plus pemetrexed subsequently received additional treatment, with 60% experiencing a greater effect after crossing over to osimertinib treatment. Therefore, while the JACAL study had only included EGFR-mutated cases and did not specifically consider older patients, we believe that osimertinib may provide good outcomes among older patients with EGFR-mutated (T790M) NSCLC.

Safety is also an important consideration in this setting, given the concerns regarding the potentially higher risk of adverse events among older patients. In the AURA3 study, it appears that Japanese patients had a higher risk of paronychia, diarrhea, and skin pruritus, although no clear increase was observed among elderly patients. However, elderly patients had a clearly increased frequency of myelosuppression events, such as anemia (75% in this study vs. 8% in AURA3 study), leukopenia (38.9% in this study vs. 8% in AURA3 study), neutropenia (41.7% in this study vs. 8% in AURA3 study), and thrombocytopenia (58.3% in this study vs. 10% in AURA3 study), although the frequencies of Grade 3–4 adverse events were generally comparable. Osimertinib has also been reported to be more frequently myelosuppressed than in other EGFR-TKI in a pivotal study [13,14]. In addition, myelosuppression was reported to be stronger in the analysis of the Japanese population [32]. Although the obvious mechanism was unclear, it was suggested that racial differences might be involved. Since myelosuppression was observed more frequently in the present study than in the aforementioned analysis of the Japanese population, caution should be exercised in the elderly Japanese. Fiala et al. reported that pre-treatment hypoalbuminemia correlated with poor prognosis in advanced NSCLC patients treated with erlotinib [33]. The present study also revealed that anorexia and exhaustion were common (30–40% of cases vs. 16-18% of cases in AURA3 study, including some Grade 3–4 cases), as well as hypoalbuminemia (75% of cases vs. N/A in AURA3 study). Therefore, careful follow-up is needed for elderly patients who are receiving osimertinib. Elevated alkaline phosphatase and creatinine values were also observed, albeit not serious cases, and related follow-up testing is also important. Cardiac adverse events, such as decreased left heart ejection fraction and QT prolongation, were observed in some cases, although only one patient experienced a Grade 3–4 cardiac adverse event. Central nervous system events, such as delirium and hallucination, may be explained by the large proportion of cases with brain metastasis (41.7%), although caution should be exercised if these events present in conjunction with sinusitis and pulmonary infection. Regarding AE by PS, no clear difference was observed between PS0 and PS1.

All-grade pneumonitis was observed in 11.1% of cases, and Grade 3–4 pneumonitis was observed in 5.6% of cases. The rates after conventional EGFR-TKI treatment were 4% in the AURA3 study and 7.3% in the Japanese subset of patients, which suggests that Japanese patients may have a higher rate of pneumonitis [13,34]. The difference between our findings and the previous findings may be related to differences in the proportions of patients with a history of smoking (69.4% for the present study, 32.2% for the AURA3 study, and 31.7% for the Japanese subset of the AURA3 population). In addition, the Japanese subset of the FLAURA study population had a higher frequency of pulmonary disorders (all grades: 12%, Grade 3 or higher: 2%); it should be noted that this is a first-line trial. Other reports have also suggested that osimertinib may be associated with an increased incidence of pulmonary disorders relative to other EGFR-TKIs [32]. Nevertheless, the odds ratio for pulmonary disorders after gefitinib treatment was 1.92-fold higher among Japanese patients who were ≥55 years old, which suggests that careful follow-up is required for patients who are ≥75 years old [35].

The present study revealed all-grade AEs in 86.1%, Grade 3 or worse AEs in 27.8%, and fatal AEs in 2.8% of the patients. These rates did not appear to be substantially elevated among elderly patients, based on results from the AURA3 study and its Japanese subgroup (all-grade: 97.8% and 100%, Grade 3 or higher: 22.6% and 31.7%, and fatal AEs: 1.4% and 0%). However, AEs leading to treatment discontinuation occurred in 12 patients (33.3%) in our study, which was more common than the rates of 6.8% in the AURA3 study and 7.3% in the Japanese subgroup. For example, we observed drug-induced lung injury in four patients (11.1%), and these patients needed to stop treatment. In addition, three patients (8.3%) discontinued treatment because of Grade 4 AEs (pulmonary infection, hallucinations, and hepatic dysfunction), although those events were judged unlikely to be associated with their treatment. One patient (2.8%) required a two-step dose reduction, and two patients (5.6%) were unable to continue the treatment protocol because of a ≥4-week treatment disruption. Treatment was also stopped in one case involving Grade 3 aspiration pneumonia, one case at the attending physician's discretion, and one case because the patient refused to continue treatment. Thus, although the safety of osimertinib outside the study protocol has not been evaluated, most of these AEs and treatment discontinuations were likely not to have been caused by a drug-induced pulmonary injury.

Most all-grade adverse events involved anorexia, fatigue, myelosuppression, and gastrointestinal symptoms. These complications were generally not serious and could be addressed using conventional management strategies. However, it is important to note that the frequency of drug-induced lung injury may increase, which highlights the importance of a careful follow-up in this population. Despite the potential need for a careful follow-up and the small sample size, which was the limitation in this study, it appears that osimertinib can be a standard treatment even for the elderly patients harboring T790M mutation.

While the present study provided encouraging data, we are conducting an additional phase II study (SPIRAL-0) to confirm the safety and efficacy of osimertinib in ≥75-year-old patients with untreated NSCLC harboring EGFR-activating mutations [36]. This may provide further information to guide the increasing use of osimertinib treatment in this setting.

**Author Contributions:** Conceptualization, J.U.; validation, J.U., A.N., and K.T.; formal analysis, K.Y.; investigation, O.H., C.S., T.A., N.H., T.I., T.T., M.K., Y.G., H.I., N.H., K.N., H.U., K.U., M.F. (Minoru Fukuda), Y.U., T.Y. (Toshihide Yokoyama), M.A., T.M. (Tadashi Mio), S.N., Y.C., N.T., Y.K., T.M. (Takako Mouri), T.Y. (Tadaaki Yamada) and M.F. (Masaki Fujita); data curation, K.Y.; writing—original draft preparation, A.N.; writing—review and editing, J.U.; supervision, K.T.; project administration, J.U.; funding acquisition, J.U. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by AstraZeneca. K.K., grant number ESR-15-11419.

**Acknowledgments:** We thank all of the patients who participated in this study, as well as their families. We also thank the Clinical Research Support Center Kyushu for managing the study.

**Conflicts of Interest:** K. Takayama received grants from Chugai-Roche Co., Ono Pharmaceutical Co. and personal fees from AstraZeneca Co., MSD-Merck Co., Eli Lilly Co., Boehringer-Ingelheim Co., Daiichi Sankyo Co. and Chugai-Roche Co. outside of the submitted work. K. Yoshimura received personal fees from AstraZeneca Co. and Chugai-Roche Co. outside of the submitted work. T. Ishizuka received grants from Boehringer-Ingelheim Co., Bayer Co, MSD Co., Astellas Co., Ono Pharmaceutical Co., Pfizer Co., Eli Lilly Co, Novartis Pharma K.K., Mochida Pharmaceutical Co and personal fees from AstraZeneca Co., Boehringer-Ingelheim Co., Novartis Pharma Co., and GlaxoSmithKline Co. outside of the submitted work. M. Fukuda received grants from AstraZeneca Co. and Eli Lilly Co. outside of the submitted work. M. Fujita received grants and personal fees from AstraZeneca Co. T. Yokoyama received personal fees from AstraZeneca Co., MSD-Merck Co., Eli Lilly Co., Boehringer-Ingelheim Co., Chugai-Roche Co. outside of the submitted work. J. Uchino received grants from Eli Lilly Japan K.K. and AstraZeneca Co. that are outside of the submitted work. The other authors have no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
