A Phase III Study to Compare the Efficacy and Safety of Paclitaxel Versus Irinotecan in Patients with Metastatic or Recurrent Gastric Cancer Who Failed in First-line Therapy (KCSG ST10-01).

LESSONS LEARNED: Irinotecan could not be proven noninferior to paclitaxel as a second-line treatment for patients with metastatic or recurrent gastric cancer.The failure to demonstrate noninferiority may have been a result of insufficient patient enrollment.Both agents were tolerable but showed different toxicity profiles.
BACKGROUND: This phase III study compared the efficacy and safety of paclitaxel versus irinotecan in patients with metastatic or recurrent gastric cancer (MRGC) who had experienced disease progression following first-line chemotherapy.
METHODS: Patients were randomized to receive either paclitaxel (70 mg/m2; days 1, 8, 15, every 4 weeks) or irinotecan (150 mg/m2 every other week). The primary endpoint was progression-free survival (PFS).
RESULTS: This study was stopped early due to low accrual rate. A total of 112 patients were enrolled; 54 were allocated to paclitaxel and 58 to irinotecan. Median PFS for the paclitaxel and irinotecan groups was 3.5 and 2.1 months, respectively (hazard ratio [HR], 1.27; 95% confidence interval [CI], 0.86-1.88; p = .234). Noninferiority of irinotecan to paclitaxel was not proved because the upper boundary of the 95% CI (1.88) exceeded the predefined upper margin of noninferiority (1.32). Median overall survival (OS) was 8.6 months in the paclitaxel group and 7.0 months in the irinotecan group (HR, 1.39; 95% CI, 0.91-2.11; p = .126). Among toxicities greater than or equal to grade 3, neutropenia (11.5%) was the most common, followed by peripheral neuropathy (7.7%) in the paclitaxel group, and neutropenia (34.5%) followed by nausea, vomiting, and anemia (8.6%, respectively) in the irinotecan group.
CONCLUSION: Although paclitaxel showed numerically longer PFS and OS compared with irinotecan, this was statistically insignificant. Both irinotecan and paclitaxel are valid second-line treatment options in MRGC. © AlphaMed Press; the data published online to support this summary is the property of the authors.

RCT Entities:   Population Interventions Outcomes

Discussion

Second‐line chemotherapy for metastatic or recurrent gastric cancer (MRGC) has been shown to improve survival in previous clinical trials. Although taxane or irinotecan are commonly used as second‐line chemotherapy for MRGC, few previous studies have directly compared the efficacy between taxane and irinotecan. In our study, therefore, we compared the efficacy and safety of irinotecan and paclitaxel as second‐line therapy in MRGC. PFS and OS were not statistically different (Figs. 1 and 2). However, noninferiority of irinotecan compared with paclitaxel could not be confirmed because of low patient enrollment. Toxicity profiles of irinotecan and paclitaxel were different.

Figure 1.

Progression‐free survival.

Abbreviations: CI, confidence interval; HR, hazard ratio; PFS, progression‐free survival.

Figure 2.

Overall survival.

Abbreviations: CI, confidence interval; HR, hazard ratio; OS, overall survival.

Like our study, a previous Japanese phase III trial (WJOG 4007) compared the efficacy of paclitaxel versus irinotecan. The WJOG 4007 study was conducted to verify the hypothesis that irinotecan has superior OS to paclitaxel, and the authors concluded that both irinotecan and paclitaxel are reasonable second‐line treatment options for MRGC because no significant difference in OS was observed. However, strictly speaking, the noninferiority of one agent to the other was also not proved in the WJOG 4007, considering the study design. When results of WJOG 4007 and our studies are compared, an interesting finding is observed simultaneously; paclitaxel showed numerically longer PFS (3.6 vs. 2.3 months [WJOG 4007; p = .33]; 3.5 vs. 2.1 months [our study; p = .234]) and OS (9.5 vs. 8.4 months [WJOG 4007; p = .38]; 8.6 vs. 7.0 months [our study; p = .126]) compared with irinotecan. As the difference in survival outcomes was not statistically significant in both WJOG 4007 and our studies, the observation of possible superiority of paclitaxel over irinotecan can only be considered hypothesis‐generating.

Patient enrollment was less than expected and is the most important limitation in interpreting study results. During the study period, results of WJOG 4007 were reported, and this drove investigators to be less interested in our study and thus to enroll patients less often.

In conclusion, noninferiority of irinotecan compared with paclitaxel could not be proven. The hazard ratio of PFS crossed the boundary needed to prove noninferiority. Although paclitaxel showed numerically longer PFS and OS compared with irinotecan, this was statistically insignificant. Therefore, both irinotecan and paclitaxel are thought to be valid second‐line treatment options in MRGC. Considering different toxicity profiles and treatment schedule, the choice of chemotherapeutic agents must be based on medical condition and compliance of patients.

Trial Information

Gastric cancer

Metastatic/advanced

1 prior regimen

Phase III

Randomized

Overall response rate

Safety

Active but results overtaken by other developments

Drug Information for Phase III Control

Paclitaxel

Taxol

Boryung Pharmacetutical

Small molecule

Microtubule‐targeting agent

70 mg/m2

IV

Paclitaxel (Taxol; 70 mg/m2) mixed with 100–250 mL of normal saline was administered intravenously over 1 hour on days 1, 8, and 15, every 4 weeks. The 4‐week schedule was considered one cycle. Premedication for paclitaxel using corticosteroid and antihistamine was conducted following each institution's policy.

Drug Information for Phase III Experimental

Irinotecan

Campto

CJ HealthCare Corp.

Small molecule

Topoisomerase I

150 mg/m2

IV

Irinotecan (Campto; 150 mg/m2) mixed with 100–250 mL of 5% dextrose water was administered intravenously over 1 hour on days 1 and 15, every 4 weeks. The 4‐week schedule was considered one cycle. In patients developing irinotecan‐induced cholinergic symptoms (abdominal pain and diarrhea), atropine was used as premedication.

Patient Characteristics for Phase III Control

38

16

All patients had metastatic or recurrent disease.

Median (range): 58.5 (38–82)

Median (range): 1 (1)

0 — 0

1 — 52

2 — 2

3 — 0

Unknown — 0

Additional information is shown in Table 1. During the period from February 2011 through January 2015, a total of 116 patients were screened, and 112 patients were enrolled at 16 sites in Korea. During the enrollment period, patient accrual rate was low, and we decided to stop this study early even though the target number of patients was 520.

Table 1.

Baseline characteristics

Unless otherwise noted, each p value was calculated by chi‐square test.

t test.

No patients had a grade 0 of ECOG PS.

Fisher's exact test.

Time interval between two lines of treatment was defined as duration from the start of the first‐line chemotherapy to the date of randomization in this trial.

Abbreviation: ECOG PS, Eastern Cooperative Oncology Group performance status.

Well differentiated 4

Moderately differentiated 10

Poorly differentiated 38

Undifferentiated 1

Cannot be assessed 1

Patient Characteristics for Phase III Experimental

40

18

All patients had metastatic or recurrent disease.

Median (range): 59 (38–77)

Median (range): 1 (1)

0 — 0

1 — 56

2 — 2

3 — 0

Unknown — 0

Additional information is shown in Table 1.

Well differentiated 34

Moderately differentiated 16

Poorly differentiated 5

Undifferentiated 2

Cannot be assessed 1

Primary Assessment Method for Phase III Control

Paclitaxel arm

54

54

52

54

RECIST 1.1

n = 1 (2.6%)

n = 5 (13.2%)

n = 16 (42.1%)

n = 14 (36.8%)

n = 2 (5.3%)

3.5 months, CI: 2.2–4.7

8.6 months, CI: 7.1–10.0

Primary Assessment Method for Phase III Experimental

Irinotecan arm

58

58

57

58

RECIST 1.1

n = 1 (2.3%)

n = 5 (11.4%)

n = 15 (34.1%)

n = 16 (36.4%)

n = 7 (15.9%)

2.1 months, CI: 1.4–2.8

7.0 months, CI: 5.6–8.4

Serious Adverse Events (Paclitaxel)

Serious Adverse Events (Irinotecan)

Assessment, Analysis, and Discussion

Study terminated before completion

Did not fully accrue

Active but results overtaken by other developments

Gastric cancer (GC) is a major cause of cancer‐related death, with more than 720,000 deaths worldwide [4]. In Korea, where GC is one of the most common types of cancer, GC is estimated to be the fourth and fifth most common cause of death in male and female cancer patients, respectively [5]. In general, fluoropyrimidine‐based doublet chemotherapy, or triplet chemotherapy for highly selected patients, is widely used as palliative first‐line therapy in patients with metastatic or recurrent GC (MRGC) [6], [7]. If the tumor shows human epidermal growth factor 2 amplification, addition of trastuzumab to fluoropyrimidine plus platinum is now the standard first‐line treatment [8]. After failure of first‐line chemotherapy, second‐line therapy with docetaxel or irinotecan showed improved overall survival (OS) compared with best supportive care in previous phase III trials [9], [10], [11].

Paclitaxel, a microtubule stabilizing agent that inhibits depolymerization during cell division, has been tested as a second‐line chemotherapy for MRGC in previous single‐arm studies [12]. Overall response rate (ORR) was 16%–24%, with median OS of 5–8 months [1], [2], [3], [13], [14], [15]. Irinotecan inhibits topoisomerase I from unwinding DNA strands during DNA replication and has also been examined as a second‐line therapy in previous single‐arm studies, which showed 12%–20% of ORR and about 5 months of OS [16], [17]. These efficacy data seemed to be comparable between irinotecan and paclitaxel in the second‐line setting for MRGC, and both agents have been widely used in clinics in Korea. As there were no data from randomized clinical trials that directly compared the efficacy and safety of paclitaxel and irinotecan as second‐line therapy in MRGC, we designed and conducted this phase III trial.

In this phase III trial, patients with histologically confirmed metastatic or recurrent gastric adenocarcinoma were eligible if they were older than 18 years and had disease progression to the palliative first‐line chemotherapy. If patients experienced recurrence during or within 6 months after the completion of adjuvant chemotherapy following curative surgery, they were allowed to be enrolled to this trial. Other eligibility criteria were Eastern Cooperative Oncology Group performance status of 0–2; evaluable tumor lesions with or without measurable target lesions based on RECIST, version 1.1; adequate organ functions including bone marrow, kidney, and liver; and more than 16 weeks of expected survival. Patients were excluded if they had received more than one line of prior chemotherapy or had been exposed to taxane or irinotecan prior to this trial.

This was an investigator‐initiated, multicenter, randomized, phase III trial conducted at 16 centers in Korea. The study protocol was approved by the institutional review board of each participating institution and the Korean Cancer Study Group (KCSG; trial number, KCSG ST10‐01). This trial was registered with ClinicalTrials.gov (NCT01224652). The randomization was performed centrally at the KCSG data center using the method of permuted block randomization. Patients were randomly assigned in a ratio of 1:1 to either paclitaxel or irinotecan.

Paclitaxel (Taxol; 70 mg/m2) was administered intravenously on days 1, 8, and 15, every 4 weeks. Irinotecan (Campto; 150 mg/m2) was administered intravenously on days 1 and 15, every 4 weeks. For both agents, the 4‐week schedule was considered one cycle. Predefined dose reduction or delay were conducted to manage treatment‐related toxicity. If neutropenia or thrombocytopenia of greater than or equal to grade 3 or clinically significant nonhematologic toxicities developed, dose reduction of paclitaxel (60 mg/m2) and irinotecan (120 mg/m2) was performed. If clinically significant hematologic or nonhematologic toxicities developed again despite dose reduction, no more dose reduction was conducted and the study treatment was permanently withdrawn. If the administration of study treatment was delayed more than 4 weeks due to delayed recovery from toxicities, the study treatment was also permanently withdrawn. The study treatment was continued until disease progression, death, development of unacceptable toxicity, or a patient's refusal of further therapy.

Tumor response evaluation using computed tomography was performed every 8 weeks (windows, ±7 days) based on RECIST (version 1.1). Hematologic laboratory test (complete blood count with differentiation) was repeated every week during the first cycle and then checked before the administration of study drugs (day 1, 8, and 15 for paclitaxel; day 1 and 15 for irinotecan). Chemistry was performed on the first day of each cycle. Adverse event was assessed at every visit according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 4.0).

Between February 2011 and January 2015, a total of 116 patients were screened, and 112 patients were enrolled at 16 sites in Korea. During the enrollment period, the accrual rate was low, and the result of a randomized phase III trial (WJOG4007), which had a similar design to our study, was reported in 2013 [18]. Therefore, we decided to stop this study early even though the target number of patients was 520. Among 112 patients, 54 were allocated to paclitaxel and 58 to irinotecan. Thus, full analysis set consisted of 54 in the paclitaxel group and 58 in the irinotecan group (n = 112). Of these patients, two in the paclitaxel group and one in the irinotecan group did not receive the allocated treatment because of consent withdrawal before the first dose. Therefore, safety analysis set included 52 in the paclitaxel group and 57 in the irinotecan group (Fig. 3; CONSORT diagram).

Figure 3.

Consolidated Standards of Reporting Trials diagram. Abbreviations: CR, complete response; ITT, intention‐to‐treat.

Table 1 summarized baseline characteristics across two treatment groups. The median time interval from the start of the first‐line chemotherapy to the date of randomization for the study treatment in all 112 patients was 7.1 months (range, 1.0–37.7). Overall, each variable was well balanced between two groups. All patients had received fluoropyrimidine‐based first‐line chemotherapy prior to the enrollment into this study. Although there seemed to be a little imbalance in first‐line chemotherapy regimens between the two groups, this was not statistically significant. In both groups, oxaliplatin‐based doublet chemotherapy was the most common treatment used as the first‐line therapy.

The data cutoff date was December 4, 2015. Median treatment duration in paclitaxel and irinotecan groups were 10.5 (range, 0–90.3) and 6.2 (range, 0–138.3) weeks, respectively. Mean dose intensities (± standard deviation) were 46.2 (±10.8) mg/m2/week in the paclitaxel group and 60.9 (±15.3) mg/m2/week in the irinotecan group, respectively.

Adverse events were observed in 96.3% of all the study population. Severe adverse events (grade ≥3) occurred in 32.7% of patients in the paclitaxel group and 45.6% in the irinotecan group (p = .177). Table 2 summarized adverse events according to the grade in each group. In the paclitaxel group, the most common adverse event was peripheral neuropathy (51.9%), followed by anemia (40.4%) and fatigue (40.4%). Neutropenia and anorexia were also frequent. Among adverse events greater than grade 3, neutropenia was the most common (11.5%). Peripheral neuropathy of greater than or equal to grade 3 was observed in 7.7% of the patients. In the irinotecan group, the most common adverse event was neutropenia (48.3%), followed by nausea (46.6%) and anorexia (44.8%). Diarrhea, anemia, and fatigue were also frequent. The most common severe adverse event (greater than or equal to grade 3) was neutropenia (34.5%). Peripheral neuropathy was more common in the paclitaxel group compared with the irinotecan group, whereas diarrhea was more prevalent in the irinotecan group. The irinotecan group also showed more frequent grade 3 or 4 neutropenia compared with the paclitaxel group.

Table 2.

Adverse events

Only adverse events observed more than 10% in any groups were listed with the proportion (%) of patients.

Irinotecan did not show statistically different progression‐free survival (PFS) compared with paclitaxel. Median PFS was 2.1 months in the irinotecan group and 3.5 months in the paclitaxel group, respectively (hazard ratio [HR], 1.27; 95% confidence interval [CI], 0.86–1.88; p = .234; Fig. 1). Noninferiority of irinotecan could not be confirmed because the CI exceeded the limit of predefined noninferiority margin of 1.32. Median OS was 7.0 months in the irinotecan group and 8.6 months in the paclitaxel group. The difference was not statistically significant (HR, 1.39; 95% CI, 0.91–2.11; p = .126; Fig. 2). Of 54 patients in the paclitaxel group, 30 (56%) received poststudy treatment, and irinotecan‐containing regimens were most commonly used (83%; 25/30). Of 58 patients in the irinotecan group, 36 (62%) received poststudy chemotherapy, and these patients received taxane‐based regimens most commonly (81%; 29/36). Thus, a total of 66 patients (59%) received at least one line of treatment after end of the study treatment. Response evaluation was conducted on patients with at least one measurable lesion (38 of 54 in the paclitaxel group and 44 of 58 in the irinotecan group). Response rate was also similar between two treatment groups. ORRs of paclitaxel group and irinotecan group were 15.8% and 13.6%, respectively (p = .355; Table 3).

Table 3.

Response rate

Each p value was calculated by chi‐square test.

Abbreviations: DCR, disease control rate; ORR, overall response rate.

The role of second‐line chemotherapy for MRGC has been shown in recent phase III trials. Irinotecan was found to be effective as a second‐line chemotherapy in German AIO trial despite low accrual [10]. A large phase III trial from Korea established the role of second‐line chemotherapy, where irinotecan or docetaxel was chosen according to physician's discretion. In this trial, second‐line chemotherapy showed superior OS compared with best supportive care alone (5.3 vs. 3.8 months [median]; HR, 0.657; p = .007) [9]. In the COUGAR‐02 trial, second‐line docetaxel showed improved OS compared with active symptom control group (5.2 vs. 3.6 months [median]; p = .01), although docetaxel was associated with more toxicity [11].

In our study (KCSG ST10‐01), we compared the efficacy and safety of irinotecan and paclitaxel as second‐line therapy in MRGC. However, noninferiority of irinotecan compared with paclitaxel could not be confirmed in our study. The most crucial reason for this is low patient enrollment, which was translated into lower power to test the hypothesis. Like our study, a previous Japanese phase III trial (WJOG 4007) compared the efficacy of paclitaxel versus irinotecan as a second‐line chemotherapy in MRGC [18]. The WJOG 4007 study was conducted to verify the hypothesis that irinotecan has superior OS to paclitaxel, and the authors concluded that both irinotecan and paclitaxel are reasonable second‐line treatment options for MRGC because no statistically significant difference in OS was observed between paclitaxel and irinotecan. However, strictly speaking, the noninferiority of one agent to the other was not proved in the WJOG 4007, considering the study design. When results of KCSG ST10‐01 and WJOG 4007 studies are compared, an interesting finding is observed simultaneously in both studies: Paclitaxel showed numerically longer PFS (3.6 vs. 2.3 months in WJOG 4007 [p = .33]; 3.5 vs. 2.1 months in our study [p = .234]) and OS (9.5 vs. 8.4 months in WJOG 4007 [p = .38]; 8.6 vs. 7.0 months in our study [p = .126]) compared with irinotecan, although statistically insignificant. In our study, paclitaxel showed comparable PFS and OS to those reported in recent phase III studies (2.9–3.6 months of PFS and 6.9–9.5 months of OS) [18], [19], [20]. As the difference in survival outcomes was not statistically significant in both WJOG 4007 and our studies, the observation of possible superiority of paclitaxel over irinotecan is just hypothesis‐generating. All toxicity profiles of irinotecan and paclitaxel were different. Among toxicities of greater than or equal to grade 3, neutropenia (11.5%) was the most common, followed by peripheral neuropathy (7.7%) in the paclitaxel group, and neutropenia (34.5%) followed by nausea, vomiting, and anemia (8.6%, respectively) in the irinotecan group. These toxicity profiles were consistent with previous reports and manageable [9], [10], [14], [17], [18]. Taken together, we authors agree that both irinotecan and paclitaxel are reasonable second‐line treatment options in MRGC.

During our study period, ramucirumab, a monoclonal antibody inhibiting vascular endothelial growth factor receptor 2, was approved for second‐line treatment as monotherapy or in combination with paclitaxel based on two pivotal phase III trials [19], [21]. In the REGARD trial, ramucirumab monotherapy showed longer OS compared with placebo (5.2 vs. 3.8 months [median]; HR, 0.776; p = .047). Median PFS was also improved with ramucirumab (2.1 vs. 1.3 months [median]; HR, 0.483; p < .0001) [21]. Likewise, in the RAINBOW trial, the combination of ramucirumab and paclitaxel showed superior PFS (4.4 vs. 2.9 months [median]; HR, 0.635; p < .0001) and OS (9.6 vs. 7.4 months [median]; HR, 0.807; p = .017) compared with paclitaxel alone [19]. However, in the preplanned subgroup analysis in the both studies, the OS in Asian patients were not statistically different between each treatment group [19], [21], although the subgroup analysis in the RAINBOW trial showed the superiority of PFS in the ramucirumab plus paclitaxel group compared with paclitaxel plus placebo group even in Asian patients [19]. Furthermore, ramucirumab‐related adverse events such as gastrointestinal perforation or proteinuria should not be ignored, although the incidence was not frequent. Thus, cytotoxic chemotherapy such as paclitaxel or irinotecan is still a viable option to treat patients with MRGC in the second‐line setting.

Clinical trials comparing ramucirumab plus irinotecan versus irinotecan or comparing ramucirumab plus irinotecan versus ramucirumab plus paclitaxel are not expected to be conducted in the future, considering the similar efficacy of paclitaxel and irinotecan and proven superior efficacy of ramucirumab plus paclitaxel to paclitaxel alone. Instead, comparison between ramucirumab with FOLFIRI (irinotecan, leucovorin, and 5‐fluorouracil) and ramucirumab with paclitaxel is now ongoing (NCT03081143). This study would be another indirect indicator to see if irinotecan has a different efficacy over paclitaxel when combined with ramucirumab.

Far fewer patients enrolled than expected, limiting our interpretation of study results. During the study period, the outcome of WJOG 4007 was reported, and this drove investigators to be less interested in this study and thus to enroll patients much less. Furthermore, confirmative landmark trials established the evident role of ramucirumab as a second‐line treatment in MRGC [19], [21]. In Korea, many kinds of clinical trials are conducted for MRGC. This often results in enrollment competition between various trials that have similar eligibility criteria. It is estimated that many investigators allocated patients to other clinical trials because neither paclitaxel nor irinotecan were novel investigational drugs.

In the paclitaxel group, a total of 11 serious adverse events were reported in 7 patients. Gastric hemorrhage (tumor bleeding) developed in two patients; one patient developed two events of gastric hemorrhage (grade 2 [one event] and grade 3 [one event], respectively). Vomiting developed in two patients (grade 2 [one event] and grade 3 [one event], respectively).

In the irinotecan group, a total of 15 serious adverse events were reported in 13 patients. Anorexia, fever, and sepsis were reported in four, two, and two patients, respectively. One patient with sepsis died from it.