Study selection
Through a comprehensive search, we found 213 articles in PubMed and 7 trials in ClinicalTrials.gov. After duplicate, there were 73 articles and 1 trial left. By screening the titles and abstracts, we excluded 44 articles on the basis of exclusion criteria. Then we viewed the full texts of remained studies, and finally, 13 articles and 1 trial were involved according to the inclusion and exclusion criteria[14-27]. The overall filter procedures and results were shown in Figure 1.
Study characteristics
The basic information of eligible studies was listed on Table 1. The trials were conducted from 2010 to 2019, including 5 phase Ⅰ trials, 1 phase Ⅰ-Ⅱ trial, 6 phase Ⅱ trials, 1 phase Ⅲ trial and 1 unspecified trial. Among them, patients in 2 phase Ⅰ trials administered mogamulizumab intravenously at a dose of 0.01, 0.1, 0.5, 1.0 mg/kg[18, 22]. Patients in 3 phase Ⅰ trials and 2 phase Ⅱ trials received mogamulizumab in combination with other drugs[16, 24-27]. And the rest received mogamulizumab monotherapy at a dose of 1.0mg/kg[14-17, 19-21, 23]. There were total 1290 patients enrolled in these studies, including patients with ATL, PTCL, CTCL, ect.
Overall adverse events analysis
The safety data, grade ≥3 or all-grade AEs we extracted, were used to calculate the AEs rate to assess the safety of mogamulizumab. The details of the results were presented in Table 2, 3. In all eligible trials administered mogamulizumab monotherapy, we divided these trials into low dose group (mogamulizumab≤0.1 mg/kg), medium dose group (0.5 mg/kg) and high dose group (1.0 mg/kg) in accordance with the dose. In low dose group, lymphopenia was the most common all-grade AEs and grade ≥3 AEs with the highest rate of 0.700 (95% CI: 0.375-0.900) and 0.401 (95% CI: 0.158-0.705), respectively. In medium dose group, the most common all-grade AEs were leukopenia and lymphopenia with the same rate of 0.875 (95% CI: 0.463-0.983) while leukopenia (0.767, 95% CI: 0.337-0.955) was the only grade ≥3 AEs. In high dose group, the common all-grade AEs were lymphopenia (0.805, 95% CI: 0.432-0.957), infusion reaction (0.607, 95% CI: 0.062-0.973), fever (0.472, 95% CI: 0.116-0.859), rash (0.407, 95% CI: 0.210-0.639) and chills (0.401, 95% CI: 0.129-0.751), while lymphopenia (0.648, 95% CI: 0.482-0.787) was the most common grade ≥3 AEs. The rest of all-grade and grade ≥3 AEs were happened relatively less. In the trials administered mogamulizumab in combination with other drugs, the most common all-grade AEs were neutropenia (0.812, 95% CI: 0.035-0.998), anaemia (0.687, 95% CI: 0.017-0.996), lymphopenia (0.619, 95% CI: 0.007-0.997) and gastrointestinal disorder (0.599, 95% CI: 0.001-0.999). The lymphopenia (0.568, 95% CI: 0.004-0.998) was the most common grade ≥3 AEs while other grade ≥ 3 AEs were relatively rare.
Overall efficacy analysis
The pooled ORR rate, mean OS and mean PFS were used to measure the efficacy of mogamulizumab in treating cancers. For monotherapy, nine trials[14-17, 19-23] were included in the ORR analysis, and 4 articles[17, 20, 21, 23] were incorporated in the mean PFS analysis. According to our analysis, the pooled ORR rate was 0.430 (95% CI: 0.393-0.469) (Fig. 3 A). The median PFS varied from 0.93 to 7.7 months and the pooled mean PFS was 1.060 months (95% CI: 1.043-1.077 Z=125.452, p=0.000) (Fig. 4). Besides, we performed further analyses to evaluate the efficacy between mogamulizumab and other chemotherapeutics. In controlled trials, the HRs for PFS in 2 control trials was 0.53 and 0.71 with a total HR of 0.56 (95% CI: 0.45-0.71, I-squared=0.0%, p=0.348) (Fig. 2), indicating a longer PFS in mogamulizumab group. For mogamulizumab in combination with other drugs, the pooled ORR rate was 0.203 (95% CI: 0.022-0.746) (Fig. 3 B). Then we performed subgroup analyses to identify the pooled PFS and OS of patients with non-small cell lung cancer in combination therapies. The pooled PFS and OS were 2.435 months (95% CI: 1.752-3.119, Z=6.982, p=0.000) and 6.519 months (95% CI: 5.523-7.514, Z=12.836, p=0.000), respectively (Tab. 4).
Assessment of study quality and publication bias
We used Review Manager 5.3 (Copenhagen, Sweden) to measure quality assessment of involved studies. Figure 5 indicated the risk of bias graph and risk of bias summary of all those eligible trials. Overall, the quality of the studies was satisfactory.