Cohort outcomes
22 NSCLC patients with EGFR mutations were qualified for the study between January 2019 to December 2020. The demographic and clinical characteristics are listed in Table 1 and S3. The minimum follow-up for the cohorts of patients was months (defined as the time from the last enrolled patient’s first treatment to the database lock). Median age was 58 years (range, 42–74 years), 50% of patients were male, 81.82% had stage IV+ (including IV, IVA, IVB and IVC) disease at study entry. All patients had received prior TKI, including 6 patients whose treatment had been combined with chemotherapy, 2 with anti-angiogenesis, 3 with radiotherapy and 1 with anti-PD1 treatment. The main EGFR mutations were 19Del (63.64%), L858R (13.64%), and 20ins (9.09%). Fourteen patients received “ICI + Chemotherapy” (63.64%), 5 patients received “ICI monotherapy” (22.73%), 2 patients received “ICI + Chemotherapy + Anti-angiogenesis” (9.09%) and 1 patient received “ICI + Anti-angiogenesis” (4.54%). Sixteen patients had discontinued treatment due to SD (72.73%). Median PFS was 4.9 (range, 0.75 ~ 9 months) (Fig. 1).
Table 1
Patient characteristics and clinical outcomes in our cohorts
Characteristic
|
Patients (N = 22)
|
Median age (range) -yr
|
58 (42–74)
|
Age group - no. (%)
|
|
< 50 yr
|
5 (22.73)
|
50–60 yr
|
8 (36.36)
|
≥ 60 yr
|
9 (40.91)
|
Male sex - no. (%)
|
11 (50.00)
|
Disease stage - no. (%)
|
|
IIIA
|
1 (4.54)
|
IIIB
|
2 (9.09)
|
IIIC
|
1 (4.54)
|
IV
|
6 (27.27)
|
IVA
|
3 (13.64)
|
IVB
|
8 (36.36)
|
IVC
|
1 (4.54)
|
Treatment history - no. (%)
|
|
TKIs
|
20 (90.90)
|
Radiotherapy
|
3 (13.64)
|
Chemotherapy
|
6 (27.27)
|
Anti-angiogenesis
|
2 (9.09)
|
Anti-PD1
|
1 (4.54)
|
EGFR mutation status - no. (%)
|
|
19Del
|
14 (63.64)
|
18-G719,18-G709
|
1 (4.54)
|
20ins
|
2 (9.09)
|
L858A
|
1 (4.54)
|
L858R
|
3 (13.64)
|
L861Q
|
1 (4.54)
|
Treatment strategies - no. (%)
|
|
ICI monotherapy
|
5 (22.73)
|
ICI + Chemotherapy
|
14 (63.64)
|
ICI + Chemotherapy + Anti-angiogenesis
|
2 (9.09)
|
ICI + Anti-angiogenesis
|
1 (4.54)
|
Outcome - no. (%)
|
|
PD
|
1 (4.54)
|
PR
|
5 (22.73)
|
SD
|
16 (72.73)
|
PFS (range) - month
|
4.9 (0.75, 9)
|
According to treatment strategies, in ICI monotherapy group, the median PFS was 2.3 months (range, 1.3 ~ 6 months); and in ICI + Chemotherapy, the median PFS was 5 months (range, 0.75 ~ 9 months) (Log rank P = 0.037) (Figure S1). The ORR was 40.0% (11.7%~76.9%) and 14.3% (4.0%~40.0%) in ICI monotherapy and ICI + Chemotherapy group, respectively.
Literature search
As shown in the PRISMA flow chart (Fig. 2), according to the pre-established screening procedures, 1540 and 14 potentially relevant articles were identified in respective databases and other sources. After reviewing the titles and abstracts, 48 articles were excluded. After retrieval of full-text articles, 182 articles were further excluded due to: i) duplicated publications, ii) lack of relevant subpopulations, iii) incomplete data, or iv) others (e.g. non-English language). Finally, 15 articles, published between 2014 and 2021, were considered eligible for this present study.
Study Characteristics
Table 2 lists the main characteristics of the studies presently selected. All studies enrolled patients within the past decade, although most were published in the last three years. Among the 15 included published articles, 10 were retrospective studies (Garassino et al. 2018; Hastings et al. 2019; Sakamoto et al., 2019; Yamada, 2019; Oya and Kuroda 2020; Chen et al. 2021; Deng et al. 2021; Ichihara et al. 2021; Masuda et al. 2021; Ito et al. 2022), 1 phase I trial (Gettinger et al. 2018), 1 phase I/II trial (Yang et al. 2019), 2 phase II trials (Jiang et al. 2021; Hayashi et al. 2022), and 1 real-word study (Bylicki et al. 2020). All included studies were of high-quality with scores ranging from 7 to 10 according to the NOS criteria.
Table 2
Baseline characteristics of including studies for meta-analysis
Study
|
Research
|
Phase
|
Period
|
Cite
|
Treatment
|
Sample
|
Gender
|
Age
|
EGFR mutation
|
Score
|
Marina Chiara Garassino, 2018
|
/
|
Retrospective cohort
|
2015.05-2016.04
|
153
|
Nivolumab (3 mg/kg every 2 weeks)
|
102
|
44/58
|
65 (40–83)
|
/
|
9
|
Scott Gettinger, 2018
|
CheckMate 012
|
Phase I
|
2016.04-2017.09
|
/
|
Nivolumab (3 mg/kg every 2 weeks) Plus Erlotinib
|
21
|
8/13
|
63(41–80)
|
Exon 19 deletion (12); Exon 21 L858R (9); T790M (6); S7681 (1)
|
10
|
Hiroaki Sakamoto 2019
|
|
Retrospective cohort
|
2014.09-2017.01
|
1
|
Nivolumab (22 patients), Pembrolizumab (2 patients)
|
24
|
5/19
|
68 (39‑82)
|
Exon 19 deletion (14); Exon 21 L858R (7); ALK (1); ROS-1 (1); RET (1)
|
8
|
James Chih-Hsin Yang, 2019
|
KEYNOTE-021
|
Open-label, phase 1/2
|
2014.04-2015.11
|
multy
|
Pembrolizumab (2mg/kg every 3 weeks) + erlotinib (150mg)
|
12
|
6/6
|
59.5 (49–74)
|
T790M
|
10
|
K. Hastings, 2019
|
/
|
Retrospective cohort
|
/
|
4
|
ICI
|
75
|
/
|
/
|
/
|
7
|
Tadaaki Yamada, 2019
|
/
|
Retrospective cohort
|
2016.02-2018.05
|
6
|
Nivolumab (3 mg/kg every 2 weeks) and Pembrolizumab (1200 mg every 3 weeks)
|
27
|
8/19
|
67.0 (37.0–82.0)
|
Exon 19 deletion (8); Exon 21 L858R (12); G719X (4); Exon20 ins (3)
|
8
|
Eiki Ichihara, 2020
|
/
|
Retrospective cohort
|
2015.12-2018.05
|
1
|
ICI
|
58
|
31/27
|
68 (37–87)
|
Exon 19 deletion (33); Exon 21 L858R (18); others (7)
|
8
|
Ken Masuda, 2020
|
/
|
Multicenter retrospective observational study
|
2017.03-2018.12
|
1
|
Nivolumab or Pembrolizumab
|
35
|
13/22
|
63 (37–85)
|
Exon 19 deletion (21); Exon 21 L858R (8); others (6)
|
7
|
Olivier Bylicki, 2020
|
IMAD study
|
Real-world
|
/
|
20
|
ICI
|
42
|
/
|
/
|
Exon 19 deletion (20); Exon 21 L858R (14); G719X (3); L747P (1); Double(4)
|
7
|
Yuko Oya, 2020
|
/
|
Retrospective cohort
|
2015.01-2018.09
|
1
|
Nivolumab (3mg/kg every 2 weeks), Pembrolizumab (200mg/body every 3 weeks)
|
47
|
/
|
/
|
/
|
8
|
Haiyi Deng, 2021
|
/
|
Retrospective cohort
|
2019.02-2020.08
|
2
|
ICIs combined with single-drug chemotherapy
|
8
|
4/4
|
66 (53–73)
|
Exon 19 deletion (5); Exon 21 L858R (3)
|
8
|
Hidetoshi Hayashi, 2021
|
/
|
Prospective, randomized phase II trial
|
2016.04-2019.06
|
37
|
Nivolumab (3 mg/kg) on day 1 every 2 weeks until disease progression or the development
|
52
|
24/28
|
70.5 (51–84)
|
Exon 19 deletion (24); Exon 21 L858R (28); bExon-18 G719X plus exon-20 S768I (1)
|
9
|
Takashi Ito, 2021
|
/
|
Retrospective cohort
|
2015.12-2020.03
|
3
|
ICI(Nivolumab/Permbrolizumab/Atezolizumab)
|
25
|
15/10
|
67 (38–80)
|
Exon 19 deletion (11); Exon 21 L858R (6); other (4)
|
9
|
Tao Jiang, 2021
|
/
|
Multicenter phase-II trial
|
2018.04-2019.03
|
multicenter
|
Toripalimab plus pemetrexed and carboplatin
|
40
|
19/21
|
57.5 (19–73)
|
Exon 19 deletion (23); Exon 21 L858R (17)
|
9
|
Ya Chen, 2021
|
/
|
Retrospective cohort
|
2017.12-2020.10
|
1
|
Pembrolizumab monotherapy
|
32
|
48.38
|
62(39–80)
|
Exon 19 deletion (11); Exon 21 L858R (6); other (4)
|
9
|
|
|
|
Pembrolizumab combined with chemotherapy
|
26
|
|
|
|
Pembrolizumab combined with anlotinib
|
28
|
According to the main types of therapeutic drugs, the treatment strategies could be divided into three groups, such as ICI monotherapy (“ICI group”), combined ICI and TKIs (“ICI + TKI group”), and, combined ICI and chemotherapy (“ICI + chemotherapy group”). As shown in Tables 2 and 3, the most common treatment strategies were ICIs, applied in 11 studies. ICI + TKI group were administered in 3 of the studies, ICI + chemotherapy group was utilized in 3 studies; among the included studies; most patients were at stage of III or IV when enrolled.
Table 3
Disease characteristics of including studies for meta-analysis
Study
|
Histology
|
Disease stage
|
ECOG PS score
|
TKI
|
Metastatic site
|
Marina Chiara Garassino, 2018
|
/
|
/
|
0 [42 (41.6%)], 1 [52 (51.0%)], 2 [7 (6.9%)], unknown [1 (1.0%)]
|
/
|
Brain [44(43.1%)], liver [32(31.4%)], bone [47(46.1%)]
|
Scott Gettinger, 2018
|
AD: [20 (95.2%)], mixed: [1 (4.8%)]
|
ΙΙΙB [1 (4.8%)], ΙV [20 (95.2%)]
|
/
|
Y
|
/
|
Hiroaki Sakamoto, 2019
|
AD: [23 (95.8%)], SCC: [1 (4.2%)]
|
ΙΙΙB [2 (8.3%)], ΙV [12 (50.0%)], recurrence [10 (41.7%)]
|
0–1 [20 (83.3%)], 2 [4 (16.7%)]
|
/
|
/
|
K. Hastings, 2019
|
/
|
/
|
/
|
/
|
/
|
James Chih-Hsin Yang, 2019
|
/
|
ΙΙΙA [1 (8.3%)], ΙΙΙB [1 (8.3%)], ΙV [10 (83.3%)]
|
/
|
Y
|
Metastatic [6 (50%)]
|
Tadaaki Yamada, 2019
|
AD [26 (96.3%)], LCNEC [1 (3.7%)]
|
ΙV [18 (66.7%)], recurrence [9 (33.3%)]
|
0–1 [23 (85.2%)], 2 [4 (14.8%)]
|
/
|
Brain [11(40.7%)], liver [4(14.8%)], bone [12(44.4%)]
|
Eiki Ichihara, 2020
|
AD [54 (93.10%)], squamous [4 (6.90%)]
|
ΙΙΙ, ΙV
|
0–1 [50 (86.21%)], 2–4 [8 (13.79%)]
|
/
|
/
|
Olivier Bylicki
|
/
|
/
|
0 [6 (11.5%)], 1 [37 (73.0%)], 2 [6 (11.5%)], 3 [2 (4%)]
|
/
|
/
|
Ken Masuda, 2020
|
PD-L1 high: AD [16 (94.12%)], squamous [0]; others [1 (5.9%)]; PD-L1 low: AD [18 (100.0%)], squamous [0]; others [0]
|
PD-L1 high: ΙΙΙ [3 (17.6%)], ΙV [9 (52.9%)], recurrence [5 (29.4%)]; PD-L1 low: ΙΙΙ [2 (11.1%)], ΙV [12 (66.7%)], recurrence [4 (22.2%)]
|
PD-L1 high: 0 ,1 [14 (82.4%)], 2 [3 (17.6%)]; PD-L1 low: 0 ,1 [16 (88.9%)], 2 [2 (11.1%)]
|
/
|
/
|
Yuko Oya, 2020
|
AD [47 (100.0%)]
|
ΙΙΙ, ΙV, recurrence
|
/
|
/
|
/
|
Haiyi Deng, 2021
|
/
|
ΙΙΙB [1 (12.5%)], ΙV [7 (87.5%)]
|
0 [2 (25%)], 1 [6 (75%)]
|
/
|
Brain [3(37.5%)], liver [3(37.5%)]
|
Hidetoshi Hayashi, 2021
|
AD [52 (100.0%)]
|
ΙΙΙB [2 (3.8%)], ΙV [41 (78.8%)], recurrence [9 (17.3%)]
|
0 [17 (32.7%)], 1 [35 (67.3%)]
|
/
|
Brain [18(34.6%)]
|
Takashi Ito, 2021
|
AD [52 (100.0%)]
|
ΙΙΙB or ΙV [15 (60.0%)], recurrence [10 (40.0%)]
|
/
|
/
|
CNS [7 (30.4%)], others [16 (69.6%)]
|
Tao Jiang, 2021
|
/
|
/
|
0 [2 (5%)], 1 [38 (95%)]
|
/
|
CNS [6 (15%)]
|
Ya Chen, 2021*
|
/
|
/
|
/
|
Y
|
Brain [17(19.8%)]
|
AD: Adenocarcinoma; CNS: Central Nervous System; ECOG PS: Eastern Cooperative Oncology Group performance status; LCNEC: Large Cell Neuroendocrine Carcinoma; SCC: Squamous Cell Carcinoma |
*26 patients received TKI treatment. |
Short-term PFS in patients with EGFR-mutation
Thirteen studies reported 6-month PFS data for EGFR-mutation, which involved a total of 612 patients. The pooled and calculated 6-month PFS ratio was 0.23 (95%CI: 0.17 ~ 0.31, I2 = 791%, p < 0.01) (Fig. 3). According to the treatment strategies, subgroup analysis showed that in the ICI + chemotherapy group presented the best PFS (ICI group: 0.18 (0.15 ~ 0.22), I2 = 30%, P = 0.12; ICI + TKI group: 0.32 (0.20 ~ 0.46), I2 = 17%, p = 0.276; ICI + chemotherapy group: 0.53 (0.23 ~ 0.81), I2 = 78%, p = 0.01).
Thirteen studies reported 12-month PFS data of EGFR-mutation, which involved a total of 600 patients. The pooled and calculated 12-month PFS ratio was 0.15 (95%CI: 0.10 ~ 0.22, I2 = 67%, p < 0.01) (Fig. 4). According to the treatment strategies, subgroup analysis showed that patients in the ICI + TKI group and ICI + chemotherapy group presented the similarly PFS (ICI group: 0.12 (0.09 ~ 0.15), I2 = 7%, p = 0.38; ICI + TKI group: 0.26 (0.04 ~ 0.76), I2 = 85%, p < 0.01; ICI + chemotherapy group: 0.23 (0.14 ~ 0.34), I2 = 13%, p = 0.32).
Short-term survival in patients with EGFR-mutation
A total of ten studies reported 6-month OS data of patients with EGFR-mutation, which included a total of 520 patients. The pooled and calculated 6-month OS ratio was 0.71 (95%CI: 0.62 ~ 0.78, I2 = 70%, p < 0.01) (Figure S2). According to the treatment strategy, subgroup analysis showed that patients in ICI + TKI group presented the highest OS (ICI group: 0.66 (0.57 ~ 0.74), I2 = 70%, p < 0.01; ICI + TKI group: 0.86 (0.66 ~ 0.96), I2 = 36%, p = 0.21; ICI + chemotherapy group: 0.82 (0.36 ~ 0.97), I2 = 75%, p = 0.05).
Eleven studies reported 12-month OS data of EGFR-mutation patients, which included a total of 532 patients. The pooled and calculated 12-month OS ratio was 0.51 (95%CI: 0.42 ~ 0.61, I2 = 73%, p < 0.01) (Figure S3). According to the treatment strategy, subgroup analysis showed that patients in the ICI + TKI and ICI + chemotherapy groups presented with similar OS (ICI group: 0.48 (0.39 ~ 0.58), I2 = 73%, p < 0.01; ICI + TKI group: 0.63 (0.35 ~ 0.85), I2 = 74%, p = 0.02; ICI + chemotherapy group: 0.59 (0.12 ~ 0.94), I2 = 90%, p < 0.01).
mPFS, mOS and ORR
Twelve studies reported mPFS data of EGFR-mutation patients. The pooled mPFS was 2.81 (95%CI: 1.98 ~ 3.98) months) (Figure S4). According to the treatment strategy, subgroup analysis showed that patients in ICI + chemotherapy group presented the highest mPFS (ICI group: 2.00 (1.60 ~ 2.49) months, I2 = 86%, p < 0.01; ICI + TKI group: 4.11 (2.25 ~ 7.51) months, I2 = 0%, p = 0.46; ICI + chemotherapy group: 5.87 (4.55 ~ 7.57) months, I2 = 79%, p < 0.01). Seven studies reported mOS. The pooled mOS was 13.01 (95%CI: 11.23 ~ 15.07) months) (Figure S5). According to the treatment strategy, subgroup analysis showed that patients in the ICI + TKI and ICI + chemotherapy groups showed similar OS (ICI group: 11.30 (8.43 ~ 15.14), I2 = 51%, p = 0.14; ICI + TKI group: 15.97 (10.45 ~ 24.42), no heterogeneity; ICI + chemotherapy group: 14.92 (10.39 ~ 21.42), no heterogeneity).
Meta-regression and Sensitivity Analysis
Meta-regression analysis was performed to investigate possible sources of heterogeneity among the included studies. The screened sources were: i) year of publication, ii) number of centers, iii) number of participants, iv) treatment strategies, v) article types, and vi) NOS scoring. The results indicated that the major factor which may explain the heterogeneity was the treatment strategy (P < 0.05 in 6-month OS, 6-month PFS and 12-month PFS analysis). These results are shown in Table S4. All sensitivity analyses associated with the meta-analysis performed in this study suggested stable results (Figure S6-S9).
Publication Bias
Funnel plots were generated for each particular outcome, to asymmetrically assess any publication bias. Hence, we evaluated the putative bias of the pooled rates and CIs by Begg’s and Egger’s tests. In most of the clinical outcomes, the publication bias was P > 0.05, according to both analytical tests (Table S5-S6 and Figure S10-S13).
Adverse events
In order to clearly show the safety of the treatment for NSCLC patients with EGFR mutation, SAE were pooled and illustrated by the forest plot. As Figure S14 showed that, seven studies reported SAE data of patients with EGFR-mutation. The pooled AE rate was 0.17 (95%CI: 0.06 ~ 0.39). According to the treatment strategy, subgroup analysis showed that patients in ICI + chemotherapy group presented the highest AE (ICI group: 0.08 (0.05 ~ 0.12), I2 = 0%, p = 0.84; ICI + TKI group: 0.20 (0.09 ~ 0.38), I2 = 11%, p = 0.29; ICI + chemotherapy group: 0.48 (0.15 ~ 0.83), I2 = 74%, p = 0.05).