Abstract

What Is Known? and Objective. Immune checkpoint inhibitors (ICIs) play an important role in various cancers. The efficacy and safety of rechallenge with ICIs after immune-related adverse events (irAEs) were not well known. Accumulating studies report inconsistent findings. Thus, we conducted an updated meta-analysis by including more studies. Methods. We searched PubMed, Web of Science, Embase, and Cochrane Library for studies reporting the rechallenge of ICIs after irAEs. The evaluation outcomes included the incidence of irAEs, objective response rate (ORR), and disease control rate (DCR). Results and Discussion. A total of 896 ICI rechallenge cases from 24 studies were included. Compared to the initial treatment with ICIs, rechallenge showed a higher incidence of all-grade irAEs (OR, 2.78; 95% CI, 1.51–5.10; ) and high-grade irAEs (OR, 1.88; 95% CI, 1.27–2.78; ), but ORR (OR, 1.01; 95% CI, 0.55–1.84; ) and DCR (OR, 1.21; 95% CI, 0.68–2.15; ) were not further improved after the rechallenge of ICIs. What Is New? and Conclusion. More studies are included in this paper to compare and analyze the efficacy and safety of ICIs after rechallenge, so as to update the previous meta-analyses, and finally get different conclusions from the previous meta-analyses in terms of safety. Our results suggest that rechallenged ICIs after irAEs showed similar efficacy and lower safety than initial ICIs. However, these results need to be further verified by high-quality studies with large samples. In addition, we added subgroup analysis not available in previous meta-analyses to explore the association of cancer type, age, and gender factors with the incidence of irAE after ICI rechallenge.

1. Introduction

In recent years, immune checkpoint inhibitors (ICIs) directed targeting programmed cell death protein-1 (PD-1)/ programmed cell death protein ligand-1 (PD-L1) and cytotoxic T-lymphocyte antigen-4 (CTLA-4) become emerging therapies to improve the survival of patients with malignancies [1, 2]. Currently, ICIs have shown unprecedented clinical efficacy in a variety of tumors such as melanoma, lymphoma, lung cancer, uroepithelial carcinoma, and gastrointestinal tract tumors [37]. However, the treatment of ICIs is a “double-edged sword,” as it may cause excessive enhancement of immune response or immune imbalance, resulting in immune-related adverse events (irAEs) in all major systems of the body, which can theoretically occur in any tissue and organ. Common irAEs are fatigue, pruritus, diarrhea, and rash [8], with the incidence of fatigue ranging from 16% to 20% and the probability of pruritus and rash being approximately 10.6% and 9.3%. Most of these adverse reactions are mild and manageable and resolve on their own after discontinuation without special treatment, but there are still some serious and even fatal toxic reactions, such as immune-related pneumonia, immune-related neurotoxicity, and fatal diarrhea, which can lead to treatment discontinuation. irAE occurrence is varied due to different ICIs. The incidence of grade 3 or 4 irAEs associated with anti-PD-1/PD-L1 and anti-CTLA-4 monotherapy is approximately 14% [9] and 23% [10], respectively; yet, severe irAEs after combination therapy are up to 53% [11].

When irAEs are fully recovered, the question of whether to rechallenge is crucial. Practical guidelines of irAE management are based on clinical observations and expert consensus, and the possibility of rechallenge is not discussed. When ICIs are discontinued due to irAEs, once irAEs are fully recovered, restarting ICIs may control tumor growth; on the other hand, this may also increase the risk of the same or different irAEs [12, 13]. Several recent studies have shown that the rechallenge of ICIs is safe, effective, and reasonable [1420]. In the study by Santini et al., patients with no observed partial responses prior to irAEs had PFS and OS that lasted longer after rechallenge with ICIs and may benefit [18]. According to Plazy et al., irAEs at resumption were severe (grade 3-4) in 18%– 62% of cases, which were really similar or even less severe than what was reported during the first ICI, suggesting that toxicity during ICI resumption seems manageable once irAEs have recovered [21]. However, other studies suggest that the incidence of irAEs after rechallenged ICIs is even higher [22, 23]. For example, in the study by Pollack et al., patients who discontinued CTLA-4/PD-1 blockade for severe irAEs had relatively high rates of recurrent or distinct toxicities with anti-PD-1 resumption [22].

Recently, several meta-analyses about the safety of ICI rechallenge have been reported, but the results are not completely consistent. Several new cohort studies are exploring the safety of ICI rechallenge [24]. Therefore, we further conducted an updated meta-analysis to clarify the safety and efficacy of rechallenged ICIs and provide an objective reference for the clinical reuse of ICIs.

2. Materials and Methods

2.1. Literature Search Strategy

PubMed, Embase, Web of Science, and Cochrane Library were searched to retrieve relevant studies published from the database inception to January 20, 2022. The key retrieve terms in the search strategy included immune checkpoint inhibitors (anti-PD-1, anti-PD-L1, and anti-CTLA-4), specific ICI names (nivolumab, pembrolizumab, ipilimumab, atezolizumab, avelumab, durvalumab, and cemiplimab), and some expressions related to “rechallenge” (retreat, readministrate, restart, reinitiate, resume, and reinduce).

2.2. Inclusion and Exclusion Criteria

Studies were considered eligible if all of the following criteria were included: (1) the study subjects were cancer patients with a clear diagnosis, (2) patients were treated with ICIs, and (3) the same or a different type of ICIs were resumed after a previous interruption for irAEs. Exclusion criteria were as follows: (1) studies that did not meet the inclusion criteria, (2) review and meta-analyses, case reports, editorials, and letters to the editors, (3) the number of cases was less than 10, and (4) duplicate studies and incomplete or unusable original study data. Two researchers independently screened the title and abstract of each search retrieved to identify all studies that might meet the inclusion criteria and then read the full text of all potentially eligible studies for further screening. The two researchers solved any discrepancies in study selection through discussion and in-depth reading of the search, consulting a noninterested third party for judgment when necessary.

2.3. Data Collection and Quality Assessment

All data were collected independently by two noninterfering investigators following a predetermined procedure, and extracted data included (1) literature-related information: author, publication year, study design, cancer type, total study population, type of initial and rechallenged ICIs, rechallenge ratios, the type and number of occurrence of initial and rechallenged irAEs, and the number of occurrence of initial and rechallenged low-grade irAEs (grade 1-2) and high-grade irAEs (≥grade 3), respectively, and (2) study event indicators: objective response rate (ORR) and disease control rate (DCR). The same two independent researchers assessed the methodological quality of all included studies using the Newcastle-Ottawa Scale (NOS) criteria [25].

2.4. Outcomes

Safety assessment included the incidence of all-grade rechallenged irAEs and the incidence of high-grade rechallenged irAEs. According to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE) version 5, the severity of irAEs was classified as grades 1–5, with grade ≥3 being high-grade irAEs and grade 1 or 2 being low-grade irAEs. Efficacy assessment included ORR and DCR after ICI rechallenge. ORR was defined as the rate of patients who had a complete response or partial response, while DCR was defined as the rate of patients who had a complete response, partial response, or stable disease.

2.5. Statistical Analysis

We employed Review Manager 5.4 (Cochrane Community, London, UK) for statistical analyses and plotting. Synthesis of all-grade and high-grade rechallenged irAEs, ORR, and DCR was conducted via a meta-analysis using pooled odds ratios (ORs), with 95% confidence intervals (CIs) calculated via the Mantel–Haenszel model. Since the vast majority of the studies included in this meta-analysis were retrospective studies, considering significant heterogeneity, we used the random-effects model with the Mantel–Haenszel model and then validated it by the Q-test and I-squared (I2) test. When or I2 < 50%, no heterogeneity or mild heterogeneity among studies could be considered, and when or I2 ≥ 50%, statistical heterogeneity among studies was suggested. The Z test was used to determine whether the combined effect sizes were statistically different, and when , it was suggested that the combined effect sizes of multiple studies were statistically significant; conversely, when , it was suggested that they were not statistically significant. In addition, we performed subgroup analyses using accessible data (here mainly for cancer types) and sensitivity analyses by sequentially omitting one study to determine the stability of the combined results.

3. Results

3.1. Eligible Studies and Characteristics

We retrieved a total of 403 relevant articles, 18 duplicates were removed, and 292 were excluded after the title and abstract review for not meeting the selection criteria. 93 were left for full-text reading, and 69 were removed after reading the full text; finally, 24 cohort studies [14, 18, 22, 23, 2644] containing 896 patients were included for qualitative and meta-analysis. The literature screening process is shown in Figure 1. Characteristics of the included studies are reported in Table 1.

3.2. Safety Analysis

Sixteen studies [18, 2630, 33, 34, 3643] were included in the safety analysis, fifteen [18, 2630, 33, 34, 36, 3843] of which were used for the combined analysis of differences in the incidence of all-grade irAEs, and eleven [18, 2729, 33, 36, 37, 3941, 43] were used for the analysis of differences in the incidence of high-grade irAEs. The results showed ICI rechallenge was associated with a significantly higher incidence of all-grade irAEs than initial ICIs (OR, 2.78; 95% CI, 1.51–5.10; ; I2 = 86%) (see Figure 2(a)), as well as a significantly higher risk of high-grade irAEs (OR, 1.88; 95% CI, 1.27–2.78; ; I2 = 23%) (see Figure 2(b)).

3.3. Efficacy Analysis

Seven studies [18, 23, 27, 28, 36, 37, 40] were included for the pooled analysis of ORR with 529 patients. Four [23, 28, 36, 37] of them were included for the pooled analysis of ORR with 332 patients. The results illustrate that neither ORR (OR, 1.01; 95% CI, 0.55–1.84; ; I2 = 60%) (see Figure 3(a)) nor DCR (OR, 1.21; 95% CI, 0.68–2.15; ; I2 = 20%) (see Figure 3(b)) was further improved in the ICI rechallenge population.

3.4. Subgroup Analysis

The results of subgroup analyses for cancer types are displayed in Figures 4(a) and 4(b). Cohorts enrolled with melanoma showed no significant difference in the incidence of all-grade irAEs and high-grade irAEs in ICI rechallenge. Patients with NSCLC had a higher incidence of all-grade rechallenged irAEs (OR, 3.22; 95% CI, 1.13–9.14; ; I2 = 78%) (see Figure 4(a)), but no significant difference existed in high-grade irAEs (OR, 1.29; 95% CI, 0.38–4.45; ; I2 = 47%) (see Figure 4(b)).

Further subgroup analysis was performed for age and gender stratification. The results showed that patients with a median age <65 years had a higher incidence of high-grade rechallenged irAEs (OR, 2.08; 95% CI, 1.22–3.54; ; I2 = 30%), whereas no significant difference existed in patients with a median age ≥65 years (OR, 1.78; 95% CI, 0.84–3.76; ; I2 = 25%) (see Figure 5(a)). On the other hand, in the predominantly male study population (male ≥50%), patients had a higher incidence of high-grade rechallenged irAEs (OR, 2.38; 95% CI, 1.45–3.91; ; I2 = 20%), whereas no significant difference existed in the predominantly female study population (female ≥50%) (OR, 1.34; 95% CI, 0.72–2.50; ; I2 = 0%) (see Figure 5(b)).

3.5. Sensitivity Analysis

In the sensitivity analysis, the pooled results for all-grade irAEs, high-grade irAEs, ORR, and DCR remained stable, regardless of which study was removed, indicating a robust correlation of the combined results.

4. Discussion

To the best of our knowledge, this updated meta-analysis is the latest and relatively comprehensive study on the safety and efficacy of ICI rechallenge in people who have discontinued ICI treatment due to irAEs previously. Our findings suggest that, in terms of safety, ICI rechallenge patients had overall higher risks of all-grade and high-grade irAEs than initial ICIs. In the previous meta-analysis, Zhao et al. [46] found a higher incidence of all-grade irAEs after ICI rechallenge, but there was no significant difference in the incidence of high-grade irAEs. This differs from our results. The discrepancy may be due to relatively large cohort studies enrolled in our safety analysis (n = 11 vs. n = 6). Inno et al. [47] conducted a meta-analysis of ICI rechallenge that showed a similar incidence of all-grade and high-grade irAEs between initial and rechallenged ICIs, which included all patients who discontinued ICIs due to disease progression and toxicity. Our results differ from those of that study because we focused on the population who discontinued ICIs due to irAEs. In terms of efficacy, there was no significant difference in ORR and DCR between initial and rechallenged ICIs, which is consistent with the study published by Zhao et al. [46]. Given the limited pooled sample size, large prospective clinical trials are required to confirm the effect of ICI rechallenge treatment in various cancer types.

Relatively large heterogeneity was presented in the analysis of all-grade irAE incidence. To explore the source of heterogeneity, we further performed subgroup analysis according to cancer types. The results showed no significant difference in the incidence of all-grade and high-grade irAEs in ICI rechallenge based on melanoma. We found patients with NSCLC had a higher incidence of all-grade irAEs, but no significant difference existed in high-grade irAEs after ICI rechallenge. In each subgroup, heterogeneity remained, indicating that the cancer type was not the source of heterogeneity. Other potential sources of this heterogeneity may be due to different intervals from initial irAEs to rechallenged ICIs and dosing regimens among studies. Unfortunately, the available data for these variables were incomplete and could not be used here. A prior study on irAEs showed the recurrence of high-grade irAEs after ICI reintroduction is associated with a longer interval between initial irAEs and rechallenged ICIs and the initial use of anti-CTLA-4 antibodies [20, 46].

In addition, a recent review indicated that the higher risk of irAEs in the initial ICI treatment population is associated with age <60 years, high body mass index, women on CTLA-4 and men on PD-1/PD-L1 agents, and chronic smokers [48]. To explore whether these factors have the same impact on rechallenged irAEs, we also stratified age and gender into subgroup analysis based on currently available data. The results prompt that the predominantly male study population (male ≥50%) with a median age <65 years appears to be associated with a higher risk of high-grade rechallenged irAEs, while in the predominantly female study population (female ≥50%) with a median age ≥65 years, the incidence of high-grade irAEs is similar between the initial and rechallenged ICI groups. However, this conclusion has certain limitations due to the limited number of the included articles.

Our meta-analysis has several limitations: First, the number of prospective studies is small, which may have had some impact on the quality of the research results. Second, not all studies reported OS or PFS, so in our meta-analysis, we evaluated the efficacy of ICI rechallenge in terms of ORR and DCR. Third, the source of heterogeneity among the studies could not be addressed. We performed subgroup analysis to decrease heterogeneity; however, the overall heterogeneity was not changed after subgroup analysis. Therefore, these results should be interpreted with caution, and additional well-designed studies are still needed to assess the safety and efficacy of ICI rechallenge in follow-up.

5. Conclusion

More studies are included in this paper to compare and analyze the efficacy and safety of ICIs after rechallenge, so as to update the previous meta-analysis, and finally get different conclusions from the previous meta-analyses in terms of safety. Our study suggested that patients who rechallenged ICIs after irAEs were associated with similar efficacy and lower safety. Therefore, based on the conclusion of our current study, for cancer patients who have previously discontinued ICIs for irAEs, the option of rechallenged ICIs should be carefully weighed against benefits and risks after taking into account the patient’s disease status, the response obtained, and the type and grade of adverse events previously experienced. In addition, more large-scale prospective studies are needed to validate our conclusion and further elucidate the role of ICIs after disease progression or toxicity.

Data Availability

The data supporting this meta-analysis are from previously reported studies, which have been cited. All data generated or analyzed are included in this article. Further information can be available from the corresponding author upon reasonable request.

Disclosure

Jiaqin Cai and Wenhua Wu are listed as joint first authors.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Authors’ Contributions

JC and XW participated in the design of the study. JC and WW retrieved and selected the articles, analyzed and interpreted the data, and wrote the manuscript. WW and XW solved all disagreements and revised the manuscript. JZ, GZ, and HS supervised the study. All the authors contributed to the article and approved the final manuscript. Jiaqin Cai and Wenhua Wu contributed equally to this work.

Acknowledgments

The authors thank the databases of PubMed, Embase, Web of Science, and Cochrane Library for providing a source of data for the study. This study was supported in part by grants from the Natural Science Foundation of Fujian (no. 2022J011004), the Fujian Provincial Health Technology Project (no. 2022CXB001), the Natural Science Foundation of Fujian, China (no. 2021J01397), and the Fujian Provincial Health Technology Project (no. 2022GGA010).