Literature meta-analysis about the efficacy of anti-programmed death protein 1 and anti-programmed death ligand 1 re-challenge in cancer patients

Background Immune checkpoint inhibitor (ICPis) re-challenge could be an attractive therapeutic option considering its good safety profile. However, little data is available regarding anti-PD-1/anti-PD-L1 retreatment. We conducted a meta-analysis focusing on outcomes of solid cancer patients performing this strategy. Methods Fourteen full papers involving 74 patients were included. Individual data about best response or progression-free survival (PFS) upon the first and second course of anti-PD-1/ anti-PD-L1 were collected. Results Non-small-cell lung and represented Higher objective response (46% versus and disease control rates versus obtained compared to No association between responses obtained upon who if

2 Abstract Background Immune checkpoint inhibitor (ICPis) re-challenge could be an attractive therapeutic option considering its good safety profile. However, little data is available regarding anti-PD-1/anti-PD-L1 retreatment.
We conducted a meta-analysis focusing on outcomes of solid cancer patients performing this strategy.

Methods
Fourteen full papers involving 74 patients were included. Individual data about best response or progression-free survival (PFS) upon the first and second course of anti-PD-1/ anti-PD-L1 were collected.

Conclusion
Anti-PD-1/anti-PD-L1 re-challenge showed interesting clinical activity in selected patients, mainly in those achieving a long-term response upon the first ICPi course, that do not discontinue therapy because of disease progression, or that are able to keep a treatment-free period.

Background
Immune checkpoint inhibitors (ICPis) have completely changed the treatment algorithm of several cancer types because of the impressive results obtained in this field. Some patients achieve a long-3 term clinical benefit from this type of drugs. However, patients eventually discontinue ICPi due to disease progression or toxicities, as well as due to trial designs imposing discontinuation after a given treatment period. Primarily for those patients who achieve a long-term response without clinicallymeaningful toxicities, ICPi re-challenge could represent an attractive therapeutic option, as the only possible alternative to chemotherapy. This strategy has already entered clinical practice in advanced melanoma patients who are permitted to being re-challenged with an anti-cytotoxic T-lymphocyte antigen 4 (anti-CTLA-4), as well as to being shifted across anti-programmed cell death 1 (anti-PD-1)/anti-programmed cell death ligand 1 (anti-PD-L1) and anti-CTLA-4. This strategy has been supported by several studies that specifically addressed this issue in melanoma patients, reporting promising results concerning ICPi re-challenge efficacy, along with a low toxicity profile (1)(2)(3)(4)(5)(6)(7)(8) However, most of these trials included selected patients exhibiting significant clinical benefits, without severe undesirable events upon the first ICPi course. Consequently, in spite of no widelyrecognized recommendations being available concerning this strategy, the National Comprehensive Cancer Network (NCCN) guidelines allow for considering ICPi re-induction in progressing patients with an initial response or disease stabilization lasting ≥ 3 months (9). Much less is known about the efficacy and safety of re-challenge with anti-PD-1/anti-PD-L1 agents. The CheckMate-153 trial explored the clinical benefits of a fixed-duration nivolumab in second-line versus continuous treatment in advanced non-small-cell lung cancer (NSCLC) patients. In this trial, 39 patients in the fixed-duration arm progressed during the surveillance period and were retreated with the same anti-PD-1. The median time between progression and nivolumab reinduction was 0.6 months, and the median duration of retreatment 3.8 months (10). Moreover, the Keynote-010 trial selected PD-L1positive (≥ 1%), pretreated, advanced NSCLC patients to receive two different schedules of pembrolizumab or docetaxel. In this trial, 14 patients received a second course of pembrolizumab, with the majority of them (78%) showing either partial response or stable disease (11,12). Recently, the UNIVOC study retrospectively analysed more than 1500 NSCLC patients that received an ICPi retreatment after a discontinuation period from nivolumab of at least 6 weeks. The median overall survival (OS) from re-challenge was 15 months for patients receiving a second course of PD-1 4 inhibitor after a treatment-free interval and 18.4 months for those who performed an intercalated chemotherapy. Interestingly, median OS was significantly longer in patients with an initial nivolumab treatment duration longer than 3 months (13). However, no information were reported about the nivolumab discontinuation reason, patients 'clinical features such as the ECOG Performance Status (PS) and the retreatment tolerance. Consequently, it was not possible to know how selecting patients to these strategy according to clinical characteristics. Lastly, two large studies that investigated the safety of resuming anti-PD-1 agents in cancer patients were recently reported (14,15). According to these studies, 50-55% of patients experienced an immune-related adverse event (irAE) of any grade upon anti-PD1 resumption. Despite these encouraging data, the usefulness of anti-PD-1 and anti-PD-L1 re-challenge is still being debated because of the lack of prospective studies that specifically address this topic. While attempting to further investigate this issue, we have reviewed the literature and selected case reports and cases series that reported outcomes of adult solid cancer patients rechallenged with an anti-PD-1 or anti-PD-L1 agent during their disease history. Individual patient data were meta-analyzed regarding efficacy outcomes to possibly identify potential clinical features associated with greater clinical benefit.

Methods
PubMed and Google Scholar were searched for clinical trials, case series, and case reports containing data on solid cancer patients who were retreated with an anti-PD-1/anti-PD-L1. Key words used included "rechallenge", "reintroduction", "retreatment", "immune checkpoint inhibitors", "immune checkpoint blockades", and "immunotherapy". We considered only English-written articles reporting information on best response or progression-free survival (PFS) achieved upon either a first or second ICPi course. We did not find any prospective trial published on this issue and we identified 14 full papers (five case reports and nine case series) reporting the outcomes of 74 patients treated according to this strategy ( Fig. 1, Table 1) (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29). Data search ended the 25th January 2020. We then meta-analyzed the individual data of these patients.

Patient's general features
The patients' clinical characteristics are described in Table 2. The mean age was 63 years , and most of the patients were male (n=45, 61%). NSCLC and melanoma were the most common tumors (n=39, 53% and n=25, 34%, respectively), while 10 patients (13%) had other cancer types.
Notably, three patients exhibited clear cell renal cell carcinomas, two urothelial carcinomas, two colorectal cancers, one breast cancer, one Merkel carcinoma, and one head and neck carcinoma.
Patients received a median of two systemic treatments before the first ICPi, which was discontinued in most of cases because of either disease progression (n=56, 76%), toxicity (n=9, 12%), or per protocol for patients included in clinical trials (n=9, 12%). Approximately 72% of patients received at least 1 systemic agent with or without a local treatment, such as radiotherapy or surgery, between the two ICPis courses (Table S1).

Progression-free survival under re-challenge
Paired PFS1 and PFSR were evaluable in 61 patients. As expected, the PFS1 was longer than the PFSR  (Fig. 4). Shorter cut-offs were also tested (3, 6, and 9 months), whereas no statistically significant differences were obtained. As expected, a longer PFSR was achieved in patients who discontinued the first anti-PD-1/anti-PD-L1 agent due to toxicity or per protocol as compared to those who experienced disease progression ( (Fig. 4). Similar results were obtained, without considering the eight patients receiving ipilimumab (anti-CTLA-4) between the two ICPis (data not shown).

Re-challenge safety profile
Concerning toxicities, a higher percentage of all grade adverse events occurred during the first anti-PD-1/anti-PD-L1 course, as compared to re-challenge (n = 58, 78% versus n = 43, 63%, p = 8.10 -3 ) (Table 3). Rash and diarrhea were the most commonly reported adverse events.  representing an attractive and certainly less toxic alternative to chemotherapy. In our analysis, no increased toxicities were registered upon the ICPi re-challenge, in line with previous studies (1,(5)(6)(7)(8)38). Nevertheless, no information pertaining to AE grades was available in most papers considered for this study.
The other primary issue we wished to investigate was as follows: which patients should be candidates for this strategy. Melanoma patients who achieve an objective response lasting ≥ 3 months with ICPi without experiencing Grade III or IV AEs can be considered for re-challenge with anti-CTLA-4 or a sequential treatment with another ICPi. Here, we consistently found a better PFS upon re-challenge for patients who achieved a longer PFS during the first ICPi course, especially if ≥ 12 months.
Moreover, patients who did not receive any treatment between the two ICPis courses and those who discontinued the first course without experiencing disease progression seem to be the best candidates. Conversely, the best response obtained in regard to the first ICPi was unable to predict the response upon re-challenge. These results are not surprising, given that patients with these characteristics are likely to have a more favorable prognosis with a more indolent disease. In addition, we cannot exclude the possibility that re-challenge did not specifically change their disease history, which would have been favorable even with another type of treatment. Nevertheless, the re-challenge strategy resulted in interesting clinical benefits in these selected patients.