Assessment of the clinical effectiveness and safety of immune checkpoint inhibitors for glioblastoma: A systematic review

Glioblastoma (GBM) is a malignant primary brain tumor which is commonly found in humans. Conventional therapeutic approaches for GBM offer only a short median overall survival (MOS), thereby accentuating its poor clinical outcome. Despite the promising potential shown by use of immune checkpoint inhibitor (ICI) in animal models to treat GBM, hu - man trials result remains inconclusive. This systematic review evaluated the safety and clinical efficacy of ICIs in GBM patients. A literature search from Embase and MEDLINE (Ovid) was completed in October 2023. A total of 10 suitable articles, which encompass 168 patients (164 recurrent and 4 newly-diagnosed GBM), are taken into account. 3 studies assessed the OS, 7 studies assessed the PFS and/or response assessment in neuro-oncology (RANO) criteria and 8 studies assessed the safety, tolerability and/or adverse events. These studies show that the range of MOS of ICI treatments was between 2.6-10.4 months (MOS using current standard treatment for recurrent GBM = 3.5-12.5 months). The median PFS and the median period until patients reach partial response score based on RANO criteria are ranging from 1.5 months to 4.6 months (PFS using current standard treatment is 5.5 months). In conclusion, ICIs are safe in patients with GBM. Re - ported adverse effects only include mild fatigue, headache, hyperglycemia, and diarrhea. However, ICIs display subopti - mal clinical efficacy compared to conventional GBM treatments. Therefore, further research is needed in order to im - prove the clinical efficacy of ICIs.


INTRODUCTION
Glioblastoma (GBM) is a primary brain tumor in humans which happens to be the most malignant and most commonly found.It is hypothesized that the tumor originates from pluripotent stem cells located within the central nervous system vascular niches [1,2].A multicenter analysis conducted by Dobes et al. in Australian teaching hospitals found that GBM makes up almost 30% of all primary brain tumors [3].In 2005, the European Organization for Research and Treatment of Cancer (EORTC) and the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG) recommended the first-line therapy regimen for newly diagnosed GBM: maximal safe tumor resection followed by concurrent temozolomide (TMZ) and conventional fractionated radiotherapy (RT) and subsequent adjuvant TMZ.However, due to its ability to develop its own blood supply, GBM is characterized by poor prognosis; patients only have a median overall survival (MOS) of 12-18 months following these conventional therapies and tend to succumb to relapses [1][2]4,5].Thus, in order to improve the clinical outcome of GBM patients, a novel therapeutic strategy against GBM is desperately needed.
A growing number of studies have started evaluating immunotherapy as a novel therapeutic approach against GBM.The notion of using immunotherapy in GBM treatment lies upon its ability to modulate the body's immune system to target and effectively kill cancerous cells [2].There are different types of immunotherapy, but one that might be effective against GBM is immune checkpoint inhibitor (ICI) therapy.Previous studies have shown that tumor cells can evade killing by our immune system by stimulating certain immune checkpoints located on the surface of immune cells like T cells.These activated immune checkpoints will inhibit regulatory pathways which eventually dampen immune response towards the tumor cell, leading to its survival.ICI drugs are able to block immune checkpoints such as PD-MRI1/PD-L1 and CTLA-4/CD-80/CD-86 [6].
Although administration of ICIs into animal models has proven to be safe and effective [7][8][9], such evidence is still limited in human trials [10][11][12][13].Hence, this systematic review attempts to evaluate the safety and clinical efficacy of ICIs in human patients in line with evidence-based medicine.To the best of our knowledge, this is the first analysis conducted to thoroughly evaluate the use of ICIs in treating patients with GBM.

Review design
The protocol of this review was prepared before the review is started and used as a strict guideline throughout the review.The reporting of this review was conducted in accordance to PRISMA guideline [14].

Data sources and search strategy
A literature search of two databases; Embase and MEDLINE (Ovid), was completed in October 2023.The keywords used were "glioblastoma" in conjunction with "immune checkpoint inhibitors" and did not restricted to "clinical efficacy" nor "safety" for more extensive findings.The search was limited to the studies in English, yet there were no boundaries on the year of publication.Secondary literatures were excluded.Both randomized controlled trials (RCTs) and non-randomized studies of interventions (NRSIs) were included in the study for a more extensive finding.Studies which have no available full-text report were not to be looked further.The established articles from the two databases were then be evaluated upon the title and abstract.Articles that failed to fulfil all the eligibility criteria will be excluded.Suitable study reports that fulfil the criteria were taken into deeper analysis.Supplementary Table 1 will exhibits the full search strategy.

Inclusion and exclusion criteria
Table 1 underneath epitomized the inclusion and exclusion criteria.The inclusion and exclusion criteria are made based on PICO components.However, because majority of cancer-related clinical trial phase I and II were done in single-arm approach, no comparator group will be included in these eligibility criteria.

Data extraction and data analysis
Articles were obtained from Embase and MED-LINE (Ovid) and were exported directly to a reference management software, EndNote X9.2.Data extraction was completed by one author and then reviewed by second author.If there was disagreement between authors, it was resolved by consensus.Duplicates were removed automatically, using the "Remove Duplicate" function, and also manually.Titles and abstracts were then screened to retrieve articles that fulfilled all eligibility criteria.Data to be included in the systematic review was then extracted and recorded in Microsoft Excel 2011; recorded data was: first author, year of study, study type, study's setting, population, number of subjects, intervention, outcome results and conclusion.

Quality and risk of bias assessments
Study quality assessments were done using standard ized critical appraisal tools and risk of bias tools.Quality and risk of bias assessments were completed independently by two reviewers.If there was disagreement between reviewers, all reviewers will inspect and discuss the disagreements tho rough ly.There were constant agreements in the final discus sion.RCT studies were appraised using the PEDro scale, NRSIs were assessed using the Newcastle-Ottawa scale (NOS) checklist and case reports were assessed using the JBI critical appraisal tool checklist for case reports [15][16][17].Furthermore, Cochrane risk of bias in non-randomized studies -of intervention (ROBINS-I) and risk of bias tool for randomized trials (RoB 2) were used to assess NRSIs and RCTs, respectively [18,19].

Assessment of heterogeneity
Assessment of heterogeneity was conducted using the quantitative comparison of the length of MOS and PFS; and qualitative comparison between the types and number of adverse events, between ICIs and current standard treatment for GBM.Assessment of heterogeneity is described using narrative approach.

Search findings
Initial electronic search was run on MEDLINE (Ovid) and Embase, which resulted in a total of 182 articles.Following duplicates removal, 166 unique articles were found.Based on title and abstract screening, 138 of these articles were excluded.The re maining 28 records were further assessed for eligibility.Of these, only 10 studies fulfilled the in clusion and exclusion criteria and were included in qualitative synthesis.For the full list of included and excluded studies, please refer to Supplementary Table 3.The com plete PRISMA flowchart is shown in Figure 1.

Quality and risk of bias assessments of included studies
Three critical appraisal tools were utilized to perform the methodological quality assessments.2 retrospective and 4 prospective cohort studies were critically appraised using Newcastle Ottawa Scale (NOS); 2 case reports were critically appraised using JBI critical appraisal tool for case reports; 2 RCTs were critically appraised using PEDro scale.Risk of Bias was calculated using ROBINS-I for 8 studies and RoB 2 for 2 studies.The result of quality and risk of bias assessment are fully displayed in Supplementary Table 7-11.Following thorough examination, 9 included studies have low risk of bias and 1 NRSI study has moderate risk of bias with moderate risk in 'measurement of outcome' and 'missing data' domain.months.The usual MOS survival for patients with recurrent GBM using current standard first-line therapy (surgery, radiation and chemotherapy) ranges from 3.5 to 12.5 months.Therefore, it can be deduced that the number is slightly lower for patients with ICI drugs [2,[21][22][23].

Summary of measured outcomes
Tumor Progression: Tumor progression could be reported based on PFS itself or based on the radiological criteria of Response Assessment in Neuro-Oncology (RANO) criteria.There were 7 studies that reported median PFS and/or RANO criteria.6

DISCUSSION
ICIs are proven to be effective in animal models with GBM [7][8][9], but remain inconclusive based on human clinical trials [10][11][12][13].This review aims to give a conclusive evidence on the clinical efficacy and safety of ICIs in GBM patients [10].Studies were included with 9 deemed to have low risk of bias [2,[21][22][23][24][25][26][27][28][29].This review summarized the clinical efficacy and safety of ICIs; clinical efficacy were measured with the length of OS, PFS and RANO criteria.The discussion of this review compared ICIs with current standard first-line treatment for GBM patients.Further plausible explanation of the results and impact on clinical practice were also made.
Current standard therapy for GBM includes multi disciplinary approach with maximal tumor resection, radiotherapy and adjuvant TMZ.Most patients would eventually experience tumor progression and passed away.The length of MOS of patients with recurrent GBM are reported to be 3.5 to 12.5 months [30].Combination therapy of standard treatment with the addition of bevacizumab are reported to be able to prolong the PFS, but its correlation with prolonged MOS remain to be inconclusive [30].

Clinical efficacy
Overall survival (OS) Based on the included studies, the length of MOS of patients with ICI treatment ranges from 2.6 to 10.4 months [2,[21][22][23]30].Blumenthal et al. reported the MOS of recurrent patient to be 2.6 months (0.4-11.6 months) after the regimen started [21].The result of the study shows that the MOS of ICI are comparatively the same with a slight tendency to be on the lower side.Before ICI regimen, the participants have previously had previous regimens, with the mean number of 2 regimens (1)(2)(3)(4)(5)(6).The conclusion of the study shows that most patients experienced no significant response to pembrolizumab [21].Interestingly, Cloughesy et al. found that neoadjuvant treatment of pembrolizumab before surgery had a statistically significant greater MOS compared to adjuvant treatment or treatment after surgery (HR 0.39 [95% CI= 0.17-9.94,p= 0.04, log-ranked test).The MOS of participants with adjuvant pembrolizumab were 228 days (7.6 months), meanwhile participants in neoadjuvant arm had the MOS of 417 days (13.9 months) [22].
Schalper et al. reported the MOS of patients using nivolumab to be 7.3 months (5.4-7.9 months) [2].Based on the study, an intriguing discovery was found on newly-diagnosed GBM patients.There were 2 newly-diagnosed GBM patients and they were alive after a long-term follow up of 28.5 and 33 months [2].Therefore, further investigations of nivolumab on newly-diagnosed GBM are highly recommended [2].Omuro et al. measured 3 different dosage of the combination therapy of nivolumab and ipilimumab and found out that the patient group that had the longest MOS had the dosage of 3mg/kg every 2 weeks of nivolumab with no administration of ipilimumab [23].
After analysis, it can be concluded that the MOS of ICI are not satisfying.All of the included studies reported a slightly lower MOS than the current standard treatment.The results among studies were deemed to be homogenous; that is lower than the standard treatment.
Byron et al. examined the PFS of nivolumab and found out to be 195 days (using radiological screening) and 100 days (without radiological screening) [24].Another study that compared the neoadjuvant and adjuvant treatment of pembrolizumab shows that neoadjuvant pembrolizumab has longer PFS of 99.5 days compared to adjuvant with 72.5 days [22].Omuro et al. reported that combination therapy of nivolumab and ipilimumab has the median PFS of between 1.5 and 2.1 months [23].Furthermore, Blumenthal et al. reported the median PFS of patients using nivolumab is 4.1 months (2.8-5.5 months) [2].
Moreover, Qin et al. reported the median PFS of nivolumab + pembroliumab is 4.6 months.Qin et al. divided the participants into two arms; first group was deemed to have potential beneficial biomarkers through MRI lived up to 6.5 months, while those who did not have potential beneficial biomarfers lived up to 2.7 months [28].However, the study done by Qin was observed to have moderate risk of bias due to bias in missing data and measurement of outcome.
A case series done by Ranjan et al. reported 4 recurrent GBM patients that were treated with ICI [27].One patient with tumor located at left frontal region was treated with combination therapy of nivolumab and TMZ.Patients had no tumor progression until 2 months, before the tumor started to regrow on her left frontal lobe near the left ventricle.Clinical deteriorations started to occur at 3.5 months but can be managed with the administration of dexamethasone.The tumor remains stable on the twelfth month [27].Second patient had GBM in right temporal region; he was treated with nivoumab + TMZ and had PFS of 8 weeks; before a new tumor started to progress in the right sylvian fissure.At 10 months, progression started to be seen through radiological imaging, but it turns out to be immunotherapy-related pseudo progression [27].Third patient had the tumor at the right temporal region and was treated with ipilimumab + TMZ.The patient had disease progression at 8.5 months and remains stable up until 19 months after the initiation of the regimen [17].Fourth patient had GBM located in the left temporal lobe.He was given a combination therapy of ipilimumab + TMZ.Tumor did not progress up until the ninth month; clinically manifested with focal seizures.Maximal tumor resection then conducted 2 weeks later with adjuvant therapy of low-dose bevacizumab and nivolumab.Patients remain stable 21 months after the diagnosis of recurrent GBM [27].
Carter et al. assessed recurrent GBM patients that underwent combination therapy of ipilimumab + bevacizumab with RANO criteria.The results show us that, after 12 weeks, 31% of participants were stable, 31% had partial progression and 38% had disease progression [25].
Lukas et al. examined the length of progression free in patients that underwent the therapy of IV atezolizumab 1200mg every 3 weeks.The study found out that the PFS is 1.2 months (0.7-10.7 months) [26].
All of the included studies reported the PFS of ICI medications were lower than the PFS of current standard treatment (5.5 months).Therefore, it can be concluded that the progression free survival of ICIs were lower than current standard treatment; and the results of studies were deemed to be homogenous.

Plausible factors that lead to suboptimal efficacy
We found that vast majority of studies exploring the effect of ICI on GBM patients showed suboptimal clinical efficacy.It is suggested that ICI are not yet applicable to be used in patients with GBM.There are numbers of plausible factors that may explain this phenomenon.One of the reasons includes the poor ability of ICI to penetrate the blood-brain barrier (BBB), thus affecting drug delivery.Aside from the BBB, it is known that GBM may induce an immunosuppressive environment of the brain, which is indicated by low number of tumor-infiltrating lymphocytes (TILs) and low neoantigen burden.Hence, new strategies of treatment that may increase immune and antitumor response in the tumor microenvironment are needed.Some examples include molecular screening prior to ICI treatment and combination therapy with other types of treatment.

Safety
A single study declared that the ipilimumab and bevacizumab combination was well tolerated [25], with fatigue (40%) and diarrhea (30%) as the most common adverse effect.An intracerebral hemorrhage was reported in one patient and two patients had pulmonary emboli that were associated with the disease.ICIs were discontinued due to presence of grade 2 rash in one patient and grade 2 arthritis in another single patient.6 patients experienced diarrhea which was well controlled using corticosteroid [21][22][23][25][26][27]29].
One study reported muscle weakness, headache, and hyperglycemia as the most common treatment-related adverse events (TRAEs) on pembrolizumab [22].However, the study stated that pembrolizumab was predominantly well tolerated.10 out of 16 patients in the neoadjuvant arm experienced grade 3-4 adverse events that may or may not be attributable to the treatment, with treatment discontinued in two patients due to grade 3 pneumocystis and grade 4 elevations in alanine transferase.Another study on nivolumab also suggested it to be safe and well tolerated, with a low incidence of TRAEs.In addition, out of 3 newly diagnosed GBM patients treated with nivolumab, 2 patients survive for 28 and 33 months [2].
In comparison to nivolumab + ipilimumab, nivolumab alone is better tolerated, as its tolerability is impacted by the ipilimumab dose.With diarrhea and fatigue as the most usual adverse reactions; TRAEs shown in 10% of patients with nivolumab treatment, while 23% in patients treated with nivolumab and ipilimumab combination [23].
Another research with recurrent GBM patients stated that atezolizumab was safe and well tolerated.Treatment-related events occurred in 10 patients (63%) but without any grade 4-5 TRAEs.Each of the patients that died in this study was due to GBM, irrelevant to drug-related events [26].
All of the studies reported similar findings that most ICIs are well tolerated and have milder adverse events in comparison to standard treatment.

Limitations of the study and future recommendations
A number of limitations are attributed to this systematic review.Firstly, there is a limited amount of studies regarding GBM patients with ICIs treatment.Moreover, the majority of published studies are phase I and II clinical trials, which only have a relatively small number of participants with no comparison arm.Thirdly, most of patients recruited in the studies have recurrent GBM (164/168), which may yield different outcomes compared to patients with newly diagnosed GBM.Fourthly, meta-analysis was not done due to reasons.Consequently, the conclusion from this analysis cannot be overstated and on further investigation, the conclusion from this analysis could be substantially altered.In order to completely understand the clinical efficacy and safety of ICI drugs in the medical field, Author highly suggests having additional research with regards to the clinical efficacy and safety of ICI drugs in patients with GBM.

CONCLUSION
To conclude, this review presents that ICIs can be safely applied in GBM patients, with fatigue and diarrhea as the most frequent adverse events.However, ICI drugs have a moderately lower clinical efficacy in comparison to the current standard treatment, marked by lower MOS and PFS in all included studies.
Therefore, ICI drugs can be an alternative treatment for GBM patients who display severe adverse reactions to conventional therapeutic approaches (chemotherapy, radiation, and surgery).
Future research should also assess the influence on newly diagnosed GBM patients, since the majority of the participants in these studies are recurrent GBM patients.Following administration of ICIs, patients with newly diagnosed GBM may exhibit different outcomes, compared to those with recurrent GBM.Moreover, further research is needed in order to improve the clinical efficacy of ICIs.

Conflict of interest: none declared
Financial support: The authors receive no financial support for this review from any funding agencies.

FIGURE 1 .
FIGURE 1. PRISMA flowchart.182 records were obtained from Embase and MEDLINE (Ovid).10 studies were included in the qualitative synthesis Overall Survival: There are four studies that assessed the OS of patients treated with ICIs.Blumenthal et al. and Cloughesy et al. studied pembrolizumab, Schalper et al. studied nivolumab, while Omuro et al. studied nivolumab and its combination with ipilimumab [2,21-23].The MOS from these studies range from as low as 2.6 months to as high as 10.4

TABLE 1 .
Full search strategy on Embase and MEDLINE (Ovid) -Risk of bias across studies15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies).-Additionalanalyses16Describemethods of additional analyses (e.g., sensitivity or subgroup analyses, metaregression), if done, indicating which were pre-specified.-RESULTSStudyselection17Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram.

TABLE 3 .
List of included and excluded studies with full-text available *N/R, not reported; AE, adverse events; GBM, glioblastoma; OS, overall survival; PFS, progression free survival; RANO, response assessment in neuro-oncology; TRAE, treatment-related adverse events SUPPLIMENTARY

TABLE 7 .
Methodological quality assessment of the included cohort studies using Newcastle Ottawa Scale

TABLE 8 .
Methodological quality assessment of the included RCTs using PEDro scale

TABLE 9 .
Methodological quality assessment of the included case reports using JBI

TABLE 10 .
Results of risk of bias assessment for non-randomized studies of interventions (ROBINS-I)