Immune checkpoint inhibitors for unresectable malignant pleural mesothelioma

ABSTRACT

Unresectable malignant pleural mesothelioma (MPM) is an aggressive disease with a 5-year survival rate of approximately 10%. Recent data suggest that MPM is an immunologically active tumor, in which checkpoint inhibition through the blockade of the anti-cytotoxic T lymphocyte antigen-4 (-CTLA-4) or anti-programmed cell death 1 (PD-1) could play a major therapeutic role. Initially, clinical trials evaluated immune checkpoint inhibitors (ICIs) in the salvage setting after platinum-based chemotherapy with mixed results in terms of efficacy. More recently, the combination of the anti-CTLA-4 agent ipilimumab plus the anti-PD-1 agent nivolumab was tested in the front-line setting, and reported a superior survival as compared to platinum/pemetrexed. While other clinical trials ore ongoing in order to investigate ICIs for MPM, it seems now evident that we have entered a new “era” for the treatment of MPM. In the future, a few issues need to be solved with regard to the use of ICIs for MPM. Among them, there is the identification of biomarkers of sensitivity to immunotherapy that may help enrich the patient population who could benefit the most from treatment, while avoiding for some other patients the potential occurrence of immune-related side effects from therapies that are anticipated to be ineffective.

KEYWORDS:

• Immune checkpoint inhibitors
• immunotherapy
• malignant pleural mesothelioma
• PD-(L)1

1. Introduction

Malignant pleural mesothelioma (MPM) is a rare form of cancer that affects the pleural surface. Environmental exposure to asbestos, mostly occurring in the workplace, accounts for the majority of MPMs, with a typical latency of 20–50 years existing between fiber exposure and manifestation of disease.¹

Regardless of the site of origin (pleura, pericardium or peritoneum) malignant mesothelioma is considered to be a rare cancer, whose incidence and prevalence largely vary based on the geographical region, and reflect the different occupational exposure to asbestos in a given country as well as the control measures that were taken or not in order to limit this exposure.² In the European Union, where the incidence has been steadily increasing in the last twenty years with an expected peak around 2020–2025, the average incidence and prevalence reported by RARECARE in the period 2000–2007 are 21 and 36 per million inhabitants/year, respectively.^2,3 By contrast, better data are reported for USA by the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute (NCI), which suggests a decrease in the incidence of mesothelioma since year 2014, with an average incidence for the period 1975–2017 of 9.6 per million inhabitants/year.³

MPM is a cancer with an aggressive phenotype, whose prognosis has been only marginally improved by treatment.⁴ A few studies have identified how survival is largely affected by the histologic variant: epithelioid histology, which is the most common variant, is associated with the best survival, while non-epithelioid forms (i.e. sarcomatoid and mixed) carry a worse prognosis.^5,6 In addition to the histologic variant, other factors that have been associated with a better prognosis are female sex, younger age, early stage, and multimodal-based therapy including surgery and chemotherapy.^6,7 Nevertheless, little therapeutic progress has been made in the treatment of MPM in the last decades, and the 5-year survival rate of affected patients is only 10%.⁸

When it comes to unresectable MPM, chemotherapy with cisplatin plus the third generation anti-folate agent pemetrexed has represented a standard first-line treatment since 2003, a year in which the seminal work by Vogelzang et al. showed that this combination regimen was able to significantly improve survival by 2.8 months (from 9.3 months to 12.1 months) as compared with cisplatin alone.⁹ Since then, a series of phase 2 studies suggested that carboplatin could be a fine substitute for cisplatin, especially in unfit patients because of factors such as performance status ≥2, older age, and comorbidities.^10–12 More recently, the MAPS study reported that for patients with no contraindications to an anti-angiogenic therapy (e.g. substantial cardiovascular comorbidity, uncontrolled hypertension, bleeding or clotting risk), the addition of the anti-VEGF agent bevacizumab to cisplatin/pemetrexed significantly extends survival versus cisplatin/pemetrexed alone (18.8 months versus 16.1 months, respectively, HR = 0.77, P =.016).^13,14

In view of the above, it is evident that, despite active treatment, most patients with unresectable MPM will eventually succumb to the disease, which implies that novel therapeutic strategies are urgently needed. For these patients, unleashing the immune response against cancer through the use of monoclonal antibodies named immune checkpoint inhibitors (ICIs) could represent a winning approach. The mechanism of action of ICIs can be schematically exemplified as follows: on one hand, ICIs may act at a central level by blocking the interaction between the anti-cytotoxic T lymphocyte antigen-4 (CTLA-4) on T cells and B7 on antigen presenting cells, on the other, they may act peripherally by preventing the interaction between the programmed cell death 1 (PD-1) on activated T cells and programmed cell death ligand-1 (PD-L1) on tumor cells.

In the present review, we will briefly discuss the biology of mesothelioma with focus on the rationale behind the use of ICIs as a therapeutic strategy. In the second part, we will summarize the results of clinical studies that have been so far employing ICIs for the treatment of unresectable MPM, finally attempting at shedding light on the unsolved and unmet needs with regard to the application of ICIs in MPM.

2. Biological rationale for the use of ICIs in MPM

Similar to other malignancies, immune evasion has been postulated to be a predominant hallmark of cancer also for MPM. In fact, several studies have indicated that the tumor microenviroment of MPM reflects that of an immunologically active tumor, in which the type and the density of lymphocytes infiltration may influence the overall prognosis.^15–18 On one hand, increased levels of CD4⁺ T and CD20⁺ B lymphocytes have been associated with a better prognosis, while on the other, a high infiltration of CD8⁺ lymphocytes into the tumor has been found to correlate with worse survival.^15,17 By contrast, a study of 275 MPM showed that high intratumoral levels of immunosuppressive cells such as regulatory T and myeloid-derived suppressor cells were significantly associated with a shorter survival.¹⁸ Other studies have suggested that the presence of certain immune cells as well as the expression of a given immunophenotype in pleural fluids obtained from patients with MPM may have prognostic significance similarly to the immunophenotype of the tumor microenviroment.^19,20

Despite this immunologically active tumor microenviroment, PD-L1 expression by immunohistochemistry is not common in MPM.^21–25 In a study conducted in 329 patients, PD-L1 positivity was found in 41.7% of cases, although only 9.6% of tumor samples were considered to be highly positive (i.e. moderate/intense staining of at least 50% of tumor cells).²⁴ Of note, PD-L1 expression varies according to histology, with the highest expression present in the non-epithelioid subtype in which it associates with increased immunological infiltrates and poor prognosis.^17,21–25 On this basis, it can be suggested that PD-L1 as a mechanism of immune evasion that leads to the formation MPM seems to be particularly relevant for non-epithelioid histology. Interestingly, PD-L1 expression has been associated with a worse prognosis in MPM patients treated with chemotherapy.²²

On the other hand, immunological infiltrates in MPM often show infiltrating T cells with an immunosuppressive phenotype, especially in PD-L1-positive tumors. in a cohort of 38 MPMs subjected to flow cytometry Awad and colleagues found that PD-L1-positive tumors were more likely to express CD4⁺ cells with the regulatory T-cell (Treg) marker FOXP3⁺ and CD8⁺ cells with co-expression of PD-1 and the inhibitory receptor TIM-3.²⁶ The latter aspect is particularly relevant as immunotherapeutic drugs targeting TIM-3 are currently in clinical trials for advanced solid malignancies either as monotherapy or in combination with agents that block PD-1 or PD-L1.²⁷ More recently, the work by Klampatsa et al. extended these findings by analyzing fresh tumor samples of 22 MPM cases and demonstrated low CD8⁺ tumor infiltrating lymphocytes, whose function was moderately to severely suppressed.²⁸ In addition, similarly to Awad et al., they found that hypofunction of tumor infiltrating lymphocytes was associated with higher numbers of CD4⁺ Tregs, which also co-expressed the inhibitory receptor TIGIT in approximately three quarter of cases. Again, these findings provide a strong rationale for the ongoing clinical trials that are testing anti-TIGIT monoclonal antibodies in advanced malignancies, including MPM.²⁷

Unlike other immunologically active malignancies, MPM has a low tumor mutational burden (average of 24 plus or minus 11 protein-coding alterations per sample), which might imply less neoatinges able to induce an active T-cell response.²⁹ Similarly, a very low frequency of mismatch repair deficiency, ranging from 0% to 1.8%, has been reported in MPM, suggesting again that this is a tumor with little formation of neoantigens.^30,31 However, MPM often harbors chromosomal rearrangements, novel gene fusions, short deletions, and other genomic alterations that are difficult to capture by whole-exome sequencing used by Bueno et al. in their study in order to assess the tumor mutational burden.^29,32,33 A study by Thapa et al. identified a median of 147 copy number aberrations (range 3–866) reflecting a high genomic instability.²⁴ Importantly, a higher percent of genomic alterations did not correlate with PD-L1 expression, and genomic aberrations were found to be more frequent in epithelioid histology. Therefore, it can be hypothesized that different histologic variants correspond to different immunophenotypes, with the epithelioid subtype possessing neoantigenic potential due to increased genomic instability, and not because of a high rate of point mutations as assessed by whole genome sequencing.

3. ICIs in the salvage setting

A standard therapy for MPM in the salvage setting after platinum-based regimen is still lacking. The most adopted treatment options include single-agent chemotherapy with vinorelbine, gemcitabine, or retreatment with pemetrexed. These choices are based mainly on phase 2 or retrospective studies which reported an overall response rate (ORR) of 0–19%, a disease control rate (DCR) of 38–84%, a progression free survival (PFS) of 1.6–3.8 months, with overall survival (OS) ranging from 2.5 up to 12 months in the most recent trials.^34–40 In this context of unmet therapeutical needs, a large body of clinical trials have investigated the potential use of ICIs in second or further-lines (Tables 1 and 2).

Table 1. Studies evaluating immune-checkpoint inhibitors as single agent in salvage setting in mesothelioma

AGENT/Study	Design/target	Sample (% Ep.)	PD-L1 status	Line	ORR	DCR	PFS	OS
TREMELIMUMAB MESOT-TREM-2008^Citation41	Phase 2/CTLA-4	29 (86%)	n.a.	2nd	7%	31%	6.2 mo	10.7 mo
TREMELIMUMAB MESOT-TREM-2012^Citation42	Phase 2/CTLA-4	29 (73%)	n.a.	2nd	3% (irOR 4%)	38%	irPFS: 6.2 mo	11.3 mo
TREMELIMUMAB vs PLACEBO DETERMINE^Citation43	Randomized phase 2b/CTLA-4	571 (83%)	n.a.	2nd-3rd	5% vs 1%	28% vs 22%	2.8 mo vs 2.7 mo	7.7 mp vs 7.3 mo HR 0.92 (95% CI 0.76–1.12, p =.41)
NIVOLUMAB NivoMes^Citation44	Phase 2/PD-1	34 (82%)	≥1% in 27%; 0 in 71%, 1 pt. n.a.	2nd	26%	47%	2.6 mo	11.8 mo
NIVOLUMAB MERIT^Citation45	Phase 2/PD-1	34 (79%)	≥1% in 59%, 0 in 35%, 6% n.a.	2nd-3rd	29% (BICR)	68% (BICR)	6.1 mo (BICR)	17.3 mo
NIVOLUMABMAPS2^Citation46	Phase 2/PD-1	63 (83%)	≥1% in 30%, 0 in 49%, 21% n.a.	2nd-3rd	17% (BICR)	40% (BICR)	4.0 mo (BICR)	11.9 mo
NIVOLUMAB vs PLACEBOCONFIRM^Citation47	Phase 3/PD-1	221 (88%)	≥1% in 37%	2nd-3rd	n.r.	n.r.	3.0 mo	9.2 mo
PEMBROLIZUMAB KEYNOTE-028^Citation48	Phase 1b/PD-1	25 (72%)	≥1%	8% 1st, 92% ≥ 2nd	20%	72%	5.4 mo	18 mo
PEMBROLIZUMAB^Citation49	Phase 2/PD-1	64 (77%)	n.a.	2nd-3rd	19%	66%	4.5 mo	11.5 mo
PEMBROLIZUMAB^Citation50	Retrospective/PD-1	93 (73%)	<5% in 42%, ≥ 5% in 21%, n.a. in 29%	Any (52% 2nd)	18%	48%	3.1 mo	7.2 mo
PEMBROLIZUMAB vs 2nd LINE CHEMOTHERAPY PROMISE-meso^Citation40	Randomized phase 3/PD-1	144 (89%)	≥1% in 65% (of 102 available)	2nd	22% vs 6% (BICR)	45% vs 38% (BICR)	2.5 mo vs 3.4 mo HR 1.06 (95% CI 0.73– 1.53, p =.76) (BICR)	10.7 mo vs 12.4 mo
AVELUMAB JAVELIN Solid tumor^Citation52	Phase 1b/PD-L1	53 (81%)	≥1% in 42%, 0 in 40%, 19% n.a.	≥2nd	9%	58%	4.1 mo	10.7 mo

BICR: blinded independent central review; DCR: disease control rate; Ep: epithelioid; irOR: immune-related objective response; mirPFS: median immune-related progression free survival; mo: months; mOS: median overall survival; PFS: median progression free survival; n.a.: not assessed; n.r. not reported; ORR: overall response rate.

Table 2. Studies evaluating combination treatments with immune-checkpoint inhibitors in second and further lines in mesothelioma

AGENT/Study	Design/target	Sample (% Ep.)	PD-L1 status	Line	ORR	DCR	mPFS	mOS
DURVALUMAB and TREMELIMUMAB NIBIT-MESO-1^Citation56	Phase 2/PD-L1 and CTLA-4	40 (80%)	≥1% in 50%, 0 in 45%, 5% n.a.	1st-2nd (30%-70%)	25%	63%	5.7 mo	16.6 mo
DURVALUMAB and TREMELIMUMAB^Citation57	Phase 2/PD-L1 and CTLA-4	19	≥1% in 37%, 0 in 53%, 11% n.a.	≥2nd	5%	52%	2.8 mo	7.8 mo
NIVOLUMAB and IPILIMUMAB INITIATE^Citation58	Phase 2/PD-1 and CTLA-4	35 (86%)	≥1% in 43%, 0 in 54%, 3% n.a.	≥2nd	38%(at 6 months)	50%(at 6 months)	6.2 mo	Not reached (1yOS 64%)
NIVOLUMAB and IPILIMUMAB MAPS2^Citation46	Randomized phase 2/PD-1 and CTLA-4	62 (85%)	≥1% in 35%, 0 in 44%, 21% n.a.	2nd-3rd	30% (BICR)	52% (BICR)	5.6 mo (BICR)	15.9 mo

BICR: blinded independent central review; DCR: disease control rate; Ep: epithelioid; mOS: median overall survival; PFS: median progression free survival; mo: months; n.a.: not assessed; n.r. not reported; ORR: overall response rate.

3.1. ICIs as single agent

The CTLA-4 inhibitor tremelimumab has been the first ICI evaluated as single agent in 2nd/3rd line setting of MPM. After preliminary results of some activity, especially in the small subpopulation of non-epithelioid mesothelioma,^41,42 a phase 2b randomized trial did not show any survival benefit with this drug compared to placebo.⁴³

Three phase 2 trials (the Dutch NivoMes, the Japanese MERIT study, and the French MAPS2), with overall 131 patients, evaluated nivolumab as a PD-1 inhibitor in MPM salvage setting.^44–46 The ORR ranged from 19% to 29% and it was higher among patients with PD-L1 expression in MERIT and MAPS2 studies.^45,46 Moreover, a trend for better OS and PFS was described in patients with PD-L1 ≥ 1% in MERIT trial. In the same study, a sub-analysis by histologic type exhibited also a trend for prolonged OS and PFS for patients with non-epithelioid mesothelioma, taking into account the small size of this subgroup (7 patients).⁴⁵ Based on the MERIT trial, nivolumab is approved in Japan as a salvage therapy in MPM. However, MAPS2 trial reported a suspected hyperprogression in 10% of treated patients, with no clear correlation with the histologic type.⁴⁶ More recently, the phase 3 ‘CONFIRM’ trial randomized patients with relapsed MPM after at least one prior line of chemotherapy to nivolumab or placebo, the primary end-point being OS.⁴⁷ Patients on the nivolumab arm experienced a significantly longer survival than placebo, with a median OS of 9.2 months for nivolumab versus 6.6 months for placebo (HR 0.72, 95% CI 0.55–0.94, P =.018). Similarly, also PFS was improved in the experimental arm, with a median PFS of 3.0 months for nivolumab versus 1.8 months for placebo (HR 0.61, 95% CI 0.48–0.77, P <.001).

Three clinical trials and one retrospective study have reported an ORR of approximately 20% and a DCR ranging from 45% to 72% with the PD-1 inhibitor pembrolizumab, similar to what observed with nivolumab.^40,48–50 In the retrospective study, higher ORR and PFS were achieved in both the subpopulation with non-epithelioid mesothelioma and the subgroup with PD-L1 expression ≥5%.⁵⁰ However, despite an improvement in ORR, PROMISE-meso study, a randomized trial comparing pembrolizumab to single agent chemotherapy in second line, failed to demonstrate any benefit in terms of PFS or OS, also after adjusting for crossover.⁴⁰ Differently from previous data, no differences in ORR, PFS or OS were detected in a subgroup analysis by PD-L1 status.^40,49,50,51

Avelumab has been the first monoclonal antibody targeting PD-L1 which was evaluated in MPM. In JAVELIN Solid Tumor phase 1b trial, avelumab showed similar DCR and PFS compared to anti-PD-1 agents in the same setting (Table 1).⁵²

Overall, toxicity was expectable and manageable. Treatment-related adverse events (TRAEs) occurred in 76–89% of patients treated with nivolumab, 64–69% with pembrolizumab and in 81% with avelumab. Patients suffering from grade ≥3 TRAEs were 14–32% with nivolumab, 19–20% with pembrolizumab and 9% with avelumab. Immune related adverse events (IrAEs) ranged from 6% to 23%. Drug discontinuation due to adverse events was required in 5–12% of patients treated with nivolumab, 0–7% with pembrolizumab and 26% with avelumab.^{40,44,45,48,51}

3.2. ICIs as combination

As already shown in other solid tumors,^53,54,55 four phase 2 trials in MPM provided results from a combination strategy with anti-CTLA-4 and anti-PD-L1 checkpoint inhibitors. The NIBIT-MESO-1 study firstly evaluated the association of durvalumab (anti-PD-L1) and tremelimumab (anti-CTLA-4), mostly in the second line setting. The treatment consisted of an induction phase with four doses of both drugs, followed by nine cycles of maintenance with durvalumab. ORR and DCR, according to modified RECIST, were 25% and 63%, respectively with durable responses (13.8 months as median duration of response).⁵⁶ No significant association between outcome and baseline PD-L1 expression was observed. Notably, a similar study evaluating the same drug combination was suspended at the time of interim analysis because of an ORR of 5% reached in the first 19 enrolled.⁵⁷ The combination of ipilimumab and nivolumab was tested in two different studies. The Dutch single arm INITIATE trial firstly evaluated ipilimumab 1 mg/kg every 6 weeks up to four cycles and nivolumab 240 mg biweekly until progression or toxicity among 34 evaluable patients with relapsed/refractory MPM. At 6 months, ORR was 38% and DCR was 50%. In this trial PD-L1 expression on tumor cells was significantly associated with clinical benefit (DCR > 6 months, longer PFS and OS.⁵⁸ The aforementioned MAPS2 trial was a randomized, non-comparative phase 2 trial that allocated 125 patients to receive either nivolumab 3 mg/kg monotherapy every two weeks or ipilimumab 1 mg/kg every 6 weeks and nivolumab 3 mg/kg biweekly, both until progression or unacceptable toxicity.⁴⁶ The latter cohort of patients achieved an ORR of 28% and a DCR of 52%. While responses seemed related to PD-L1 expression in the arm of patients treated with nivolumab as single agent (2 PD-L1 negative and 20 PD-L1 positive patients achieved a partial response, while 26 PD-L1 negative and 6 PD-L1 positive patients had progressive disease), the addition of CTLA-4 inhibitor blunted this imbalance in the cohort of patients treated with ipilimumab/nivolumab (10 PD-L1 negative and 9 PD-L1 positive patients achieved a partial response, while 12 PD-L1 negative and 4 PD-L1 positive patients had progressive disease) (Table 2).

The safety profile was consistent with previously described data, with increased toxicity from ICIs combination as compared to monotherapy. TRAEs were reported in more than 93–94% of patients with ipilimumab/nivolumab (34–37% ≥ grade 3) and in 75% of cases with tremelimumab/durvalumab (18% ≥ grade 3).^46,56,58 More in detail, Calabrò et al. registered IrAEs of any grade in 75% of patients, while Disselhorst et al. reported that 23% of patients needed corticosteroids for the management of immune-related toxicity.^56,58 Of note, the discontinuation rate was 3–8% when the induction phase with a CTLA-4 inhibitor was administered for a limited period of time,^56,58 and up to 21% in MAPS2 trial where ipilimumab was continued every 6 weeks until progression.⁴⁶

4. ICI in the first line setting

Table 3 lists the study that have evaluated ICI in the first line setting. One area for ICI testing in unresectable MPM is that of adding immunotherapy to platinum/pemetrexed as first line treatment. ‘DREAM’ was a phase 2 trial in which patients with untreated unresectable MPM received cisplatin/pemetrexed plus the anti-PD-L1 antibody durvalumab.⁵⁹ In case of cisplatin-related toxicity, replacement of cisplatin with carboplatin was permitted. After documenting no dose-limiting toxicity in a safety run-in cohort of 6 patients, a total of 54 patients were treated with this combination. The primary efficacy end-point of the study, which was showing a PFS rate at 6 months greater than 45%, was met, as it was found to be 57%. The other parameters of activity reported an ORR of 48%, a median PFS of 6.9 months, and a median OS of 18.4 months. Regardless of severity, the majority of treatment-related AEs were attributed to chemotherapy. On the other hand, grade 3 and 4 IrAEs were reported in 15% of patients (8 patients), of whom 7 patients required high-dose steroids and/or other immunosuppressive drugs. Recently, the results of ‘PrE0505ʹ, a twin phase 2 study which had the same study design of ‘DREAM’ trial have been made available.⁶⁰ In total 55 patients were enrolled, and the results in terms of activity were similar as ORR, PFS-6, median PFS, and median OS were 56.4%, 69.1%, 6.7 months, and 20.4 months. Both trials did not report an association between PD-L1 status and efficacy of treatment. Similarly, no correlation between tumor mutational burden and treatment efficacy was seen in ‘PrE0505ʹ. Remarkably, no grade 3 and 4 immune-related AEs were reported in ‘PrE0505ʹ. Taken together, both studies showed that the combination of platinum/pemetrexed plus durvalumab is active and safe in unresectable, advanced MPM. Currently, a phase 3 study of standard cisplatin/pemetrexed chemotherapy with or without pembrolizumab is ongoing (NCT02784171). Of note, this trial will also address the activity of the chemo-immunotherapy combination in the non-epithelioid variant of MPM, which consisted only 17% and 25.5% in the ‘DREAM’ and ‘PrE0505ʹ trials, respectively.^59,60

Table 3. Studies evaluating treatment with immune-checkpoint inhibitors in first line in mesothelioma

AGENT/Study	Design/target	Sample (% Ep.)	PD-L1 status	ORR	DCR	mPFS	mOS
DURVALUMAB/CISPLATIN/PEMETREXED DREAM*^Citation59	Phase 2/PD-L1	54 (83%)	≥1% in 50%, 0 in 44.4%, n.a. in 5.6%	48%	87%	6.9 mo	18.4 mo
DURVALUMAB/CISPLATIN/PEMETREXED PrE0505*^Citation60	Phase 2/PD-L1 and	55 (74.5%)	≥1% in 40%, 0 in 35.5%, n.a. in 25.5%	56.4%	96.4%	6.7 mo	20.4 mo
IPILIMUMAB/NIVOLUMAB vs CISPLATIN or CARBOPLATIN/PEMETREXED CheckMate 743^Citation61	Randomized phase 3/CTLA-4 and PD-1	605 (25%)	≥1% in 74%, 0 in 22.5%, n.a. in 3%	40% vs 43%	76.6% vs 85.1%	6.8 movs 7.2 mo HR 1.00 (95% CI 0.82–1.21, p = not significant)	18.1 mo vs 14.1 mo HR 0.74 (95% CI 0.60–0.91, p =.0020)

*Carboplatin was substituted if cisplatin was contraindicated, or due to toxicity during treatment.

DCR: disease control rate; Ep: epithelioid; mOS: median overall survival; PFS: median progression free survival; mo: months; n.a.: not assessed; ORR: overall response rate.

Based on the interesting results obtained in the second-line setting, a chemotherapy-free strategy of ipilimumab/nivolumab has been tested in the first treatment of MPM. CheckMate 743 was a randomized phase 3 study in which 605 untreated patients with a performance status of 0 or 1 were randomized to either ipilimumab/nivolumab (303 patients) for up to 2 years or standard platinum/pemetrexed (302 patients) for up to 6 cycles, the primary endpoint being OS.⁶¹ At a median follow-up of 29.7 months, the authors reported that the ipilimumab/nivolumab arm a had a significantly superior OS of 18.1 months versus 14.1 months for chemotherapy (HR 0.74, 96.6% CI 0.60–0.91, P =.002). Of note, the survival curves started to separate at around 6 months since treatment initiation, with a difference which appeared to be more prominent over the time (1 and 2-year survival were 68% and 41% for ipilimumab/nivolumab versus 58% and 27% for chemotherapy, respectively). Remarkably, although a survival benefit was noted both in epithelioid and non-epithelioid histology, this was much more pronounced in the non-epithelioid variant (HR 0.86, 95% CI 0.69–1.08 and HR 0.46, 95% CI 0.31–0.68 for epithelioid and non-epithelioid histologies, respectively). This difference was mainly due to the poor performance of the non-epithelioid histology in the chemotherapy arm (median OS 8.8 months versus 16.5 months, respectively), while patients in the ipilimumab/nivolumab arm performed similarly regardless of the histologic variant (median OS 18.7 months versus 18.1 months for the epithelioid versus the non-epithelioid arms, respectively). Taken together, these data suggest that the ipilimumab/nivolumab could be used as a new standard of care in the first-line treatment of patients with unresectable, advanced MPM, and that this immunotherapy combination is particularly appealing for patients with non-epithelioid histology, in which the expected prognosis with standard chemotherapy is dismal. However, a few considerations should be made about the results of this study. First, data on the percentage of PD-L1 expression according to the histologic variant were not made available, and it is unclear to what extent an imbalance in PD-L1 expression between the epithelioid and non-epithelioid variant could contribute to the overall survival difference. Secondly, as expected, the combination of ipilimumab/nivolumab resulted into a higher incidence of TRAEs leading to discontinuation of any component of the regimen (23% versus 16%) as well as greater percentage of serious treatment-related adverse events (21% versus 8%) compared with chemotherapy. Nevertheless, based on the results of CheckMate 743 the ipilimumab/nivolumab combination regimen has been approved by FDA in October 2020 as first line treatment of patients with unresectable malignant mesothelioma.

5. Concluding remarks

Based on the above mentioned trial results, ICI treatment may soon represent a new standard of care for the medical management of MPM. However, a few issues need to be clarified before the full implementation of ICI therapy in the first line setting.

Optimal patients’ selection based on predictive biomarkers is key for the success of treatment. Correlation of ICI efficacy with PD-L1 expression has been contradictory so far. That is because PD-L1 is not commonly expressed in MPM, except for the non-epithelioid variant, in which PD-L1 may retain some predictive role. However, PD-L1 is a fallacious biomarker as it is subjected to spatial (i.e. depending on the site of tissue collection) and temporal (i.e. depending on the time of tissue collection) heterogeneity.⁶² In addition, different clones have been employed for the measurement of PD-L1 status across different studies, which represents a further bias to the reproducibility of data that correlate PD-L1 expression with sensitivity to ICI treatment. In the future, it is reasonable to foresee the use of composite biomarkers beyond PD-L1 and TMB, including but not limited to the assessment of the immunophenotype of tumor microenviroment. Also, it should be taken into account that different histologic variants correspond to peculiar molecular alterations as well as different immunohenotypes, which suggest that predictive biomarker to ICI treatment may vary according to the histologic subtype.

Finally, it is evident that not all patients with MPM may require immediate treatment, owing to the presence of indolent disease in some cases. Identifying the patients in which treatment can be postponed, thus sparing potentially severe immune-related adverse events is crucial during daily clinical practice.

In conclusion, a new era has begun for the treatment of MPM. While ICI treatment represents only one aspect of the application of immunotherapy for MPM, the future will consist in the evaluation of ICI and other immunomodulatory agents in the earlier stage disease as well as in combination regimens in order to fulfill the final goal of obtaining long-term survival in the face of an acceptable toxicity for our patients.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.