The Value of Pemetrexed for the Treatment of Malignant Pleural Mesothelioma: A Comprehensive Review

Clinical Efficacy

Pemetrexed. One phase III RCT (17) was identified which established CISPEM as the preferred regimen in the treatment of non-resectable MPM (2-4, 9-11). The risk of bias has previously been assessed in a NICE appraisal (15, 16). The authors found the lack of a double-blind design and the reported ineligibility of approximately 20% of the 574 consenting patients in the trial to be a potential risk of bias. Initial severe pemetrexed-induced toxicity, particularly neutropenia, febrile neutropenia, and diarrhea, led to the introduction of supplementary treatment with folic acid and vitamin B12, which substantially reduced grade 3/4 toxicity (except for cisplatin-induced dehydration). Supplementation was given to about 75% of patients (17). Patients responded considerably better to CISPEM than to cisplatin (Table I). Although patients were treated with a median number of six cycles, 87% of those who responded did so within the first four cycles (9). For CISPEM, the median time-to-progression (TTP) was 6.1 months as compared to 3.9 months for cisplatin alone. The median overall survival of vitamin-supplemented patients treated with CISPEM was also longer than of patients treated with cisplatin only (13.3 vs. 10.0 months), which was supported by the results of a long-term survival update in 2005 (41). However, the addition of pemetrexed to cisplatin resulted in a substantial increase of serious (grade 3/4) hematological and non-hematological toxicities. Nausea, vomiting and fatigue were the most common serious adverse events related to CISPEM, affecting up to 15% of patients. Up to 5% suffered from diarrhea, dehydration and stomatitis (17). In a recent non-comparative phase II study, the addition of bevacizumab to CISPEM resulted in a 40% response rate (RR) and 14.8-month median overall survival (Table I)(31). Grade 3/4 toxicities included neutropenia in 11%, fatigue in 8%, hypertension in 6%, and venous thromboembolism in 13% of patients.

For patients with unresectable MPM unfit for CISPEM combination therapy, the use of pemetrexed monotherapy might be a treatment option (2-4, 10, 11). One phase II study was identified that evaluated first-line use of pemetrexed and resulted in a 14% RR and a median overall survival of 10.7 months (25). In patients supplemented with folic acid and vitamin B12 the median overall survival amounted to 13 months (25).

CARPEM. With respect to non-hematological toxicity (nausea, vomiting, renal function), carboplatin is generally better-tolerated than cisplatin (2-4). Carboplatin is therefore considered an alternative for patients intolerant to cisplatin (2, 10, 11). For this combination, one recently published phase II RCT was identified in which CARPEM was compared with the combination of CISGEM. RRs were 78.9% in the CARPEM group compared with 47.6% in the CISGEM group. Cumulative survival proportion at 18 months was 57.8% and 41%, respectively. Most common grade 3/4 toxicities were leukopenia (15% vs. 38%), thrombocytopenia (10% vs. 24%), and nausea/vomiting (0% vs. 33%). Two Italian phase II studies evaluated carboplatin in combination with pemetrexed and reported RRs of 19% and 25% (19, 20). The median TTP and overall survival were 6.5-8 months and 12.7-14 months, respectively (Table I). Serious grade 3/4 toxicities consisted of hematological side-effects, particularly anemia and neutropenia, in up to 25% of patients using CARPEM. Serious non-hematological toxicities, nausea, vomiting, diarrhea and conjunctivitis, were observed in about 5% of patients (19, 20). Similar trends were reported in patients older than 70 years of age compared with those under 70; older patients, however, experienced greater hematologic toxicity (42). The results of a recent phase II trial (n=62) with CARPEM were in line with those of the Italian trials (Table I) (23).

CISPEM vs. CARPEM. The large non-randomized open-label, under the manufacturer-sponsored International EAP study in European (36, 37) and US patients (35) compared CISPEM with CARPEM. The efficacy of CARPEM was similar to that of CISPEM; for both doublets, the median TTP was between seven and eight months, the median overall survival ranged from 12 to 14 months and the 1-year survival rate was close to 60% (Table I). In European patients, the most common grade 3/4 toxicities were neutropenia (24% vs. 36%), leukopenia (13% vs. 21%), anemia (7% vs. 14%), and thrombocytopenia (5 vs. 14%), indicating that CARPEM induces somewhat greater hematological toxicity than CISPEM. However, nausea and vomiting occurred in about 3% of patients treated with either CISPEM or CARPEM (36). The incidence of grade 3/4 toxicity was similar in the German EAP study, with neutropenia occurring in 29% vs. 37%, leukopenia in 22% vs. 22%, anemia in 16% vs. 20%, and thrombocytopenia in 13% vs. 17%, for patients treated with CISPEM or CARPEM (37). The most common investigator-reported hematological toxicities in the EAP study including US patients treated with CISPEM were anemia (2.3%) and thrombocytopenia (1.3%). Neutropenia was reported in <1% of patients (35).

Raltitrexed. In one small open phase III RCT in 250 chemotherapy-naïve patients with good performance status, the effect of the combination CISRAL was compared to that of cisplatin monotherapy (18). The results of this trial were of limited power. An Independent Data Monitoring Committee recommended to extend patient inclusion after two interim analyses, but this could not be realized due to lack of funding. The lack of a double blind design may have introduced bias in investigator assessments. The median overall survival was 11.4 months for patients treated with CISRAL as compared to 8.8 months for those treated with cisplatin. However, improvements in RR and progression-free survival (PFS) were not significant (Table I). Serious hematological toxicity was limited to neutropenia occurring in 16% of patients using CISRAL compared to 8% of patients using cisplatin. Leukopenia and anemia occurred in fewer than 10% of patients in both arms. Serious non-hematological toxicity was largely restricted to nausea (CISRAL, 14% vs. cisplatin, 10%), vomiting (13% vs. 7%) and fatigue (12% vs. 6%).

Vinorelbine. One phase III RCT was identified in which the combination of BSC with chemotherapy was investigated (6). A total of 409 patients with MPM were randomly assigned to three treatment groups: BSC-only, BSC plus mitomycin-vinblastine-cisplatin (MVP), or BSC plus vinorelbine. The results of the trial may have been influenced by selection bias. The trial aimed to recruit 840 patients. However, patient recruitment was slower than required, and therefore the study design was changed into a comparison of two groups: BSC-only versus BSC plus either one of the chemotherapeutic treatments. The methods for randomization were not clearly explained and the lack of a double blind design may also have introduced bias. Treatment with BSC plus vinorelbine resulted in a 31% RR (vs. 14% for BSC) and a median overall survival of 9.5 months (vs. 7.6 months for BSC) (Table I). There was no evidence of a survival benefit with BSC plus MVP. The only comparisons showing evidence of increased side-effects were for alopecia (BSC plus vinorelbine, 6% vs. BSC, 0%), lethargy (45% vs. 30%), and hematological toxicity, which became apparent in the latter half of the trial (36% vs. 0%). No significant differences between treatments were recorded at 6 months (6).

One recent non-comparative Danish phase II trial (N=54) was identified which showed that treatment with cisplatin in combination with weekly-dosed vinorelbine resulted in a 29.6% RR and a median survival and median TTP of 16.8 months and 7.2 months, respectively (Table I) (26). The most common grade 3/4 hematological toxicity was leukopenia (48%), and the most common non-hematological grade 3/4 toxicities were nausea (13%), neurotoxicity (11%), and toxicities such as tiredness or constipation in 9%. In another phase II study, first-line treatment with vinorelbine monotherapy resulted in a 24% RR and a median overall survival of 10.6 months (27).

Gemcitabine. Eight phase II studies were identified concerning the effect of gemcitabine (21, 22, 24, 28-30, 33, 34), out of which two were randomized (33, 34). The combination CISGEM resulted in an RR ranging from 12 to 48%, a median TTP of 5.8-6.4 months and a median survival of 9.5-11.2 months (Table I) (22, 28-30). Serious toxicity was predominantly hematological. A recently performed phase II study on CISGEM resulted in a 50% RR and a median PFS and median survival of 8 and 17 months, respectively (Table I)(24). Neutropenia (23%) was the most common grade 3/4 hematological toxicity. CISGEM, alone or in combination with bevacizumab, has also been studied in an open, placebo-controlled phase II RCT (33). The addition of bevacizumab (BEVCISGEM) did not improve the treatment result of CISGEM but the 15-month duration of the median overall survival for both regimens is among the longest recorded (Table I). The incidence of grade 3/4 toxicity was not significantly different between CISGEM and BEVCISGEM, with neutropenia occurring in 40% vs. 42%, anemia in 15% vs. 4% and thrombocytopenia in 25% vs. 38% of patients, respectively.

Gemcitabine has also been studied in combination with carboplatin (21). Overall, 26% of the patients had partial responses but none of the patients had a complete response. Treatment resulted in a median PFS and overall survival of 9.2 and 15.2 months, respectively (Table I). The most commonly reported grade 3/4 toxicity was thrombocytopenia, occurring in 34% of patients (21).

Pemetrexed-gemcitabine. Attempts to identify less toxic but equally effective combinations of agents such as antimetabolites, antracyclines and topoisomerase inhibitors have failed (2-4). We identified one study exploring the activity of pemetrexed in combination with gemcitabine, with gemcitabine administered on days 1 and 8 plus pemetrexed on day 8 (n=56) vs. pemetrexed on day 1 (n=52) (32). Treatment resulted in 26% and 17% RR, a median TTP of 4.3 and 7.4 months, and a median survival of 8.1 and 10.1 months, respectively (Table I). Incidence of grade 4 neutropenia was 25.0% and 29.4%, grade 4 thrombocytopenia 14.3% and 3.9%, grade 3/4 anemia 5.4% and 5.9%, and grade 3/4 fatigue 23.2% and 15.7%, respectively.

Second-line chemotherapy. Pemetrexed has also been evaluated as second-line treatment in patients previously treated with chemotherapy (2-4, 10, 11, 43). We identified one phase III RCT in which pemetrexed plus BSC was compared with BSC-alone in 243 pemetrexed-naïve patients with chemotherapy-relapsed MPM (44). Tumor response and PFS were improved with pemetrexed but active treatment did not influence overall survival. Subgroup analysis showed that patients who responded to first-line chemotherapy tended to have a longer survival when treated with pemetrexed in second line (43, 44). Trial results and observational study data also suggest that patients may benefit from re-treatment with pemetrexed (alone or in combination with cisplatin/carboplatin) with acceptable toxicity (2-4, 43, 45, 46). A recently published open-label phase III RCT investigated the combination of BSC with thalidomide in patients with unresectable MPM or peritoneal mesothelioma who had completed first-line chemotherapy (47). A total of 222 patients were randomly assigned to receive either maintenance thalidomide combined with BSC, or BSC-alone. Treatment with thalidomide plus BSC resulted in a median TTP of 3.6 months and a median overall survival of 10.6 months (vs. 3.5 and 12.9 months for BSC, respectively). Grade 3/4 toxicity occurred in 39% for the thalidomide group (vs. 28% for BSC). Phase II trial results of patients out of which many were previously treated with pemetrexed, with vinorelbine-only and gemcitabine in combination with vinorelbine or epirubicine yielded RRs between 10 and 16% and a median overall survival of around 10 months (2-4, 43).

Quality of Life

Symptoms such as cough, dyspnea, (chest) pain, night sweats, tiredness and weight loss, as well as rapid disease progression, severely impair the health-related QoL of patients with advanced MPM (1, 2, 6, 14). Palliative treatment is predominantly aimed at disease stabilization and symptom control but adverse effects associated with chemotherapy should not compromise the treatment aim (1-4, 6). In a large phase III trial comparing BSC plus chemotherapy with BSC, chemotherapy was associated with a slight decrease of pain and increase of lethargy, whereas the scores for physical functioning and global QoL remained essentially unchanged regardless of whether chemotherapy was given or not (6). In several trials, treatment was reported to induce a better control of MPM-associated symptoms, dyspnea and pain in particular (2-4, 17, 18). However, to a considerable extent, this effect was offset by chemotherapy-induced fatigue, gastrointestinal symptoms, nausea and stomatitis (1, 2, 14, 17, 18). In the phase III trials comparing CISPEM and CISRAL with cisplatin, it also appeared that combination treatment did not result in clinically relevant improvements of major symptom scores (pain, dyspnea, fatigue, anorexia and cough) and global QoL. On the other hand, the addition of pemetrexed or raltitrexed to cisplatin had no detrimental effect on QoL but resulted in a stabilization of QoL over the course of treatment (17, 18, 48, 49). In the light of the small but clinically relevant gain in overall survival, this might be considered a valuable treatment effect. No RCT demonstrated the impact of pemetrexed as second-line treatment on survival or QoL (2).

Economic Evaluations

In 2007, the NICE published its first full guidance for the treatment of MPM and the use of pemetrexed based on a systematic review of clinical data (16) and an economic evaluation (15). In 2010, a revised version was published (9). The revised appraisal considered three cost-effectiveness studies. Two of these were submitted by the manufacturer and comprised a pharmacoeconomic analysis of CISPEM versus cisplatin (model 1), and CISPEM versus BSC plus MVP and/or BSC plus vinorelbine (with or without platinum) (model 2). Data on the clinical effectiveness of the comparator treatments of model 2 were gathered from market research surveys commissioned by the manufacturer (9). It is unclear if or to what extent these data correspond with the results of a UK RCT in which these alternative regimens were compared to BSC (6). At present the treatment options of model 2 are not recommended (2, 9-11), while in the revised appraisal, there is no distinction between the effects and costs contribution of vinorelbine and CISVIN (9). The data of model 2 are, therefore, considered largely irrelevant. Model 1 (CISPEM vs. cisplatin) was also evaluated by a NICE assessment group (9). The revised NICE appraisal contains an updated version of this analysis with respect to costs and estimates of cost-effectiveness (9). According to the manufacturer, the incremental cost-effectiveness ratio (ICER) per quality-adjusted life-year (QALY) gained was £47,567 (€63,974). Taking into account the lower QoL of patients with MPM towards the end of their life, the NICE panel used a lower (mean) utility value than the manufacturer, who did not include disease progression in its models. The NICE panel's estimate of the ICER per QALY gained amounted to £37,700 (€50,704). In the case that, in addition to a 500-mg vial, pemetrexed would also be available in a 100-mg vial the ICER would amount to £34,500 (€46,400) per QALY gained. These estimates are below the UK threshold for cost-effectiveness of about £40,000 (nearly €50,000) (15). Moreover, on considering that MPM is a rare and aggressive malignant disease, that 100 mg vials would become available, treatment for most patients would be limited to four cycles and QoL benefits might have been underestimated in the pharmaco-economic analysis in 2010, the NICE recommended CISPEM as an option for first-line treatment of patients with non-resectable, advanced MPM and good condition (9). In the Netherlands and several other EU countries, pemetrexed is also recommended for treatment of this patient group (2, 50).

The acquisition price is an important factor underlying cost-effectiveness estimates. Following introduction in the Netherlands in 2005, the list price of a 500 mg vial rose from €1,617 to €1,733 (ex. VAT) in 2009 (51, 52). As the result, on the basis of the mean body surface (1.7 m²: dose of 850 mg) the costs of a three-week cycle amounted to €3,234. In 2012, largely as the result of governmental price control measures, the maximum price per vial was lowered to €1,334, making the costs per cycle €2,668 (53). Since pemetrexed is now also available in 100 mg vials [€288 (ex. VAT)], as anticipated by the NICE (9), the use of these vials would lower the theoretical costs of treatment per cycle to €2,380. Accordingly, at present the ICER per QALY gained would be considerably lower than the 2010 NICE estimate of £34,500 (€46,400).

The effect of the acquisition price is also illustrated by the recent comparison of CISPEM with CISRAL sponsored by the UK holder of the raltitrexed marketing authorization (40). Raltitrexed is not available in the Netherlands, but on the basis of list prices in 2009, the costs of raltitrexed were found to be lower than those of pemetrexed by 266% in the UK to around 500% in France and Spain, and to nearly 750% in Italy (40). On the basis of RCT data and assumptions essentially similar to those underlying the recent NICE evaluation (9), it was concluded that as compared to CISRAL, CISPEM offered marginally lower QALY gained at a substantially higher total cost and was economically inferior to CISRAL with an ICER of £13,454 (€16,757) per QALY gained as compared to cisplatin.