METHODS

Patients

In collaborating centres all new patients with a diagnosis of mesothelioma were registered with the MRC Clinical Trials Unit. Registered patients were eligible for randomisation if they fulfilled the following criteria:

• microscopically and immunohistochemically confirmed malignant pleural mesothelioma, including epithelial and other histological types;

• any symptomatic pleural effusion under control by drainage, pleurodesis, or pleurectomy;

• a CT scan within 1 month before randomisation (preferably after pleurodesis);

• if mesothelioma resected, two CT scans 6 weeks apart showing assessable stable or progressive disease;

• no previous chemotherapy for mesothelioma;

• no other disease likely to interfere with protocol treatments or comparisons;

• WHO performance status 0–2;¹⁰

• white blood cell count >3 × 10⁹/l, neutrophil count >1.5 × 10⁹/l, platelet count >100 × 10⁹/l, and no clinical evidence of infection;

• medically fit to receive chemotherapy;

• informed consent form signed following full discussion of a patient information sheet describing the study design and stating that the doctor would discuss with the patient which comparison was most suitable;

• quality of life forms completed before patient told the treatment allocated.

Multicentre Research Ethics Committees (MREC) approval of the protocol was obtained. Confirmation of Local Research Ethics Committee’s (LREC) approval was required before a centre could start registering patients.

Treatment groups

Clinicians were encouraged to randomise eligible patients between all three treatment groups: (1) ASC only; (2) ASC+MVP; (3) ASC+N, but were allowed to offer patients randomisation between any two. This choice was permitted to ensure a good measure of the acceptability of the ASC only arm to both patients and clinicians.

Active symptom control (ASC)

The essential elements were as follows:

• Regular follow up in a specialist clinic by an identified physician or team.

• Structured assessments at every clinic visit of physical, psychological, and social problems with appropriate treatment or other action.

• Rapid involvement of additional specialists such as a pain relief service, specialist palliative care team, medical social worker, or physiotherapist. (It was recommended that patients with pain not easily controlled by a combination of slow release morphine and co-analgesics be referred to a pain relief service.)

• Parallel nursing support from a named specialist nurse or similar person.

ASC could include treatment with palliative radiotherapy and steroids.

Mitomycin, vinblastine, and cisplatin (MVP)

Mitomycin 8 mg/m² by bolus intravenous injection, vinblastine 6 mg/m² (maximum dose 10 mg) by bolus intravenous injection, and cisplatin 50 mg/m² by intravenous infusion over 4 hours were given every 21 days for a total of four cycles with standard hydration and anti-emetics.

Vinorelbine (N)

Six intravenous injections at weekly intervals followed by a 2 week break before a further six injections at weekly intervals of vinorelbine 30 mg/m² (maximum dose 60 mg) were given, with standard anti-emetics.

QL questionnaires

All patients completed EORTC QLQ-C30+LC13¹¹ and FACT-L¹² QL questionnaires. To avoid bias that may arise if one questionnaire is always completed before the other, the sequence in which they were completed was randomised by centre, the sequence remaining the same for every assessment of all patients from that centre.

Assessments

Pretreatment assessments included a CT scan of the thorax; the QL questionnaires had to be completed before the patients were told their treatment group. During the treatment period, when MVP and N patients were receiving chemotherapy, all patients were assessed clinically every 3 weeks. Thereafter, patients were assessed every 8 weeks until death. QL questionnaires were completed immediately before all assessments. A CT scan was performed at the end of chemotherapy or 3–4 months after randomisation in the ASC only group.

Statistical methods

The outcome measures were the acceptability of the study design to patients, suitability of the EORTC QLQ-C30+LC13 and FACT-L QL questionnaires for mesothelioma, and palliation. Acceptability of the design was measured by the proportions of registered patients accepting randomisation and of randomised patients accepting ASC only as one of their randomisation options. The suitability of the QL questionnaires was assessed by compliance with their use and patient preference: patients were asked to state their preferred QL questionnaire but were not asked their reasons. Palliation was measured in terms of WHO performance status¹⁰ and severity of chest pain, breathlessness, anorexia, and sweating (present “not at all” = 0, “a little” = 1, “quite a bit” = 2, or “very much” = 3). Palliation was defined as: (1) improvement (change from a score of 2 or 3 to 0 or 1), (2) control (a score of 1 not getting worse), and (3) prevention (a score of 0 being maintained)¹³ from baseline to 16 weeks after randomisation, based on clinicians’ assessments. The study was not powered to compare tumour response or survival between treatment groups; these outcome measures are therefore not reported but are included in the phase III trial.

RESULTS

Acceptability of the trial design

From September 2000 to September 2001, 242 patients were registered from 46 UK centres. Of the 197 eligible for randomisation, 109 from 35 centres were randomised, 55% (95% CI 48 to 62) of those eligible and 45% of those registered. The main reasons for non-randomisation were ineligibility and patient refusal (table 1).

The randomisation options are shown in table 2. A third of patients were randomised to the recommended option (between all three treatment groups), but the option included ASC only in 52 (48%) of the 109 patients (95% CI 38 to 57), these 52 being randomised from 20 centres. With an eventual rate of randomisation including ASC only of 8.7 patients per month (104 per year), this translates into an estimated rate of 312 (104 × 3) per year from the 60 centres that expressed an interest in collaborating in the phase III trial.

The characteristics of the patients at registration or randomisation according to their randomisation group are shown in table 3. The three groups of patients (randomised with ASC only option, randomised without ASC only option, not randomised) were very similar except that there was a tendency for the group randomised without the ASC only option to be younger than the group with this option (median age 61 and 66 years, respectively).

Comparison of the two QL questionnaires

Compliance with the completion of the QL questionnaires was similar (table 4) but with a higher rate of missing responses per question with the FACT-L (9.1%) than with the EORTC QLQ-C30+LC13 (6.9%). Questionnaire preference was expressed on 386 forms. Although on 291 (75%) occasions patients expressed no preference between the questionnaires, on 75 occasions (19%) they preferred the EORTC QLQ-C30+LC13 questionnaire and on 20 occasions (5%) the FACT-L questionnaire was preferred.

Palliation of symptoms

The analysis of palliation of symptoms was based on patients assessed before treatment and with more than one subsequent assessment during the first 16 weeks after randomisation, or who died during the 16 weeks but with at least one assessment after randomisation. Patients treated with ASC only were compared with those treated with ASC plus any chemotherapy (table 5). There is a suggestion that palliation was better with chemotherapy than without, but there is evidence that clinically useful levels of palliation were achieved with ASC only (28% for chest pain, 11% for breathlessness, 44% for anorexia, and 67% for sweating attacks). It is important to note, however, that the study was not powered to make reliable comparisons between treatment groups. All the patients had WHO performance status grades 0, 1, or 2 before treatment. In both patient groups these levels changed very little during the first 16 weeks (details not shown).

DISCUSSION

This feasibility study has shown that, in the treatment of malignant pleural mesothelioma, randomisation to ASC with or without chemotherapy was sufficiently acceptable to patients and clinicians for a large multicentre randomised trial comparing ASC only versus ASC+MVP versus ASC+N to be feasible. Of the patients eligible for randomisation, 55% were randomised between all three or any two of the treatment groups, half of whom accepted a randomisation option including ASC only. Indeed, intake to a multicentre cancer research UK trial (MS01) in which all patients are being randomised between all three treatment groups is now underway and is currently attracting support from 52 centres. At the time we made our grant application for the phase III trial, 60 centres that expressed an interest estimated that between them they could randomise more than 400 patients per year. Allowing for overestimation, we plan to recruit a total of 840 patients (280 per treatment group) during 4 years, including those already randomised between the three treatment groups in the present feasibility study. The current mean rate of accrual is three patients per centre per year.

Crucial to the study of the possible roles of chemotherapy in the treatment of mesothelioma, a cancer relatively resistant to chemotherapy, is the study of quality of life as reported by patients. Only in this way can the possible benefits and adverse effects be adequately assessed. An important finding from the present feasibility study is therefore the suitability of the EORTC QLQ-C30+LC13 questionnaire for use in patients with pleural mesothelioma, and this instrument is being used in the subsequent MS01 phase III trial.

The study was not powered to make reliable comparisons between the treatment groups, but an important finding is that clinically relevant levels of palliation were achieved with ASC only. It is therefore important to establish whether chemotherapy improves palliation compared with ASC.

After the intake to the MS01 trial had started, results of the randomised trial comparing cisplatin with or without pemetrexed (ALIMTA) in 456 patients were presented at the 2002 Annual Conference of the American Society of Clinical Oncology and appeared to show a survival advantage for pemetrexed + cisplatin over cisplatin alone.¹⁴ The question therefore arose whether intake to the MS01 trial should continue. We decided that it should for the following reasons. Firstly, the pemetrexed trial has only been presented and the findings were not included in the published abstract; details of its design, methods, analyses, and findings are therefore not yet publicly available and have not been peer reviewed. Secondly, pemetrexed is not yet available in the UK and is unlikely to be so for at least 2 years. Thirdly, the data presented showed that the pemetrexed trial design was changed during patient recruitment because of high toxicity and death rates in the pemetrexed arm. Full vitamin B12 and folate supplementation was introduced part way through for all subsequent patients. This reduced the toxicity and death rates and the planned number of patients was increased. The survival difference in the fully vitamin supplemented patients was of borderline significance (p = 0.051) and eight patients were excluded from this analysis (that is, it was not done by intention to treat). Fourthly, the median age was lower (61 years) and the performance status of the patients higher (only ECOG 0–1 eligible) than those typically seen in patients in the UK and, indeed, in the current feasibility study. It is therefore uncertain whether similar results would be achievable in UK practice. Finally, it is questionable whether single drug cisplatin, in a moderately high dose (75 mg/m²) for up to six cycles, is an appropriate control: it achieved, as expected, a low response rate (17% according to non-standard criteria); it gave rise to 6% grade 3 or 4 vomiting and 2% drug related deaths; and it was associated with increasing pain and dyspnoea during treatment. It could well be that, at least in terms of quality of life, ASC as given in the present study is superior, especially in the light of the palliation it achieved.¹⁵ A recent systematic review of published phase II studies concluded that the role of chemotherapy for unresectable malignant mesothelioma is unclear but that the combination of cisplatin and doxorubicin should be considered as a control for randomised trials.¹⁶ This recommendation was, however, based solely on response rates.

Confidential interim analyses from the MS01 trial will continue to be regularly reviewed by its independent data monitoring and ethics committee. Moreover, this committee is also reviewing results of the pemetrexed trial as they emerge and can recommend changes to the MS01 protocol if and when it considers these to be ethically desirable.