Hemithoracic Irradiation With Proton Therapy in Malignant Pleural Mesothelioma

Study design: Randomised phase III clinical trial for patients with unilateral MPM.

Primary endpoint: Progression free survival (PFS) and overall survival (OS), defined as the time from randomisation to the date of progression and death from any cause.

Secondary Endpoints: Safety and Tolerability, Health related Quality of Life (QOL): EuroQoL EQ-5D-3L, Locoregional Control.

Randomisation and stratification: 1:1 randomisation. Patients with be stratified for histology (epithelioid versus non-epithelioid), potential PBT centre (UCLH or The Christie)

, laterality (left or right sided) and time since diagnosis (<1 year or > 1 year)

Treatment:

Experimental Arm: Patients in the experimental arm will receive PBT to the hemithorax to a dose of 50Gy in 25 fractions with a boost to 60Gy for the visible tumour (gross tumour volume-GTV). Treatment is given daily Monday-Friday over 5 weeks. Following completion of treatment in the experimental arm patients will have 2 years of follow-up from time of randomisation at the local recruiting/referring centre.

Control Arm:

The patients in the control arm would be under standard of care surveillance i.e. "watch and wait", with no treatment or other intervention. Patients will have 2 years of follow-up from time of randomisation at the local recruiting/referring centre. If the disease progresses, the patient will receive SOC treatment i.e. immunotherapy with nivolumab and ipilimumab, or chemotherapy at the clinician's discretion.

Statistical analysis plan:

The sample size is 148 patients (74 patients per arm). This is to detect a OS hazard ratio of 0.58, equivalent to an improvement in 2-year OS from 30% to 50%, with 85% power and 5% two-sided alpha. Recruitment to complete in 3 years across 20 UK centres with 2 years of additional follow-up and up to 5% dropout. Interim analyses for OS efficacy will be performed when 50, 75 and 110 patients have been randomised at around 1.5, 2.0 and 2.5 years respectively. Using a fixed-sequence approach, a difference for OS will only be tested if the co-primary endpoint of PFS is statistically significant (p<0.05); N=148 will provide >85% power to detect a PFS hazard ratio of 0.58 accounting for up to 10% dropout.