Objectives

Chemotherapy and surgery achieve a 5-year event-free survival of 60-70% in localized osteosarcoma, but little additional progress has been made since the 1980s. All recent large randomized trials in localized osteosarcoma have failed to further improve long-term outcomes.

Clinical research in osteosarcoma is hampered by its rarity and by a limited pipeline of new agents. Globally, only 4 randomized trials with a survival endpoint are currently recruiting patients with localized osteosarcoma (last verified on 12 June 2018): 3 in China and 1 in Japan.

To accelerate the pace of clinical research in localized osteosarcoma, we proposed the establishment of a trial infrastructure (platform) which combines two concepts:

• A trial design allowing the addition of new arms and the removal of arms for futility, within the same trial;
• The testing of drugs already approved for other indications in 2 phases, a screening phase & a confirmation phase.

We have named the proposed trial PRESTO (Platform for the Rapid Evaluation of Several Treatments in Osteosarcoma). Our objective was now to gather feedback about the necessary criteria and the main issues to be overcome to be able to conduct such a multi-arm multi-stage (MAMS) platform trial in localized osteosarcoma.

Methods

A successful implementation of a MAMS platform trial is dependent on five key criteria:

1. The feasibility of sufficient patient accrual
2. A surrogate endpoint to guide the selection of first phase interventions
3. Interventions with a positive risk/benefit ratio for a population undergoing curative treatment
4. The existence of a network of osteosarcoma centres
5. The availability of large scale funding

An initial assessment of these criteria was performed.

Results

For criteria 1, we summed up the number of patients randomized in trials performed by various collaborative groups in localized osteosarcomas and found a capacity of randomizing 340 patients a year, within EURAMOS countries, France, Italy & Latin-America.

For criteria 2, we identified possible surrogate candidates for each of the key treatment period:

• good histological response (GHR) for the neoadjuvant period,
• circulating tumour cells for the perioperative period,
• early DFS for both the adjuvant and the maintenance period.

Using a frequentist approach, aiming at improving GHR from 50% to 70% would require 53 patients per arm for a 80% power and a relaxed alpha of 20%. Another 214 patients in each arm would need to be included to evaluate the effect of the intervention on EFS with a power of 80%, an alpha of 5% and a target HR of 0.70, assuming accrual over 3 years and 3 years of additional follow-up.