Modelling of dose-escalated proton beam therapy for Locally Advanced Pancreatic Cancer using simulated phantoms

This study explores the feasibility of dose-escalated proton beam therapy (dPBT) for Locally Advanced Pancreatic Cancer (LAPC) patients by modelling common patient scenarios using current clinically-adopted practices.

Five patient datasets were used as simulation phantoms, each with six tumour sizes, to systematically simulate treatment scenarios typical in LAPC patients. Using the Raystation treatment planning system, robustly-optimised dPBT and stereotactic ablative radiotherapy (SABR) treatment plans were created with a 5 mm margin allowing for intra- and inter-fraction anatomical changes, following clinically-adopted protocols. Safe dose-escalation feasibility is assessed with dose metrics, tumour control (TCP) and normal tissue complication probabilities (NTCP) for average and worst-case intra-fraction motion scenarios. Significance testing was performed using a paired student's t-test.

Dose-escalation feasibility is largely dependent on tumour size and proximity to critical structures. Minimal therapeutic benefit was observed for patients with >4.5cm tumours, however for tumours ≤4.5cm dPBT TCPs of 45%-90% compared to SABR TCPs of 10%-40% (p<0.05). The worst-case scenario dPBT TCP was comparable to SABR. Hypofractioned dPBT further improved this result to >90% (p<0.05) for tumours ≤4.5cm.

Safe dPBT is feasible for patients with targets up to 4.5 cm in size and see a significant therapeutic benefit compared to the current standard of care in SABR. A patient-specific approach should be taken based on tumour size and surrounding anatomy.