Oral Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2017

Improving radiotherapy quality through collaboration (#96)

Colin Hornby 1 , Amanda Smith 2 , Pete Griffin 3 , Vanessa Panettieri 3 , Prabhakar Ramachandrin 2 , Phillip Moloney 4 , Flavio Nelli 4 , Adam Chapman 1
  1. Victorian Dept of Health & Human Services, Melbourne, VIC, Australia
  2. Peter MacCallum Cancer Centre, Moorabbin, Victoria
  3. William Buckland Radiotherapy Centre, Prahran, Victoria
  4. Andrew Love Cancer Centre, Geelong, Victoria

Aim:

In 2016 the Victorian Department of Health & Human Services facilitated the purchase of knowledge based planning (KBP) software. This software produces and uses models constructed from previously treated patient radiotherapy plans to predict and then generate optimised treatments by utilising the knowledge of what is possible embedded in those original plans.

The aim of this project was to enable a collaboration of Victorian public radiotherapy providers to develop tumour stream specific models from pooled data for implementation at all participating sites to reduce plan variation and improve overall plan quality.

Methods:

8 public rural and metropolitan radiotherapy departments participated in the Victorian KBP project. Radiation therapists and medical physicists from each site were trained in model building. Tumour specific models for Prostate and Anal cancers were constructed using Victorian patient data sourced from all sites.

Department specific maximum organ-at-risk doses were ascertained and harmonised to ensure the models would be able to produce clinically acceptable plans. A separate set of previously treated cases was used to validate that a model could produce equivalent or better quality treatment plans compared to the original.

Results:

Data from KBP produced treatment plans were compared to the original clinically implemented plans. Overall, organ-at-risk doses were equivalent or lower when produced with the KBP software. In prostate the average dose reduction for bladder V50 was 2.2Gy and V60 was 0.80Gy while the average dose reduction for rectum V50 was 1.50Gy and V75 was 0.6Gy, however the dose reductions did vary according to department.

Conclusion:

This project has successfully implemented the same KBP models for Prostate cancer across 8 collaborating radiotherapy sites enabling the same plan to be produced for a given patient data set irrespective of treatment location. On average the KBP plans produced were of higher quality than those previously locally developed.