Inter-observer variation in outlining of pre-trial test case in the SCOPE1 trial: a United Kingdom definitive chemoradiotherapy trial for esophageal cancer

Gwynne, S., Spezi, E. and Staffurth, J. , et al. (2011) Inter-observer variation in outlining of pre-trial test case in the SCOPE1 trial: a United Kingdom definitive chemoradiotherapy trial for esophageal cancer. International Journal of Radiation Oncology*Biology*Physics, 81 (2), S67-S68. (doi:10.1016/j.ijrobp.2011.06.135).

Abstract

Purpose/objective(s): to assess the variation in target volume outlining of a mid esophagus tumor within a UK dCRT trial.

Materials/methods: the National Cancer Research Institute SCOPE1 trial is a UK Phase 3 RCT of dCRT with capecitabine and cisplatin with or without cetuximab. It is funded by CRUK, sponsored by Velindre NHS Trust and run out of the Wales Cancer Trials Unit. RT trials quality assurance (RTTQA) is being run by the UK's national RTTQA group. A CR UK funded pre-trial RTTQA program was developed in collaboration with the NCRI RT QA group to ensure high quality consistent target delineation. This included a detailed RT protocol, an electronic outlining educational package and a mid esophageal tumor test case that was sent to each investigator to outline. The test case outline of each investigator was assessed by the Chief Investigator by comparison with a pre-defined gold standard (GS), and detailed feedback given, prior to being approved to recruit patients into SCOPE1. GTV was grown to produce a PTV using a 3 cm sup-inf margin and 1.5 cm ant-post and left-right margins. Investigator GTV (iGTV) and PTV (iPTV) outlines were received from 50 investigators from 34 UK centers. The structure sets were imported into CERR to perform a morphometric assessment of outlines against pre defined GS-GTV and GS-PTV, respectively. Analysis was carried out using different metrics including GTV/PTV length and volume. We compared each iGTV and iPTV quantitatively to the GS volumes using the Jaccard conformity index (JCI) and qualitatively.

Results: mean iGTV length was 8.17 cm (range, 6.60-9.60, SD 0.43; GS-GTV 7.8) and mean volume 48.34 cc (29.32-69.90, SD 9.0; GS-GTV 39.15). Mean iGTV-JCI was 0.66 (0.49-0.75) with 24% of outliners achieving a JCI of ≥0.7. Mean iPTV volume was 386.73 cc (298.93-518.64, SD 45; GS-PTV 334.29) and mean iPTV-JCI was 0.79 (0.64-0.86) with 94% of outliners achieving a CI of ≥0.7 (p < 0.0001). All volumes were within acceptable limits with no major protocol deviations seen. Anatomically, areas of outlining variation were inclusion of azygous vein, bronchial tissue, pericardial sac/mediastinal fat or failure to include the whole esophageal wall.

Conclusions: this is the first analysis of variation of GTV outlining within an esophageal cancer RT trial. Our results show that despite a detailed RT protocol and an educational CDROM wide inter-investigator variation in GTV outlining remains, which is partly due to radiological misinterpretation of normal anatomy. We have shown that the PTV outlines show greater conformity to the GS than the GTV and further work will be conducted to see if this is purely due to the growth to PTV. Further training or input from radiologist during outlining may reduce this variation.

This record has no associated files available for download.

Contributors

Author: C. Hurt

Author: G. Griffiths

Download statistics