External validation of the lumbosacral plexus‐contouring protocol developed by Yi et al. (IJROBP 2012; 84: 376–82) for pelvic malignancies

Abstract

To evaluate interobserver variability in contouring lumbosacral plexuses (LSP) using the protocol developed by Yi et al. (IJROBP 2012; 84: 376-82) and to review LSP dosimetries for conventional radiotherapy and intensity-modulated radiotherapy (IMRT) for pelvic cancers. Using the above-mentioned protocol, seven outliners independently contoured LSPs of 10 consecutive patients (five patients treated with conventional radiotherapy and five with IMRT). Interobserver variability was reviewed visually by using planning axial CT images and anteroposterior digitally reconstructed radiographs. Dosimetries of LSPs were also calculated and compared. There was a notable learning curve for each outliner; duration to outline the first patient was 45-185 minutes, versus 15-50 minutes after six patients. We found significant interobserver variability among outliners below the level of the S2 nerve roots. The LSP volumes (mean volume range of 40.9-58.4 cc) were smaller than those described in the atlas paper (71-138 cc). The mean values of mean dose, maximum dose, V40 Gy, V50 Gy and V55 Gy, respectively, for patients treated with conventional radiotherapy versus those treated with IMRT were 35.5 Gy versus 33.6 Gy, 52.2 Gy versus 52.2 Gy, 61.3% versus 54.4%, 14.9% versus 18.8% and 0% versus 2.5%. We conclude that the protocol developed by Yi et al. is a useful set of guidelines but suggest that additional at-risk components of the LSP also be contoured. We recommend that radiation oncologists practise 'nerve-sparing' radiotherapy by contouring LSPs, especially when using IMRT. We propose the term 'lumbosacral plexus regions' (LSPRs) to highlight the fact that LSPs are not always radiologically visible, only the regions where they are likely to be present.