Patient position verification in magnetic-resonance imaging only radiotherapy of anal and rectal cancers

Abstract

Background and Purpose: Magnetic resonance (MR)-only treatment pathways require either the MR-simulation or synthetic-computed tomography (sCT) as an alternative reference image for cone beam computed tomography (CBCT) patient position verification. This study assessed whether using T2 MR or sCT as CBCT reference images introduces systematic registration errors as compared to CT for anal and rectal cancers. Materials and Methods: A total of 32 patients (18 rectum,14 anus) received pre-treatment CT-and T2 MRsimulation. Routine treatment CBCTs were acquired. sCTs were generated using a validated research model. The local clinical registration protocol, using a grey-scale registration algorithm, was performed for 216 CBCTs using CT, MR and sCT as the reference image. Linear mixed effects modelling identified systematic differences between modalities. Results: Systematic translation and rotation differences to CT for MR were − 0.3 to + 0.3 mm and − 0.1 to 0.4 • for anal cancers and − 0.4 to 0.0 mm and 0.0 to 0.1 • for rectal cancers, and for sCT were − 0.4 to + 0.8 mm, − 0.1 to 0.2 • for anal cancers and − 0.6 to + 0.2 mm, − 0.1 to + 0.1 • for rectal cancers. Conclusions: T2 MR or sCT can successfully be used as reference images for anal and rectal cancer CBCT position verification with systematic differences to CT <±1 mm and <±0.5 • . Clinical enabling of alternative modalities as reference images by vendors is required to reduce challenges associated with their use.

Cone beam computed tomography image guidance within a magnetic resonance imaging-only planning workflow

“…Other studies have assessed MRI-CBCT matching using automatic registration algorithms. Edmund et al assessed automatic registration for prostate with a resulting 2 mm difference amongst the matching cohort, while Bird et al, assessed a single observer automatic registration for rectal and anal canal regions [17] , [24] . Bird et al, reported a < 1 mm difference and < 0.4° in automatic registration differences [17] .…”

“…Edmund et al assessed automatic registration for prostate with a resulting 2 mm difference amongst the matching cohort, while Bird et al, assessed a single observer automatic registration for rectal and anal canal regions [17] , [24] . Bird et al, reported a < 1 mm difference and < 0.4° in automatic registration differences [17] . While the reported difference was lower for Bird et al, than in this study, the use of automatic registration is not necessarily reflective of clinical practice, as clinical decisions and adjustments to the image matching need to be made for anatomical changes daily, therefore, these studies do not take into account the differences imparted from MRI image interpretation in image guidance.…”

“…There is limited work around MRI to CBCT image guidance, with publications mostly focusing on sCT-CBCT image guidance [13] , [14] , [15] , [16] . Two studies, which focused on MRI-CBCT image guidance, have assessed the soft tissue match accuracy for prostate treatments, while another study focused on anal canal and rectum, utilising only one matcher with automatic registration [17] , [18] , [19] . While the results of these studies appear promising, additional work on this topic has identified that a training package reduces the inter-observer variability in MRI-CBCT matching, to align with the variability of CT-CBCT matching values [20] .…”