Session Item

Prostate radiotherapy treatment planning optimally involves acquisition of a CT and MR in the radiotherapy planning position. MR-only radiotherapy planning exploits the benefits of MRI soft tissue delineation, whilst negating the registration inaccuracies caused by using both MRI and CT for planning and reduces the number of hospital visits. Fiducial markers have conventionally been used in prostate radiotherapy to reduce on-treatment image matching variability. However, this is an invasive procedure for the patient, and presents technical difficulties in an MR-only pathway as fiducial markers are difficult to visualise on MRI. To the best of our knowledge, this is the first study to compare inter-observer variability of soft-tissue matching to fiducial matching in an MR-only prostate pathway.

Four therapeutic radiographers with MR prostate CBCT matching experience reviewed first fraction CBCTs for 25 patients. The CBCT was compared to a T2 weighted multi-echo sequence for the soft tissue match, the CBCT was compared to a T1 weighted MRI  sequence for the fiducial match. The MRI sequences were acquired in the same session. Fiducial markers were not visible on T2 MR sequence and radiographers performed the soft-tissue match first, ensuring the fiducial marker match did not influence the soft-tissue match. Inter-observer variability was quantified using the inter-observer error (mean of the per-patient standard deviation in radiographer matches) and 95% limits of agreement from a modified Bland-Altman analysis. Accuracy of the soft tissue match was quantified by calculating the difference from the fiducial match.

Limits of agreement on the MR soft tissue match were 0.15 mm, 0.40 mm, 0.35 mm and fiducial match 0.25 mm, 0.36 mm, 0.25 mm (lateral, longitudinal, vertical). Inter-observer error (±standard deviation) on the MR soft tissue match were 0.06(±0.05) mm, 0.11(±0.40) mm, 0.07(±0.35) mm and fiducial match 0.07(±0.11) mm, 0.11(±0.15) mm, 0.08(±0.07) mm (lateral, longitudinal, vertical). The difference of the soft tissue match from the fiducial match was 0.03(±0.11) mm, 0.01(±0.27) mm, 0.01(±0.19) mm (lateral, longitudinal, vertical). A paired t-test showed p = .072 , p = .73, p = .66 (lateral, longitudinal, vertical). The lateral p-value is approaching statistical significance with the soft tissue match showing less variation than the fiducial marker match.

In this study there was no statistically significant difference between soft tissue matching and fiducial matching in prostate radiotherapy and suggest the routine use of fiducial markers is reviewed. Soft tissue image matching removes the need of an invasive procedure for patients and this data supports extending MR-only radiotherapy pathways to other pelvic tumour sites.