Radiological work‐up for proton beam therapy of uveal melanoma patients

Abstract

AbstractPurposeIn ocular proton beam therapy (PBT), 2.5 mm tantalum clips are surgically sutured near the tumour border for clinical target volume marking. The clip‐tumour distances are assessed peroperative using transillumination, indentation and or fundoscopy. To support the clip localisation, target definition and 3D tumour modelling, a dedicated MRI‐protocol was developed and evaluated.Methods23 consecutive uveal melanoma (UM) patients, planned for PBT, were scanned on a 3T MRI. MRI based clip‐tumour distances were compared with the peroperative measurements. For clips with more than 1mm difference, the most likely origin of this discrepancy was discussed in a multidisciplinary setting.ResultsIn 39/87 (45%) of the clips, both measurements differed more than 1 mm. For the majority of these clips, the MRI showed a larger clip‐tumour distance (n = 30). Largest differences were observed in flat UM (n = 13) where an average absolute difference of 3.9 mm was found. Here, MRI was generally considered less reliable as the extend of flat tumours is difficult to determine on MRI. In 15% of the clips, the discrepancy was attributed to a complex tumour geometry. Here an average absolute difference of 3.2 mm was observed. In these cases 3D imaging provided valuable additional information for the treatment planning. In 10% of the clips, the anterior localized tumour casted a shadow, resulting in an overestimation of the tumour dimensions during surgery, leading to an average absolute difference of 3.7 mm. For these patients, MR‐imaging was considered more reliable as obtaining correct peroperative measurements was difficult.ConclusionsMRI and peroperative measurements were equal in 55% of the cases. For flat UM, MRI generally underestimates tumour dimensions. For complex and anterior UM, MRI‐based clip‐tumour measurements are more reliable, which could result in more accurate target definition, thereby possibly reducing toxicity and improving probability to retain vision.