Organ motion in pancreatic radiotherapy Quantifying motion for pancreatic radiotherapy margin calculation Gillian Whitfield a,1 , Pooja Jain a,1 , Melanie Green a , Gillian Watkins a , Ann Henry a , Julie Stratford b , Ali Amer c , Thomas Marchant c , Christopher Moore c , Patricia Price d,⇑ a Academic Radiation Oncology, The University of Manchester; b Wade Centre for Radiotherapy Research; c Christie Medical Physics and Engineering, The Christie Hospital NHS Foundation Trust, Manchester, UK; d Department of Surgery and Cancer, Hammersmith Hospital, London, UK article info Article history: Received 31 January 2011 Received in revised form 13 February 2012 Accepted 17 February 2012 Available online 10 March 2012 Keywords: Pancreatic radiotherapy Organ motion Cone-beam CT Treatment margins abstract Background and purpose: Pancreatic radiotherapy (RT) is limited by uncertain target motion. We quanti- fied 3D patient/organ motion during pancreatic RT and calculated required treatment margins. Materials and methods: Cone-beam computed tomography (CBCT) and orthogonal fluoroscopy images were acquired post-RT delivery from 13 patients with locally advanced pancreatic cancer. Bony setup errors were calculated from CBCT. Inter- and intra-fraction fiducial (clip/seed/stent) motion was deter- mined from CBCT projections and orthogonal fluoroscopy. Results: Using an off-line CBCT correction protocol, systematic (random) setup errors were 2.4 (3.2), 2.0 (1.7) and 3.2 (3.6) mm laterally (left–right), vertically (anterior–posterior) and longitudinally (cranio- caudal), respectively. Fiducial motion varied substantially. Random inter-fractional changes in mean fidu- cial position were 2.0, 1.6 and 2.6 mm; 95% of intra-fractional peak-to-peak fiducial motion was up to 6.7, 10.1 and 20.6 mm, respectively. Calculated clinical to planning target volume (CTV–PTV) margins were 1.4 cm laterally, 1.4 cm vertically and 3.0 cm longitudinally for 3D conformal RT, reduced to 0.9, 1.0 and 1.8 cm, respectively, if using 4D planning and online setup correction. Conclusions: Commonly used CTV–PTV margins may inadequately account for target motion during pan- creatic RT. Our results indicate better immobilisation, individualised allowance for respiratory motion, online setup error correction and 4D planning would improve targeting. Ó 2012 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 103 (2012) 360–366 Radiotherapy (RT) for pancreatic cancer is currently limited by target volume movement and the associated risk of excessive nor- mal tissue irradiation, thereby inhibiting dose escalation. The upper abdomen is difficult to immobilise and the pancreas is sub- ject to motion induced by internal organ variability, principally respiration [1–5]. Physiological changes during RT may also impact target motion during a RT course. To compensate for unquantified geometric uncertainties in tar- get volume position, generic margins are commonly applied to the clinical target volume (CTV) to form a planning target volume (PTV). To support increasingly conformal RT techniques that facili- tate dose escalation, quantification of pancreatic motion and patient positioning is required. Previous studies have provided some information regarding pancreatic/tumour movement [1,4–10], with only limited data derived from studies carried out during RT [5,9,11–14]. Only one prior study has used cone-beam CT (CBCT)/ kilovoltage (kV) X-ray imaging [13], and none have comprehen- sively investigated patient motion plus inter- and intra-fraction tar- get motion, for objective margin calculation. We used on-treatment CBCT and orthogonal kilovoltage (kV) fluoroscopy to quantify 3-dimensional (3D) target motion in pa- tients undergoing pancreatic RT, allowing subsequent calculation of treatment margins to incorporate the observed group motion. Materials and methods This was a prospective observational imaging study, approved by the South Manchester Local Research Ethics Committee (03/SM/ 554). The primary objectives were to quantify daily setup errors and inter- and intra-fraction motion of pancreatic fiducials to assess target motion during RT. The secondary objective was to use the data obtained to calculate CTV–PTV margins for radical RT. Patients with locally advanced pancreatic cancer due to undergo radical che- mo-RT treatment were eligible. Good performance status and abil- ity to comply with the imaging protocol were prerequisite. Conformal RT was planned and delivered to patients positioned supine on a lung board, with arms immobilised above the head. Following oral gastrograffin, a free-breathing helical planning CT scan, with a 5 mm slice interval, was acquired in the treatment po- sition from L4/5 to just above the dome of the diaphragm. CT data were transferred to a Pinnacle™ (Philips Medical systems, Wiscon- 0167-8140/$ - see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2012.02.012 ⇑ Corresponding author. Address: Department of Surgery and Cancer, Imperial College Healthcare NHS Trust, Hammersmith Hospital, Du Cane Road, London W12 ONN, UK. E-mail address: p.price@imperial.ac.uk (P. Price). 1 Joint first authors. Radiotherapy and Oncology 103 (2012) 360–366 Contents lists available at SciVerse ScienceDirect Radiotherapy and Oncology journal homepage: www.thegreenjournal.com