ESTRO 37 Abstract book

ESTRO 37

S598

comparable with those of PTV-CT for the majority of cases studied.

observer. Median CIs among all ob are 0.45, 0.54 and 0.74 for CT, T2w and DWI, respectively. The average of the median values for pairs of ob are 0.62, 0.53 and 0.73 respectively for the three modalities. There is a statistically significant difference between the CIs (p=0.007) and also the relationship between CIs and the modalities tends to significance (p=0.062). Mean difference of urinary bladder volume between CT and MRI was 25 cm 3 but this did not affect CIs. Conclusion MRI, mostly DWI series, showed high agreement among different ob and therefore it could be useful for the definition of GTV in LARC. Studies with a larger number of pts are needed to confirm whether such agreement is significant. The best imaging technique for the definition of the true tumor volume is still to be determined. PO-1066 Delineation uncertainty and parotid gland doses and estimated NTCPs in head and neck proton therapy G. Engeseth 1 , M. Brydøy 1 , J. Dale 1 , C. Boer 1 , H. Gripsgård 1 , J. Moi 1 , L. Muren 2 , C. Stokkevåg 1 1 Haukeland University Hospital, Department of Oncology and Medical Physics, Bergen, Norway 2 Aarhus University Hospital, Department of Medical Physics, Aarhus, Denmark Purpose or Objective Proton therapy may be beneficial for head and neck cancer patients, mainly due to the potential reduction in doses to organs at risk (OAR). Late side effects like xerostomia have been associated with parotid gland (mean) dose. Treatment plans are typically optimised to spare the parotid gland and therefore depends on accurate volume delineations. The aim of this study was to investigate the influence of observer variation in parotid gland delineation on the parotid gland dose, as well as the estimated normal tissue complication probabilities (NTCPs) for xerostomia in intensity modulated proton therapy (IMPT) plans. Material and Methods Six observers (two physicians and four RTTs) delineated the parotid glands according to consensus guidelines (Radiother Oncol 2015; 117:83-90) in five head and neck cancer patients. Delineation variation was evaluated using two different quantities. The Concordance Index (CI) was defined as the ratio between the intersection and the union of two parotid gland volumes; a reference volume (Observer 1) and the corresponding parotid gland volume from each of the observers. We also calculated the OAR ratio, defined as the ratio between the median parotid gland volume per observer and the median parotid gland volume across all observers. IMPT plans using a simultaneous integrated boost technique were generated using the volume definitions from Observer 1. The prescribed dose was 56-70 Gy(RBE).The mean doses for left and right parotid glands from each observer were calculated as well as the corresponding NTCPs for patient rated xerostomia, using a published NTCP model (Radiother Oncol 2012;105:101-106). Results The inter-observer variation largely varied between patients, resulting in differences between parotid gland delineations ranging from 5 to 10 cm 3 for the left parotid gland, and from 5 to 13 cm 3 for the right parotid gland (Figure 1). The OAR ratios ranged from 0.92 to 1.17, while the CI ranged from 0.66 and 0.82 (Table I). The differences in the mean parotid gland dose due to delineation variation, ranged from 0.7 to 4.7 Gy(RBE) for

PO-1065 Role of CT-MRI co-registration in tumor delineation for preoperative radiotherapy of rectal cancer E. Palazzari 1 , M. Lupattelli 2 , C. Galuppo 3 , M. Napoletano 4 , A. Podlesko 5 , V. Bini 6 , M. Iacco 7 , C. Fulcheri 7 , G. Verzini 2 , A. De Paoli 1 , C. Aristei 8 1 Centro di Riferimento Oncologico IRCCS Aviano, Radiation Oncology Department, Aviano, Italy 2 Santa Maria della Misericordia Hospital, Radiation Oncology Department, Perugia, Italy 3 Santa Maria della Misericordia Hospital, Radiology Department, Perugia, Italy 4 Perugia University, Radiology Department, Perugia, Italy 5 Perugia University, Radiation Oncology Department, Perugia, Italy 6 Perugia University, Internal Medicine- Endocrine and Metabolic Science Section, Perugia, Italy 7 Santa Maria della Misericordia Hospital, Medical Physics Department, Perugia, Italy 8 Santa Maria della Misericordia Hospital and Perugia University, Radiation Oncology Department, Perugia, Italy Purpose or Objective Intensification of preoperative radio-chemotherapy in locally advanced rectal cancer (LARC) is gaining progressive interest with a view to increasing response rates and to developing new organ preservation approaches. Dose escalation programs are providing promising results but they require optimization of boost volume definition. T2 weighted (T2w) magnetic resonance (MRI) imaging is the gold standard for rectal cancer staging and the diffusion-weighted sequence (DWI) allows for the best definition of the tumor. We compared the rectal gross tumor volume (GTV) delineated on computed-tomography (CT), T2w MRI and DWI MRI images to evaluate inter- and intra-observer agreement. Material and Methods LARC patients (pts) underwent CT without contrast medium for radiotherapy treatment planning and subsequently T2w and DWI (b = 1500 s / mmq) MRI axial sequences. CT and MRI were acquired in the prone position and pts had the same bladder preparation for both procedures. Rigid CT-MRI co-registration of image series was obtained. Four independent observers (ob) with different skill levels and experience delineated the GTV on CT, T2w and DWI MRI images. Conformity index (CI) was calculated between each ob-pair per patient per technique and between each technique-pair per patient per ob. Reliability between techniques and between ob was assessed using the intraclass correlation coefficient (ICC). Comparisons within and between groups were analyzed by mixed repeated-measures ANOVA, with Geisser-Greenhouse adjustments for non-sphericity with Bonferroni’s post hoc correction. Results Twenty consecutive pts were included. CT, T2w MRI and DWI MRI mean GTV in cm 3 were 43.01, 34.98 and 36.10 respectively for ob one; 42.88, 34.39 and 41.89 for ob two; 47.20, 36.39 and 36.70 for ob three; 40.41, 33.67 and 34.15 for ob four. ICC among ob for CT, T2W MRI and DWI MRI volumes were 0.76, 0.81 and 0.94 respectively and this reliability were confirmed by the comparison between ob-pairs. Volumes are significantly related to imaging technique (p=0.0003) with CT volumes larger than T2w and DWI and they do not depend on the