7th ICHNO Abstract book

7th ICHNO 7 th ICHNO Conference International Conference on innovative approaches in Head and Neck Oncology 14 – 16 March 2019 Barcelona, Spain __________________________________________________________________________________________ page 65

PO-125 Visibility study for GTVp delineation for Head and Neck following recent consensus guidelines S. Whittam 1 , B. Al-Qaisieh 1 , R. Prestwich 2 , K. Cardale 2 , S. Ramasamy 2 , P. Murray 2 , K. Dyker 2 , M. Sen 2 1 Leeds Cancer Centre NHS, Medical Physics, Leeds, United Kingdom; 2 Leeds Cancer Centre NHS, Clinical Oncology, Leeds, United Kingdom Purpose or Objective Delineation of treatment volumes is one source of uncertainties in radiotherapy. New international guidelines published in 2017 provide a consensus on outlining primary clinical target volume (CTVp) for laryngeal, hypopharyngeal, oropharyngeal and oral cavity squamous cell carcinoma. The new guidelines give recommendation on methods of creating and editing CTVp. Clinical implementation of the new guidelines has the potential to reduce CTVp margin without impact on outcome, while reducing toxicity and improving quality of life for patients. Accuracy in delineation of primary gross tumour volume (GTVp) is a key step for accurate delineation of CTVp. Accordingly prior to implementation of new guidelines locally, it is important to quantify the intra-observer variability for GTVp delineation. Material and Methods Six experienced head and neck oncologists independently segmented GTVp on CT data sets of four patients with oropharynx, larynx, or hypopharynx cancer. These patients were previously treated for radical radiotherapy and the clinically treated GTVp was taken as the gold standard volume. Data analysis was done on ImSim QA software with the following comparison metrics used: Volume difference, Centre of gravity distance (CGD), DICE index, mean distance to conformity (MDC), sensitivity and inclusion indices. Results Significant differences in GTVp were observed between oncologists. Six of the 24 outlined GTVp’s were less than 10% volume difference from the gold standard with only 5 of these less than 5% volume difference. The largest volume difference was -47.3% which corresponds to an absolute volume difference of -3cm 3 . Eleven out of the 24 patients have MDC within 4mm. Seventeen of the 24 GTVp’s outlined have a CGD within 5mm of the gold standard but only 5 of these are within 2mm. Ten out of 24 of the drawn GTVp had a DICE coefficient of 0.80 or better. Eleven had a DICE coefficient of between 0.70 and 0.80, leaving just 3 with DICE coefficients less than 0.70 with the lowest DICE being 0.60. Eighteen out of 24 of the GTVp’s had larger sensitivity index than inclusion index indicating more over- contouring than under-contouring.6 out of 24 had a difference in sensitivity and inclusion index less than 0.10. Eight out of 24 had a sensitivity-inclusion difference between 0.10 and 0.20, and 6 had a sensitivity-inclusion difference above 0.20 with the highest difference being 0.42. Conclusion There was low conformity for GTVp between different clinicians. The result from this study highlights the importance of careful peer review of GTV contours. PO-126 Predictors of dose differences to swallowing OARs in patients undergoing radiotherapy for HNC. D. Noble 1 , K. Harrison 2 , M. Wilson 3 , A. Hoole 4 , S. Thomas 4 , N. Burnet 5 , R. Jena 1

1 University of Cambridge, Oncology, Cambridge, United Kingdom; 2 University of Cambridge, Cavendish Laboratory, Cambridge, United Kingdom; 3 University College London, Medical Physics and Biomedical Engineering, London, United Kingdom; 4 Cambridge University Hospitals NHS Foundation Trust, Medical Physics and Biomedical Engineering, Cambridge, United Kingdom; 5 University of Manchester, Manchester Academic Health Science Centre, Manchester, United Kingdom Purpose or Objective Toxicity following radical radiotherapy (RT) for HNC is a major clinical problem. Inter-fraction anatomical change and weight loss (WL) can lead to differences between planned (D P ) and delivered (D A ) OAR dose - thus changing toxicity risk. Understanding and predicting such dose differences can guide Adaptive RT strategies (ARTS), and mitigate this risk. Material and Methods All patients with HNC, recruited to VoxTox study (UK CRN ID 13716) were eligible. All were treated on TomoTherapy with daily IG; patients with complete kVCT and all daily MVCT datasets were included in this sub-study. Swallowing OARs (Ips & contralateral parotids – IPG & CPG, Ips & contralateral submandibular glands – ISMG & CSMG, superior & middle pharyngeal constrictors – SPCs/MPCs, oral cavity – OC and supraglottic larynx – SGL) were segmented on kVCT planning scans, according to published atlases. In-house software (CheckTomo) calculated individual dose cubes on each kV and MVCT scan. A deformable image registration (Elastix), was trained and validated following consensus guideline TG132. Defined transformation parameters permitted dose accumulation to a final D A DVH, for comparison with D P (mean doses reported). Patient data on hypothesised predictors (HPs) of dose differences, including T-stage (T0-2 vs T3-4), N- stage (N0-1 vs N2+), WL (pre-RT vs final week), and primary disease site (PDS) were collected. Univariate analysis was undertaken to examine relationships between HPs and dose differences; Bonferonni corrections were used. Results 172 patients had full datasets. PDS were: 111 oropharynx, 20 hypopharynx/larynx, 17 salivary gland/sinus/skin, 14 oral cavity, 5 nasopharynx, 5 unknown primary. Mean WL was 6.9kg (7.7%). Mean dose differences (D A – D P , 95% CI) were as follows: IPG 1.05Gy (0.85-1.25), CPG 0.51Gy (0.33-0.69), ISMG 1.01Gy (0.82-1.20), CSMG 1.02Gy (0.86- 1.19), SPC 1.02Gy (0.90-1.14), MPC 1.03Gy (0.85–1.22), OC -0.16Gy (-0.31-0), SGL 1.01Gy (0.78-1.25). Figure 1 shows comparisons of D A and D P for each structure.