First Results of COSTAR (CRUK/08/004): A randomised trial of 3-Dimensional Conformal Radiotherapy (3DCRT) vs Cochlea-Sparing Intensity Modulated Radiotherapy (CS-IMRT) in patients with parotid cancer.


Session type:

Christopher Nutting1,James Morden2,Matthew Beasley3,Shree Bhide1,Marie Emson4,Mererid Evans5,Lydia Fresco6,Dorothy Gujral1,Kevin Harrington1,Catherine Lemon7,Rekha Neupane8,Kate Newbold1,Robin Prestwich9,Martin Robinson10,Paul Sanghera11,Muthiah Sivaramalingam12,Mark Sydenham4,Emma Wells1,Steph Witts4,Emma Hall4
1Royal Marsden Hospital,2The Institute of Cancer Research,3Bristol Haematology and Oncology Centre,4Institute of Cancer Research,5Velindre Hospital,6University Hospitals Coventry and Warwickshire NHS Trust,7Mount Vernon Hospital,8Glan Clwyd Hospital,9St James Institute of Oncology,10Sheffield Teaching Hospitals NHS Foundation Trust,11Queen Elizabeth Hospital, Birmingham,12Royal Preston Hospital



30-50% of patients receiving post-operative RT for parotid cancer experience ipsilateral hearing loss. IMRT can reduce radiation dose to cochlea. COSTAR investigated the role of IMRT in reducing hearing loss in these patients.


Patients with histologically confirmed carcinoma of the parotid gland (pT1-4, pN0-3, M0) were randomized 1:1 to receive CT-planned 3DCRT or CS-IMRT. 60Gy (R0 resection) or 65Gy (R1-2) in 30 fractions were delivered over 6 weeks. Treatment allocation used minimisation, balancing for centre and planned RT dose. The primary endpoint was proportion of patients with hearing loss in the ipsilateral cochlea of ≥10dB measured by bone conduction at 4000Hz 12 months (m) after RT; compared between randomized groups by an exact test (α=0.05). Secondary endpoints (α=0.01) included hearing loss at 6 and 24m, vestibular function, acute and late toxicity, patient reported quality of life (including Glasgow Hearing Aid Benefit Profile (GHABP), time to tumour recurrence and survival.


110 patients (54 3DCRT; 56 CS-IMRT) were randomised between 2008 and 2013 from 22 UK centres. 99 (90%) patients were R1-2 (47 3DCRT; 52 CS-IMRT). Mean dose to the ipsilateral cochlea was 56.2Gy for 3DCRT and 35.7Gy for CS-IMRT, (p<0.001). 66/110 (60%) patients were evaluable for the primary endpoint. At 12ms, a loss of ≥10dB in ipsilateral bone conduction was observed in 14/35 (40%) 3DCRT and 11/31 (36%) CS-IMRT patients (p=0.80). No statistically significant differences in bone or air conduction were observed at 6 or 24m after RT nor for any GHABP initial disability or handicap subscales, vestibular function, acute or late toxicity, overall quality of life, time to tumour recurrence or survival.


IMRT reduced the radiation dose below the accepted tolerance of the cochlea. This did not lead to a statistically significant reduction in the proportion of patients with hearing loss in the ipsilateral ear at 12 months after RT.