B108: Dose-escalated radiotherapy for non small cell lung cancer: association between heart dosimetry and survival in IDEAL-CRT

Sindu Vivekanandan1,Nicholas Counsell2,Emma Parsons3,Yenting Ngai2,Laura Hughes2,Maria Hawkins1,David Landau4,John Fenwick1

1University of Oxford, Oxford, UK,2Cancer Research UK & UCL Cancer Trials Centre, London, UK,3NCRI Radiotherapy Trials Quality Assurance Group, Northwood, UK,4Kings College London and University College London, London, UK

Presenting date: Tuesday 3 November
Presenting time: 13.10-14.00

Background

Radiotherapy dose-escalation trials have achieved inconsistent overall survival (OS) levels for non-small cell lung cancer (NSCLC). RTOG-0617 reported poorer OS for 74Gy than for 60Gy. IDEAL-CRT, a phase I/II isotoxic dose-escalation trial (Sponsor:UCL & funder:CRUK(C13530/A10424)) delivered mean and maximum doses of 67.5Gy and 73Gy in 30 fractions over 40 days, equivalent to 69Gy and 75.6Gy in 2Gy fractions (EQD2) and reported better OS. We investigate associations between OS, and whole heart and left ventricle (LV) dosimetry.

Method

For 80 of 82 IDEAL-CRT patients, whole heart and LV dose volume histograms (DVHs) were extracted using Computational Environment for Radiotherapy Research (CERR). Since prescribed doses-per-fraction varied, physical DVHs were converted to EQD2s using linear-quadratic equation with ?/?=3Gy. Patient-to-patient heart and LV DVH variability was represented using five Varimax-rotated principal components (PCs) explaining 95% of total variance. OS was modelled from start of treatment using Cox regression.

Results

On univariate analysis larger planning target volumes (PTVs) (HR=1.002, 95%CI:1.000-1.003; p=0.026), greater Heart-PC3 (HR=1.257, 95%CI:1.070-1.478; p=0.005) and greater LV-PC4 (HR=1.257, 95%CI:0.998-1.584; p=0.052) are associated with worse OS. Heart-PC3 represents heart volumes receiving doses of 65-75Gy, and LV-PC4 represents LV volumes receiving 1-5Gy. Although OS improved with increasing prescribed radiation dose (HR=0.951, 95%CI: 0.883-1.024; p=0.185), this trend did not reach statistical significance. The estimates of these effects did not change markedly after adjusting for dose, PTV, Heart-PC3 and LV-PC4 in multivariate analysis, and the relationship between greater Heart-PC3 and worse OS remained highly significant (HR=1.228, 95%CI:1.039-1.452; p=0.016).

Conclusion

We found a strong association between heart volumes receiving 65-75Gy and poorer OS. There was some evidence of a negative association between LV volumes receiving 1-5Gy and OS. Potential survival gains achievable through tumour dose-escalation may be offset by associated heart dose increases, therefore further studies are required to improve understanding of heart substructures radiation effects for cardiac-sparing dose-escalated NSCLC treatments.