Margriet Kwint

Prognostic value of volumetric changes during concurrent chemoradiation 93 5 median GTV-volume at baseline and a lower proportion of adenocarcinomas was seen in subgroup 3. In multivariate Cox analyses, WHO-PS (HR:2.93), age (HR:1.56) and GTV-volume of the primary tumor (HR:1.42) were significantly associated with OS, and WHO-PS (HR:2.12), nodal-stage (HR:1.85) and GTV-volume of the primary tumor (HR:1.47) were significantly associated with PFS ( Table 2 ). Surprisingly none of GTV-change parameters were significantly associated with treatment outcome. When the 3-subgroup classification derived from LCMM was added to the multivariate baseline model for OS, the Harrell’s c-statistic did not substantially change (baselinemodel c-statistics = 0.624; baselinemodel+ LCMM subgroup c-statistics = 0.628). Kaplan Meier plots of OS, PFS and LRC of the subgroups are shown in Supplement Figure 3 . In our cohort, the median decrease of tumor volume was 18.5% (IQR 8.9 - 32.5) for rel_GTV first-last and 11.9% (IQR 4.4-18.2) for rel_GTV 11 -21. No significant different OS, PFS and LRC rates were found for rel_GTV first-last or_ rel_GTV 11-21 ( Table 2 ). Median cut-off values of tumor volume change from literature were higher; 39.3% (22) and 22.8% (23), respectively. Nonetheless, no significant OS differences were found for those cut-off values either (HR: 1.10, p-value 0.571 & HR: 1.32, p-value 0.227 respectively). Within the pathology subgroup analysis a borderline significant association was found only for rel_GTV first-last in the AC-subgroup (HR:1.50, p-value:0.064). None of the other GTV-change parameters were significantly associated with OS, PFS or LRC in either subgroup ( Table 3 ).