Inter-observer and intra-observer variability in reporting vaginal dose points for cervical cancer in high-dose-rate...

The ICRU 89 recommends reporting a set of vaginal dose points for cervical cancer treatments in order to quantify the goodness of implant. This vaginal dose reporting method for combined external beam radiotherapy and brachytherapy has been adopted by the EMBRACE II study protocol. Large variations in dose between patients and centers have been reported. The aim of this study was to determine possible discrepancies with consensus observers from the same institution. Therefore, the inter- and intra-observer variability were analyzed.

For five patients, five experienced observers reported dose at the proposed vaginal points twice. The effect of inter- and intra-observer variations on total dose was analyzed by estimating biologically equivalent dose EQD₂ (α/β = 3 Gy). Coefficient of variation (CV) was used to provide a measure of data dispersion as a proportion to the mean.

The maximum inter-observer deviation among all patients and all points ranged from 0.5 Gy to 24.1 Gy in EQD₂. The higher inter-observer discrepancies were found at points at 3 o'clock and at 6 o'clock, with respect to ovoids. In case of the maximum intra-observer deviation, it ranged from 0.5 Gy to 14.2 Gy, with higher deviation points at 12 o'clock and 9 o'clock, with respect to ovoids.

There is a need to ensure consistency in vaginal points reporting. The impact of the dosimetric inter- and intra-observer variability should also be considered when dealing with dose tolerances and limits due to the potential dose gradient.