Session Item

Prostate planning has substantially changed over the last decade. This work aimed to assess the radiobiological impact of rectum toxicity for all the changes implemented at our department in prostate treatment planning during the past 15 years. We chose the normal tissue complication probability (NTCP) as a predictor of the radiobiological effect of rectum toxicities.

A total of 1764 prostate patients referred for intended curative radiotherapy were included. This series included patients treated from 2014 to 2021 at our department. Patients with nodal involvement were excluded from this study. During this period, several treatment planning changes occurred: we introduced 6-field IMRT in 2013, VMAT in 2017, gEUD optimization in 2019, and RapidPlan in 2020. During this time, PTV coverage dosimetric constraints remained the same.

We automatically extracted all the DVHs for all patients using the Eclipse Scripting API. All 1764 DVH were then re-normalized to a prescription dose of 70 Gy in EQD2. For each patient, we extracted the following rectum dose-volume values: V10, V20, V30, V40, V50, V60, V70, V72, and V75.

NTCP was evaluated from the empirical Lyman-Kutcher-Burman (LKB) model. For the rectum, the following complication endpoints were selected: rectal bleeding (G1&2), proctitis (G1&2), stool frequency (G1&2), loose stools (G1&2), rectal urgency (G1&2), rectal bleeding (G2), proctitis (G2), loose stools (G2), rectal urgency (G2). For these complication endpoints, we selected a corresponding set of parameters (n, m, TD50) from Gulliford [1]. We also evaluated the “Late toxicity G>= 2” taking the set of parameters (n, m, TD50) offered by the QUANTEC publication [2].

Dose-volume constraints and complications endpoints were graphically represented as a function of time using descriptive statistics (boxplot method).

Figure 1 shows a boxplot representation of the evolution of rectum dose-volume constraints over the years. No changes are visible for the introduction of IMRT and VMAT. However, the introduction of the gEUD optimization method had a noticeable impact, more clearly visible for the higher dose constraints. In addition, the introduction of the Rapidplan substantially decreased variability while maintaining the improvement already achieved with the introduction of the gEUDs.

Figure 2 shows another boxplot representation of the evolution of the prediction of rectum endpoints complications over time. As before, no appreciable changes appeared during the introduction of IMRT and VMAT. However, a visible impact occurred when the introduction of gEUD and, subsequently, Rapidplan.

The present study demonstrates the impact of gEUD optimization and Rapidplan on all rectum endpoint complications.

Surprisingly, the introduction of IMRT and VMAT had no radiobiological impact on rectum toxicity.