Session Item

Conserving surgery followed by external beam radiotherapy is considered the "Gold Standard" for early stage of breast cancer. This approach may introduce an excess risk for second cancer induction due to breast exposure to therapeutic doses. The aim of this study was to estimate radiotherapy-induced lung cancer risk and to propose simplified models useful in clinical routine to reduce it in a preventive way during the treatment planning phase.

Using the Schneider Mechanistic Model, radiotherapy-induced lung cancer risk for breast cancer has been estimated for 288 patients (aged between 30 and 70 years) treated with Three-Dimensional Conformal Radiation Therapy and Standard Fractioned (3D-CRT SF) at Humanitas - Istituto Clinico Catanese (H-ICC) (Catania, Italy). Organ Equivalent Dose (OED), Excess Absolute Risk (EAR), Lifetime Attributable Risk (LAR) and Relative Risk (RR) values has been calculated implementing a Script (C# language) through the Varian Eclipse Scripting Application Programming Interface (ESAPI). Statistical parameters have been provided by several sources: H-ICC, Istituto Nazionale di Statistica (ISTAT) and Integrated Cancer Registry CT-ME-EN. Using a C++ code, simulations have been performed on the whole statistical sample imposing an attained patient’s age (age_a) equal to 75 Y and an age of the patient during exposure (age_e) varying between 30 and 70 years in steps of 5 years. In order to minimise the difference between LAR from Schneider model (LAR_Schneider) and LAR from simplified model a parameter optimisation process has been performed, i.e., a minimization of Mean Square Error (MSE). 

The first step was to fit linearly LAR values as a function of OED for i-th age_e (i = 30, 35, …70) (1), where a_i is the i-th angular coefficient for the i-th age_e. he next step was to relate the growth rate of the LAR to the time range age_e-age_a (fig.).The analytic relationship (R²=0.99) found was (1). Substituting (2) to (1) we obtained the simplified model named OSM (3). OSM optimised parameters obtained after optimisation process were: A=-0.340 [(1/10000 P) Gy^-1], B=-11.688 [Y], C=144.557 [Gy^-1]. The percentage differences (Δ%) between LAR_Schneider and LAR_OSM values were < 2%. A linear relation (R²=0.96) was found between OED and V4 (% lung volume absorbing a dose of 4 Gy) (4). Substituting (4) to (3) we obtained the simplified model named VSM (5). VSM optimised parameters obtained, after optimisation process were: A=-0.300 [(1/10000 P) Gy^-1], B=-11.387 [Y], C=147.601 [Gy^-1], M=0.830 [(1/10000 P) Gy^-1], N =0.069 [Y]. The percentage differences (Δ%) between LAR_Schneider and LAR_VSM values were < 5%.

This study provided three different tools for risk calculation: Eclipse™ script, OSM and VSM. These, in different ways, allow the medical physicist to quickly obtain LAR values for each treatment plan, which is why they could easily be used in clinical practice.