Session Item

To evaluate an in-house developed patient-reported outcome measures (PROM) platform to assess on-treatment radiation related toxicities of breast cancer patients.

Starting in November 2020, every eligible patient with breast cancer was enrolled on a voluntary basis into the IRB-approved Patient Experience Data in Radiation Oncology (PEDRO) study (EK 2184/2019). Patients received hypofractionated whole-breast irradiation to 40 Gy in 15 fractions, followed by a risk-adapted sequential boost of 10 Gy in four fractions if indicated. Consenting participants answered a predefined set of questions from the PRO-CTCAE catalogue. Surveys were carried out weekly on tablets in the out-patient clinic for the duration of the treatment. The PROM application, the app-server and database infrastructure were all in-house developed. The PROM data were directly linked to relevant treatment data from the oncology information system. Differences between the two cohorts (w/o boost) were analyzed for significant differences with respect to the endpoints: itching, radiation skin reaction, skin darkening, and breast swelling and tenderness. 

A total of 330 patients with breast cancer participated in this real-world PROM setting (40% of all patients with breast cancer treated in this period), resulting in 1527 individual completed questionnaires. 177 patients (54%) were treated with a sequential boost, 153 patients (46%) received no boost. The maximum reported side effects at the respective last fractions were significantly higher in the boosted group for different clinical endpoints (Fig. 1): itching (p<0.001), radiation skin reaction (p<0.001), skin darkening (p<0.001), and breast swelling and tenderness (p=0.05). These differences generally disappeared at the sensitivity analysis, when the surveys at 40 Gy were compared for both groups. Baseline PROMs were also equal among groups (p=0.40, p=0.42, p=0.88, p=0.73, respectively). An analysis of the significance in differences with growing sample size (Fig. 2) suggested that dichotomy stabilized at around 100 patients for radiation skin reaction and skin darkening, as well as 200 patients for itching. Data suggests that the reporting of breast swelling and tenderness requires larger sample sizes (>330). 

Fig1: Comparison patients that received boost (green) with patients without boost (red) with respect to the maximum reported value for the endpoints (dashed = quatiles). 

Fig2: Change of p-value (boost vs. no boost) over increasing number of sample size.

PROMs were successfully implemented into clinical routine and yielded a surprisingly high sensitivity detecting discrete dose-dependent radiation reactions in the breast-irradiation PEDRO pilot module. The extension of this in-house developed tool is prone to provide valuable information on subtle radiation induced toxicities and will help guide knowledge-based improvements of treatment planning and monitor changes in treatment schedules such as e.g., extreme hypofractionation.