ESTRO 2025 - Abstract Book

S155

Invited Speaker

ESTRO 2025

Abstract:

A quality improvement project may improve safety, side effects, effectiveness, survivorship and the experience of care. Some of the aspects are interlinked. Care pathways in radiation oncology are invariable very complex and this makes tangible results difficult. The multidisciplinary team nature adds to the complexity and the human factors need to be understood. For a project, select a relevant endpoint, a measurement tool, a process or change to improve the outcome and finally implement the changes. As with all complex pathways, re-measuring the outcome repeatedly gives feedback, helps to check effectiveness and maintains changes. Potential topics are waiting times to start treatment, patient selection for radical treatments (TMN stage, performance status and frailty, comorbidity), local and metastatic relapse rates, survival, acute and late toxicity. Specific radiotherapy aspects are contouring reviews, radiation doses to planning target volume and organs at risk, use of treatment technology (image guidance) and adherence published guidelines. The multidisciplinary nature requires a good functioning team, but we can learn from big datasets such as the National Clinical Audits in the UK, compare the local practice with guidelines and understand the unwanted variation in the local context. A literature review found 26 studies specific to radiotherapy published in 2000-24: 15 evaluated waiting times, 5 contouring and 6 the management of acute symptoms. As a radiotherapy community, we have a limited knowledge of quality improvements. Purpose/Objective : As an example for a quality improvement project in prostate radiotherapy, we used the results of the National Prostate Cancer Audit (NPCA). Patients treated with radical radiotherapy in 2016 and 2017 in a large cancer centre in Northwest England were included in the cohort. The NPCA provided the patient list with comorbidity scores (Charlston index) and deprivation (IMD5) scores. Patients who underwent any investigation or treatment for bowel (GI) symptoms during 2 years after radiotherapy were included in the toxicity group. The study aimed to identify any risk factors for GI toxicity. Material/Methods : In the Centre, eight clinicians treat PC patients with radical radiotherapy. We used a generic treatment protocol, standardised contouring guidelines and regular contouring peer review. We had standardised dose prescription for prostate only (60Gy/20 fraction) and prostate / pelvic node (72Gy / 50Gy/ 32 fractions) radiotherapy, volumetric arc therapy planning class solutions in ARIA and uniform Organ-At-Risk constraints. The prostate only group had fiducials and daily image guidance with kV imaging. Cone beam CT was used for patients with large shifts. The prostate / pelvic node group had an off line no action 3 Cone beam CT image guidance protocol. For the quality improvement project, we identified all patients with toxicity and checked the clinical records. For the radiotherapy planning analysis, the dose volume histograms (DVH) were adjusted to 2-Gy equivalent schedules (EQD2) using α/β = 3. Training sets and testing sets were obtained by a stratified 70:30 split. The odds ratios of toxicity were calculated with logistic regression and restricted cubic spline regression. The optimal regression model was used to create a risk stratification tool with ‘low-risk’ and ‘high-risk’ constraints determined with Receiver Operative Characteristic analysis. Results : We included 802 patients; 39 had bowel toxicity (4.9%). Baseline characteristics were similar. Prostate pelvic node patients (25/300) were more likely to develop bowel toxicity (RR: 3.15, 95% CI: 1.58-5.66), compared to prostate only group (13/492). Patients with bowel toxicity had statistically significant higher relative rectal volumes irradiated at all dose levels with logistic and restricted cubic spline regression but were only significant increased for bowel from V40-50 with logistic regression and restricted cubic spline regression. Restricted cubic spline regression was used to generate a rectal DVH risk stratification tool with statistically significant outcomes in the unseen testing group (Low risk group RR: 0.13, (CI: 0.02-0.70); high-risk group RR: 5.39 (CI: 2.07-14.0)). Conclusions : We demonstrated a low GI toxicity rate in a large UK based cancer centre. Deprivation score and comorbidity score was not linked to a higher risk if bowel toxicity. Bowel toxicity was associated with high risk localised prostate cancer, the prostate/ pelvic node schedule and increased rectal doses (all dose levels) and bowel doses (V40-V50).