Normal Tissue Complication Probability (NTCP) Modelling of Severe Acute Mucositis using a Novel Oral Mucosal Surface
J A Dean1, L C Welsh2, K H Wong2
1Joint Department of Physics at The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, London, UK.
Clinical Oncology (Royal College of Radiologists (Great Britain))
|January 7, 2017
Summary
The oral cavity organ at risk (OAR) is simpler and performs as well as the mucosal surface OAR for normal tissue complication probability (NTCP) models of severe acute mucositis. Minimizing intermediate and high radiation doses to the oral mucosa is recommended.
Area of Science:
- Radiation oncology
- Medical physics
- Clinical decision support
Background:
- Severe acute mucositis is a common side effect of radiotherapy, impacting patient quality of life.
- Accurate normal tissue complication probability (NTCP) models are crucial for radiotherapy planning and clinical decision-making.
- Previous models utilized the oral cavity organ at risk (OAR), but a novel mucosal surface OAR was proposed for improved accuracy.
Purpose of the Study:
- To compare the performance of NTCP models for severe acute mucositis using either the oral cavity OAR or the novel oral mucosal surface OAR.
- To identify the optimal OAR for developing robust predictive models of radiotherapy-induced mucositis.
- To investigate the association between radiation dose parameters and the occurrence of severe acute mucositis.
Main Methods:
- Two types of OARs (oral cavity and mucosal surface) were used to generate predictive models for severe acute mucositis.
- Penalised logistic regression and random forest classification (RFC) were employed for model development.
- Internal validation utilized 100-iteration stratified shuffle split cross-validation with multiple performance metrics.
- RFC feature importance was used to explore associations between treatment covariates and mucositis.
Main Results:
- Both penalised logistic regression and RFC models using the oral cavity OAR demonstrated performance comparable to models using the mucosal surface OAR.
- Similar associations between radiation dose metrics and severe mucositis were observed for both OARs.
- The volume of the oral cavity or mucosal surface receiving intermediate and high radiation doses showed the strongest correlation with severe mucositis.
Conclusions:
- The simpler oral cavity OAR is recommended over the mucosal surface OAR for developing NTCP models of severe acute mucositis.
- Minimizing the volume of oral mucosa exposed to intermediate and high radiation doses should be prioritized in radiotherapy planning.
- These findings can help optimize radiotherapy strategies to reduce the incidence and severity of mucositis.
