Artificial intelligence based treatment planning of radiotherapy for locally advanced breast cancer.

BACKGROUND AND PURPOSE: Treatment planning of radiotherapy for locally advanced breast cancer patients can be a time consuming process. Artificial intelligence based treatment planning could be used as a tool to speed up this process and maintain plan quality consistency. The purpose of this study was to create treatment plans for locally advanced breast cancer patients using a Convolutional Neural Network (CNN). MATERIALS AND METHODS: Data of 60 patients treated for left-sided breast cancer was used with a training, validation and test split of 36/12/12, respectively. The in-house built CNN model was a hierarchically densely connected U-net (HD U-net). The inputs for the HD U-net were 2D distance maps of the relevant regions of interest. Dose predictions, generated by the HD U-net, were used for a mimicking algorithm in order to create clinically deliverable plans. RESULTS: Dose predictions were generated by the HD U-net and mimicked using a commercial treatment planning system. The predicted plans fulfilling all clinical goals while showing small (≤0.5 Gy) statistically significant differences (p < 0.05) in the doses compared to the manual plans. The mimicked plans show statistically significant differences in the average doses for the heart and lung of ≤0.5 Gy and a reduced D2% of all PTVs. In total, ten of the twelve mimicked plans were clinically acceptable. CONCLUSIONS: We created a CNN model which can generate clinically acceptable plans for left-sided locally advanced breast cancer patients. This model shows great potential to speed up the treatment planning process while maintaining consistent plan quality.

Generalizability assessment of head and neck cancer NTCP models based on the TRIPOD criteria.

This review aimed to provide an overview of the level of maturity of normal tissue complication probability (NTCP) models for head and neck cancer (HNC) patients. A systematic literature review was performed to retrieve NTCP models for HNC toxicities. Patient population characteristics, NTCP model and the predictors, treatment technique and endpoint definition were extracted per article. Models were then scored based on the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) consensus guidelines to evaluate their generalizability. 335 articles on photon and proton therapy of HNC were identified and 52 relevant articles were further analyzed. Eighteen articles on xerostomia and sticky saliva (TRIPOD types 1a-2b: 15; TRIPOD type 3: 1; TRIPOD types 4a: 1 & 4b:1), thirteen articles on dysphagia and tube feeding dependence (TRIPOD types 1a-2b: 7; TRIPOD type 3: 2; TRIPOD types 4a:2 & 4b:2), five articles on oral mucositis (TRIPOD types 1a-2b: 4; TRIPOD type 4b: 1), seven articles on hypothyroidism (TRIPOD types 1a-2b: 4; TRIPOD type 3: 1; TRIPOD types 4a: 1 & 4b:1), four articles on hearing loss and tinnitus (TRIPOD type 1a: 4) and ten articles on esophagitis (TRIPOD types 1a-2b: 9; TRIPOD type 4a: 1) were included. External validation studies of HNC NTCP models are scarce. Moreover, the majority of them were validating a model developed by the same researchers. Only 2 independent external validation studies were found. There is a strong need to publish external validation studies to get more mature NTCP models applicable in clinical practice.