OBJECTIVES: The growing number of patients treated with radiotherapy requires a profound change in the management of radiation oncology units to offer their patients an advanced service while maximizing their human and machine resources. This was one of the challenges that Radiotherapy Department of CHU Liege Hospital wanted to resolve during HosmartAI project. Staff members of CHU de Liège weigh about ten variables of each patient according to their own oncological and psychosocial context before planning their radiotherapy treatment. These variables continue to increase due to the rise in personalized medicine in oncology treatments. The objective of the study was to identify the proper KPIs in order to establish an AI algorithm for optimizing patient scheduling at the radiotherapy department.

METHODS: A literature review was performed followed by a detailed reconding and classification of physical scheduling procedures to identify and map the most important parameters (KPIs) to be used in the AI algorithm. Upon implementation of the AI algorithm an incremental analysis will be performed to estimate the difference of the current practice versus the new AI scheduling.

RESULTS: The current manual scheduling reported the following performance:15days protocol for breast and lung tumour irradiation, in accordance with the respective guidelines and 1 irradiation for urgent bone metastasis treatment. The number of variables considered currently are 10. At present 3 employees perform the appointment co-ordination and the time spent for each appointment scheduling is between 16 and 55 minute depending on the planning irradiation for one or multiple tumours. The total number of treatments/sessions for 2022 were 2.768 (584 for breast cancer and 357 for lung cancer). The expected AI algorithm will take into consideration 20 variables by decreasing the time of manual planning.

CONCLUSIONS: The new promising AI algorithm will pave the way for the optimal and efficient use of resources.