Process mining is a relatively new method that connects data science and process modelling. In this paper we apply process mining on clinical oncological data with the purpose of studying survival outcomes and chemotherapy treatment decision in a real-world cohort of small cell lung cancer patients treated at Karolinska University Hospital (Stockholm, Sweden).

The study presents an analysis of treatment pathways and associated outcomes in subgroups of SCLC patients treated with chemotherapeutic agents.The chosen methodology was a pipeline that directly extracted treatment processes from the patient data, thus defining graphs from treatment and prognostic outcomes. 

Results showed that process mining has potential for enriching the analysis of oncological cohorts by developing an increased understanding of the underlying treatment pathways and decision points. In addition to the rapid processing and the impactful visualisation, the technique can be used to inform longitudinal modelling of disease progression and subsequent impact on treatment decisions, as well as patient outcomes. The ready-to-use longitudinal models directly extracted from real-world data constitute suitable objects for the design of multi-state survival models, or other process model solutions, such as causal networks.  

The study provides insights into the role of process mining in oncology to study prognosis and survival outcomes, and additional indications on how to develop this further in the future.