The influence of age in oesophageal cancer treatment decisions: a machine-learning approach

Abstract

Background: although patient age plays a crucial role in determining curative treatment options for Oesophageal cancer (OC), the exact extent of its influence is not well defined. We used a computational machine learning (ML) approach to model the impact of age in combination with other key decision drivers, such as tumour and patient characteristics, on the likelihood of receiving different types of curative treatment.

Methods: retrospective analysis of 399 OC patients undergoing curative treatment between 2010-2020 at our tertiary unit. A random forests (RF) classifier model was trained to predict curative treatment decisions for OC patients (neoadjuvant chemotherapy (NACT) and surgery, neoadjuvant chemoradiotherapy (NACRT) and surgery or surgery alone). Variable importance and Partial Dependence analyses were used to assess the importance of age in the model, its influence on treatment decisions, and how that relationship was affected with other decision driver co-variates.

Results: variable importance analysis confirmed age as most important in the RF model, (26% of total model performance). Partial dependence analysis demonstrate that predicted base probabilities for receiving surgery and NACT changed significantly in older patients. Patients above 70 years had a substantially higher probability of receiving curative surgery and lower probability of NACT. Moreover, the probability of receiving Surgery and NACT is driven by disease characteristics (T and N staging) in patients < 70 years but age becomes increasingly important in predicting a surgical decision >70 years. The base probability of surgery and NACRT decisions was also influenced by performance status and age, but only age for NACT patients.

Conclusion: we have successfully applied ML modelling combined with partial dependence analysis to delineate the relationship between age and OC MDT treatment decisions. Age heavily influences curative decisions in OC patients but plays a greater role in patients with specific tumour characteristics. This study provides the basis for exploring subconscious decision drivers and allows teams to address any health inequality.

More information

Identifiers

ORCID for Navamayooran Thavanesan: orcid.org/0000-0002-7127-9606

ORCID for Zoe S. Walters: orcid.org/0000-0002-1835-5868

ORCID for Tim Underwood: orcid.org/0000-0001-9455-2188

ORCID for Ganesh Vigneswaran: orcid.org/0000-0002-4115-428X

Catalogue record

Export record