Evaluation of machine learning methods for the retrospective detection of ovarian cancer recurrences from chemotherapy data.

Coles, AD.; McInerney, CD.; Zucker, K.; Cheeseman, S.; Johnson, OA.; Hall, G.

Evaluation of machine learning methods for the retrospective detection of ovarian cancer recurrences from chemotherapy data.

All Authors

Coles, AD.

McInerney, CD.

Zucker, K.

Cheeseman, S.

Johnson, OA.

Hall, G.

LTHT Author

Zucker, Kieran
Cheeseman, Sue
Hall, Geoff

LTHT Department

Oncology
Leeds Cancer Centre
Research & Innovation
Real World Evidence Alliance
Digital Health
Chief Clinical Information Officer

Publication Date

2024

Item Type

Journal Article

Subject

ARTIFICIAL INTELLIGENCE , NEOPLASM RECURRENCE, LOCAL , CHEMOTHERAPY , HEALTH AND CARE RECORDS , MACHINE LEARNING

Abstract

Background: Cancer recurrences are poorly recorded within electronic health records around the world. This hinders research into the efficacy of cancer treatments. Currently, the retrospective identification of recurrence/progression diagnosis dates is achieved by staff who manually review patients' health records. This is expensive, time-consuming, and inefficient. Machine Learning models may expedite the review of health records and facilitate the assessment of alternative cancer therapies. Materials and methods: This paper evaluates the use of four machine learning models (random forests, conditional inference trees, decision trees, and logistic regression) in identifying proxy dates of epithelial ovarian cancer recurrence/progression from chemotherapy data, in 531 patients at Leeds Teaching Hospital Trust. Results: The random forest achieved the highest F1 score of 0.941 (95% confidence interval 0.916-0.968) when identifying recurrence events. Both the random forest and decision tree models' classifications closely conform to chart-reviewed time to next treatment, serving as a surrogate for recurrence-free survival. Additionally, all models reached an F1 score >0.940 when identifying patients whose cancer recurred/progressed. Conclusions: Our models proficiently identify both proxy dates for recurrence/progression diagnoses and patients whose cancer recurred/progressed. Considering the similar performance of the random forest and decision tree, model preference should be determined by the interpretability required to assist chart review and the ease of implementation into existing architecture.