Machine Learning-Based Risk Prediction of Critical Care Unit Admission for Advanced Stage High Grade Serous Ovarian Cancer Patients Undergoing Cytoreductive Surgery: The Leeds-Natal Score.

Alexandros Laios,Diederick De Jong,David Nugent,Yong Sheng Tan,Amudha Thangavelu,Gwendolyn Saalmink,Angelika Kaufmann,Racheal Johnson,Georgios Theophilou,Albina Zubayraeva,Kassio Michell Gomes De Lima,Raissa Vanessa De Oliveira Silva,Richard Hutson,Daniel Lucas Dantas De Freitas,Tim Broadhead,Mohammed Otify

doi:10.3390/jcm11010087

Abstract

Achieving complete surgical cytoreduction in advanced stage high grade serous ovarian cancer (HGSOC) patients warrants an availability of Critical Care Unit (CCU) beds. Machine Learning (ML) could be helpful in monitoring CCU admissions to improve standards of care. We aimed to improve the accuracy of predicting CCU admission in HGSOC patients by ML algorithms and developed an ML-based predictive score. A cohort of 291 advanced stage HGSOC patients with fully curated data was selected. Several linear and non-linear distances, and quadratic discriminant ML methods, were employed to derive prediction information for CCU admission. When all the variables were included in the model, the prediction accuracies were higher for linear discriminant (0.90) and quadratic discriminant (0.93) methods compared with conventional logistic regression (0.84). Feature selection identified pre-treatment albumin, surgical complexity score, estimated blood loss, operative time, and bowel resection with stoma as the most significant prediction features. The real-time prediction accuracy of the Graphical User Interface CCU calculator reached 95%. Limited, potentially modifiable, mostly intra-operative factors contributing to CCU admission were identified and suggest areas for targeted interventions. The accurate quantification of CCU admission patterns is critical information when counseling patients about peri-operative risks related to their cytoreductive surgery.

Highlights

Cancer of the fallopian tube, ovary, or peritoneum (EOC) is the leading cause of death from gynaecological malignancy in the western world [1]
The details of baseline characteristics and operative factors are displayed in Table 1 to enable the better understanding of this group of patients undergoing cytoreductive surgery and potentially requiring Critical Care Unit (CCU) admission
We demonstrated the utility of linear discriminants (LDA, QDA) of clinical characteristics in model performance with maximum accuracy reaching 98% and the lowest boundary of performance being no worse than 68%

Summary

Introduction

Cancer of the fallopian tube, ovary, or peritoneum (EOC) is the leading cause of death from gynaecological malignancy in the western world [1]. Over 70% of women diagnosed with EOC have advanced stage disease at presentation (International Federation Obstetrics and Gynaecology FIGO stage 3–4) [1]. The gynaecological oncology community is recognizing high grade serous ovarian cancer (HGSOC), as yet the most prevalent, as a single clinical entity. For such an aggressive subtype, identification of novel, ideal biomarkers that strongly correlates with the stage and may be effective for early diagnosis, remains critical [2]. When the cancer is at an advanced stage, the debulking surgery often requires prolonged surgical times and possible multi-visceral resections, necessitating Critical Care Unit (CCU) support, and prolonged hospitalization [3]. Risk-prediction models of severe post-operative complications after ovarian cytoreductive surgery have been thoroughly proposed [4]

Objectives

Methods

Results

Discussion

Conclusion