Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models

Background: Systemic anticancer therapy (SACT) near the end of life (EOL) reduces the quality of the patient’s remaining life without clinical benefit. Studies investigating machine learning models for predicting cancer mortality to guide treatment decisions have primarily focused on specific types...

Full description

Saved in:
Bibliographic Details
Main Authors: S. Bjerregaard-Michelsen, L.Ø. Poulsen, A. Bjerrum, M. Bøgsted, C. Vesteghem
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:ESMO Real World Data and Digital Oncology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2949820125000359
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849324172820348928
author S. Bjerregaard-Michelsen
L.Ø. Poulsen
A. Bjerrum
M. Bøgsted
C. Vesteghem
author_facet S. Bjerregaard-Michelsen
L.Ø. Poulsen
A. Bjerrum
M. Bøgsted
C. Vesteghem
author_sort S. Bjerregaard-Michelsen
collection DOAJ
description Background: Systemic anticancer therapy (SACT) near the end of life (EOL) reduces the quality of the patient’s remaining life without clinical benefit. Studies investigating machine learning models for predicting cancer mortality to guide treatment decisions have primarily focused on specific types of cancer. This study aimed to evaluate the ability of a pan-cancer model to generalize across 10 cancer types when predicting short-term mortality. Patients and methods: This study included patients with advanced cancer who were referred to the Department of Oncology at Aalborg University Hospital and died between January 2008 and December 2021 (N = 8690). Clinical data were used to train, validate, and test a pan-cancer model and 10 single-cancer models based on the eXtreme Gradient Boosting (XGBoost) algorithm. The average precision (AP) of the pan-cancer and single-cancer models was assessed and compared. Furthermore, explainable AI with Shapley additive explanations (SHAP) was used to evaluate shared prognostic information across the cancer types. Results: The mean AP increased from 0.51 using the single-cancer models to 0.56 using the pan-cancer model to predict short-term mortality (random baseline 0.12). Important features identified by SHAP were shared across cancer types, indicating shared predictors of 30-day mortality. The most important features for predicting 30-day mortality were plasma albumin level, white blood cell count, and lactate dehydrogenase levels. Conclusion: A pan-cancer model enhanced the performance of short-term mortality estimates compared with models based on single cancer types. Thus, including multiple cancer types in predictive modeling in oncology could be advantageous when shared predictors are expected across cancer types.
format Article
id doaj-art-a8c77ccefa424e769407c7a0d7778714
institution Kabale University
issn 2949-8201
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series ESMO Real World Data and Digital Oncology
spelling doaj-art-a8c77ccefa424e769407c7a0d77787142025-08-20T03:48:47ZengElsevierESMO Real World Data and Digital Oncology2949-82012025-06-01810014610.1016/j.esmorw.2025.100146Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer modelsS. Bjerregaard-Michelsen0L.Ø. Poulsen1A. Bjerrum2M. Bøgsted3C. Vesteghem4Center for Clinical Data Science, Department of Clinical Medicine, Aalborg University and Research, Education and Innovation, Aalborg University Hospital, Aalborg, Denmark; Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, Denmark; Correspondence to: Signe Bjerregaard-Michelsen, Center for Clinical Data Science, Department of Clinical Medicine, Aalborg University and Research, Education and Innovation, Aalborg University Hospital, Søndre Skovvej 15, 9000 Aalborg, Denmark.Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, Denmark; Department of Oncology, Aalborg University Hospital, Aalborg, DenmarkDepartment of Oncology, Copenhagen University Hospital, Rigshospitalet, Copenhagen, DenmarkCenter for Clinical Data Science, Department of Clinical Medicine, Aalborg University and Research, Education and Innovation, Aalborg University Hospital, Aalborg, Denmark; Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, DenmarkCenter for Clinical Data Science, Department of Clinical Medicine, Aalborg University and Research, Education and Innovation, Aalborg University Hospital, Aalborg, Denmark; Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, DenmarkBackground: Systemic anticancer therapy (SACT) near the end of life (EOL) reduces the quality of the patient’s remaining life without clinical benefit. Studies investigating machine learning models for predicting cancer mortality to guide treatment decisions have primarily focused on specific types of cancer. This study aimed to evaluate the ability of a pan-cancer model to generalize across 10 cancer types when predicting short-term mortality. Patients and methods: This study included patients with advanced cancer who were referred to the Department of Oncology at Aalborg University Hospital and died between January 2008 and December 2021 (N = 8690). Clinical data were used to train, validate, and test a pan-cancer model and 10 single-cancer models based on the eXtreme Gradient Boosting (XGBoost) algorithm. The average precision (AP) of the pan-cancer and single-cancer models was assessed and compared. Furthermore, explainable AI with Shapley additive explanations (SHAP) was used to evaluate shared prognostic information across the cancer types. Results: The mean AP increased from 0.51 using the single-cancer models to 0.56 using the pan-cancer model to predict short-term mortality (random baseline 0.12). Important features identified by SHAP were shared across cancer types, indicating shared predictors of 30-day mortality. The most important features for predicting 30-day mortality were plasma albumin level, white blood cell count, and lactate dehydrogenase levels. Conclusion: A pan-cancer model enhanced the performance of short-term mortality estimates compared with models based on single cancer types. Thus, including multiple cancer types in predictive modeling in oncology could be advantageous when shared predictors are expected across cancer types.http://www.sciencedirect.com/science/article/pii/S2949820125000359pan-cancer modelsshort-term mortalitycounterproductive treatmentmachine learning
spellingShingle S. Bjerregaard-Michelsen
L.Ø. Poulsen
A. Bjerrum
M. Bøgsted
C. Vesteghem
Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
ESMO Real World Data and Digital Oncology
pan-cancer models
short-term mortality
counterproductive treatment
machine learning
title Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
title_full Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
title_fullStr Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
title_full_unstemmed Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
title_short Machine learning for prediction of 30-day mortality in patients with advanced cancer: comparing pan-cancer and single-cancer models
title_sort machine learning for prediction of 30 day mortality in patients with advanced cancer comparing pan cancer and single cancer models
topic pan-cancer models
short-term mortality
counterproductive treatment
machine learning
url http://www.sciencedirect.com/science/article/pii/S2949820125000359
work_keys_str_mv AT sbjerregaardmichelsen machinelearningforpredictionof30daymortalityinpatientswithadvancedcancercomparingpancancerandsinglecancermodels
AT løpoulsen machinelearningforpredictionof30daymortalityinpatientswithadvancedcancercomparingpancancerandsinglecancermodels
AT abjerrum machinelearningforpredictionof30daymortalityinpatientswithadvancedcancercomparingpancancerandsinglecancermodels
AT mbøgsted machinelearningforpredictionof30daymortalityinpatientswithadvancedcancercomparingpancancerandsinglecancermodels
AT cvesteghem machinelearningforpredictionof30daymortalityinpatientswithadvancedcancercomparingpancancerandsinglecancermodels