Abstract
Infections have become the major cause of morbidity and mortality among patients with chronic lymphocytic leukemia (CLL) due to immune dysfunction and cytotoxic CLL treatment. Yet, predictive models for infection are missing. In this work, we develop the CLL Treatment-Infection Model (CLL-TIM) that identifies patients at risk of infection or CLL treatment within 2 years of diagnosis as validated on both internal and external cohorts. CLL-TIM is an ensemble algorithm composed of 28 machine learning algorithms based on data from 4,149 patients with CLL. The model is capable of dealing with heterogeneous data, including the high rates of missing data to be expected in the real-world setting, with a precision of 72% and a recall of 75%. To address concerns regarding the use of complex machine learning algorithms in the clinic, for each patient with CLL, CLL-TIM provides explainable predictions through uncertainty estimates and personalized risk factors.
Highlights
Infections have become the major cause of morbidity and mortality among patients with chronic lymphocytic leukemia (CLL) due to immune dysfunction and cytotoxic CLL treatment
While infection is the major cause of death in CLL1, no prognostic index for prediction of infection in CLL has been presented prior to our recent work[36]
A first step in attempting to change the ramifications of immune dysfunction in CLL is to identify high-risk patients prior to any infection or CLL treatment[5]. We addressed this unmet need by developing an explainable machine learning model based on data from 4,149 patients diagnosed with CLL in Denmark between 2004 and 2017
Summary
Infections have become the major cause of morbidity and mortality among patients with chronic lymphocytic leukemia (CLL) due to immune dysfunction and cytotoxic CLL treatment. We developed the CLL Treatment-Infection Model (CLL-TIM), to select patients for a randomized clinical trial (PreVent-ACaLL, NCT03868722), investigating whether three months of venetoclax and acalabrutinib combination therapy can improve the natural history of immune dysfunction due to CLL To our knowledge, this is the first time a machine learning model will be used for patient selection in a randomized clinical trial. As a result of our multiple-outlooks approach, we were able to substantially increase the number of patients detected as highrisk, and even under high rates of missing data, provided predictions for all CLL patients By modeling both infectious disease and CLL treatment events as an outcome, we establish a link between immune dysfunction and progressive disease in CLL, and demonstrate the complexity and non-linearity of risk factors contributing to immune dysfunction and treatment need. Through our online version of CLL-TIM, CLL-TIM.org, we provide explainable predictions by accompanying them with uncertainty estimates and personalized risk factors driving a given patient’s predicted risk
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
