Prediction of Neutropenic Events in Chemotherapy Patients: A Machine Learning Approach

The Use of Artificial Intelligence in Predicting Chemotherapy-Induced Toxicities in Metastatic Colorectal Cancer: A Data-Driven Approach for Personalized Oncology

BACKGROUND

Machine learning models learn about general behavior from data by finding the relationships between features. Our purpose was to develop a predictive model to identify and predict which subset of colorectal cancer patients are more likely to experience chemotherapy-induced toxicity and to determine the specific attributes that influence the presence of treatment-related side effects.

METHODS

The predictor was general toxicity, and for the construction of our data training, we selected 95 characteristics that represent the health state of 74 patients prior to their first round of chemotherapy. After the data were processed, Random Forest models were trained to offer an optimal balance between accuracy and interpretability.

RESULTS

We constructed a machine learning predictor with an emphasis on assessing the importance of numerical and categorical variables in relation to toxicity.

CONCLUSIONS

The incorporation of artificial intelligence in personalizing colorectal cancer management by anticipating and overseeing toxicities more effectively illustrates a pivotal shift towards more personalized and precise medical care.

Improving management of febrile neutropenia in oncology patients: the role of artificial intelligence and machine learning

INTRODUCTION

Artificial intelligence (AI) and machine learning (ML) have the potential to revolutionize the management of febrile neutropenia (FN) and drive progress toward personalized medicine.

AREAS COVERED

In this review, we detail how the collection of a large number of high-quality data can be used to conduct precise mathematical studies with ML and AI. We explain the foundations of these techniques, covering the fundamentals of supervised and unsupervised learning, as well as the most important challenges, e.g. data quality, 'black box' model interpretation and overfitting. To conclude, we provide detailed examples of how AI and ML have been used to enhance predictions of chemotherapy-induced FN, detection of bloodstream infections (BSIs) and multidrug-resistant (MDR) bacteria, and anticipation of severe complications and mortality.

EXPERT OPINION

There is promising potential of implementing accurate AI and ML models whilst managing FN. However, their integration as viable clinical tools poses challenges, including technical and implementation barriers. Improving global accessibility, fostering interdisciplinary collaboration, and addressing ethical and security considerations are essential. By overcoming these challenges, we could transform personalized care for patients with FN.

Predicting Severe Haematological Toxicity in Gastrointestinal Cancer Patients Undergoing 5-FU-Based Chemotherapy: A Bayesian Network Approach

PURPOSE

Severe toxicity is reported in about 30% of gastrointestinal cancer patients receiving 5-Fluorouracil (5-FU)-based chemotherapy. To date, limited tools exist to identify at risk patients in this setting. The objective of this study was to address this need by designing a predictive model using a Bayesian network, a probabilistic graphical model offering robust, explainable predictions.

METHODS

We utilized a dataset of 267 gastrointestinal cancer patients, conducting preprocessing, and splitting it into TRAIN and TEST sets (80%:20% ratio). The RandomForest algorithm assessed variable importance based on MeanDecreaseGini coefficient. The bnlearn R library helped design a Bayesian network model using a 10-fold cross-validation on the TRAIN set and the aic-cg method for network structure optimization. The model's performance was gauged based on accuracy, sensitivity, and specificity, using cross-validation on the TRAIN set and independent validation on the TEST set.

RESULTS

The model demonstrated satisfactory performance with an average accuracy of 0.85 (±0.05) and 0.80 on TRAIN and TEST datasets, respectively. The sensitivity and specificity were 0.82 (±0.14) and 0.87 (±0.07) for the TRAIN dataset, and 0.71 and 0.83 for the TEST dataset, respectively. A user-friendly tool was developed for clinical implementation.

CONCLUSIONS

Despite several limitations, our Bayesian network model demonstrated a high level of accuracy in predicting the risk of developing severe haematological toxicity in gastrointestinal cancer patients receiving 5-FU-based chemotherapy. Future research should aim at model validation in larger cohorts of patients and different clinical settings.

Risk of chemotherapy-induced febrile neutropenia in intermediate-risk regimens: Clinical and economic outcomes of granulocyte colony-stimulating factor prophylaxis

BACKGROUND: Chemotherapy-induced neutropenia increases the risk of febrile neutropenia (FN) and infection with resultant hospitalizations, with substantial health care resource utilization (HCRU) and costs. Granulocyte-colony stimulating factor (GCSF) is recommended as primary prophylaxis for chemotherapy regimens having more than a 20% risk of FN. Yet, for intermediate-risk (10%-20%) regimens, it should be considered only for patients with 1 or more clinical risk factors (RFs) for FN. It is unclear whether FN prophylaxis for intermediate-risk patients is being optimally implemented. OBJECTIVE: To examine RFs, prophylaxis use, HCRU, and costs associated with incident FN during chemotherapy. METHODS: This retrospective study used administrative claims data for commercial and Medicare Advantage enrollees with nonmyeloid cancer treated with intermediate-risk chemotherapy regimens during January 1, 2009, to March 31, 2020. Clinical RFs, GCSF prophylaxis, incident FN, HCRU, and costs were analyzed descriptively by receipt of primary GCSF, secondary GCSF, or no GCSF prophylaxis. Multivariable Cox regression analysis was used to examine the association between number of RFs and cumulative FN risk. RESULTS: The sample comprised 13,937 patients (mean age 67 years, 55% female). Patients had a mean of 2.3 RFs, the most common being recent surgery, were aged 65 years or greater, and had baseline liver or renal dysfunction; 98% had 1 or more RFs. However, only 35% of patients received primary prophylaxis; 12% received secondary prophylaxis. The hazard ratio of incident FN was higher with increasing number of RFs during the first line of therapy, yet more than 54% of patients received no prophylaxis, regardless of RFs. Use of GCSF prophylaxis varied more by chemotherapeutic regimen than by number of RFs. Among patients treated with rituximab, cyclophosphamide, hydroxydaunorubicin hydrochloride (doxorubicin hydrochloride), vincristine, and prednisone, 76% received primary prophylaxis, whereas only 22% of patients treated with carboplatin/paclitaxel received primary prophylaxis. Among patients with a first line of therapy FN event, 78% had an inpatient stay and 42% had an emergency visit. During cycle 1, mean FN-related coordination of benefits-adjusted medical costs per patient per month ($13,886 for patients with primary prophylaxis and $18,233 for those with none) were driven by inpatient hospitalizations, at 91% and 97%, respectively. CONCLUSIONS: Incident FN occurred more often with increasing numbers of RFs, but GCSF prophylaxis use did not rise correspondingly. Variation in prophylaxis use was greater based on regimen than RF number. Lower health care costs were observed among patients with primary prophylaxis use. Improved individual risk identification for intermediate-risk regimens and appropriate prophylaxis may decrease FN events toward the goal of better clinical and health care cost outcomes. DISCLOSURES: This work was funded by Sandoz Inc., which participated in the design of the study, interpretation of the data, writing and revision of the manuscript, and the decision to submit the manuscript for publication. The study was performed by Optum under contract with Sandoz Inc. The author(s) meet criteria for authorship as recommended by the International Committee of Medical Journal Editors. The authors received no direct compensation related to the development of the manuscript. Dr Li is an employee of Sandoz Inc. Drs Bell and Lal and Mr Peterson-Brandt were employees of Optum at the time of the study. Ms Anderson and Dr Aslam are employees of Optum. Dr Lyman has been primary investigator on a research grant from Amgen to their institution and has consulted for Sandoz, G1 Therapeutics, Partners Healthcare, BeyondSpring, ER Squibb, Merck, Jazz Pharm, Kallyope, Teva; Fresenius Kabi, Seattle Genetics, and Samsung.

Prediction of Multiple Clinical Complications in Cancer Patients to Ensure Hospital Preparedness and Improved Cancer Care

Reliable and rapid medical diagnosis is the cornerstone for improving the survival rate and quality of life of cancer patients. The problem of clinical decision-making pertaining to the management of patients with hematologic cancer is multifaceted and intricate due to the risk of therapy-induced myelosuppression, multiple infections, and febrile neutropenia (FN). Myelosuppression due to treatment increases the risk of sepsis and mortality in hematological cancer patients with febrile neutropenia. A high prevalence of multidrug-resistant organisms is also noted in such patients, which implies that these patients are left with limited or no-treatment options amidst severe health complications. Hence, early screening of patients for such organisms in their bodies is vital to enable hospital preparedness, curtail the spread to other weak patients in hospitals, and limit community outbreaks. Even though predictive models for sepsis and mortality exist, no model has been suggested for the prediction of multidrug-resistant organisms in hematological cancer patients with febrile neutropenia. Hence, for predicting three critical clinical complications, such as sepsis, the presence of multidrug-resistant organisms, and mortality, from the data available from medical records, we used 1166 febrile neutropenia episodes reported in 513 patients. The XGboost algorithm is suggested from 10-fold cross-validation on 6 candidate models. Other highlights are (1) a novel set of easily available features for the prediction of the aforementioned clinical complications and (2) the use of data augmentation methods and model-scoring-based hyperparameter tuning to address the problem of class disproportionality, a common challenge in medical datasets and often the reason behind poor event prediction rate of various predictive models reported so far. The proposed model depicts improved recall and AUC (area under the curve) for sepsis (recall = 98%, AUC = 0.85), multidrug-resistant organism (recall = 96%, AUC = 0.91), and mortality (recall = 86%, AUC = 0.88) prediction. Our results encourage the need to popularize artificial intelligence-based devices to support clinical decision-making.