Automated prediction of low ferritin concentrations using a machine learning algorithm

Abstract

Objectives Computational algorithms for the interpretation of laboratory test results can support physicians and specialists in laboratory medicine. The aim of this study was to develop, implement and evaluate a machine learning algorithm that automatically assesses the risk of low body iron storage, reflected by low ferritin plasma levels, in anemic primary care patients using a minimal set of basic laboratory tests, namely complete blood count and C-reactive protein (CRP). Methods Laboratory measurements of anemic primary care patients were used to develop and validate a machine learning algorithm. The performance of the algorithm was compared to twelve specialists in laboratory medicine from three large teaching hospitals, who predicted if patients with anemia have low ferritin levels based on laboratory test reports (complete blood count and CRP). In a second round of assessments the algorithm outcome was provided to the specialists in laboratory medicine as a decision support tool. Results Two separate algorithms to predict low ferritin concentrations were developed based on two different chemistry analyzers, with an area under the curve of the ROC of 0.92 (Siemens) and 0.90 (Roche). The specialists in laboratory medicine were less accurate in predicting low ferritin concentrations compared to the algorithms, even when knowing the output of the algorithms as support tool. Implementation of the algorithm in the laboratory system resulted in one new iron deficiency diagnosis on average per day. Conclusions Low ferritin levels in anemic patients can be accurately predicted using a machine learning algorithm based on routine laboratory test results. Moreover, implementation of the algorithm in the laboratory system reduces the number of otherwise unrecognized iron deficiencies.

Classification and Explanation of Iron Deficiency Anemia from Complete Blood Count Data Using Machine Learning

“…It was found that having a low blood level of hemoglobin (LBXHGB), older age (RIDAGEYR), higher RDW (LBXRDW), being female and pregnant, and having lower values of MCH, MCV, HCT, lymphocytes, and monocytes contribute to the prediction of IDA class, with relative importances. While all these variable contributions are consistent with the known literature [6][7][8]27,28] the contribution of lower lymphocytes and monocytes to IDA could be mediated by the effect of inflammation on serum ferritin levels [5,39,41].…”

“…The published ML models for anemia classification all use different data than used here or perform a task other than discriminating IDA. For example, they rely on image data, or they use CBC variables to identify genetic disorders related to hemoglobin [10][11][12][13][14][15][16][17][18][19][20][27][28][29][30]. Since the cause of anemia is multifactorial [4], identifying concurrent iron deficiency is required to guide iron therapy.…”

“…In the few other studies that include IDA classification, either serum iron parameters or serum ferritin was used as features along with CBC data [23][24][25][26][27]. However, a recent study reported the prediction of low ferritin levels or IDA among adult anemic subjects (more than 18 years of age) in referred lab tests with 90-98% specificity and sensitivity, using a random forest algorithm [28]. However, as suggested in a recent review [29], these models need to be validated in larger datasets and across larger age and gender subgroups to show robust generalizability.…”

Clinlabomics: leveraging clinical laboratory data by data mining strategies

“…Clinlabomics also can predict many diseases, including cancer [66][67][68][69]. As everyone knows, diabetes is a global epidemic, chronic and incurable and long-term exposure to hyperglycemia can cause chronic damage to various tissues [70].…”

Artificial intelligence in clinical chemistry – Boon or a bane

“…7 Kurstjens et al, has developed an algorithm to predict low ferritin levels based on data available from 3,797 primary care anemic patients. 8 Finally, there is an urgent need to integrate the expertise available in clinical laboratory with AI experts for the needed digital transformation to occur in the near future. It is appropriate to manage the huge data generated by the clinical laboratory to extend the additional benefits to patient care.…”