Artificial intelligence-driven diagnosis of β-thalassemia minor & iron deficiency anemia using machine learning models

Background

Iron deficiency anemia (IDA) and b-thalassemia minor (BTM) are the two most common causes of microcytic anemia, and although these conditions do not share many symptoms, differential diagnosis by blood tests is a time-consuming and expensive process. CBC can be used to diagnose anemia, but without advanced techniques, it cannot differentiate between iron deficiency anemia and BTM. This makes the differential diagnosis of IDA and BTM costly, as it requires advanced techniques to differentiate between the two conditions. This study aims to develop a model to differentiate IDA from BTM using an automated machine-learning method using only CBC data.

Methods

This retrospective study included 396 individuals, consisting of 216 IDAs and 180 BTMs. The work was divided into three parts. The first section focused on the individual effects of hematological parameters on the differentiation of IDA and BTM. The second part discusses traditional methods and discriminant indices used in diagnosis. In the third section, models developed using artificial neural networks (ANN) and decision trees are analysed and compared with the methods used in the first two sections.

Machine learning models can predict the presence of variants in hemoglobin: artificial neural network-based recognition of human hemoglobin variants by HPLC

Objectives

This article presents the use of machine learning techniques such as artificial neural networks, K-nearest neighbors (KNN), naive Bayes, and decision trees in the prediction of hemoglobin variants. To the best of our knowledge, this is the first study using machine learning models to predict suspicious cases with HbS or HbD Los Angeles carriers state.

Methods

We had a dataset of 238 observations, of which 128 were HbD carriers, and 110 were HbS carriers. The features were age, sex, RBC, Hb, HTC, MCV, MCH, RDW, serum iron, TIBC, ferritin, HbA2, HbF, HbA0, retention time (RT) of the abnormal peak, and the area under the peak of the abnormal peak. KNN, naive Bayes, decision tree models, and artificial neural network models were trained. Model performances were estimated using 7-fold cross-validation.

Results

When RT, the key point of differentiation used in high-performance liquid chromatography (HPLC), was included as a feature, all models performed well. When RT was excluded (eliminated), the deep learning model performed the best (Accuracy: 0.99; Specificity: 0.99; Sensitivity: 0.99; F1 score: 0.99), while the naive Bayes model performed the worst (Accuracy: 0.94; Specificity: 0.97; Sensitivity: 0.90; F1 score: 0.93).

Conclusions

Deep learning and decision tree models have demonstrated high performance and have the potential to be integrated into medical laboratory work practices as a tool for hemoglobinopathy detection. These outcomes suggest that when machine learning models are fed enough data, they can detect a wide range of hemoglobin variants. However, more comprehensive studies with data from a larger number of patients and hemoglobinopathies will be useful for validating our models.

IVS-II-16 (G>C) (HBB: c.315+16G>C) or IVS-II-666 (C>T) (HBB: c.316-185C>T) Mutations Trigger an Hb S (HBB: c.20A>T)/β+-Thalassemia Phenotype in an Hb S Trait Patient

Sickle cell trait is a medical condition caused by the presence of both mutant Hb S (HBB: c.20A>T) and normal Hb A alleles. Although sickle cell trait is typically considered to be asymptomatic and benign, genetic modifiers and mutations can lead to severe clinical complications. In this study, the possible pathogenicity of the IVS-II-16 (G>C) (HBB: c.315+16G>C) and IVS-II-666 (C>T) (HBB: c.316-185C>T) mutations, which are considered to be neutral polymorphisms, and the association between the Hb S mutation are presented. To the best of our knowledge, these polymorphisms have not been previously reported in any sickle cell trait patient, and no relevant studies have been conducted. We recently studied a 40-year-old woman (proband), diagnosed to be an Hb S/β-thalassemia (β-thal) carrier. β-Globin mutations were analyzed using a DNA sequencer based on the Sanger method. The HbVar and ClinVar databases show IVS-II-16 and IVS-II-666 to be intronic mutations. However, statements in these data banks contradict our findings. In the present study, a transfusion-dependent Hb S patient, behaving as an Hb S/β-thal case due to these mutations, was reported. These mutations have not been previously reported in an Hb S patient. Although the IVS-II-16 and IVS-II-666 mutations were previously reported as benign, they converted the Hb S phenotype to transfusion-dependent Hb S/β-thal when combined with Hb S. In this regard, IVS-II-16 and IVS-II-666 mutations may not be innocent, as previously thought.

Difficulties in the diagnosis of Hb S/Beta thalassemia: Really a mild disease?

Background

HbS/b cases having clinical, hematologic and electrophoretic similarities cannot be sufficiently distinguished from sickle cell anemia cases and are misdiagnosed as sickle cell anemia. This study will investigate the congruence between the HPLC thalassemia scanning tests and the laboratory findings compared to the DNA sequence analysis results of the patients diagnosed with SCA between 2016 and 2020. This study also aims to indicate the current status to accurately diagnose sickle cell anemia and HbS/b in the light of hematologic, electrophoretic and molecular studies.

Methods

Fourteen patients who were diagnosed with SCA in hospitals at different cities in Turkey and followed by the Thalassemia Diagnosis, Treatment and Research Center, Muğla Sıtkı Koçman University were included in this retrospective study. The socio-demographic characteristics, hemogram, hemoglobin variant analysis results and DNA chain analysis results of the patients were taken from the database of the centre and then examined. The informed consents were taken from the patients. The patients were administered a survey containing questions about transfusion history and diagnostic awareness. The Beta-Thalassemia mutations were analysed using a DNA sequencer (Dade Behring, Germany) based on the Sanger method.

Results

According to the DNA sequence analysis, the results of these patients diagnosed with SCA in hospitals in different cities of Turkey were the following: of 14 patients, 8 had HbS/b0, and HbS/b+ and one had HbS carrier, and one had Hb-O, and three had SCA. The patient with HbS carrier status also contains three additional mutations, all of which are heterozygous. We discovered that although two of three mutations, which are c.315+16G>C and c.316-185C>T, are previously reported as benign, at least one of the two mentioned mutations, when combined with HbS, causes transfusion-dependent HbS/b.

Conclusions

Briefly, HbSS and HbS/b thalassemia genotypes cannot be definitely characterized by electrophoretic and hematologic data, resulting in misdiagnosis. c.315+16G>C and c.316-185C>T are previously reported as benign; at least one of the two mentioned mutations, when combined with HbS, causes transfusion-dependent HbS/b. In undeveloped or some developing countries, molecular diagnosis methods and genetic analyses cannot be used. If mutation analyses could be performed, then such differential diagnosis errors would reduce. However, if mutation analysis cannot be performed, other methods such as HPLC, capillary electrophoresis absolutely be sought to have insight into the parental carriage status.