Volume 60, Issue 12 - Special Issue: Artificial Intelligence and Big Data in Laboratory Medicine; Issue Editors: Andrea Padoan and Mario Plebani

Artificial intelligence plays an important role in contemporary medicine. In this short note, we emphasize that philosophy played a role in the development of artificial intelligence. We argue that research in computability theory, the theoretical foundation of modern computer science, was motivated by a philosophical question: can we characterize precisely the class of problems that can be solved by algorithms? We suggest that reflecting on the connection between philosophy and artificial intelligence helps realize that philosophical and scientific progress are connected.

The contribution of laboratory medicine in delivering value-based care depends on active cooperation and trust between pathologist and clinician. The effectiveness of medicine more in general depends in turn on active cooperation and trust between clinician and patient. From the second half of the 20th century, the art of medicine is challenged by the spread of artificial intelligence (AI) technologies, recently showing comparable performances to flesh-and-bone doctors in some diagnostic specialties. Being the principle source of data in medicine, the laboratory is a natural ground where AI technologies can disclose the best of their potential. In order to maximize the expected outcomes and minimize risks, it is crucial to define ethical requirements for data collection and interpretation by-design, clarify whether they are enhanced or challenged by specific uses of AI technologies, and preserve these data under rigorous but feasible norms. From 2018 onwards, the European Commission (EC) is making efforts to lay the foundations of sustainable AI development among European countries and partners, both from a cultural and a normative perspective. Alongside with the work of the EC, the United Kingdom provided worthy-considering complementary advice in order to put science and technology at the service of patients and doctors. In this paper we discuss the main ethical challenges associated with the use of AI technologies in pathology and laboratory medicine, and summarize the most pertaining key-points from the guidelines and frameworks before-mentioned.

During the last few years, clinical laboratories have faced a sea change, from facilities producing a high volume of low-cost test results, toward a more integrated and patient-centered service. Parallel to this paradigm change, the digitalization of healthcare data has made an enormous quantity of patients’ data easily accessible, thus opening new scenarios for the utilization of artificial intelligence (AI) tools. Every day, clinical laboratories produce a huge amount of information, of which patients’ results are only a part. The laboratory information system (LIS) may include other “relevant” compounding data, such as internal quality control or external quality assessment (EQA) results, as well as, for example, timing of test requests and of blood collection and exams transmission, these data having peculiar characteristics typical of big data, as volume, velocity, variety, and veracity, potentially being used to generate value in patients’ care. Despite the increasing interest expressed in AI and big data in laboratory medicine, these topics are approaching the discipline slowly for several reasons, attributable to lack of knowledge and skills but also to poor or absent standardization, harmonization and problematic regulatory and ethical issues. Finally, it is important to bear in mind that the mathematical postulation of algorithms is not sufficient for obtaining useful clinical tools, especially when biological parameters are not evaluated in the appropriate context. It is therefore necessary to enhance cooperation between laboratory and AI experts, and to coordinate and govern processes, thus favoring the development of valuable clinical tools.

Objectives The development and use of artificial intelligence (AI) methodologies, especially machine learning (ML) and deep learning (DL), have been considerably fostered during the ongoing coronavirus disease 2019 (COVID-19) pandemic. Several models and algorithms have been developed and applied for both identifying COVID-19 cases and for assessing and predicting the risk of developing unfavourable outcomes. Our aim was to summarize how AI is being currently applied to COVID-19. Methods We conducted a PubMed search using as query MeSH major terms “Artificial Intelligence” AND “COVID-19”, searching for articles published until December 31, 2021, which explored the possible role of AI in COVID-19. The dataset origin (internal dataset or public datasets available online) and data used for training and testing the proposed ML/DL model(s) were retrieved. Results Our analysis finally identified 292 articles in PubMed. These studies displayed large heterogeneity in terms of imaging test, laboratory parameters and clinical-demographic data included. Most models were based on imaging data, in particular CT scans or chest X-rays images. C-Reactive protein, leukocyte count, creatinine, lactate dehydrogenase, lymphocytes and platelets counts were found to be the laboratory biomarkers most frequently included in COVID-19 related AI models. Conclusions The lion’s share of AI applied to COVID-19 seems to be played by diagnostic imaging. However, AI in laboratory medicine is also gaining momentum, especially with digital tools characterized by low cost and widespread applicability.

The current gold standard for COVID-19 diagnosis, the rRT-PCR test, is hampered by long turnaround times, probable reagent shortages, high false-negative rates and high prices. As a result, machine learning (ML) methods have recently piqued interest, particularly when applied to digital imagery (X-rays and CT scans). In this review, the literature on ML-based diagnostic and prognostic studies grounded on hematochemical parameters has been considered. By doing so, a gap in the current literature was addressed concerning the application of machine learning to laboratory medicine. Sixty-eight articles have been included that were extracted from the Scopus and PubMed indexes. These studies were marked by a great deal of heterogeneity in terms of the examined laboratory test and clinical parameters, sample size, reference populations, ML algorithms, and validation approaches. The majority of research was found to be hampered by reporting and replicability issues: only four of the surveyed studies provided complete information on analytic procedures (units of measure, analyzing equipment), while 29 provided no information at all. Only 16 studies included independent external validation. In light of these findings, we discuss the importance of closer collaboration between data scientists and medical laboratory professionals in order to correctly characterise the relevant population, select the most appropriate statistical and analytical methods, ensure reproducibility, enable the proper interpretation of the results, and gain actual utility by using machine learning methods in clinical practice.

Objectives Turnaround time (TAT) is an essential performance indicator of a medical diagnostic laboratory. Accurate TAT prediction is crucial for taking timely action in case of prolonged TAT and is important for efficient organization of healthcare. The objective was to develop a model to accurately predict TAT, focusing on the automated pre-analytical and analytical phase. Methods A total of 90,543 clinical chemistry samples from Erasmus MC were included and 39 features were analyzed, including priority level and workload in the different stages upon sample arrival. PyCaret was used to evaluate and compare multiple regression models, including the Extra Trees (ET) Regressor, Ridge Regression and K Neighbors Regressor, to determine the best model for TAT prediction. The relative residual and SHAP (SHapley Additive exPlanations) values were plotted for model evaluation. Results The regression-tree-based method ET Regressor performed best with an R 2 of 0.63, a mean absolute error of 2.42 min and a mean absolute percentage error of 7.35%, where the average TAT was 30.09 min. Of the test set samples, 77% had a relative residual error of at most 10%. SHAP value analysis indicated that TAT was mainly influenced by the workload in pre-analysis upon sample arrival and the number of modules visited. Conclusions Accurate TAT predictions were attained with the ET Regressor and features with the biggest impact on TAT were identified, enabling the laboratory to take timely action in case of prolonged TAT and helping healthcare providers to improve planning of scarce resources to increase healthcare efficiency.

Objectives Automated machine learning (AutoML) tools can help clinical laboratory professionals to develop machine learning models. The objective of this study was to develop a novel formula for the estimation of urine osmolality using an AutoML tool and to determine the efficiency of AutoML tools in a clinical laboratory setting. Methods Three hundred routine urinalysis samples were used for reference osmolality and urine clinical chemistry analysis. The H2O AutoML engine completed the machine learning development steps with minimum human intervention. Four feature groups were created, which include different urinalysis measurements according to the Boruta feature selection algorithm. Method comparison statistics including Spearman’s correlation, Passing–Bablok regression analysis were performed, and Bland Altman plots were created to compare model predictions with the reference method. The minimum allowable bias (24.17%) from biological variation data was used as the limit of agreement. Results The AutoML engine developed a total of 183 ML models. Conductivity and specific gravity had the highest variable importance. Models that include conductivity, specific gravity, and other urinalysis parameters had the highest R 2 (0.70–0.83), and 70–84% of results were within the limit of agreement. Conclusions Combining urinary conductivity with other urinalysis parameters using validated machine learning models can yield a promising surrogate. Additionally, AutoML tools facilitate the machine learning development cycle and should be considered for developing ML models in clinical laboratories.

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.

Objectives Mass spectrometry-based steroidomics combined with machine learning (ML) provides a potentially powerful approach in endocrine diagnostics, but is hampered by limitations in the conveyance of results and interpretations to clinicians. We address this shortcoming by integration of the two technologies with a laboratory information management systems (LIMS) model. Methods The approach involves integration of ML algorithm-derived models with commercially available mathematical programming software and a web-based LIMS prototype. To illustrate clinical utility, the process was applied to plasma steroidomics data from 22 patients tested for primary aldosteronism (PA). Results Once mass spectrometry data are uploaded into the system, automated processes enable generation of interpretations of steroid profiles from ML models. Generated reports include plasma concentrations of steroids in relation to age- and sex-specific reference intervals along with results of ML models and narrative interpretations that cover probabilities of PA. If PA is predicted, reports include probabilities of unilateral disease and mutations of KCNJ5 known to be associated with successful outcomes of adrenalectomy. Preliminary results, with no overlap in probabilities of disease among four patients with and 18 without PA and correct classification of all four patients with unilateral PA including three of four with KCNJ5 mutations, illustrate potential utility of the approach to guide diagnosis and subtyping of patients with PA. Conclusions The outlined process for integrating plasma steroidomics data and ML with LIMS may facilitate improved diagnostic-decision-making when based on higher-dimensional data otherwise difficult to interpret. The approach is relevant to other diagnostic applications involving ML.

Objectives The present study was conducted to improve the performance of predictive methods by introducing the most important factors which have the highest effects on the prediction of esophageal varices (EV) grades among patients with cirrhosis. Methods In the present study, the ensemble learning methods, including Catboost and XGB classifier, were used to choose the most potent predictors of EV grades solely based on routine laboratory and clinical data, a dataset of 490 patients with cirrhosis gathered. To increase the validity of the results, a five-fold cross-validation method was applied. The model was conducted using python language, Anaconda open-source platform. TRIPOD checklist for prediction model development was completed. Results The Catboost model predicted all the targets correctly with 100% precision. However, the XGB classifier had the best performance for predicting grades 0 and 1, and totally the accuracy was 91.02%. The most significant variables, according to the best performing model, which was CatBoost, were child score, white blood cell (WBC), vitalism K (K), and international normalized ratio (INR). Conclusions Using machine learning models, especially ensemble learning models, can remarkably increase the prediction performance. The models allow practitioners to predict EV risk at any clinical visit and decrease unneeded esophagogastroduodenoscopy (EGD) and consequently reduce morbidity, mortality, and cost of the long-term follow-ups for patients with cirrhosis.

Objectives The aim of the study was to implement a non-invasive model to predict ascites grades among patients with cirrhosis. Methods In the present study, we used modern machine learning (ML) methods to develop a scoring system solely based on routine laboratory and clinical data to help physicians accurately diagnose and predict different degrees of ascites. We used ANACONDA3–5.2.0 64 bit, free and open-source platform distribution of Python programming language with numerous modules, packages, and rich libraries that provide various methods for classification problems. Through the 10-fold cross-validation, we employed three common learning models on our dataset, k-nearest neighbors (KNN), support vector machine (SVM), and neural network classification algorithms. Results According to the data received from the research institute, three types of data analysis have been performed. The algorithms used to predict ascites were KNN, cross-validation (CV), and multilayer perceptron neural networks (MLPNN), which achieved an average accuracy of 94, 91, and 90%, respectively. Also, in the average accuracy of the algorithms, KNN had the highest accuracy of 94%. Conclusions We applied well-known ML approaches to predict ascites. The findings showed a strong performance compared to the classical statistical approaches. This ML-based approach can help to avoid unnecessary risks and costs for patients with acute stages of the disease.

Objectives All patients with cirrhosis should be periodically examined for esophageal varices (EV), however, a large percentage of patients undergoing screening, do not have EV or have only mild EV and do not have high-risk characteristics. Therefore, developing a non-invasive method to predict the occurrence of EV in patients with liver cirrhosis as a non-invasive method with high accuracy seems useful. In the present research, we compared the performance of several machine learning (ML) methods to predict EV on laboratory and clinical data to choose the best model. Methods Four-hundred-and-ninety data from the Liver and Gastroenterology Research Center of Shahid Beheshti University of Medical Sciences in the period 2014–2021, were analyzed applying models including random forest (RF), artificial neural network (ANN), support vector machine (SVM), and logistic regression. Results RF and SVM had the best results in general for all grades of EV. RF showed remarkably better results and the highest area under the curve (AUC). After that, SVM and ANN had the AUC of 98%, for grade 3, the SVM algorithm had the highest AUC after RF (89%). Conclusions The findings may help to better predict EV with high precision and accuracy and also can help reduce the burden of frequent visits to endoscopic centers. It can also help practitioners to manage cirrhosis by predicting EV with lower costs.

Objectives Hepatocellular carcinoma (HCC) is difficult to diagnose early and progresses rapidly, making it one of the most deadly malignancies worldwide. This study aimed to evaluate whether salivary glycopattern changes combined with machine learning algorithms could help in the accurate diagnosis of HCC. Methods Firstly, we detected the alteration of salivary glycopatterns by lectin microarrays in 118 saliva samples. Subsequently, we constructed diagnostic models for hepatic cirrhosis (HC) and HCC using three machine learning algorithms: Least Absolute Shrinkage and Selector Operation, Support Vector Machine (SVM), and Random Forest (RF). Finally, the performance of the diagnostic models was assessed in an independent validation cohort of 85 saliva samples by a series of evaluation metrics, including area under the receiver operator curve (AUC), accuracy, specificity, and sensitivity. Results We identified alterations in the expression levels of salivary glycopatterns in patients with HC and HCC. The results revealed that the glycopatterns recognized by 22 lectins showed significant differences in the saliva of HC and HCC patients and healthy volunteers. In addition, after Boruta feature selection, the best predictive performance was obtained with the RF algorithm for the construction of models for HC and HCC. The AUCs of the RF-HC model and RF-HCC model in the validation cohort were 0.857 (95% confidence interval [CI]: 0.780–0.935) and 0.886 (95% CI: 0.814–0.957), respectively. Conclusions Detecting alterations in salivary protein glycopatterns with lectin microarrays combined with machine learning algorithms could be an effective strategy for diagnosing HCC in the future.

Artificial intelligence (AI) is a branch of computer science that includes research in robotics, language recognition, image recognition, natural language processing, and expert systems. AI is poised to change medical practice, and oncology is not an exception to this trend. As the matter of fact, lung cancer has the highest morbidity and mortality worldwide. The leading cause is the complexity of associating early pulmonary nodules with neoplastic changes and numerous factors leading to strenuous treatment choice and poor prognosis. AI can effectively enhance the diagnostic efficiency of lung cancer while providing optimal treatment and evaluating prognosis, thereby reducing mortality. This review seeks to provide an overview of AI relevant to all the fields of lung cancer. We define the core concepts of AI and cover the basics of the functioning of natural language processing, image recognition, human-computer interaction and machine learning. We also discuss the most recent breakthroughs in AI technologies and their clinical application regarding diagnosis, treatment, and prognosis in lung cancer. Finally, we highlight the future challenges of AI in lung cancer and its impact on medical practice.

Objectives Delta check (DC) is widely used for detecting sample mix-up. Owing to the inadequate error detection and high false-positive rate, the implementation of DC in real-world settings is labor-intensive and rarely capable of absolute detection of sample mix-ups. The aim of the study was to develop a highly accurate DC method based on designed deep learning to detect sample mix-up. Methods A total of 22 routine hematology test items were adopted for the study. The hematology test results, collected from two hospital laboratories, were independently divided into training, validation, and test sets. By selecting six mainstream algorithms, the Deep Belief Network (DBN) was able to learn error-free and artificially (intentionally) mixed sample results. The model’s analytical performance was evaluated using training and test sets. The model’s clinical validity was evaluated by comparing it with three well-recognized statistical methods. Results When the accuracy of our model in the training set reached 0.931 at the 22nd epoch, the corresponding accuracy in the validation set was equal to 0.922. The loss values for the training and validation sets showed a similar (change) trend over time. The accuracy in the test set was 0.931 and the area under the receiver operating characteristic curve was 0.977. DBN demonstrated better performance than the three comparator statistical methods. The accuracy of DBN and revised weighted delta check (RwCDI) was 0.931 and 0.909, respectively. DBN performed significantly better than RCV and EDC. Of all test items, the absolute difference of DC yielded higher accuracy than the relative difference for all methods. Conclusions The findings indicate that input of a group of hematology test items provides more comprehensive information for the accurate detection of sample mix-up by machine learning (ML) when compared with a single test item input method. The DC method based on DBN demonstrated highly effective sample mix-up identification performance in real-world clinical settings.

Objectives Artificial intelligence (AI) models are increasingly being developed for clinical chemistry applications, however, it is not understood whether human interaction with the models, which may occur once they are implemented, improves or worsens their performance. This study examined the effect of human supervision on an artificial neural network trained to identify wrong blood in tube (WBIT) errors. Methods De-identified patient data for current and previous (within seven days) electrolytes, urea and creatinine (EUC) results were used in the computer simulation of WBIT errors at a rate of 50%. Laboratory staff volunteers reviewed the AI model’s predictions, and the EUC results on which they were based, before making a final decision regarding the presence or absence of a WBIT error. The performance of this approach was compared to the performance of the AI model operating without human supervision. Results Laboratory staff supervised the classification of 510 sets of EUC results. This workflow identified WBIT errors with an accuracy of 81.2%, sensitivity of 73.7% and specificity of 88.6%. However, the AI model classifying these samples autonomously was superior on all metrics (p-values<0.05), including accuracy (92.5%), sensitivity (90.6%) and specificity (94.5%). Conclusions Human interaction with AI models can significantly alter their performance. For computationally complex tasks such as WBIT error identification, best performance may be achieved by autonomously functioning AI models.

Objectives Patient-based real-time quality control (PBRTQC) has gained attention as an alternative/integrative tool for internal quality control (iQC). However, it is still doubted for its performance and its application in real clinical settings. We aim to generate a newly and easy-to-access patient-based real-time QC by machine learning (ML) traceable to standard reference data with assigned values by National Institute of Metrology of China (NIM), and to compare it with PBRTQC for clinical validity evaluation. Methods For five representative biochemistry analytes, 1,195 000 patient testing results each were collected. After data processing, independent training and test sets were divided. Machine learning internal quality control (MLiQC) was set up by Random Forest in ML and was validated by way of both metrology algorithm traceability and 4 PBRTQC methods recommended by IFCC analytical working group. Results MLiQC were established. As an example of albumin (ALB) at the critical bias, the uncertainty of MLiQC was 0.14%, which was evaluated by standard reference data produced by NIM. Compared with four optimal PBRTQC methods at critical bias, the average of the number of patient samples from a bias introduced until detected (ANPed) of MLiQC averagely decreased from 600 to 20. The median and 95 quantiles of NPeds (MNPed and 95NPed) of MLiQC were superior to all optimal PBRTQCs above 90% for all test items. Conclusions MLiQC is highly superior to PBRTQC and well-suited in real settings. The validation of the model from two aspects of algorithm traceability and clinical effectiveness confirms its satisfactory performance.

Objectives To evaluate the ability of an artificial intelligence (AI) model to predict the risk of cancer in patients referred from primary care based on routine blood tests. Results obtained with the AI model are compared to results based on logistic regression (LR). Methods An analytical profile consisting of 25 predefined routine laboratory blood tests was introduced to general practitioners (GPs) to be used for patients with non-specific symptoms, as an additional tool to identify individuals at increased risk of cancer. Consecutive analytical profiles ordered by GPs from November 29th 2011 until March 1st 2020 were included. AI and LR analysis were performed on data from 6,592 analytical profiles for their ability to detect cancer. Cohort I for model development included 5,224 analytical profiles ordered by GP’s from November 29th 2011 until the December 31st 2018, while 1,368 analytical profiles included from January 1st 2019 until March 1st 2020 constituted the “out of time” validation test Cohort II. The main outcome measure was a cancer diagnosis within 90 days. Results The AI model based on routine laboratory blood tests can provide an easy-to use risk score to predict cancer within 90 days. Results obtained with the AI model were comparable to results from the LR model. In the internal validation Cohort IB, the AI model provided slightly better results than the LR analysis both in terms of the area under the receiver operating characteristics curve (AUC) and PPV, sensitivity/specificity while in the “out of time” validation test Cohort II, the obtained results were comparable. Conclusions The AI risk score may be a valuable tool in the clinical decision-making. The score should be further validated to determine its applicability in other populations.

Objectives The Italian Society of Clinical Biochemistry and Clinical Molecular Biology (SIBioC) Big Data and Artificial Intelligence (BAI) Working Group promoted a survey to frame the knowledge, skills and technological predisposition in clinical laboratories. Methods A questionnaire, focussing on digitization, information technology (IT) infrastructures, data accessibility, and BAI projects underway was sent to 1,351 SIBioC participants. The responses were evaluated using SurveyMonkey software and Google Sheets. Results The 227 respondents (17%) from all over Italy (47% of 484 labs), mainly biologists, laboratory physicians and managers, mostly from laboratories of public hospitals, revealed lack of hardware, software and corporate Wi-Fi, and dearth of PCs. Only 25% work daily on clouds, while 65%—including Laboratory Directors—cannot acquire health data from sources other than laboratories. Only 50% of those with access can review a clinical patient’s health record, while the other access only to laboratory information. The integration of laboratory data with other health data is mostly incomplete, which limits BAI-type analysis. Many are unaware of integration platforms. Over 90% report pulling data from the Laboratory Information System, with varying degrees of autonomy. Very few have already undertaken BAI projects, frequently relying on IT partnerships. The majority consider BAI as crucial in helping professional judgements, indicating a growing interest. Conclusions The questionnaire received relevant feedback from SIBioC participants. It highlighted the level of expertise and interest in BAI applications. None of the obstacles stands out more than the others, emphasising the need to all-around work: IT infrastructures, data warehouses, BAI analysis software acquisition, data accessibility and training.