Machine learning applications for prediction of relapse in childhood acute lymphoblastic leukemia

Machine learning in paediatric haematological malignancies: a systematic review of prognosis, toxicity and treatment response models

BACKGROUND

Machine Learning (ML) has demonstrated potential in enhancing care in adult oncology. However, its application in paediatric haematological malignancies is still emerging, necessitating a comprehensive review of its capabilities and limitations in this area.

METHODS

A literature search was conducted through Ovid. Studies included focused on ML models in paediatric patients with haematological malignancies. Studies were categorised into thematic groups for analysis.

RESULTS

Twenty studies, primarily on leukaemia, were included in this review. Studies were organised into thematic categories such as prognoses, treatment responses and toxicity predictions. Prognostic studies showed AUC scores between 0.685 and 0.929, indicating moderate-high predictive accuracy. Treatment response studies demonstrated AUC scores between 0.840 and 0.875, reflecting moderate accuracy. Toxicity prediction studies reported high accuracy with AUC scores from 0.870 to 0.927. Only five studies (25%) performed external validation. Significant heterogeneity was noted in ML tasks, reporting formats, and effect measures across studies, highlighting a lack of standardised reporting and challenges in data comparability.

CONCLUSION

The clinical applicability of these ML models remains limited by the lack of external validation and methodological heterogeneity. Addressing these challenges through standardised reporting and rigorous external validation is needed to translate ML from a promising research tool into a reliable clinical practice component.

IMPACT

Key message: Machine Learning (ML) significantly enhances predictive models in paediatric haematological cancers, offering new avenues for personalised treatment strategies. Future research should focus on developing ML models that can integrate with real-time clinical workflows. Addition to literature: Provides a comprehensive overview of current ML applications and trends. It identifies limitations to its applicability, including the limited diversity in datasets, which may affect the generalisability of ML models across different populations.

IMPACT

Encourages standardisation and external validation in ML studies, aiming to improve patient outcomes through precision medicine in paediatric haematological oncology.

Bridging the gap: understanding contemporary autopsies in acute leukemia by comparing ante-mortem and post-mortem profiles

This study investigates acute myeloid leukemia/lymphoblastic leukemia (AML/ALL) through a 14-year analysis (2009-2022) of 46 autopsied cases (age >12 years). B-ALL was the dominant subtype (34.8%). Liver and spleen were the common sites of active leukemia (63% cases). Symptoms like dyspnea and altered sensorium associated significantly with heart (p = .031) and brain leukostasis (p = .006). Measurable residual disease (MRD) negativity correlated with disease-free status outside the bone marrow, while MRD-positive cases displayed leukemic infiltrates. Infections were identified in 23 autopsied cases, notably linked to post-induction and post-transplant fatalities. Surprisingly, 18 of these 23 cases had unexpected infections mainly fungal (13 cases) with Aspergillus species as the most common. Diagnostic discrepancies were identified in 48% of cases. Malignant infiltration (46%) and infections (25%) were the leading causes of death. This research sheds light on leukemia in extra-medullary tissues, uncovers novel clinical-pathological associations, and highlights overlooked therapy side effects, offering insights for future case management.

The new era of immunological treatment, last updated and future consideration of CAR T cell-based drugs

Cancer treatment is one of the fundamental challenges in clinical setting, especially in relapsed/refractory malignancies. The novel immunotherapy-based treatments bring new hope in cancer therapy and achieve various treatment successes. One of the distinguished ways of cancer immunotherapy is adoptive cell therapy, which utilizes genetically modified immune cells against cancer cells. Between different methods in ACT, the chimeric antigen receptor T cells have more investigation and introduced a promising way to treat cancer patients. This technology progressed until it introduced six US Food and Drug Administration-approved CAR T cell-based drugs. These drugs act against hematological malignancies appropriately and achieve exciting results, so they have been utilized widely in cell therapy clinics. In this review, we introduce all CAR T cells-approved drugs based on their last data and investigate them from all aspects of pharmacology, side effects, and compressional. Also, the efficacy of drugs, pre- and post-treatment steps, and expected side effects are introduced, and the challenges and new solutions in CAR T cell therapy are in the last speech.

Longitudinal clinical data improve survival prediction after hematopoietic cell transplantation using machine learning

ABSTRACT

Serial prognostic evaluation after allogeneic hematopoietic cell transplantation (allo-HCT) might help identify patients at high risk of lethal organ dysfunction. Current prediction algorithms based on models that do not incorporate changes to patients' clinical condition after allo-HCT have limited predictive ability. We developed and validated a robust risk-prediction algorithm to predict short- and long-term survival after allo-HCT in pediatric patients that includes baseline biological variables and changes in the patients' clinical status after allo-HCT. The model was developed using clinical data from children and young adults treated at a single academic quaternary-care referral center. The model was created using a randomly split training data set (70% of the cohort), internally validated (remaining 30% of the cohort) and then externally validated on patient data from another tertiary-care referral center. Repeated clinical measurements performed from 30 days before allo-HCT to 30 days afterwards were extracted from the electronic medical record and incorporated into the model to predict survival at 100 days, 1 year, and 2 years after allo-HCT. Naïve-Bayes machine learning models incorporating longitudinal data were significantly better than models constructed from baseline variables alone at predicting whether patients would be alive or deceased at the given time points. This proof-of-concept study demonstrates that unlike traditional prognostic tools that use fixed variables for risk assessment, incorporating dynamic variability using clinical and laboratory data improves the prediction of mortality in patients undergoing allo-HCT.

Machine-Learning Algorithms Using Systemic Inflammatory Markers to Predict the Oncologic Outcomes of Colorectal Cancer After Surgery

BACKGROUND

This study aimed to investigate the clinical significance of machine-learning (ML) algorithms based on serum inflammatory markers to predict survival outcomes for patients with colorectal cancer (CRC).

METHODS

The study included 941 patients with stages I to III CRC. Based on random forest algorithms using 15 compositions of inflammatory markers, four different prediction scores (DFS score-1, DFS score-2, DFS score-3, and DFS score-4) were developed for the Yonsei cohort (training set, n = 803) and tested in the Ulsan cohort (test set, n = 138). The Cox proportional hazards model was used to determine correlation between prediction scores and disease-free survival (DFS). Harrell's concordance index (C-index) was used to compare the predictive ability of prediction scores for each composition.

RESULTS

The multivariable analysis showed the DFS score-4 to be an independent prognostic factor after adjustment for clinicopathologic factors in both the training and test sets (hazard ratio [HR], 8.98; 95% confidence interval [CI] 6.7-12.04; P < 0.001 for the training set and HR, 2.55; 95% CI 1.1-5.89; P = 0.028 for the test set]. With regard to DFS, the highest C-index among single compositions was observed in the lymphocyte-to-C-reactive protein ratio (LCR) (0.659; 95% CI 0.656-0.662), and the C-index of DFS score-4 (0.727; 95% CI 0.724-0.729) was significantly higher than that of LCR in the test set. The C-index of DFS score-3 (0.725; 95% CI 0.723-0.728) was similar to that of DFS score-4, but higher than that of DFS score-2 (0.680; 95% CI 0.676-0.683).

CONCLUSIONS

The ML-based approaches showed prognostic utility in predicting DFS. They could enhance clinical use of inflammatory markers in patients with CRC.

A machine learning tool for the diagnosis of SARS-CoV-2 infection from hemogram parameters

Monocytes and neutrophils play key roles in the cytokine storm triggered by SARS-CoV-2 infection, which changes their conformation and function. These changes are detectable at the cellular and molecular level and may be different to what is observed in other respiratory infections. Here, we applied machine learning (ML) to develop and validate an algorithm to diagnose COVID-19 using blood parameters. In this retrospective single-center study, 49 hemogram parameters from 12,321 patients with clinical suspicion of COVID-19 and tested by RT-PCR (4239 positive and 8082 negative) were analysed. The dataset was randomly divided into training and validation sets. Blood cell parameters and patient age were used to construct the predictive model with the support vector machine (SVM) tool. The model constructed from the training set (5936 patients) achieved an accuracy for diagnosis of SARS-CoV-2 infection of 0.952 (95% CI: 0.875-0.892). Test sensitivity and specificity was 0.868 and 0.899, respectively, with a positive (PPV) and negative (NPV) predictive value of 0.896 and 0.872, respectively (prevalence 0.50). The validation set model (4964 patients) achieved an accuracy of 0.894 (95% CI: 0.883-0.903). Test sensitivity and specificity was 0.8922 and 0.8951, respectively, with a positive (PPV) and negative (NPV) predictive value of 0.817 and 0.94, respectively (prevalence 0.34). The area under the receiver operating characteristic curve was 0.952 for the algorithm performance. This algorithm may allow to rule out COVID-19 diagnosis with 94% of probability. This represents a great advance for early diagnostic orientation and guiding clinical decisions.

A pilot study of implication of machine learning for relapse prediction after allogeneic stem cell transplantation in adults with Ph-positive acute lymphoblastic leukemia

The posttransplant relapse in Ph-positive ALL increases the risk of death. There is an unmet need for instruments to predict the risk of relapse and plan prophylaxis. In this study, we analyzed posttransplant data by machine learning algorithms. Seventy-four Ph-positive ALL patients with a median age of 30 (range 18-55) years who previously underwent allo-HSCT, were retrospectively enrolled. Ninety-three percent of patients received prophylactic/preemptive TKIs after allo-HSCT. The values of the BCR::ABL1 level at serial assessments and over variables were collected in specified intervals after allo-HSCT. They were used to model relapse risk with several machine-learning approaches. GBM proved superior to the other algorithms and provided a maximal AUC score of 0.91. BCR::ABL1 level before and after allo-HSCT, prediction moment, and chronic GvHD had the highest value in the model. It was shown that after Day + 100, both error rates do not exceed 22%, while before D + 100, the model fails to make accurate predictions. As a result, we determined BCR::ABL1 levels at which the relapse risk remains low. Thus, the current BCR::ABL1 level less than 0.06% in patients with chronic GvHD predicts low risk of relapse. At the same time, patients without chronic GVHD after allo-HSCT should be classified as high risk with any level of BCR::ABL1. GBM model with posttransplant laboratory values of BCR::ABL1 provides a high prediction of relapse after allo-HSCT in the era of TKIs prophylaxis. Validation of this approach is warranted.

Artificial Intelligence-Assisted Diagnostic Cytology and Genomic Testing for Hematologic Disorders

Artificial intelligence (AI) is a rapidly evolving field of computer science that involves the development of computational programs that can mimic human intelligence. In particular, machine learning and deep learning models have enabled the identification and grouping of patterns within data, leading to the development of AI systems that have been applied in various areas of hematology, including digital pathology, alpha thalassemia patient screening, cytogenetics, immunophenotyping, and sequencing. These AI-assisted methods have shown promise in improving diagnostic accuracy and efficiency, identifying novel biomarkers, and predicting treatment outcomes. However, limitations such as limited databases, lack of validation and standardization, systematic errors, and bias prevent AI from completely replacing manual diagnosis in hematology. In addition, the processing of large amounts of patient data and personal information by AI poses potential data privacy issues, necessitating the development of regulations to evaluate AI systems and address ethical concerns in clinical AI systems. Nonetheless, with continued research and development, AI has the potential to revolutionize the field of hematology and improve patient outcomes. To fully realize this potential, however, the challenges facing AI in hematology must be addressed and overcome.

The development of machine learning in bariatric surgery

Background

Machine learning (ML), is an approach to data analysis that makes the process of analytical model building automatic. The significance of ML stems from its potential to evaluate big data and achieve quicker and more accurate outcomes. ML has recently witnessed increased adoption in the medical domain. Bariatric surgery, otherwise referred to as weight loss surgery, reflects the series of procedures performed on people demonstrating obesity. This systematic scoping review aims to explore the development of ML in bariatric surgery.

Methods

The study used the Preferred Reporting Items for Systematic and Meta-analyses for Scoping Review (PRISMA-ScR). A comprehensive literature search was performed of several databases including PubMed, Cochrane, and IEEE, and search engines namely Google Scholar. Eligible studies included journals published from 2016 to the current date. The PRESS checklist was used to evaluate the consistency demonstrated during the process.

Results

A total of seventeen articles qualified for inclusion in the study. Out of the included studies, sixteen concentrated on the role of ML algorithms in prediction, while one addressed ML's diagnostic capacity. Most articles (n = 15) were journal publications, whereas the rest (n = 2) were papers from conference proceedings. Most included reports were from the United States (n = 6). Most studies addressed neural networks, with convolutional neural networks as the most prevalent. Also, the data type used in most articles (n = 13) was derived from hospital databases, with very few articles (n = 4) collecting original data via observation.

Conclusions

This study indicates that ML has numerous benefits in bariatric surgery, however its current application is limited. The evidence suggests that bariatric surgeons can benefit from ML algorithms since they will facilitate the prediction and evaluation of patient outcomes. Also, ML approaches to enhance work processes by making data categorization and analysis easier. However, further large multicenter studies are required to validate results internally and externally as well as explore and address limitations of ML application in bariatric surgery.

Leukemia can be Effectively Early Predicted in Routine Physical Examination with the Assistance of Machine Learning Models

Objectives. The diagnosis of leukemia relies very much on the results of bone marrow examinations, which is never generally performed in routine physical examination. In many rural areas even community hospitals and primary care clinics, the lack of hematological specialist and facility does not allow a definite diagnosis of leukemia. Thus, there will be a significant benefit if machine learning (ML) models could help early predict leukemia using preliminary blood test data in a routine physical examination in community hospitals to save time before a definite diagnosis. Methods. We collected the routine physical examination data of 1230 newly diagnosed leukemia patients and 1300 healthy people. We trained and tested 3 machine learning (ML) models including linear support vector machine (LSVM), random forest (RF), and XGboost models. We not only examined the accordance between model results and statistical analysis of the input data but also examined the consistency of model accuracy scores and relative importance order of model factors with regard to different input data sets and different model arguments to check the applicability of both the models and the input data. Results. Generally, the RF and XGboost models give more identical, consistent, and robust relative importance order of factors that is also accordant with the statistical analysis, while the LSVM gives much different and nonsense orders for different inputs. Results of the RF and XGboost models show that (1) generally, the models achieve accuracy scores above 0.9, indicating effective identification of leukemia, and (2) the top three factors that contribute most to the identification of leukemia include red blood cell (RBC), hematocrit (HCT), and white blood cell (WBC), while the other factors contribute relatively less. Conclusions. This study shows a feasible case example for early identification of leukemia using routine physical examination data with the assistance of ML models, which can be conveniently, cheaply, and widely applied in community hospitals or primary care clinics to save time before definite diagnosis; however, more studies are still needed to validate the applicability of more ML models to a larger variety of input data sets.

An efficient computer vision-based approach for acute lymphoblastic leukemia prediction

Leukemia (blood cancer) diseases arise when the number of White blood cells (WBCs) is imbalanced in the human body. When the bone marrow produces many immature WBCs that kill healthy cells, acute lymphocytic leukemia (ALL) impacts people of all ages. Thus, timely predicting this disease can increase the chance of survival, and the patient can get his therapy early. Manual prediction is very expensive and time-consuming. Therefore, automated prediction techniques are essential. In this research, we propose an ensemble automated prediction approach that uses four machine learning algorithms K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Random Forest (RF), and Naive Bayes (NB). The C-NMC leukemia dataset is used from the Kaggle repository to predict leukemia. Dataset is divided into two classes cancer and healthy cells. We perform data preprocessing steps, such as the first images being cropped using minimum and maximum points. Feature extraction is performed to extract the feature using pre-trained Convolutional Neural Network-based Deep Neural Network (DNN) architectures (VGG19, ResNet50, or ResNet101). Data scaling is performed by using the MinMaxScaler normalization technique. Analysis of Variance (ANOVA), Recursive Feature Elimination (RFE), and Random Forest (RF) as feature Selection techniques. Classification machine learning algorithms and ensemble voting are applied to selected features. Results reveal that SVM with 90.0% accuracy outperforms compared to other algorithms.

Forecasting determinants of recurrence in lung cancer patients exploiting various machine learning models

Lung cancer recurrence seems to be the most leading cause of death as well as deterioration of lifespan. Proper assessment of the probability of recurrence in early-stage lung cancer is necessary to push up the treatment progress. We therefore employed machine-learning technologies to forecast post-operative recurrence risks using 174 lung cancer patient records. Six classification algorithms logistic regression, SVM, decision tree classification, random forest classification, XGBoost and lightGBM were used to predict the cancer recurrence. The patient samples were divided into training and test group with the split ratio of 3:1 for model generation and the accuracy were validated using k-fold cross-validation method. It is worth noting that the logistic regression model outperformed all the models in both training (Accuracy = 0.82) and test set (Accuracy = 0.79) on k-fold validation. Further, the optimal features (n = 7) identified using the RFE method is certainly helpful to improve the model in a high precision. The imperative risk factors associated with recurrence were identified using three feature selection methods. Importantly, our research showed that age is an important prognostic factor to be considered during the recurrence prediction. Indeed, severe concern on the identified risk factors combined with predictive models assists the physician to reduce the cancer recurrence rate in patients with lung cancer.

A Catalogue of Machine Learning Algorithms for Healthcare Risk Predictions

Extracting useful knowledge from proper data analysis is a very challenging task for efficient and timely decision-making. To achieve this, there exist a plethora of machine learning (ML) algorithms, while, especially in healthcare, this complexity increases due to the domain's requirements for analytics-based risk predictions. This manuscript proposes a data analysis mechanism experimented in diverse healthcare scenarios, towards constructing a catalogue of the most efficient ML algorithms to be used depending on the healthcare scenario's requirements and datasets, for efficiently predicting the onset of a disease. To this context, seven (7) different ML algorithms (Naïve Bayes, K-Nearest Neighbors, Decision Tree, Logistic Regression, Random Forest, Neural Networks, Stochastic Gradient Descent) have been executed on top of diverse healthcare scenarios (stroke, COVID-19, diabetes, breast cancer, kidney disease, heart failure). Based on a variety of performance metrics (accuracy, recall, precision, F1-score, specificity, confusion matrix), it has been identified that a sub-set of ML algorithms are more efficient for timely predictions under specific healthcare scenarios, and that is why the envisioned ML catalogue prioritizes the ML algorithms to be used, depending on the scenarios' nature and needed metrics. Further evaluation must be performed considering additional scenarios, involving state-of-the-art techniques (e.g., cloud deployment, federated ML) for improving the mechanism's efficiency.

An interpretable stacking ensemble learning framework based on multi-dimensional data for real-time prediction of drug concentration: The example of olanzapine

Background and Aim: Therapeutic drug monitoring (TDM) has evolved over the years as an important tool for personalized medicine. Nevertheless, some limitations are associated with traditional TDM. Emerging data-driven model forecasting [e.g., through machine learning (ML)-based approaches] has been used for individualized therapy. This study proposes an interpretable stacking-based ML framework to predict concentrations in real time after olanzapine (OLZ) treatment.

Methods: The TDM-OLZ dataset, consisting of 2,142 OLZ measurements and 472 features, was formed by collecting electronic health records during the TDM of 927 patients who had received OLZ treatment. We compared the performance of ML algorithms by using 10-fold cross-validation and the mean absolute error (MAE). The optimal subset of features was analyzed by a random forest-based sequential forward feature selection method in the context of the top five heterogeneous regressors as base models to develop a stacked ensemble regressor, which was then optimized via the grid search method. Its predictions were explained by using local interpretable model-agnostic explanations (LIME) and partial dependence plots (PDPs).

Results: A state-of-the-art stacking ensemble learning framework that integrates optimized extra trees, XGBoost, random forest, bagging, and gradient-boosting regressors was developed for nine selected features [i.e., daily dose (OLZ), gender_male, age, valproic acid_yes, ALT, K, BW, MONO#, and time of blood sampling after first administration]. It outperformed other base regressors that were considered, with an MAE of 0.064, R-square value of 0.5355, mean squared error of 0.0089, mean relative error of 13%, and ideal rate (the percentages of predicted TDM within ± 30% of actual TDM) of 63.40%. Predictions at the individual level were illustrated by LIME plots, whereas the global interpretation of associations between features and outcomes was illustrated by PDPs.

Conclusion: This study highlights the feasibility of the real-time estimation of drug concentrations by using stacking-based ML strategies without losing interpretability, thus facilitating model-informed precision dosing.

Survival Prediction of Children Undergoing Hematopoietic Stem Cell Transplantation Using Different Machine Learning Classifiers by Performing Chi-Square Test and Hyperparameter Optimization: A Retrospective Analysis

Bone marrow transplant (BMT) is an effective surgical treatment for bone marrow-related disorders. However, several associated risk factors can impair long-term survival after BMT. Machine learning (ML) technologies have been proven useful in survival prediction of BMT receivers along with the influences that limit their resilience. In this study, an efficient classification model predicting the survival of children undergoing BMT is presented using a public dataset. Several supervised ML methods were investigated in this regard with an 80-20 train-test split ratio. To ensure prediction with minimal time and resources, only the top 11 out of the 59 dataset features were considered using Chi-square feature selection method. Furthermore, hyperparameter optimization (HPO) using the grid search cross-validation (GSCV) technique was adopted to increase the accuracy of prediction. Four experiments were conducted utilizing a combination of default and optimized hyperparameters on the original and reduced datasets. Our investigation revealed that the top 11 features of HPO had the same prediction accuracy (94.73%) as the entire dataset with default parameters, however, requiring minimal time and resources. Hence, the proposed approach may aid in the development of a computer-aided diagnostic system with satisfactory accuracy and minimal computation time by utilizing medical data records.

An ensemble learning with active sampling to predict the prognosis of postoperative non-small cell lung cancer patients

Background

Lung cancer is the leading cause of cancer death worldwide. Prognostic prediction plays a vital role in the decision-making process for postoperative non-small cell lung cancer (NSCLC) patients. However, the high imbalance ratio of prognostic data limits the development of effective prognostic prediction models.

Methods

In this study, we present a novel approach, namely ensemble learning with active sampling (ELAS), to tackle the imbalanced data problem in NSCLC prognostic prediction. ELAS first applies an active sampling mechanism to query the most informative samples to update the base classifier to give it a new perspective. This training process is repeated until no enough samples are queried. Next, an internal validation set is employed to evaluate the base classifiers, and the ones with the best performances are integrated as the ensemble model. Besides, we set up multiple initial training data seeds and internal validation sets to ensure the stability and generalization of the model.

Results

We verified the effectiveness of the ELAS on a real clinical dataset containing 1848 postoperative NSCLC patients. Experimental results showed that the ELAS achieved the best averaged 0.736 AUROC value and 0.453 AUPRC value for 6 prognostic tasks and obtained significant improvements in comparison with the SVM, AdaBoost, Bagging, SMOTE and TomekLinks.

Conclusions

We conclude that the ELAS can effectively alleviate the imbalanced data problem in NSCLC prognostic prediction and demonstrates good potential for future postoperative NSCLC prognostic prediction.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12911-022-01960-0.

A Review of Artificial Intelligence Applications in Hematology Management: Current Practices and Future Prospects

Background

Machine learning (ML) and deep learning (DL) methods have recently garnered a great deal of attention in the field of cancer research by making a noticeable contribution to the growth of predictive medicine and modern oncological practices. Considerable focus has been particularly directed toward hematologic malignancies because of the complexity in detecting early symptoms. Many patients with blood cancer do not get properly diagnosed until their cancer has reached an advanced stage with limited treatment prospects. Hence, the state-of-the-art revolves around the latest artificial intelligence (AI) applications in hematology management.

Objective

This comprehensive review provides an in-depth analysis of the current AI practices in the field of hematology. Our objective is to explore the ML and DL applications in blood cancer research, with a special focus on the type of hematologic malignancies and the patient’s cancer stage to determine future research directions in blood cancer.

Methods

We searched a set of recognized databases (Scopus, Springer, and Web of Science) using a selected number of keywords. We included studies written in English and published between 2015 and 2021. For each study, we identified the ML and DL techniques used and highlighted the performance of each model.

Results

Using the aforementioned inclusion criteria, the search resulted in 567 papers, of which 144 were selected for review.

Conclusions

The current literature suggests that the application of AI in the field of hematology has generated impressive results in the screening, diagnosis, and treatment stages. Nevertheless, optimizing the patient’s pathway to treatment requires a prior prediction of the malignancy based on the patient’s symptoms or blood records, which is an area that has still not been properly investigated.

Impact on outcomes of mixed chimerism of bone marrow CD34+ sorted cells after matched or haploidentical allogeneic stem cell transplantation for myeloid malignancies

Allogeneic hematopoietic stem cell transplantation (Allo-HSCT), proposed to patients with high-risk myeloid malignancies, may ultimately fail because of disease relapse. Bone marrow (BM) CD34⁺ cells in Allo-HSCT recipients can be either re-emerging recipient malignant cells or donor cells attesting of hematopoietic reconstitution. In this context, investigating donor/recipient chimerism in the population of BM CD34⁺ sorted cells (BM-CD34⁺SC) was performed in 261 Allo-HSCT recipients (matched n = 145, haploidentical n = 65, matched unrelated n = 51) with myeloid malignancies. BM-CD34⁺SC chimerism was compared to that of whole peripheral blood (PB) cells as well as other Allo-HSCT-related parameters, and impact on relapse and survival was assessed. Thresholds of 98% donor cells for PB and 90% for BM-CD34⁺SC were found to allow relapse prediction. This was completed by the application of machine learning tools to explore the predictive value of these parameters in multidimensional models with repeated iterations. BM-CD34⁺SC mixed chimerism stood out with all these methods as the most robust predictor of relapse with a significant impact on disease-free and overall survivals even after haploidentical Allo-HSCT and/or PTCY administration. This marker therefore appears to be of great interest for the decision of preemptive treatment to avoid post-transplant relapse.

Comprehensive analysis of the associations between clinical factors and outcomes by machine learning, using post marketing surveillance data of cabazitaxel in patients with castration-resistant prostate cancer

BACKGROUND

We aimed to evaluate relationships between clinical outcomes and explanatory variables by network clustering analysis using data from a post marketing surveillance (PMS) study of castration-resistant prostate cancer (CRPC) patients.

METHODS

The PMS was a prospective, multicenter, observational study of patients with metastatic, docetaxel-refractory CRPC treated with cabazitaxel in Japan after its launch in 2014. Graphical Markov (GM) model-based simulations and network clustering in 'R' package were conducted to identify correlations between clinical factors and outcomes. Factors shown to be associated with overall survival (OS) in the machine learning analysis were confirmed according to the clinical outcomes observed in the PMS.

RESULTS

Among the 660 patients analyzed, median patient age was 70.0 years, and median OS and time-to-treatment failure (TTF) were 319 and 116 days, respectively. In GM-based simulations, factors associated with OS were liver metastases, performance status (PS), TTF, and neutropenia (threshold 0.05), and liver metastases, PS, and TTF (threshold 0.01). Factors associated with TTF were OS and relative dose intensity (threshold 0.05), and OS (threshold 0.01). In network clustering in 'R' package, factors associated with OS were number of treatment cycles, discontinuation due to disease progression, and TTF (threshold 0.05), and liver and lung metastases, PS, discontinuation due to adverse events, and febrile neutropenia (threshold 0.01). Kaplan-Meier analysis of patient subgroups demonstrated that visceral metastases and poor PS at baseline were associated with worse OS, while neutropenia or febrile neutropenia and higher number of cabazitaxel cycles were associated with better OS.

CONCLUSIONS

Neutropenia may be a predictive factor for treatment efficacy in terms of survival. Poor PS and distant metastases to the liver and lungs were shown to be associated with worse outcomes, while factors related to treatment duration were shown to positively correlate with better OS.

Research and Application of Artificial Intelligence Based on Electronic Health Records of Patients With Cancer: Systematic Review

BACKGROUND

With the accumulation of electronic health records and the development of artificial intelligence, patients with cancer urgently need new evidence of more personalized clinical and demographic characteristics and more sophisticated treatment and prevention strategies. However, no research has systematically analyzed the application and significance of artificial intelligence based on electronic health records in cancer care.

OBJECTIVE

The aim of this study was to conduct a review to introduce the current state and limitations of artificial intelligence based on electronic health records of patients with cancer and to summarize the performance of artificial intelligence in mining electronic health records and its impact on cancer care.

METHODS

Three databases were systematically searched to retrieve potentially relevant papers published from January 2009 to October 2020. Four principal reviewers assessed the quality of the papers and reviewed them for eligibility based on the inclusion criteria in the extracted data. The summary measures used in this analysis were the number and frequency of occurrence of the themes.

RESULTS

Of the 1034 papers considered, 148 papers met the inclusion criteria. Cancer care, especially cancers of female organs and digestive organs, could benefit from artificial intelligence based on electronic health records through cancer emergencies and prognostic estimates, cancer diagnosis and prediction, tumor stage detection, cancer case detection, and treatment pattern recognition. The models can always achieve an area under the curve of 0.7. Ensemble methods and deep learning are on the rise. In addition, electronic medical records in the existing studies are mainly in English and from private institutional databases.

CONCLUSIONS

Artificial intelligence based on electronic health records performed well and could be useful for cancer care. Improving the performance of artificial intelligence can help patients receive more scientific-based and accurate treatments. There is a need for the development of new methods and electronic health record data sharing and for increased passion and support from cancer specialists.

Multi-Cohort Transcriptomic Subtyping of B-Cell Acute Lymphoblastic Leukemia

RNA sequencing provides a snapshot of the functional consequences of genomic lesions that drive acute lymphoblastic leukemia (ALL). The aims of this study were to elucidate diagnostic associations (via machine learning) between mRNA-seq profiles, independently verify ALL lesions and develop easy-to-interpret transcriptome-wide biomarkers for ALL subtyping in the clinical setting. A training dataset of 1279 ALL patients from six North American cohorts was used for developing machine learning models. Results were validated in 767 patients from Australia with a quality control dataset across 31 tissues from 1160 non-ALL donors. A novel batch correction method was introduced and applied to adjust for cohort differences. Out of 18,503 genes with usable expression, 11,830 (64%) were confounded by cohort effects and excluded. Six ALL subtypes (ETV6::RUNX1, KMT2A, DUX4, PAX5 P80R, TCF3::PBX1, ZNF384) that covered 32% of patients were robustly detected by mRNA-seq (positive predictive value ≥ 87%). Five other frequent subtypes (CRLF2, hypodiploid, hyperdiploid, PAX5 alterations and Ph-positive) were distinguishable in 40% of patients at lower accuracy (52% ≤ positive predictive value ≤ 73%). Based on these findings, we introduce the Allspice R package to predict ALL subtypes and driver genes from unadjusted mRNA-seq read counts as encountered in real-world settings. Two examples of Allspice applied to previously unseen ALL patient samples with atypical lesions are included.

Data analytics and artificial intelligence in predicting length of stay, readmission, and mortality: a population-based study of surgical management of colorectal cancer

Data analytics and artificial intelligence (AI) have been used to predict patient outcomes after colorectal cancer surgery. A prospectively maintained colorectal cancer database was used, covering 4336 patients who underwent colorectal cancer surgery between 2003 and 2019. The 47 patient parameters included demographics, peri- and post-operative outcomes, surgical approaches, complications, and mortality. Data analytics were used to compare the importance of each variable and AI prediction models were built for length of stay (LOS), readmission, and mortality. Accuracies of at least 80% have been achieved. The significant predictors of LOS were age, ASA grade, operative time, presence or absence of a stoma, robotic or laparoscopic approach to surgery, and complications. The model with support vector regression (SVR) algorithms predicted the LOS with an accuracy of 83% and mean absolute error (MAE) of 9.69 days. The significant predictors of readmission were age, laparoscopic procedure, stoma performed, preoperative nodal (N) stage, operation time, operation mode, previous surgery type, LOS, and the specific procedure. A BI-LSTM model predicted readmission with 87.5% accuracy, 84% sensitivity, and 90% specificity. The significant predictors of mortality were age, ASA grade, BMI, the formation of a stoma, preoperative TNM staging, neoadjuvant chemotherapy, curative resection, and LOS. Classification predictive modelling predicted three different colorectal cancer mortality measures (overall mortality, and 31- and 91-days mortality) with 80-96% accuracy, 84-93% sensitivity, and 75-100% specificity. A model using all variables performed only slightly better than one that used just the most significant ones.

Machine learning methods to predict presence of residual cancer following hysterectomy

Surgical management for gynecologic malignancies often involves hysterectomy, often constituting the most common gynecologic surgery worldwide. Despite maximal surgical and medical care, gynecologic malignancies have a high rate of recurrence following surgery. Current machine learning models use advanced pathology data that is often inaccessible within low-resource settings and are specific to singular cancer types. There is currently a need for machine learning models to predict non-clinically evident residual disease using only clinically available health data. Here we developed and tested multiple machine learning models to assess the risk of residual disease post-hysterectomy based on clinical and operative parameters. Data from 3656 hysterectomy patients from the NSQIP dataset over 14 years were used to develop models with a training set of 2925 patients and a validation set of 731 patients. Our models revealed the top postoperative predictors of residual disease were the initial presence of gross abdominal disease on the diaphragm, disease located on the bowel mesentery, located on the bowel serosa, and disease located within the adjacent pelvis prior to resection. There were no statistically significant differences in performances of the top three models. Extreme gradient Boosting, Random Forest, and Logistic Regression models had comparable AUC ROC (0.90) and accuracy metrics (87–88%). Using these models, physicians can identify gynecologic cancer patients post-hysterectomy that may benefit from additional treatment. For patients at high risk for disease recurrence despite adequate surgical intervention, machine learning models may lay the basis for potential prospective trials with prophylactic/adjuvant therapy for non-clinically evident residual disease, particularly in under-resourced settings.

Cost-Sensitive KNN Algorithm for Cancer Prediction Based on Entropy Analysis

Early diagnosis of cancer is beneficial in the formulation of the best treatment plan; it can improve the survival rate and the quality of patient life. However, imaging detection and needle biopsy usually used not only find it difficult to effectively diagnose tumors at early stage, but also do great harm to the human body. Since the changes in a patient’s health status will cause changes in blood protein indexes, if cancer can be diagnosed by the changes in blood indexes in the early stage of cancer, it can not only conveniently track and detect the treatment process of cancer, but can also reduce the pain of patients and reduce the costs. In this paper, 39 serum protein markers were taken as research objects. The difference of the entropies of serum protein marker sequences in different types of patients was analyzed, and based on this, a cost-sensitive analysis model was established for the purpose of improving the accuracy of cancer recognition. The results showed that there were significant differences in entropy of different cancer patients, and the complexity of serum protein markers in normal people was higher than that in cancer patients. Although the dataset was rather imbalanced, containing 897 instances, including 799 normal instances, 44 liver cancer instances, and 54 ovarian cancer instances, the accuracy of our model still reached 95.21%. Other evaluation indicators were also stable and satisfactory; precision, recall, F1 and AUC reach 0.807, 0.833, 0.819 and 0.92, respectively. This study has certain theoretical and practical significance for cancer prediction and clinical application and can also provide a research basis for the intelligent medical treatment.

A New Time-Window Prediction Model For Traumatic Hemorrhagic Shock Based on Interpretable Machine Learning

ABSTRACT

Early warning prediction of traumatic hemorrhagic shock (THS) can greatly reduce patient mortality and morbidity. We aimed to develop and validate models with different stepped feature sets to predict THS in advance. From the PLA General Hospital Emergency Rescue Database and Medical Information Mart for Intensive Care III, we identified 604 and 1,614 patients, respectively. Two popular machine learning algorithms (i.e., extreme gradient boosting [XGBoost] and logistic regression) were applied. The area under the receiver operating characteristic curve (AUROC) was used to evaluate the performance of the models. By analyzing the feature importance based on XGBoost, we found that features in vital signs (VS), routine blood (RB), and blood gas analysis (BG) were the most relevant to THS (0.292, 0.249, and 0.225, respectively). Thus, the stepped relationships existing in them were revealed. Furthermore, the three stepped feature sets (i.e., VS, VS + RB, and VS + RB + sBG) were passed to the two machine learning algorithms to predict THS in the subsequent T hours (where T = 3, 2, 1, or 0.5), respectively. Results showed that the XGBoost model performance was significantly better than the logistic regression. The model using vital signs alone achieved good performance at the half-hour time window (AUROC = 0.935), and the performance was increased when laboratory results were added, especially when the time window was 1 h (AUROC = 0.950 and 0.968, respectively). These good-performing interpretable models demonstrated acceptable generalization ability in external validation, which could flexibly and rollingly predict THS T hours (where T = 0.5, 1) prior to clinical recognition. A prospective study is necessary to determine the clinical utility of the proposed THS prediction models.

Applications of Artificial Intelligence in Pediatric Oncology: A Systematic Review

PURPOSE

There is a need for an improved understanding of clinical and biologic risk factors in pediatric cancer to improve patient outcomes. Machine learning (ML) represents the application of computational inference from advanced statistical methods that can be applied to increasing amount of data available for study in pediatric oncology. The goal of this systematic review was to systematically characterize the state of ML in pediatric oncology and highlight advances and opportunities in the field.

METHODS

We conducted a systematic review of the Embase, Scopus, and MEDLINE databases for applications of ML in pediatric oncology. Query results from all three databases were aggregated and duplicate studies were removed.

RESULTS

A total of 42 unique articles that examined the applications of ML in pediatric oncology met inclusion criteria for review. We identified 20 studies of CNS tumors, 13 of solid tumors, and nine of leukemia. ML tasks included classification, prediction of treatment response, and dose optimization with a variety of methods being used including neural network, k-nearest neighbor, random forest, naive Bayes, and support vector machines. Strengths of the identified studies included matching or outperforming physician comparators via automated analysis and predicting therapeutic response. Common limitations included significant heterogeneity in reporting standards, clinical applicability, small sample sizes, and missing external validation cohorts.

CONCLUSION

We identified areas where ML can enhance clinical care in ways that may not otherwise be achievable. Although ML promises enormous potential in improving diagnostics, decision making, and monitoring for children with cancer, the field remains in early stages and future work will be aided by standards and guidelines to ensure rigorous methodologic design and maximizing clinical utility.

Clinical Data Prediction Model to Identify Patients With Early-Stage Pancreatic Cancer

PURPOSE

Pancreatic cancer is an aggressive malignancy with patients often experiencing nonspecific symptoms before diagnosis. This study evaluates a machine learning approach to help identify patients with early-stage pancreatic cancer from clinical data within electronic health records (EHRs).

MATERIALS AND METHODS

From the Optum deidentified EHR data set, we identified early-stage (n = 3,322) and late-stage (n = 25,908) pancreatic cancer cases over 40 years of age diagnosed between 2009 and 2017. Patients with early-stage pancreatic cancer were matched to noncancer controls (1:16 match). We constructed a prediction model using eXtreme Gradient Boosting (XGBoost) to identify early-stage patients on the basis of 18,220 features within the EHR including diagnoses, procedures, information within clinical notes, and medications. Model accuracy was assessed with sensitivity, specificity, positive predictive value, and the area under the curve.

RESULTS

The final predictive model included 582 predictive features from the EHR, including 248 (42.5%) physician note elements, 146 (25.0%) procedure codes, 91 (15.6%) diagnosis codes, 89 (15.3%) medications, and 9 (1.5%) demographic features. The final model area under the curve was 0.84. Choosing a model cut point with a sensitivity of 60% and specificity of 90% would enable early detection of 58% late-stage patients with a median of 24 months before their actual diagnosis.

CONCLUSION

Prediction models using EHR data show promise in the early detection of pancreatic cancer. Although widespread use of this approach on an unselected population would produce high rates of false-positive tests, this technique may be rapidly impactful if deployed among high-risk patients or paired with other imaging or biomarker screening tools.

Joint Modeling of RNAseq and Radiomics Data for Glioma Molecular Characterization and Prediction

RNA sequencing (RNAseq) is a recent technology that profiles gene expression by measuring the relative frequency of the RNAseq reads. RNAseq read counts data is increasingly used in oncologic care and while radiology features (radiomics) have also been gaining utility in radiology practice such as disease diagnosis, monitoring, and treatment planning. However, contemporary literature lacks appropriate RNA-radiomics (henceforth, radiogenomics ) joint modeling where RNAseq distribution is adaptive and also preserves the nature of RNAseq read counts data for glioma grading and prediction. The Negative Binomial (NB) distribution may be useful to model RNAseq read counts data that addresses potential shortcomings. In this study, we propose a novel radiogenomics-NB model for glioma grading and prediction. Our radiogenomics-NB model is developed based on differentially expressed RNAseq and selected radiomics/volumetric features which characterize tumor volume and sub-regions. The NB distribution is fitted to RNAseq counts data, and a log-linear regression model is assumed to link between the estimated NB mean and radiomics. Three radiogenomics-NB molecular mutation models (e.g., IDH mutation, 1p/19q codeletion, and ATRX mutation) are investigated. Additionally, we explore gender-specific effects on the radiogenomics-NB models. Finally, we compare the performance of the proposed three mutation prediction radiogenomics-NB models with different well-known methods in the literature: Negative Binomial Linear Discriminant Analysis (NBLDA), differentially expressed RNAseq with Random Forest (RF-genomics), radiomics and differentially expressed RNAseq with Random Forest (RF-radiogenomics), and Voom-based count transformation combined with the nearest shrinkage classifier (VoomNSC). Our analysis shows that the proposed radiogenomics-NB model significantly outperforms (ANOVA test, p < 0.05) for prediction of IDH and ATRX mutations and offers similar performance for prediction of 1p/19q codeletion, when compared to the competing models in the literature, respectively.

Identification of a Seven-lncRNA-mRNA Signature for Recurrence and Prognostic Prediction in Relapsed Acute Lymphoblastic Leukemia Based on WGCNA and LASSO Analyses

Abnormal expressions of long noncoding RNAs (lncRNAs) and protein-encoding messenger RNAs (mRNAs) are important for the development of childhood acute lymphoblastic leukemia (ALL). This study developed an lncRNA-mRNA integrated classifier for the prediction of recurrence and prognosis in relapsed childhood ALL by using several transcriptome data. Weighted gene coexpression network analysis revealed that green, turquoise, yellow, and brown modules were preserved across the TARGET, GSE60926, GSE28460, and GSE17703 datasets, and they were associated with clinical relapse and death status. A total of 184 genes in these four modules were differentially expressed between recurrence and nonrecurrence samples. Least absolute shrinkage and selection operator analysis showed that seven genes constructed a prognostic signature (including one lncRNA: LINC00652 and six mRNAs: INSL3, NIPAL2, REN, RIMS2, RPRM, and SNAP91). Kaplan-Meier curve analysis observed that patients in the high-risk group had a significantly shorter overall survival than those of the low-risk group. Receiver operating characteristic curve analysis demonstrated that this signature had high accuracy in predicting the 5-year overall survival and recurrence outcomes, respectively. LINC00652 may function by coexpressing with the above prognostic genes (INSL3, SNAP91, and REN) and lipid metabolism-related genes (MIA2, APOA1). Accordingly, this lncRNA-mRNA-based classifier may be clinically useful to predict the recurrence and prognosis for childhood ALL. These genes represent new targets to explain the mechanisms for ALL.

A machine learning approach to personalized dose adjustment of lamotrigine using noninvasive clinical parameters

The pharmacokinetic variability of lamotrigine (LTG) plays a significant role in its dosing requirements. Our goal here was to use noninvasive clinical parameters to predict the dose-adjusted concentrations (C/D ratio) of LTG based on machine learning (ML) algorithms. A total of 1141 therapeutic drug-monitoring measurements were used, 80% of which were randomly selected as the "derivation cohort" to develop the prediction algorithm, and the remaining 20% constituted the "validation cohort" to test the finally selected model. Fifteen ML models were optimized and evaluated by tenfold cross-validation on the "derivation cohort,” and were filtered by the mean absolute error (MAE). On the whole, the nonlinear models outperformed the linear models. The extra-trees’ regression algorithm delivered good performance, and was chosen to establish the predictive model. The important features were then analyzed and parameters of the model adjusted to develop the best prediction model, which accurately described the C/D ratio of LTG, especially in the intermediate-to-high range (≥ 22.1 μg mL⁻¹ g⁻¹ day), as illustrated by a minimal bias (mean relative error (%) =  + 3%), good precision (MAE = 8.7 μg mL⁻¹ g⁻¹ day), and a high percentage of predictions within ± 20% of the empirical values (60.47%). This is the first study, to the best of our knowledge, to use ML algorithms to predict the C/D ratio of LTG. The results here can help clinicians adjust doses of LTG administered to patients to minimize adverse reactions.

High-Dimensional Analysis of Single-Cell Flow Cytometry Data Predicts Relapse in Childhood Acute Lymphoblastic Leukaemia

Simple Summary

B-cell Acute Lymphoblastic Leukaemia is one of the most common cancers in childhood, with 20% of patients eventually relapsing. Flow cytometry is routinely used for diagnosis and follow-up, but it currently does not provide prognostic value at diagnosis. The volume and the high-dimensional character of this data makes it ideal for its exploitation by means of Artificial Intelligence methods. We collected flow cytometry data from 56 patients from two hospitals. We analysed differences in intensity of marker expression in order to predict relapse at the moment of diagnosis. We finally correlated this data with biomolecular information, constructing a classifier based on CD38 expression.

Abstract

Artificial intelligence methods may help in unveiling information that is hidden in high-dimensional oncological data. Flow cytometry studies of haematological malignancies provide quantitative data with the potential to be used for the construction of response biomarkers. Many computational methods from the bioinformatics toolbox can be applied to these data, but they have not been exploited in their full potential in leukaemias, specifically for the case of childhood B-cell Acute Lymphoblastic Leukaemia. In this paper, we analysed flow cytometry data that were obtained at diagnosis from 56 paediatric B-cell Acute Lymphoblastic Leukaemia patients from two local institutions. Our aim was to assess the prognostic potential of immunophenotypical marker expression intensity. We constructed classifiers that are based on the Fisher’s Ratio to quantify differences between patients with relapsing and non-relapsing disease. We also correlated this with genetic information. The main result that arises from the data was the association between subexpression of marker CD38 and the probability of relapse.

A stacking-based model for predicting 30-day all-cause hospital readmissions of patients with acute myocardial infarction

BACKGROUND

Acute myocardial infarction (AMI) is a serious cardiovascular disease, followed by a high readmission rate within 30-days of discharge. Accurate prediction of AMI readmission is a crucial way to identify the high-risk group and optimize the distribution of medical resources.

METHODS

In this study, we propose a stacking-based model to predict the risk of 30-day unplanned all-cause hospital readmissions for AMI patients based on clinical data. Firstly, we conducted an under-sampling method of neighborhood cleaning rule (NCR) to alleviate the class imbalance and then utilized a feature selection method of SelectFromModel (SFM) to select effective features. Secondly, we adopted a self-adaptive approach to select base classifiers from eight candidate models according to their performances in datasets. Finally, we constructed a three-layer stacking model in which layer 1 and layer 2 were base-layer and level 3 was meta-layer. The predictions of the base-layer were used to train the meta-layer in order to make the final forecast.

RESULTS

The results show that the proposed model exhibits the highest AUC (0.720), which is higher than that of decision tree (0.681), support vector machine (0.707), random forest (0.701), extra trees (0.709), adaBoost (0.702), bootstrap aggregating (0.704), gradient boosting decision tree (0.710) and extreme gradient enhancement (0.713).

CONCLUSION

It is evident that our model could effectively predict the risk of 30-day all cause hospital readmissions for AMI patients and provide decision support for the administration.

Epidemic analysis of COVID-19 in Italy based on spatiotemporal geographic information and Google Trends

Since the first two novel coronavirus cases appeared in January of 2020, the outbreak of the COVID-19 epidemic seriously threatens the public health of Italy. In this article, the distribution characteristics and spreading of COVID-19 in various regions of Italy were analysed by heat maps. Meanwhile, spatial autocorrelation, spatiotemporal clustering analysis and kernel density method were also applied to analyse the spatial clustering of COVID-19. The results showed that the Italian epidemic has a temporal trend and spatial aggregation. The epidemic was concentrated in northern Italy and gradually spread to other regions. Finally, the Google Trends index of the COVID-19 epidemic was further employed to build a prediction model combined with machine learning algorithms. By using Adaboost algorithm for single-factor modelling,the results show that the AUC of these six features (mask, pneumonia, thermometer, ISS, disinfection and disposable gloves) are all >0.9, indicating that these features have a large contribution to the prediction model. It is also implied that the public's attention to the epidemic is increasing as well as the awareness of the need for protective measures. This increased awareness of the epidemic will prompt the public to pay more attention to protective measures, thereby reducing the risk of coronavirus infection.

Utilization of machine-learning models to accurately predict the risk for critical COVID-19

Among patients with Coronavirus disease (COVID-19), the ability to identify patients at risk for deterioration during their hospital stay is essential for effective patient allocation and management. To predict patient risk for critical COVID-19 based on status at admission using machine-learning models. Retrospective study based on a database of tertiary medical center with designated departments for patients with COVID-19. Patients with severe COVID-19 at admission, based on low oxygen saturation, low partial arterial oxygen pressure, were excluded. The primary outcome was risk for critical disease, defined as mechanical ventilation, multi-organ failure, admission to the ICU, and/or death. Three different machine-learning models were used to predict patient deterioration and compared to currently suggested predictors and to the APACHEII risk-prediction score. Among 6995 patients evaluated, 162 were hospitalized with non-severe COVID-19, of them, 25 (15.4%) patients deteriorated to critical COVID-19. Machine-learning models outperformed the all other parameters, including the APACHE II score (ROC AUC of 0.92 vs. 0.79, respectively), reaching 88.0% sensitivity, 92.7% specificity and 92.0% accuracy in predicting critical COVID-19. The most contributory variables to the models were APACHE II score, white blood cell count, time from symptoms to admission, oxygen saturation and blood lymphocytes count. Machine-learning models demonstrated high efficacy in predicting critical COVID-19 compared to the most efficacious tools available. Hence, artificial intelligence may be applied for accurate risk prediction of patients with COVID-19, to optimize patients triage and in-hospital allocation, better prioritization of medical resources and improved overall management of the COVID-19 pandemic.

Using a machine learning algorithm to predict acute graft-versus-host disease following allogeneic transplantation

Acute graft-versus-host disease (aGVHD) is 1 of the critical complications that often occurs following allogeneic hematopoietic stem cell transplantation (HSCT). Thus far, various types of prediction scores have been created using statistical calculations. The primary objective of this study was to establish and validate the machine learning-dependent index for predicting aGVHD. This was a retrospective cohort study that involved analyzing databases of adult HSCT patients in Japan. The alternating decision tree (ADTree) machine learning algorithm was applied to develop models using the training cohort (70%). The ADTree algorithm was confirmed using the hazard model on data from the validation cohort (30%). Data from 26 695 HSCT patients transplanted from allogeneic donors between 1992 and 2016 were included in this study. The cumulative incidence of aGVHD was 42.8%. Of >40 variables considered, 15 were adapted into a model for aGVHD prediction. The model was tested in the validation cohort, and the incidence of aGVHD was clearly stratified according to the categorized ADTree scores; the cumulative incidence of aGVHD was 29.0% for low risk and 58.7% for high risk (hazard ratio, 2.57). Predicting scores for aGVHD also demonstrated the link between the risk of development aGVHD and overall survival after HSCT. The machine learning algorithms produced clinically reasonable and robust risk stratification scores. The relatively high reproducibility and low impacts from the interactions among the variables indicate that the ADTree algorithm, along with the other data-mining approaches, may provide tools for establishing risk score.

Identification of significant risks in pediatric acute lymphoblastic leukemia (ALL) through machine learning (ML) approach

Pediatric acute lymphoblastic leukemia (ALL) through machine learning (ML) technique was analyzed to determine the significance of clinical and phenotypic variables as well as environmental conditions that can identify the underlying causes of child ALL. Fifty pediatric patients (n = 50) included who were diagnosed with acute lymphoblastic leukemia (ALL) according to the inclusion and exclusion criteria. Clinical variables comprised of the blood biochemistry (CBC, LFTs, RFTs) results, and distribution of type of ALL, i.e., T ALL or B ALL. Phenotypic data included the age, sex of the child, and consanguinity, while environmental factors included the habitat, socioeconomic status, and access to filtered drinking water. Fifteen different features/attributes were collected for each case individually. To retrieve most useful discriminating attributes, four different supervised ML algorithms were used including classification and regression trees (CART), random forest (RM), gradient boosted machine (GM), and C5.0 decision tree algorithm. To determine the accuracy of the derived CART algorithm on future data, a ten-fold cross validation was performed on the present data set. The ALL was common in children of age below 5 years in male patients whole belonged to middle class family of rural areas. (B-ALL) was most frequent as compared with T-ALL. The consanguinity was present in 54% of cases. Low levels of platelets and hemoglobin and high levels of white blood cells were reported in child ALL patients. CART provided the best and complete fit for the entire data set yielding a 99.83% model fit accuracy, and a misclassification of 0.17% on the entire sample space, while C5.0 reported 98.6%, random forest 94.44%, and gradient boosted machine resulted in 95.61% fitting. The variable importance of each primary discriminating attribute is platelet 43%, hemoglobin 24%, white blood cells 4%, and sex of the child 4%. An overall accuracy of 87.4% was recorded for the classifier. Platelet count abnormality can be considered as a major factor in predicting pediatric ALL. The machine learning algorithms can be applied efficiently to provide details for the prognosis for better treatment outcome. Graphical Abstract Identification of significant risks in pediatric acute lymphoblastic leukemia (ALL) through machine learning (ML) approach.

Machine learning and artificial intelligence in haematology

Digitalization of the medical record and integration of genomic methods into clinical practice have resulted in an unprecedented wealth of data. Machine learning is a subdomain of artificial intelligence that attempts to computationally extract meaningful insights from complex data structures. Applications of machine learning in haematological scenarios are steadily increasing. However, basic concepts are often unfamiliar to clinicians and investigators. The purpose of this review is to provide readers with tools to interpret and critically appraise machine learning literature. We begin with the elucidation of standard terminology and then review examples in haematology. Guidelines for designing and evaluating machine-learning studies are provided. Finally, we discuss limitations of the machine-learning approach.

Microrandomized trials for promoting engagement in mobile health data collection: Adolescent/young adult oral chemotherapy adherence as an example

Long-term engagement with mobile health (mHealth) apps can provide critical data for improving empirical models for real-time health behaviors. To learn how to improve and maintain mHealth engagement, micro-randomized trials (MRTs) can be used to optimize different engagement strategies. In MRTs, participants are sequentially randomized, often hundreds or thousands of times, to different engagement strategies or treatments. The data gathered are then used to decide which treatment is optimal in which context. In this paper, we discuss an example MRT for youth with cancer, where we randomize different engagement strategies to improve self-reports on factors related to medication adherence. MRTs, moreover, can go beyond improving engagement, and we reference other MRTs to address substance abuse, sedentary behavior, and so on.

A BLSTM with Attention Network for Predicting Acute Myeloid Leukemia Patient's Prognosis using Comprehensive Clinical Parameters

The prognosis management is crucial for highrisk disease like Acute Myeloid Leukemia (AML) in order to support decisions of clinical treatment. However, the challenges of accurate and consistent forecasting lie in the high variability of the disease outcomes and the complexity of the multiple clinical measurements available over the course of the treatment. In order to capture the multi-dimensional and longitudinal aspect of these comprehensive clinical parameters, we utilize an attention-based bi-directional long shortterm memory (Att-BLSTM) network to predict AML patient's survival and relapse. Specifically, we gather a 10-year worth of real patient's clinical data including blood test, medication, HSCT status, and gene mutation information. Our proposed Att-BLSTM framework achieves 77.1% and 67.3% AUC in tasks of predicting the next 2-year mortality and disease relapse with these comprehensive clinical parameters, and our further analysis demonstrates that a next 0 to 3 months prediction performs equally well, i.e., 74.8% and 67% AUC for mortality and relapse respectively.

Artificial Intelligence based Models for Screening of Hematologic Malignancies using Cell Population Data

Cell Population Data (CPD) provides various blood cell parameters that can be used for differential diagnosis. Data analytics using Machine Learning (ML) have been playing a pivotal role in revolutionizing medical diagnostics. This research presents a novel approach of using ML algorithms for screening hematologic malignancies using CPD. The data collection was done at Konkuk University Medical Center, Seoul. A total of (882 cases: 457 hematologic malignancy and 425 hematologic non-malignancy) were used for analysis. In our study, seven machine learning models, i.e., SGD, SVM, RF, DT, Linear model, Logistic regression, and ANN, were used. In order to measure the performance of our ML models, stratified 10-fold cross validation was performed, and metrics, such as accuracy, precision, recall, and AUC were used. We observed outstanding performance by the ANN model as compared to other ML models. The diagnostic ability of ANN achieved the highest accuracy, precision, recall, and AUC ± Standard Deviation as follows: 82.8%, 82.8%, 84.9%, and 93.5% ± 2.6 respectively. ANN algorithm based on CPD appeared to be an efficient aid for clinical laboratory screening of hematologic malignancies. Our results encourage further work of applying ML to wider field of clinical practice.

Machine Learning Algorithms for Predicting the Recurrence of Stage IV Colorectal Cancer After Tumor Resection

The aim of this study is to explore the feasibility of using machine learning (ML) technology to predict postoperative recurrence risk among stage IV colorectal cancer patients. Four basic ML algorithms were used for prediction-logistic regression, decision tree, GradientBoosting and lightGBM. The research samples were randomly divided into a training group and a testing group at a ratio of 8:2. 999 patients with stage 4 colorectal cancer were included in this study. In the training group, the GradientBoosting model's AUC value was the highest, at 0.881. The Logistic model's AUC value was the lowest, at 0.734. The GradientBoosting model had the highest F1_score (0.912). In the test group, the AUC Logistic model had the lowest AUC value (0.692). The GradientBoosting model's AUC value was 0.734, which can still predict cancer progress. However, the gbm model had the highest AUC value (0.761), and the gbm model had the highest F1_score (0.974). The GradientBoosting model and the gbm model performed better than the other two algorithms. The weight matrix diagram of the GradientBoosting algorithm shows that chemotherapy, age, LogCEA, CEA and anesthesia time were the five most influential risk factors for tumor recurrence. The four machine learning algorithms can each predict the risk of tumor recurrence in patients with stage IV colorectal cancer after surgery. Among them, GradientBoosting and gbm performed best. Moreover, the GradientBoosting weight matrix shows that the five most influential variables accounting for postoperative tumor recurrence are chemotherapy, age, LogCEA, CEA and anesthesia time.

Volumetric assessment of unaffected parenchyma and Wilms' tumours: analysis of response to chemotherapy and surgery using a semi-automated segmentation algorithm in children with renal neoplasms

OBJECTIVE

To present our proof of concept with semi-automatic image recognition/segmentation technology for calculation of tumour/parenchyma volume.

METHODS

We reviewed Wilms' tumours (WTs) between 2000 and 2018, capturing computed tomography images at baseline, after neoadjuvant chemotherapy (NaC) and postoperatively. Images were uploaded into MATLAB-3-D volumetric image processing software. The program was trained by two clinicians who supervised the demarcation of tumour and parenchyma, followed by automatic recognition and delineation of tumour margins on serial imaging, and differentiation from uninvolved renal parenchyma. Volume was automatically calculated for both.

RESULTS

During the study period, 98 patients were identified. Of these, based on image quality and availability, 32 (38 affected moieties) were selected. Most patients (65%) were girls, diagnosed at age 50 ± 37 months of age. NaC was employed in 64% of patients. Surgical management included 27 radical and 11 partial nephrectomies. Automated volume assessment demonstrated objective response to NaC for unilateral and bilateral tumours (68 ± 20% and 53 ± 39%, respectively), as well as preservation on uninvolved parenchyma with partial nephrectomy (70 ± 46 cm³ at presentation to 57 ± 41 cm³ post-surgery).

CONCLUSION

Volumetric analysis is feasible and allows objective assessment of tumour and parenchyma volume in response to chemotherapy and surgery. Our data show changes after therapy that may be otherwise difficult to quantify. Use of such technology may improve surgical planning and quantification of response to treatment, as well as serving as a tool to predict renal reserve and long-term changes in renal function.

Patient-based prediction algorithm of relapse after allo-HSCT for acute Leukemia and its usefulness in the decision-making process using a machine learning approach

Although allogeneic hematopoietic stem cell transplantation (allo‐HSCT) is a curative therapy for high‐risk acute leukemia (AL), some patients still relapse. Since patients simultaneously have many prognostic factors, difficulties are associated with the construction of a patient‐based prediction algorithm of relapse. The alternating decision tree (ADTree) is a successful classification method that combines decision trees with the predictive accuracy of boosting. It is a component of machine learning (ML) and has the capacity to simultaneously analyze multiple factors. Using ADTree, we attempted to construct a prediction model of leukemia relapse within 1 year of transplantation. With the model of training data (n = 148), prediction accuracy, the AUC of ROC, and the κ‐statistic value were 78.4%, 0.746, and 0.508, respectively. The false positive rate (FPR) of the relapse prediction was as low as 0.134. In an evaluation of the model with validation data (n = 69), prediction accuracy, AUC, and FPR of the relapse prediction were similar at 71.0%, 0.667, and 0.216, respectively. These results suggest that the model is generalized and highly accurate. Furthermore, the output of ADTree may visualize the branch point of treatment. For example, the selection of donor types resulted in different relapse predictions. Therefore, clinicians may change treatment options by referring to the model, thereby improving outcomes. The present results indicate that ML, such as ADTree, will contribute to the decision‐making process in the diversified allo‐HSCT field and be useful for preventing the relapse of leukemia.

A Comparative Study of Machine Learning Algorithms in Predicting Severe Complications after Bariatric Surgery

BACKGROUND

Severe obesity is a global public health threat of growing proportions. Accurate models to predict severe postoperative complications could be of value in the preoperative assessment of potential candidates for bariatric surgery. So far, traditional statistical methods have failed to produce high accuracy. We aimed to find a useful machine learning (ML) algorithm to predict the risk for severe complication after bariatric surgery.

METHODS

We trained and compared 29 supervised ML algorithms using information from 37,811 patients that operated with a bariatric surgical procedure between 2010 and 2014 in Sweden. The algorithms were then tested on 6250 patients operated in 2015. We performed the synthetic minority oversampling technique tackling the issue that only 3% of patients experienced severe complications.

RESULTS

Most of the ML algorithms showed high accuracy (>90%) and specificity (>90%) in both the training and test data. However, none of the algorithms achieved an acceptable sensitivity in the test data. We also tried to tune the hyperparameters of the algorithms to maximize sensitivity, but did not yet identify one with a high enough sensitivity that can be used in clinical praxis in bariatric surgery. However, a minor, but perceptible, improvement in deep neural network (NN) ML was found.

CONCLUSION

In predicting the severe postoperative complication among the bariatric surgery patients, ensemble algorithms outperform base algorithms. When compared to other ML algorithms, deep NN has the potential to improve the accuracy and it deserves further investigation. The oversampling technique should be considered in the context of imbalanced data where the number of the interested outcome is relatively small.

Developing a Machine Learning System for Identification of Severe Hand, Foot, and Mouth Disease from Electronic Medical Record Data

Children of severe hand, foot, and mouth disease (HFMD) often present with same clinical features as those of mild HFMD during the early stage, yet later deteriorate rapidly with a fulminant disease course. Our goal was to: (1) develop a machine learning system to automatically identify cases with high risk of severe HFMD at the time of admission; (2) compare the effectiveness of the new system with the existing risk scoring system. Data on 2,532 HFMD children admitted between March 2012 and July 2015, were collected retrospectively from a medical center in China. By applying a holdout strategy and a 10-fold cross validation method, we developed four models with the random forest algorithm using different variable sets. The prediction system HFMD-RF based on the model of 16 variables from both the structured and unstructured data, achieved 0.824 sensitivity, 0.931 specificity, 0.916 accuracy, and 0.916 area under the curve in the independent test set. Most remarkably, HFMD-RF offers significant gains with respect to the commonly used pediatric critical illness score in clinical practice. As all the selected risk factors can be easily obtained, HFMD-RF might prove to be useful for reductions in mortality and complications of severe HFMD.