Taipei Medical University-led team developed an Artificial Intelligence based models for Screening of Blood cancers using Complete Blood Count results

Tuesday, March 17, 2020

On March 16th 2020, Nature’s Scientific Reports published “Artificial Intelligence based Models for Screening of Hematologic Malignancies using Cell Population Data.”

Scientific Reports is an online open access scientific mega journal published by Nature Research (IF 4.01). Lead author Dr. Shabbir Syed-Abdul from Taipei Medical University in collaboration with researchers from South Korea’s Konkuk University School of Medicine, Kingdom of Saudi Arabia’s King Abdullah International Medical Research Center, King Saud bin Abdulaziz University for Health Sciences, and Slovenia’s Jožef Stefan Institute developed an “Artificial Intelligence based Models for Screening of Hematologic Malignancies using Cell Population Data”

Complete blood count (CBC) or Cell Population Data (CPD) is a common blood test used to measures the amount of various types of blood cells in a sample of your blood. Blood cancers may be detected using this test if too many or too few of a type of blood cell or abnormal cells are found. A bone marrow biopsy may help confirm a diagnosis of a blood cancer. There are no tests specifically designed or endorsed for blood cancer screening, the condition may be detected through a: Chest X-ray or CAT scan – Images of the chest can sometimes reveal swollen lymph nodes or signs of infection.

As the abstract notes, “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 the Konkuk University Medical Center, Seoul. A total of (882 cases: 457 hematologic malignancy and 425 hematologic non-malignancy) were used for analysis. In this study, seven machine learning models, i.e., Stochastic Gradient Descent (SGD), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Linear model, Logistic regression, and Artificial Neural Network (ANN), were used. The authors achieved a decent result, 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.

This research presents a novel approach of using ML algorithm for screening patients with suspected hematologic malignancies versus non-malignancies using CPD that was generated by routine CBC. The conclusion of the study is that based on CPD, the ANN algorithm appears to be an efficient aid for the clinical laboratory diagnostic approach of hematologic malignancies. In the future, we are planning to apply this algorithm to the outpatient data in hematology departments.

The far-reaching implications of this study for other areas of cancer screening will surely result in a wave of research started by this global team, which is Taiwan, South Korea, Kingdom of Saudi Arabia and Slovenia researchers representing from four nationalities.

The paper closes with a call for further prospective research and trials that will be mandatory to confirm the validity of clinical AI-based models before it actually helps physicians in clinical practice, particularly in haematologic diseases.

For more information, contact Dr. Shabbir Syed-Abdul at drshabbir@tmu.edu.tw