Abstract P391: An-adcs
            2
            : A Novel Machine-Learning Model to Predict the Risk of Stroke-Associated Pneumonia

Objective: We aimed to develop and validate a machine learning-based prediction model that could assess the risk of stroke-associated pneumonia (SAP) for individual patients with acute ischemic stroke (AIS). Methods: A machine-learning model incorporating A 2 DS 2 scores and clinical features (AN-ADCS 2 ) was developed to predict the risk of SAP in patients with AIS. Two independent datasets were used for model derivation and external validation. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were estimated. The further analysis evaluated thresholds from the training set that identified patients as low-risk, intermediate-risk and high-risk, and performance at these thresholds was compared in the external validation set. Results: The AN-ADCS 2 model achieved favorable performance with a high AUC of 0.892 (95% confidence interval [CI] 0.885-0.898) in the test set and similar performance in the external validation set (AUC 0.813 [95% CI 0.812-0.814]). The AN-ADCS 2 threshold identifying low-risk was 0.03, with a NPV of 97.6% (97.2-97.9%) and sensitivity of 93.5% (92.5-94.5%). The AN-ADCS 2 threshold identifying high-risk was 0.65, with a PPV of 94.7% (93.9-95.6%) and specificity of 99.5% (99.5-99.6%). The AN-ADCS 2 model performed better than the A 2 DS 2 score (AUC 0.739, 95%CI [0.720-0.754]). Having a high risk of SAP classified by the AN-ADCS 2 was associated with unfavorable outcomes of mortality and in-hospital stroke recurrence. Conclusions: Using machine learning, the AN-ADCS 2 model provides an individualized risk prediction of SAP, which can be used as an indicator of clinical prognosis for patients with AIS.

Download Full-text

CT imaging-based machine learning model: a potential modality for predicting low-risk and high-risk groups of thymoma: “Impact of surgical modality choice”

World Journal of Surgical Oncology ◽

10.1186/s12957-021-02259-6 ◽

2021 ◽

Vol 19 (1) ◽

Author(s):

Ayten Kayi Cangir ◽

Kaan Orhan ◽

Yusuf Kahya ◽

Hilal Özakıncı ◽

Betül Bahar Kazak ◽

...

Keyword(s):

Machine Learning ◽

High Risk ◽

Risk Groups ◽

Learning Model ◽

Low Risk ◽

Pathology Report ◽

Histopathological Diagnosis ◽

Gray Level ◽

Long Run ◽

Machine Learning Model

Abstract Introduction Radiomics methods are used to analyze various medical images, including computed tomography (CT), magnetic resonance, and positron emission tomography to provide information regarding the diagnosis, patient outcome, tumor phenotype, and the gene-protein signatures of various diseases. In low-risk group, complete surgical resection is typically sufficient, whereas in high-risk thymoma, adjuvant therapy is usually required. Therefore, it is important to distinguish between both. This study evaluated the CT radiomics features of thymomas to discriminate between low- and high-risk thymoma groups. Materials and methods In total, 83 patients with thymoma were included in this study between 2004 and 2019. We used the Radcloud platform (Huiying Medical Technology Co., Ltd.) to manage the imaging and clinical data and perform the radiomics statistical analysis. The training and validation datasets were separated by a random method with a ratio of 2:8 and 502 random seeds. The histopathological diagnosis was noted from the pathology report. Results Four machine-learning radiomics features were identified to differentiate a low-risk thymoma group from a high-risk thymoma group. The radiomics feature names were Energy, Zone Entropy, Long Run Low Gray Level Emphasis, and Large Dependence Low Gray Level Emphasis. Conclusions The results demonstrated that a machine-learning model and a multilayer perceptron classifier analysis can be used on CT images to predict low- and high-risk thymomas. This combination could be a useful preoperative method to determine the surgical approach for thymoma.

Download Full-text

CT Imaging-based Machine Learning Model: A Potential Modality for Predicting Low-Risk and High-Risk Groups of Thymoma “Impact of surgical modality choice”

10.21203/rs.3.rs-139108/v1 ◽

2021 ◽

Author(s):

Ayten KAYICANGIR ◽

Kaan ORHAN ◽

Yusuf KAHYA ◽

Hilal ÖZAKINCI ◽

Betül Bahar KAZAK ◽

...

Keyword(s):

Machine Learning ◽

High Risk ◽

Risk Groups ◽

Learning Model ◽

Low Risk ◽

Pathology Report ◽

Histopathological Diagnosis ◽

Gray Level ◽

Long Run ◽

Machine Learning Model

Abstract IntroductionRadiomics has become a hot issue in the medical imaging field, particularly in cancer imaging. Radiomics methods are used to analyze various medical images, including computed tomography (CT), magnetic resonance, and positron emission tomography to provide information regarding the diagnosis, patient outcome, tumor phenotype, and the gene-protein signatures of various diseases.This study evaluated the CT radiomics features of thymomas to discriminate between low- and high-risk thymoma groups.Materials and MethodsIn total, 83 patients with thymoma were included in this study between 2004 and 2019. We used the Radcloud platform (Huiying Medical Technology Co., Ltd.) to manage the imaging and clinical data and perform the radiomics statistical analysis. The training and validation datasets were separated by a random method with a ratio of 2:8 and 502 random seeds. The histopathological diagnosis was noted from the pathology report.ResultsFour machine learning radiomics features were identified to differentiate a low-risk thymoma group from a high-risk thymoma group. The radiomics feature names were Energy, Zone Entropy, Long Run Low Gray Level Emphasis, and Large Dependence Low Gray Level Emphasis.ConclusionsThe results demonstrated that a machine-learning model and a multilayer perceptron classifier analysis can be used on CT images to predict low- and high-risk thymomas. This combination could be a useful preoperative method to determine the surgical approach for thymoma.

Download Full-text

Statistical machine learning model to predict Oncotype DX risk category in women over age 50.

Journal of Clinical Oncology ◽

10.1200/jco.2020.38.15_suppl.524 ◽

2020 ◽

Vol 38 (15_suppl) ◽

pp. 524-524

Author(s):

Kate R. Pawloski ◽

Mithat Gonen ◽

Hannah Yong Wen ◽

Audree B. Tadros ◽

Kelly Abbate ◽

...

Keyword(s):

Machine Learning ◽

High Risk ◽

Tumor Size ◽

Low Risk ◽

Risk Category ◽

Statistical Machine Learning ◽

Node Negative ◽

Machine Learning Model ◽

Er Positive ◽

Validation Set

524 Background: The 21-gene Oncotype DX Breast Recurrence Score multigene assay (RS) identifies women with ER positive, HER negative, axillary node-negative breast cancer (BC) for whom chemotherapy provides no invasive disease-free survival benefit compared to endocrine therapy alone. International adoption of RS testing is limited by cost and resource availability. We created a supervised statistical machine learning model using standard clinicopathologic data to predict RS risk category in women > 50 years old. Methods: From 2012 to 2018, women with ER positive, HER2 negative, pathologically node-negative BC of all ages were retrospectively identified from a prospective institutional database. Standard clinicopathologic data and RS were collected. Per institutional protocol, RS are ordered for all early-stage, ER positive tumors > 5 mm. Data were randomly split into training (n=3755) and validation sets (n=1609). A random forest model with 500 trees was developed on the training set, then evaluated on the validation set. Model predictors included age, tumor size, histologic subtype, hormone receptor status, lymphovascular invasion, and overall grade. The model was used to predict RS category (low risk: RS ≤ 25, high risk: RS > 25) in women > 50 years old. Results: 5364 unique tumors in 5189 women were identified. 3731 (70%) of tumors were identified in women > 50 years; median age was 63 years (IQR 57-69). In women > 50, median tumor size was 12 mm (IQR 9-17). Most tumors were invasive ductal (79%), low or intermediate grade (79%), and LVI was absent in 82% of tumors. Median ER staining by IHC was 95%; 28% of tumors had negative or weakly positive PR staining (1-20%). The model correctly classified 96.8% of patients as low risk (95% CI: 95.7-97.7). Negative predictive value for identifying low risk category was also high (92.3%, 90.7-93.6). Sensitivity for identifying high-risk women was 44.7% (37.4-52.1) and positive predictive value was 67.2% (58.2-75.3). A classification table on the validation set includes tumors with complete data available, including predictors and RS. Conclusions: Our model was highly specific (96.8%) for identifying women > 50 with RS ≤ 25 who do not benefit from adjuvant chemotherapy. This model may be utilized in lieu of RS testing if cost and availability are prohibitive. True RS > 25 was not as well predicted. The model will be refined following pathologic review of discordant cases to reduce false negatives. [Table: see text]

Download Full-text

Development of a machine learning model for predicting pediatric mortality in the early stages of intensive care unit admission

Scientific Reports ◽

10.1038/s41598-020-80474-z ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Bongjin Lee ◽

Kyunghoon Kim ◽

Hyejin Hwang ◽

You Sun Kim ◽

Eun Hee Chung ◽

...

Keyword(s):

Machine Learning ◽

Intensive Care Unit ◽

Intensive Care ◽

Validation Cohort ◽

External Validation ◽

Learning Model ◽

Derivation Cohort ◽

Icu Admission ◽

Early Stages ◽

Machine Learning Model

AbstractThe aim of this study was to develop a predictive model of pediatric mortality in the early stages of intensive care unit (ICU) admission using machine learning. Patients less than 18 years old who were admitted to ICUs at four tertiary referral hospitals were enrolled. Three hospitals were designated as the derivation cohort for machine learning model development and internal validation, and the other hospital was designated as the validation cohort for external validation. We developed a random forest (RF) model that predicts pediatric mortality within 72 h of ICU admission, evaluated its performance, and compared it with the Pediatric Index of Mortality 3 (PIM 3). The area under the receiver operating characteristic curve (AUROC) of RF model was 0.942 (95% confidence interval [CI] = 0.912–0.972) in the derivation cohort and 0.906 (95% CI = 0.900–0.912) in the validation cohort. In contrast, the AUROC of PIM 3 was 0.892 (95% CI = 0.878–0.906) in the derivation cohort and 0.845 (95% CI = 0.817–0.873) in the validation cohort. The RF model in our study showed improved predictive performance in terms of both internal and external validation and was superior even when compared to PIM 3.

Download Full-text

Development and external validation of a deep learning-based computed tomography classification system for COVID-19

10.31219/osf.io/j6xhb ◽

2021 ◽

Author(s):

Yuki KATAOKA

Keyword(s):

Machine Learning ◽

Computed Tomography ◽

Interstitial Pneumonia ◽

External Validation ◽

Usual Interstitial Pneumonia ◽

Evaluation Process ◽

High Sensitivity ◽

Learning Model ◽

Machine Learning Model ◽

Ablation Study

Rationale: Currently available machine learning models for diagnosing COVID-19 based on computed tomography (CT) images are limited due to concerns regarding methodological flaws or underlying biases in the evaluation process. Objectives: We aimed to develop and externally validate a novel machine learning model that can classify CT image findings as positive or negative for SARS-CoV-2 reverse transcription polymerase chain reaction (RT-PCR).Methods: We used 3128 images from a wide variety of two-gate data sources for the development and ablation study of the machine learning model. A total of 633 COVID-19 cases and 2295 non-COVID-19 cases were included in the study. We randomly divided cases into a development set and ablation set at a ratio of 8:2. For the ablation study, we used another dataset including 150 cases of interstitial pneumonia among non-COVID-19 images. For external validation, we used 893 images from 740 consecutive patients at 11 acute care hospitals suspected of having COVID-19 at the time of diagnosis. The dataset included 343 COVID-19 patients. The reference standard was RT-PCR.Result: In ablation study, using interstitial pneumonia images, the specificity of the model were 0.986 for usual interstitial pneumonia pattern, 0.820 for non-specific interstitial pneumonia pattern, 0.400 for organizing pneumonia pattern. In the external validation study, the sensitivity and specificity of the model were 0.869 and 0.432, respectively, at the low-level cutoff, and 0.724 and 0.721, respectively, at the high-level cutoff.Conclusions: Our machine learning model exhibited a high sensitivity in external validation datasets and may assist physicians to rule out COVID-19 diagnosis in a timely manner. Further studies are warranted to improve model specificity.

Download Full-text

Machine Learning Model for Outcome Prediction of Patients Suffering from Acute Diverticulitis Arriving at the Emergency Department—A Proof of Concept Study

Diagnostics ◽

10.3390/diagnostics11112102 ◽

2021 ◽

Vol 11 (11) ◽

pp. 2102

Author(s):

Eyal Klang ◽

Robert Freeman ◽

Matthew A. Levin ◽

Shelly Soffer ◽

Yiftach Barash ◽

...

Keyword(s):

Machine Learning ◽

Emergency Department ◽

Acute Diverticulitis ◽

External Validation ◽

White Blood Cells ◽

Learning Model ◽

Complicated Diverticulitis ◽

Gradient Boosting ◽

Proof Of Concept ◽

Machine Learning Model

Background & Aims: We aimed at identifying specific emergency department (ED) risk factors for developing complicated acute diverticulitis (AD) and evaluate a machine learning model (ML) for predicting complicated AD. Methods: We analyzed data retrieved from unselected consecutive large bowel AD patients from five hospitals from the Mount Sinai health system, NY. The study time frame was from January 2011 through March 2021. Data were used to train and evaluate a gradient-boosting machine learning model to identify patients with complicated diverticulitis, defined as a need for invasive intervention or in-hospital mortality. The model was trained and evaluated on data from four hospitals and externally validated on held-out data from the fifth hospital. Results: The final cohort included 4997 AD visits. Of them, 129 (2.9%) visits had complicated diverticulitis. Patients with complicated diverticulitis were more likely to be men, black, and arrive by ambulance. Regarding laboratory values, patients with complicated diverticulitis had higher levels of absolute neutrophils (AUC 0.73), higher white blood cells (AUC 0.70), platelet count (AUC 0.68) and lactate (AUC 0.61), and lower levels of albumin (AUC 0.69), chloride (AUC 0.64), and sodium (AUC 0.61). In the external validation cohort, the ML model showed AUC 0.85 (95% CI 0.78–0.91) for predicting complicated diverticulitis. For Youden’s index, the model showed a sensitivity of 88% with a false positive rate of 1:3.6. Conclusions: A ML model trained on clinical measures provides a proof of concept performance in predicting complications in patients presenting to the ED with AD. Clinically, it implies that a ML model may classify low-risk patients to be discharged from the ED for further treatment under an ambulatory setting.

Download Full-text

Identification of Patient Profiles with High Risk of Hospital Re-Admissions for Acute COPD Exacerbations (AECOPD) in France Using a Machine Learning Model

International Journal of Chronic Obstructive Pulmonary Disease ◽

10.2147/copd.s236787 ◽

2020 ◽

Vol Volume 15 ◽

pp. 949-962 ◽

Cited By ~ 1

Author(s):

Arnaud Cavailles ◽

Boris Melloni ◽

Stéphane Motola ◽

Florent Dayde ◽

Marie Laurent ◽

...

Keyword(s):

Machine Learning ◽

High Risk ◽

Learning Model ◽

Copd Exacerbations ◽

Machine Learning Model ◽

Patient Profiles

Download Full-text

GP.6 Improving Triaging of EEG Referrals for Rule out Infantile Spasms (ITERIS)

Canadian Journal of Neurological Sciences / Journal Canadien des Sciences Neurologiques ◽

10.1017/cjn.2021.262 ◽

2021 ◽

Vol 48 (s3) ◽

pp. S13-S13

Author(s):

D Djordjevic ◽

J Tracey ◽

M Alqahtani ◽

J Boyd ◽

C Go

Keyword(s):

Machine Learning ◽

Predictive Factors ◽

Predictive Accuracy ◽

Learning Model ◽

Infantile Spasms ◽

Low Risk ◽

Machine Learning Techniques ◽

Risk Category ◽

Point System ◽

Machine Learning Model

Background: Infantile spasms (IS) is a devastating pediatric seizure disorder for which EEG referrals are prioritized at the Hospital for Sick Children, representing a resource challenge. The goal of this study was to improve the triaging system for these referrals. Methods: Part 1: descriptive analysis was performed retrospectively on EEG referrals. Part 2: prospective questionnaires were used to determine relative risk of various predictive factors. Part 3: electronic referral form was amended to include 5 positive predictive factors. A triage point system was tested by assigning EEGs as high risk (3 days), standard risk (1 week), or low risk (2 weeks). A machine learning model was developed. Results: Most EEG referrals were from community pediatricians with a low yield of IS diagnoses. Using the 5 predictive factors, the proposed triage system accurately diagnosed all IS within 3 days. No abnormal EEGs were missed in the low-risk category. The machine learning model had over 90% predictive accuracy and will be prospectively tested. Conclusions: Improving EEG triaging for IS may be possible to prioritize higher risk patients. Machine Learning techniques can potentially be applied to help with predictions. We hope that our findings will ultimately improve resource utilization and patient care.

Download Full-text