Internal and External Validation Datasets for Use in Detecting Right Ventricular Dysfunction in Acute Pulmonary Embolism

This study included patients with acute pulmonary embolism (PE) confirmed by contrast-enhanced multidetector CT. Eligibility required echocardiography and lower-extremity venous Doppler reports. Exclusion criteria were pregnancy, missing imaging, chronic thromboembolic pulmonary hypertension, prior right heart dysfunction, and age <18 years.

Each row in the dataset provides information about a patient's characteristics. The first column shows the gender of the PE patients, the second column shows their ages, and the other columns show various characteristics used in the diagnosis of RVD. The last column shows whether the PE patients were diagnosed with RVD by the physician. This dataset can be used in the artificial intelligence models. It is a suitable dataset for the classification performance of various models.
We established two non-overlapping cohorts:
• A development set of n = 250 patients (103 positive, 147 negative)
• An external validation set of n = 113 patients (36 positive, 77 negative).
Outcome (Target Variable). We defined the target variable as Right Ventricular Involvement (RVD), coded as 1 = Presence, 0 = Absence, based on echocardiographic assessment.

Table 1. List of features indicating the range of changes in values.
№ Feature name Value range Value description
1 Sex ( 0-1 ) (1:Male, 0:Female)
2 Age ( 22-93 ) -
3 Thrombus Main Pulmonary ( 0-1 ) (1: Presence, 0: Absence)
4 Thrombus Lobe Arteries ( 0-1 ) (1: Presence, 0: Absence)
5 Thrombus Segment ( 0-1 ) (1: Presence, 0: Absence)
6 Thrombus Subsegment ( 0-1 ) (1: Presence, 0: Absence)
7 Thrombus ( 0-1 ) (1: Unilateral, 0: Bilateral)
8 DVT ( 0-1 ) (1: Presence, 0: Absence)
9 DVT Distal Ven ( 0-1 ) (1: Presence, 0: Absence)
10 DVT Proksimal Ven ( 0-1 ) (1: Presence, 0: Absence)
11 Appearance of DVT ( 0-2 ) (Absence: 0 Unilateral :1 Bilateral:
12 Comorbid Disease ( 0-1 ) (1: Presence, 0: Absence)
13 Malignancy ( 0-1 ) (1: Presence, 0: Absence)
14 Diabetes Mellitus ( 0-1 ) (1: Presence, 0: Absence)
15 Hypertension ( 0-1 ) (1: Presence, 0: Absence)
16 Coronary Artery Disease (CAD) ( 0-1 ) (1: Presence, 0: Absence)
17 COPD ( 0-1 ) (1: Presence, 0: Absence)
18 Asthma ( 0-1 ) (1: Presence, 0: Absence)
19 Cerebrovascular Occlusion ( 0-1 ) (1: Presence, 0: Absence)
20 Heart Failure ( 0-1 ) (1: Presence, 0: Absence)
*21 Right Ventricular Involvement ( 0-1 ) (1: Presence, 0: Absence)
Note: *Target variable
DVT: Deep vein thrombosis.
Please cite the following four articles when you use any of these datasets:
1) Huyut, M.T., Velichko, A., Belyaev, M. et al. Identification of right ventricular dysfunction with LogNNet based diagnostic model: A comparative study with supervised ML algorithms. Sci Rep 15, 25262 (2025). https://doi.org/10.1038/s41598-025-00274-1.
2) Huyut, M.T. et. al., "Detection of Right Ventricular Dysfunction Using LogNNet Neural Network Model Based on Pulmonary Embolism Data Set," Eastern Journal of Medicine , vol.29, no.1, pp.118-128, 2024.