This Author published in this journals
All Journal Atom Indonesia
Ha, B. N.
Unknown Affiliation

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Development of a Vietnamese PET/CT Dataset for Machine Learning-Based Analysis of Non-Small Cell Lung Cancer Images Tuan, H. Q.; Duong, T. T.; Ha, B. N.; Quyet, N. H.; Tinh, L. V.; Tuynh, C. V.; Nam, V. K.; Dao, L. T. Q.; Luong, C. V.; Linh, D. T. M.; Nhung, D. T.; Nguyen, N. D.; Trang, V. Q.
Atom Indonesia Vol 51, No 2 (2025): AUGUST 2025
Publisher : National Research and Innovation Agency

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55981/aij.2025.1645

Abstract

Positron Emission Tomography and Computed Tomography (PET/CT), a key imaging modality in nuclear medicine, Combines Anatomical (CT) and functional (PET) data for cancer diagnosis. Despite advancements in machine learning for automated medical image analysis, publicly available PET/CT datasets remain scarce, limiting Artificial Intelligence (AI) research compared to CT and MRI. This study built a publicly accessible PET/CT Vietnamese dataset for Non-Small Cell Lung Cancer (NSCLC). A total of 416 PET/CT scans were collected from three Vietnamese hospitals, including 300 NSCLC cases. Malignant FDG-sensitive lesions, identified via clinical PET/CT reports, were manually segmented in 3D (slice-by-slice) on PET images and validated by three experienced radiologists. The dataset includes both original and annotated DICOM files, along with clinical patient data. It achieved a dice similarity coefficient of 80.3 % and volume similarity of 81.9 %, demonstrating high segmentation accuracy comparable to other studies. This dataset supports AI-driven NSCLC research and contributes to global efforts in automated PET/CT analysis for nuclear medicine applications.