Performance validation of deep-learning-based approach in stool examination

Abstract Background Human intestinal parasitic infections (IPI) pose a significant global health issue caused by parasitic helminths and protozoa, affecting around 3.5 billion people worldwide, with more than 200,000 deaths annually. Despite advancements in molecular methods with higher sensitivity...

Full description

Saved in:
Bibliographic Details
Main Authors: Kristal Dale Felimon Corpuz, Teera Kusolsuk, Benjamaporn Wongphan, Putza Chonsawat, Kaung Myat Naing, Siridech Boonsang, Veerayuth Kittichai, Chia-Kwung Fan, Santhad Chuwongin, Dorn Watthanakulpanich
Format: Article
Language:English
Published: BMC 2025-08-01
Series:Parasites & Vectors
Subjects:
Online Access:https://doi.org/10.1186/s13071-025-06878-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849333628725624832
author Kristal Dale Felimon Corpuz
Teera Kusolsuk
Benjamaporn Wongphan
Putza Chonsawat
Kaung Myat Naing
Siridech Boonsang
Veerayuth Kittichai
Chia-Kwung Fan
Santhad Chuwongin
Dorn Watthanakulpanich
author_facet Kristal Dale Felimon Corpuz
Teera Kusolsuk
Benjamaporn Wongphan
Putza Chonsawat
Kaung Myat Naing
Siridech Boonsang
Veerayuth Kittichai
Chia-Kwung Fan
Santhad Chuwongin
Dorn Watthanakulpanich
author_sort Kristal Dale Felimon Corpuz
collection DOAJ
description Abstract Background Human intestinal parasitic infections (IPI) pose a significant global health issue caused by parasitic helminths and protozoa, affecting around 3.5 billion people worldwide, with more than 200,000 deaths annually. Despite advancements in molecular methods with higher sensitivity and specificity, the Kato-Katz or formalin-ethyl acetate centrifugation technique (FECT) remains the gold standard and a routine diagnostic procedure suitable for its simplicity and cost-effectiveness. However, these techniques have limitations that must be addressed. Thus, this study evaluated the performance of a deep-learning-based approach for intestinal parasite identification and compared it with that of human experts. Methods Human experts performed FECT and Merthiolate-iodine-formalin (MIF) techniques to serve as ground truth and reference for parasite species. Subsequently, a modified direct smear was conducted to gather images for the training (80%) and testing (20%) datasets. State-of-the-art models, including YOLOv4-tiny, YOLOv7-tiny, YOLOv8-m, ResNet-50, and DINOv2 (base, small, and large), were employed and were operated using in-house CIRA CORE platform. Overall performance was evaluated using confusion matrices, the metrics of which were calculated on the basis of the one-versus-rest and micro-averaging approaches. Moreover, the receiver operating characteristic (ROC) and precision-recall (PR) curves were determined for visual comparison. Lastly, Cohen’s Kappa and Bland–Altman analyses were used to statistically measure the significant differences and visualize the association levels between the human experts and the deep learning models’ classification performance in intestinal parasite identification. Results Findings demonstrated the potential of a deep-learning-based approach, particularly of models DINOv2-large (accuracy: 98.93%; precision: 84.52%; sensitivity: 78.00%; specificity: 99.57%; F1 score: 81.13%; AUROC: 0.97) and YOLOv8-m (accuracy: 97.59%; precision: 62.02%; sensitivity: 46.78%; specificity: 99.13%; F1 score: 53.33%; AUROC: 0.755; AUPR: 0.556) for their high metric values in intestinal parasite identification. Class-wise prediction showed high precision, sensitivity, and F1 scores for helminthic eggs and larvae due to more distinct morphology. Moreover, all models obtained a > 0.90 k score, which indicates a strong level of agreement compared with the medical technologists. The Bland–Altman analysis also presented the best agreement between FECT performed by medical technologist A and YOLOv4-tiny, while the MIF technique performed by medical technologist B and DINOv2-small demonstrated the best bias-free agreement, with mean differences of 0.0199 and −0.0080, and standard deviation differences of 0.6012 and 0.5588, respectively. Conclusions The results highlight the potential of integrating a deep-learning-based approach into parasite identification. The models showcased superiority in automated detection, suggesting a significant leap toward improving diagnostic procedures for IPI. This hybridization could enhance early detection and diagnosis, facilitating timely and targeted interventions to reduce the burden of IPI through more effective management and prevention strategies. Graphical Abstract
format Article
id doaj-art-44225bb66b4943afbca61b3ee2530349
institution Kabale University
issn 1756-3305
language English
publishDate 2025-08-01
publisher BMC
record_format Article
series Parasites & Vectors
spelling doaj-art-44225bb66b4943afbca61b3ee25303492025-08-20T03:45:48ZengBMCParasites & Vectors1756-33052025-08-0118112610.1186/s13071-025-06878-wPerformance validation of deep-learning-based approach in stool examinationKristal Dale Felimon Corpuz0Teera Kusolsuk1Benjamaporn Wongphan2Putza Chonsawat3Kaung Myat Naing4Siridech Boonsang5Veerayuth Kittichai6Chia-Kwung Fan7Santhad Chuwongin8Dorn Watthanakulpanich9Department of Helminthology, Faculty of Tropical Medicine, Mahidol UniversityDepartment of Helminthology, Faculty of Tropical Medicine, Mahidol UniversityHospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol UniversityHospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol UniversityCenter of Industrial Robot and Automation (CIRA), College of Advanced Manufacturing Innovation, King Mongkut’s Institute of Technology LadkrabangDepartment of Electrical Engineering, School of Engineering, King Mongkut’s Institute of Technology LadkrabangFaculty of Medicine, King Mongkut’s Institute of Technology LadkrabangDepartment of Molecular Parasitology and Tropical Diseases, School of Medicine, College of Medicine, Taipei Medical UniversityCenter of Industrial Robot and Automation (CIRA), College of Advanced Manufacturing Innovation, King Mongkut’s Institute of Technology LadkrabangDepartment of Helminthology, Faculty of Tropical Medicine, Mahidol UniversityAbstract Background Human intestinal parasitic infections (IPI) pose a significant global health issue caused by parasitic helminths and protozoa, affecting around 3.5 billion people worldwide, with more than 200,000 deaths annually. Despite advancements in molecular methods with higher sensitivity and specificity, the Kato-Katz or formalin-ethyl acetate centrifugation technique (FECT) remains the gold standard and a routine diagnostic procedure suitable for its simplicity and cost-effectiveness. However, these techniques have limitations that must be addressed. Thus, this study evaluated the performance of a deep-learning-based approach for intestinal parasite identification and compared it with that of human experts. Methods Human experts performed FECT and Merthiolate-iodine-formalin (MIF) techniques to serve as ground truth and reference for parasite species. Subsequently, a modified direct smear was conducted to gather images for the training (80%) and testing (20%) datasets. State-of-the-art models, including YOLOv4-tiny, YOLOv7-tiny, YOLOv8-m, ResNet-50, and DINOv2 (base, small, and large), were employed and were operated using in-house CIRA CORE platform. Overall performance was evaluated using confusion matrices, the metrics of which were calculated on the basis of the one-versus-rest and micro-averaging approaches. Moreover, the receiver operating characteristic (ROC) and precision-recall (PR) curves were determined for visual comparison. Lastly, Cohen’s Kappa and Bland–Altman analyses were used to statistically measure the significant differences and visualize the association levels between the human experts and the deep learning models’ classification performance in intestinal parasite identification. Results Findings demonstrated the potential of a deep-learning-based approach, particularly of models DINOv2-large (accuracy: 98.93%; precision: 84.52%; sensitivity: 78.00%; specificity: 99.57%; F1 score: 81.13%; AUROC: 0.97) and YOLOv8-m (accuracy: 97.59%; precision: 62.02%; sensitivity: 46.78%; specificity: 99.13%; F1 score: 53.33%; AUROC: 0.755; AUPR: 0.556) for their high metric values in intestinal parasite identification. Class-wise prediction showed high precision, sensitivity, and F1 scores for helminthic eggs and larvae due to more distinct morphology. Moreover, all models obtained a > 0.90 k score, which indicates a strong level of agreement compared with the medical technologists. The Bland–Altman analysis also presented the best agreement between FECT performed by medical technologist A and YOLOv4-tiny, while the MIF technique performed by medical technologist B and DINOv2-small demonstrated the best bias-free agreement, with mean differences of 0.0199 and −0.0080, and standard deviation differences of 0.6012 and 0.5588, respectively. Conclusions The results highlight the potential of integrating a deep-learning-based approach into parasite identification. The models showcased superiority in automated detection, suggesting a significant leap toward improving diagnostic procedures for IPI. This hybridization could enhance early detection and diagnosis, facilitating timely and targeted interventions to reduce the burden of IPI through more effective management and prevention strategies. Graphical Abstracthttps://doi.org/10.1186/s13071-025-06878-wIntestinal parasitic infectionSelf-supervised learningDeep learningAutomation
spellingShingle Kristal Dale Felimon Corpuz
Teera Kusolsuk
Benjamaporn Wongphan
Putza Chonsawat
Kaung Myat Naing
Siridech Boonsang
Veerayuth Kittichai
Chia-Kwung Fan
Santhad Chuwongin
Dorn Watthanakulpanich
Performance validation of deep-learning-based approach in stool examination
Parasites & Vectors
Intestinal parasitic infection
Self-supervised learning
Deep learning
Automation
title Performance validation of deep-learning-based approach in stool examination
title_full Performance validation of deep-learning-based approach in stool examination
title_fullStr Performance validation of deep-learning-based approach in stool examination
title_full_unstemmed Performance validation of deep-learning-based approach in stool examination
title_short Performance validation of deep-learning-based approach in stool examination
title_sort performance validation of deep learning based approach in stool examination
topic Intestinal parasitic infection
Self-supervised learning
Deep learning
Automation
url https://doi.org/10.1186/s13071-025-06878-w
work_keys_str_mv AT kristaldalefelimoncorpuz performancevalidationofdeeplearningbasedapproachinstoolexamination
AT teerakusolsuk performancevalidationofdeeplearningbasedapproachinstoolexamination
AT benjamapornwongphan performancevalidationofdeeplearningbasedapproachinstoolexamination
AT putzachonsawat performancevalidationofdeeplearningbasedapproachinstoolexamination
AT kaungmyatnaing performancevalidationofdeeplearningbasedapproachinstoolexamination
AT siridechboonsang performancevalidationofdeeplearningbasedapproachinstoolexamination
AT veerayuthkittichai performancevalidationofdeeplearningbasedapproachinstoolexamination
AT chiakwungfan performancevalidationofdeeplearningbasedapproachinstoolexamination
AT santhadchuwongin performancevalidationofdeeplearningbasedapproachinstoolexamination
AT dornwatthanakulpanich performancevalidationofdeeplearningbasedapproachinstoolexamination