Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning

<b>Background/Objectives:</b><i> Helicobacter pylori</i> is a major risk factor for gastric cancer. The incidence and prevalence of the pathogen are increasing worldwide, urging novel approaches to reduce detection turnaround times. <i>H. pylori</i> diagnosis reli...

Full description

Saved in:
Bibliographic Details
Main Authors: José Aneiros-Fernández, Pedro Montero Pavón, Natalia García Gómez, Rosa María Palo Prian, Ismael Sánchez García, Ana Isabel Romero Ortiz, Rodrigo López Castro, César Casado-Sánchez, Víctor Sánchez Turrión, Antonio Luna, Manuel Álvaro Berbís
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Diagnostics
Subjects:
Online Access:https://www.mdpi.com/2075-4418/15/9/1085
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850137655935762432
author José Aneiros-Fernández
Pedro Montero Pavón
Natalia García Gómez
Rosa María Palo Prian
Ismael Sánchez García
Ana Isabel Romero Ortiz
Rodrigo López Castro
César Casado-Sánchez
Víctor Sánchez Turrión
Antonio Luna
Manuel Álvaro Berbís
author_facet José Aneiros-Fernández
Pedro Montero Pavón
Natalia García Gómez
Rosa María Palo Prian
Ismael Sánchez García
Ana Isabel Romero Ortiz
Rodrigo López Castro
César Casado-Sánchez
Víctor Sánchez Turrión
Antonio Luna
Manuel Álvaro Berbís
author_sort José Aneiros-Fernández
collection DOAJ
description <b>Background/Objectives:</b><i> Helicobacter pylori</i> is a major risk factor for gastric cancer. The incidence and prevalence of the pathogen are increasing worldwide, urging novel approaches to reduce detection turnaround times. <i>H. pylori</i> diagnosis relies on histological examination of gastric biopsies, but interobserver variability considerably impacts its identification. We present an algorithm combining a feature pyramid network and a ResNet architecture for automatic and rapid <i>H. pylori</i> detection in digitized Warthin–Starry-stained gastric biopsies. <b>Methods</b>: Whole-slide images were segmented into manually annotated smaller patches and segments containing stomach tissue were analyzed for the presence of Gram-negative bacteria. Patches classified as positive were examined to confirm the presence/absence of bacteria in contact with the gastric epithelial surface (<i>H. pylori</i>). <b>Results</b>: The algorithm exhibited 0.923 average precision and 0.982 average recall. The conducted efficiency study demonstrated that algorithm utilization significantly decreased (<i>p</i> < 0.001) diagnostic turnaround times for all participants (two pathologists, a pathology resident, a pathology technician, and a biotechnologist), observing an 88.13–91.76% time reduction. Implementation of the algorithm also improved diagnostic accuracy for the resident, technician, and biotechnologist, indicating that the tool remarkably supports less experienced personnel. <b>Conclusions</b>: We believe that the incorporation of our algorithm into pathology workflows will help standardize diagnostic protocols and drastically reduce <i>H. pylori</i> diagnostic turnaround times.
format Article
id doaj-art-e2611e1ca9e84f2494a6ce131cd4becc
institution OA Journals
issn 2075-4418
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Diagnostics
spelling doaj-art-e2611e1ca9e84f2494a6ce131cd4becc2025-08-20T02:30:46ZengMDPI AGDiagnostics2075-44182025-04-01159108510.3390/diagnostics15091085Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep LearningJosé Aneiros-Fernández0Pedro Montero Pavón1Natalia García Gómez2Rosa María Palo Prian3Ismael Sánchez García4Ana Isabel Romero Ortiz5Rodrigo López Castro6César Casado-Sánchez7Víctor Sánchez Turrión8Antonio Luna9Manuel Álvaro Berbís10Department of Pathology, University Hospital Complex of Granada, 18014 Granada, SpainDepartment of Anatomical Pathology, Hospital San Juan de la Cruz, 23400 Úbeda, Jaén, SpainDepartment of Histology, Faculty of Medicine, University of Cádiz, 11003 Cádiz, SpainPathology Department, University Hospital Virgen de las Nieves, 18014 Granada, SpainCells IA Technologies, 28006 Madrid, SpainPathology Department, University Hospital Virgen de las Nieves, 18014 Granada, SpainPathology Department, San Cecilio University Clinic Hospital, 18016 Granada, SpainFaculty of Medicine, Autonomous University of Madrid, 28029 Madrid, SpainFaculty of Medicine, Autonomous University of Madrid, 28029 Madrid, SpainDepartment of Integrated Diagnostics, HT Médica, Clínica Las Nieves, 23007 Jaén, SpainCells IA Technologies, 28006 Madrid, Spain<b>Background/Objectives:</b><i> Helicobacter pylori</i> is a major risk factor for gastric cancer. The incidence and prevalence of the pathogen are increasing worldwide, urging novel approaches to reduce detection turnaround times. <i>H. pylori</i> diagnosis relies on histological examination of gastric biopsies, but interobserver variability considerably impacts its identification. We present an algorithm combining a feature pyramid network and a ResNet architecture for automatic and rapid <i>H. pylori</i> detection in digitized Warthin–Starry-stained gastric biopsies. <b>Methods</b>: Whole-slide images were segmented into manually annotated smaller patches and segments containing stomach tissue were analyzed for the presence of Gram-negative bacteria. Patches classified as positive were examined to confirm the presence/absence of bacteria in contact with the gastric epithelial surface (<i>H. pylori</i>). <b>Results</b>: The algorithm exhibited 0.923 average precision and 0.982 average recall. The conducted efficiency study demonstrated that algorithm utilization significantly decreased (<i>p</i> < 0.001) diagnostic turnaround times for all participants (two pathologists, a pathology resident, a pathology technician, and a biotechnologist), observing an 88.13–91.76% time reduction. Implementation of the algorithm also improved diagnostic accuracy for the resident, technician, and biotechnologist, indicating that the tool remarkably supports less experienced personnel. <b>Conclusions</b>: We believe that the incorporation of our algorithm into pathology workflows will help standardize diagnostic protocols and drastically reduce <i>H. pylori</i> diagnostic turnaround times.https://www.mdpi.com/2075-4418/15/9/1085artificial intelligencegastroenterology<i>Helicobacter pylori</i>deep learningwhole-slide imagingdigital pathology
spellingShingle José Aneiros-Fernández
Pedro Montero Pavón
Natalia García Gómez
Rosa María Palo Prian
Ismael Sánchez García
Ana Isabel Romero Ortiz
Rodrigo López Castro
César Casado-Sánchez
Víctor Sánchez Turrión
Antonio Luna
Manuel Álvaro Berbís
Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
Diagnostics
artificial intelligence
gastroenterology
<i>Helicobacter pylori</i>
deep learning
whole-slide imaging
digital pathology
title Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
title_full Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
title_fullStr Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
title_full_unstemmed Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
title_short Rapid and Efficient Screening of <i>Helicobacter pylori</i> in Gastric Samples Stained with Warthin–Starry Using Deep Learning
title_sort rapid and efficient screening of i helicobacter pylori i in gastric samples stained with warthin starry using deep learning
topic artificial intelligence
gastroenterology
<i>Helicobacter pylori</i>
deep learning
whole-slide imaging
digital pathology
url https://www.mdpi.com/2075-4418/15/9/1085
work_keys_str_mv AT joseaneirosfernandez rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT pedromonteropavon rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT nataliagarciagomez rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT rosamariapaloprian rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT ismaelsanchezgarcia rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT anaisabelromeroortiz rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT rodrigolopezcastro rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT cesarcasadosanchez rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT victorsanchezturrion rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT antonioluna rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning
AT manuelalvaroberbis rapidandefficientscreeningofihelicobacterpyloriiingastricsamplesstainedwithwarthinstarryusingdeeplearning