Rectal Cancer Segmentation: A Methodical Approach for Generalizable Deep Learning in a Multi‐Center Setting

ABSTRACT Noninvasive Artificial Intelligence (AI) techniques have shown great potential in assisting clinicians through the analysis of medical images. However, significant challenges remain in integrating these systems into clinical practice due to the variability of medical data across multi‐center databases and the lack of clear implementation guidelines. These issues hinder the ability to achieve robust, reproducible, and statistically significant results. This study thoroughly analyzes several decision‐making steps involved in managing a multi‐center database and developing AI‐based segmentation models, using rectal cancer as a case study. A dataset of 1212 Magnetic Resonance Images (MRIs) from 14 centers was used. The study examined the impact of different image normalization techniques, network hyperparameters, and training set compositions (in terms of size and construction strategies). The findings emphasize the critical role of image normalization in reducing variability and improving performance. Additionally, the study underscores the importance of carefully selecting network structures and loss functions based on the desired outcomes. The potential of clustering approaches to identify representative training subsets, even with limited data sizes, was also evaluated. While no definitive preprocessing pipeline was identified, several networks developed during the study produced promising results on the external validation set. The insights and methodologies presented may help raise awareness and promote more informed decisions when implementing AI systems in medical imaging.

Authors

Institutions

Publication Details

Journal
International Journal of Imaging Systems and Technology
Published
2025-04-03
DOI
https://doi.org/10.1002/ima.70076
Primary Topic
Radiomics and Machine Learning in Medical Imaging
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Rectal Cancer Segmentation: A Methodical Approach for Generalizable Deep Learning in a Multi‐Center Setting

Jovana Panić, Antonio Esposito, Giuditta Chiloiro, Gabriella Balestra et al.
International Journal of Imaging Systems and Technology
Radiomics and Machine Learning in Medical Imaging
article

Rectal Cancer Segmentation: A Methodical Approach for Generalizable Deep Learning in a Multi‐Center Setting

Jovana Panić, Antonio Esposito, Giuditta Chiloiro, Gabriella Balestra, Samanta Rosati, Arianna Defeudis, Vincenza Granata, Michele Avanzo, Giuseppe Roberto D’Agostino, Antonella Petrillo, Angelo Vanzulli, Serena Dell’Aversana, Barbara Alicja Jereczek‐Fossa, Roberta Fusco, Daniele Regge, Huong Elena Tran, Marco Gatti, Luca Sorrentino, E. Menghi, Roberto Sghedoni, Luca Boldrini, C. Fiorino, Valentina Giannini, Marianna Alessandra Gerardi, Martina Mori, Lorenzo Vassallo, Stefano Cirillo
article en

Abstract

ABSTRACT Noninvasive Artificial Intelligence (AI) techniques have shown great potential in assisting clinicians through the analysis of medical images. However, significant challenges remain in integrating these systems into clinical practice due to the variability of medical data across multi‐center databases and the lack of clear implementation guidelines. These issues hinder the ability to achieve robust, reproducible, and statistically significant results. This study thoroughly analyzes several decision‐making steps involved in managing a multi‐center database and developing AI‐based segmentation models, using rectal cancer as a case study. A dataset of 1212 Magnetic Resonance Images (MRIs) from 14 centers was used. The study examined the impact of different image normalization techniques, network hyperparameters, and training set compositions (in terms of size and construction strategies). The findings emphasize the critical role of image normalization in reducing variability and improving performance. Additionally, the study underscores the importance of carefully selecting network structures and loss functions based on the desired outcomes. The potential of clustering approaches to identify representative training subsets, even with limited data sizes, was also evaluated. While no definitive preprocessing pipeline was identified, several networks developed during the study produced promising results on the external validation set. The insights and methodologies presented may help raise awareness and promote more informed decisions when implementing AI systems in medical imaging.

International Journal of Imaging Systems and TechnologyVol. 35(3)
Vita-Salute San Raffaele University (IT), Politecnico di Torino (IT), University of Milan (IT), Agostino Gemelli University Polyclinic (IT), Azienda Ospedaliera Citta' della Salute e della Scienza di Torino (IT), Azienda Sanitaria Unità Locale di Reggio Emilia (IT), Azienda Socio Sanitaria Territoriale Grande Ospedale Metropolitano Niguarda (IT), Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IT), Osaka Health Science University (JP), Ospedale Santa Maria (IT), Centro di Riferimento Oncologico (IT), A. O. Ordine Mauriziano di Torino (IT), Istituto Nazionale Tumori IRCCS "Fondazione G. Pascale" (IT), Candiolo Cancer Institute (IT), Istituti di Ricovero e Cura a Carattere Scientifico (IT), Fondazione IRCCS Istituto Nazionale dei Tumori (IT), IRCCS Humanitas Research Hospital (IT), Istituto di Ricovero e Cura a Carattere Scientifico San Raffaele (IT), University of Turin (IT), European Institute of Oncology (IT), The University of Osaka (JP)
Ministero della Salute, Associazione Italiana per la Ricerca sul Cancro, Alleanza Contro il Cancro
Openalex Percentile: Top 90%
Radiomics and Machine Learning in Medical Imaging
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.