Feasibility of a Deep-Learning Model for Objective Visualization-Quality Assessment During Cholangioscopy-Guided Lithotripsy: A Proof-of-Concept Study.
- 1Department of Gastroenterology and Neurology, Faculty of Medicine Kagawa University Kagawa Japan.
- 2Department of Gastroenterological Surgery, Faculty of Medicine Kagawa University Takamatsu Japan.
Abstract
During peroral cholangioscopy-guided electrohydraulic lithotripsy (EHL), stone fragmentation obscures the visual field (the "snowstorm effect"), and visualization quality has been assessed only subjectively. This proof-of-concept study explored the feasibility of a deep-learning method for objectively quantifying visualization quality during EHL and, as an exploratory application, compared viscous gel immersion with saline irrigation. A ResNet-18 model was trained on expert-annotated frames (5-point scale) from three development cases. For independent validation, 120 frames from 12 previously unseen comparison cases were scored by two blinded endoscopists (inter-reader weighted κ = 0.98). The model showed only moderate agreement with their consensus (weighted κ = 0.60; accuracy 70.8%); because the validation sample was enriched for predicted-usable frames, metrics re-weighted to the natural frame distribution gave sensitivity 29.8% and specificity 93.5%, indicating systematic under-detection of usable views. Agreement did not differ systematically between gel and saline frames. The High-Quality View Rate (HQVR) was the proportion of EHL-phase frames scored Good/Excellent. In a patient-level comparison, median HQVR was 22.7% for gel versus 8.4% for saline (exact Mann-Whitney U, p = 0.15, not significant); in post hoc exploratory analyses, a threshold-independent mean predicted score differed (1.65 vs. 1.35, p = 0.005) and the longest continuous usable interval favored gel (median 24.4 vs. 5.2 s). Mild cholangitis occurred in 1/5 gel and 4/7 saline patients. Training such a model is feasible, but it is not yet a validated measurement tool, and the gel-versus-saline signal is hypothesis-generating only. Larger multi-case training and prospective validation are required. Trial Registration: N/A.
Related
- Breaking the compatibility barrier: cholangioscopy-guided laser lithotripsy during enteroscopy-assisted endoscopic retrograde cholangiopancreatography in Roux-en-Y anatomy.
- Single-Operator Cholangioscopy-Guided Electrohydraulic Lithotripsy for Difficult Common Bile Duct Stones: A Prospective Cohort Study of Efficacy, Safety, and Clinical Outcomes
- Treatment of Calcified Peripheral Artery Disease With the Shockwave Forward Peripheral Intravascular Lithotripsy System: One-Year Outcomes From the FORWARD PAD IDE and Mini S Feasibility Studies
- Current evidence of flexible ureteroscopic laser lithotripsy under local anesthesia: feasibility, safety, and clinical application.
- A Robotic Tele-Laser System for Remote Visualization and Nd:YAG Laser Triggering: A Preclinical Feasibility Study
BibTeX
@article{kamada2026feasibility,
title = {{Feasibility of a Deep-Learning Model for Objective Visualization-Quality Assessment During Cholangioscopy-Guided Lithotripsy: A Proof-of-Concept Study.}},
author = {Hideki Kamada and Kiyoyuki Kobayashi and Hiroki Yamana and Naoki Fujita and Daisuke Namima and Manabu Yamada and Syunsuke Fujita and Hironobu Suto and Keiichi Okano and Hideki Kobara},
journal = {DEN open},
year = {2026},
month = oct,
doi = {10.1002/deo2.70454},
url = {https://europepmc.org/article/MED/42898633},
note = {CC BY},
}
RIS
TY - JOUR
TI - Feasibility of a Deep-Learning Model for Objective Visualization-Quality Assessment During Cholangioscopy-Guided Lithotripsy: A Proof-of-Concept Study.
AU - Hideki Kamada
AU - Kiyoyuki Kobayashi
AU - Hiroki Yamana
AU - Naoki Fujita
AU - Daisuke Namima
AU - Manabu Yamada
AU - Syunsuke Fujita
AU - Hironobu Suto
AU - Keiichi Okano
AU - Hideki Kobara
PY - 2026
DA - 2026/10/18
JO - DEN open
DO - 10.1002/deo2.70454
UR - https://europepmc.org/article/MED/42898633
AB - During peroral cholangioscopy-guided electrohydraulic lithotripsy (EHL), stone fragmentation obscures the visual field (the "snowstorm effect"), and visualization quality has been assessed only subjectively. This proof-of-concept study explored the feasibility of a deep-learning method for objectively quantifying visualization quality during EHL and, as an exploratory application, compared viscous gel immersion with saline irrigation. A ResNet-18 model was trained on expert-annotated frames (5-point scale) from three development cases. For independent validation, 120 frames from 12 previously unseen comparison cases were scored by two blinded endoscopists (inter-reader weighted κ = 0.98). The model showed only moderate agreement with their consensus (weighted κ = 0.60; accuracy 70.8%); because the validation sample was enriched for predicted-usable frames, metrics re-weighted to the natural frame distribution gave sensitivity 29.8% and specificity 93.5%, indicating systematic under-detection of usable views. Agreement did not differ systematically between gel and saline frames. The High-Quality View Rate (HQVR) was the proportion of EHL-phase frames scored Good/Excellent. In a patient-level comparison, median HQVR was 22.7% for gel versus 8.4% for saline (exact Mann-Whitney U, p = 0.15, not significant); in post hoc exploratory analyses, a threshold-independent mean predicted score differed (1.65 vs. 1.35, p = 0.005) and the longest continuous usable interval favored gel (median 24.4 vs. 5.2 s). Mild cholangitis occurred in 1/5 gel and 4/7 saline patients. Training such a model is feasible, but it is not yet a validated measurement tool, and the gel-versus-saline signal is hypothesis-generating only. Larger multi-case training and prospective validation are required. Trial Registration: N/A.
SN - 2692-4609
ER -