Research summary, original research, 2026

Establishing discard criteria for lead aprons using deep learning-based quantification of defect area on X-ray fluoroscopic video

Hanada K, Suzuki H, Takahashi M, Fuse H, Kimura M, Sasaki K, et al.. Radiological physics and technology. 2026

Read the original paper

In plain language

In a phantom study, radiation passing through 0.25 mmPb lead aprons rose linearly with defect size, and an automated video tool helped set a discard threshold of about 2.87 cm2 for slit-like defects under endoscopic retrograde cholangiopancreatography (ERCP)-like conditions.

At a glance

Type of study
Original research, Phantom or simulation study
Who was studied
Phantom or simulation (no people)
Size
n = 18 lead aprons
Summary based on
the abstract

What it found

  • Using a 5 mSv per 3 months dose limit and a baseline transmitted dose of 2.38 mSv, the discard threshold for slit-like apron defects was 2.87 cm2.

    Key figures: 2.87 cm2

  • Transmitted dose through the apron increased linearly with defect area under all tested conditions in an ERCP-mimicking phantom setup.

  • A fine-tuned YOLOv8 segmentation model detected apron defects on fluoroscopic video with modest accuracy and processed each video in 15 seconds.

    Key figures: AP@0.5 of 0.457; 15 s processing time

  • Slit widths of 1.0, 2.0 and 3.5 mm behaved similarly in dose terms and were pooled for modelling.

    Key figures: 1.0, 2.0, and 3.5 mm

Limits to keep in mind

Phantom dosimetry with artificial slit defects under ERCP-mimicking conditions; only 18 aprons of one lead equivalence, and the detection model's precision was modest (AP@0.5 0.457). The threshold depends on assumed baseline dose and regulatory limit.

How this summary was made

ORSIF prepared this summary with AI assistance from the abstract of the paper. A program then checked that the passage supporting each finding appears word for word in that source. A person has not yet reviewed each summary, so read the original before relying on a number. ORSIF summarizes research and does not endorse any product or company. Report a problem with this summary.

Prepared November 3, 2026.