Radiation Scatter at Fluoroscopy-Guided Endoscopic Procedures: Building Toward Further Optimization

Abstract

INTRODUCTION:

Exposure to ionizing radiation by endoscopy personnel during fluoroscopy-guided procedures remains a health hazard. We aimed to evaluate a novel lead flap-enhanced artificial intelligence-enabled fluoroscopy (LE-AIF) system to minimize radiation scatter.

METHODS:

We conducted prospective study of 218 patients who underwent fluoroscopy-guided endoscopic procedures using conventional AIF or LE-AIF systems. The primary outcome measure was comparison of radiation scatter to endoscopists.

RESULTS:

Radiation scatter to endoscopists and anesthesia personnel were significantly lower using LE-AIF. On linear regression analysis, conventional AIF and fluoroscopy duration were associated with radiation scatter.

DISCUSSION:

LE-AIF significantly reduced radiation scatter at endoscopic procedures.

INTRODUCTION

Although fluoroscopy is integral to many basic and advanced gastrointestinal endoscopic procedures, it exposes patients and personnel to ionizing radiation. High-dose radiation is associated with genomic instability (1,2), and even low-dose exposure from diagnostic imaging may increase cancer mortality (3). Lifetime risk, measured in Sieverts, correlates with effective dose received. Radiation exposure therefore represents a health hazard in the endoscopy suite, warranting adherence to principle of keeping doses as low as reasonably achievable. Primary sources of exposure are incident beam affecting the patient and scatter radiation emanating from the patient, affecting surrounding personnel. Beyond protective wear and minimizing fluoroscopy time, additional strategies have been developed to minimize radiation exposure. In a previous study, we have shown that use of under and over-table scatter shields combined with a motorized variable source-to-image distance reduced scatter radiation six-fold compared with conventional C-arm systems (4). Further integration of artificial intelligence-enabled fluoroscopy (AIF), using ultrafast collimation to confine radiation to region of interest, significantly reduced both patient dose and scatter exposure to endoscopy personnel (5).

Standard fluoroscopic configurations for endoscopic procedures in conventional AIF system (FluoroShield, Omega Medical Imaging, Sanford, Florida) incorporate lead flaps suspended from the procedure table and detector panel to reduce patient-generated scatter radiation. In benchtop dosimetric modeling, we observed localized increase in radiation exposure at table height corresponding to the horizontal gap between these shields, suggesting a residual scatter pathway. To further reduce radiation scatter, we recommended supplementation of AIF system with 2 additional lead flaps: 1 positioned along physician side of the table and the other at head of the bed (anesthesia location) to specifically shield this gap (Figure 1).

Figure 1.

(a) Conventional artificial intelligence-enabled fluoroscopy system (b) lead flap-enhanced artificial intelligence-enabled fluoroscopy system. The lead flaps are outlined in yellow.

METHODS

We conducted a prospective study at Orlando Health Digestive Health Institute (Orlando, Florida), to evaluate effectiveness of proposed technical modifications. Consecutive patients aged 18 years and older requiring image-guided endoscopic procedures were alternately assigned to 1 of 2 fluoroscopy systems: conventional AIF (C-AIF) or lead flap-enhanced AIF (LE-AIF). Indications included endoscopic retrograde cholangiopancreatography, endoscopic ultrasound-guided interventions, and feeding tube or luminal stent placement. Pregnant patients and children were excluded. Fluoroscopy tables were positioned at waist height. All procedures were performed using high-reduction mode with pulsed fluoroscopy at 12.5 frames/second operated by 3 certified radiology technicians. Collimation was used for all procedures. Scatter radiation was measured using Geiger-Mueller Counter Dosimeters (RadFlash, Polimaster, Sterling, Virginia) positioned at chest level on endoscopist and anesthesia personnel. Chest level was selected based on previous data demonstrating maximal exposure at this height (4). Dosimeter readings were then tabulated at study completion. Primary outcome measure was difference in scatter radiation to endoscopists between the 2 fluoroscopy systems. All patients provided informed consent for endoscopic procedures. Because study posed minimal risk and involved no deviation from standard care, requirement for additional study-specific consent was waived by Orlando Health Institutional Review Board. Study was investigator-initiated and received no external funding.

Statistical analysis

Two-sided sample size calculation was performed to detect difference of 5 µSv in mean radiation scatter to endoscopists (assuming 10 µSv for LE-AIF and 15 µSv for C-AIF), with standard deviation of 10 for both groups, based on pilot system evaluation data. Power calculation performed at 95% power and 2-sided α = 0.05 resulted in sample size estimation at 105 patients for each cohort (PASS 15 Software, NCSS LLC, Kaysville, Utah).

Categorical data were summarized as frequencies with percentages and compared using the χ² or Fisher exact test. Continuous data were summarized as means with standard deviation and medians with interquartile range and compared using Wilcoxon rank-sum test. Multivariable linear regression and backward selection stepwise multivariable linear regression analyses (α = 0.05) were performed to identify factors associated with scatter radiation effect to endoscopists and anesthesia personnel. Clinically relevant patient demographics and procedure details were included as predictor variables. Two-sided P-values were reported for all outcome measures and statistical significance was determined at P < 0.05 (Stata 17, StataCorp LP, College Station, Texas).

RESULTS

Total of 218 patients were evaluated between December 3, 2025, and February 3, 2026, with alternate allocation to C-AIF (n = 113) or LE-AIF (n = 105) systems. Except for procedure type (more endoscopic retrograde cholangiopancreatography in LE-AIF), there was no difference in patient characteristics, procedure duration, fluoroscopy time, and patient radiation exposure (Table 1). No intraprocedural adverse events occurred in either cohort. Radiation scatter to endoscopists was reduced by 81% with LE-AIF compared with C-AIF, as measured by deep dose equivalent dosimetry (median 2.6 µSv [IQR 1.3–6.7] vs 13.8 µSv [IQR 4.0–31.4], P < 0.001). On multivariable linear regression analysis adjusting for patient and procedure details, use of LE-AIF system and fluoroscopy duration were the only independent factor associated with radiation scatter to endoscopists and anesthesia personnel (Table 2).

Table 1.

Patient characteristics, procedure details, and radiation exposure readings

Table 2.

Multivariable linear regression analysis and backward selection stepwise multivariable linear regression analysis to identify factors associated with radiation scatter effect to endoscopists and anesthesia personnel

DISCUSSION

Radiation mitigation in endoscopy has traditionally relied on operator-dependent strategies-heavy protective garments, deliberate reduction in fluoroscopy time, and physical positioning maneuvers-each of which imposes ergonomic strain or risks compromising procedural efficiency. These approaches, while necessary, are inherently reactive. There is a clear need for technology that reduces radiation exposure passively, seamlessly, and independent of operator behavior. Although AIF represented a meaningful advance, the present technical refinement further attenuated scatter radiation to near negligible levels, without affecting ergonomics, procedural performance or anesthesia access to patients. In our model, cumulative exposure was only 0.81 mSv for 105 procedures. Using median radiation scatter as reference, 7,700 fluoroscopy-guided procedures would be required to approach the recommended annual occupational limit of 20 mSv (6). Such procedural volumes per individual endoscopist are highly improbable, suggesting that radiation exposure under this configuration may become functionally inconsequential for most operators. This raises an important question: should radiation protection in endoscopy continue to rely primarily on operator burden, or should it be engineered out of the system? Limitation of the current study is that evaluation was conducted at a single institution, using fluoroscopy system from 1 manufacturer, and therefore requires validation by other investigators.

Building on these findings, next-generation fluoroscopic platform integrating AIF, optimized shielding architecture, and endoscopy-specific enhancements are currently under development. A randomized trial evaluating its clinical performance and radiation metrics is forthcoming. If validated, this paradigm shift-from behavioral mitigation to system-level elimination-has the potential to redefine radiation safety standards in fluoroscopy-guided endoscopic practice.

CONFLICTS OF INTEREST

Guarantor of the article: Shyam Varadarajulu, MD.

Specific author contributions: J.Y.B.: study design, endoscopist performing procedures in the trial, acquisition of data, analysis and interpretation of data, statistical analysis, drafting of manuscript, critical revision of manuscript. S.V.: study concept and design, endoscopist performing procedures in the trial, interpretation of data, drafting of manuscript, critical revision of manuscript. R.H.: study design, critical revision of manuscript.

Financial support: None to report.

Potential competing interests: None to report, however see below for disclosures. J.Y.B.: consultant for Olympus America Inc. and Boston Scientific Corporation. S.V.: consultant for Boston Scientific Corp., Olympus America Inc. R.H.: consultant for Boston Scientific Corp., Olympus America Inc.

ABBREVIATIONS:

AIF

artificial intelligence-enabled fluoroscopy

C-AIF

conventional artificial intelligence-enabled fluoroscopy

LE-AIF

lead flap-enhanced artificial intelligence-enabled fluoroscopy