Polyp size measurement during colonoscopy using a virtual scale: variability and systematic differences

Abstract

BACKGROUND : Accurate polyp size measurement is important for polyp risk stratification and decision-making regarding polypectomy and surveillance. Recently, a virtual scale (VS) function has been developed that allows polyp size measurement through projection of an adaptive VS onto colorectal polyps during real-time endoscopy. We aimed to evaluate the VS in terms of variability and systematic differences. METHODS : We conducted a video-based study with 120 colorectal polyps, measured by eight dedicated colorectal gastroenterologists (experts) and nine gastroenterology residents following endoscopy training (trainees). Three endoscopic measurement methods were compared: (1) visual, (2) snare and (3) VS measurement. We evaluated the method-specific variance (as measure of variability) in polyp size measurements and systematic differences between these methods. RESULTS : Variance in polyp size measurements was significantly lower for VS measurements compared to visual and snare measurements for both experts (0.52 vs. 1.59 and 1.96, p < 0.001) and trainees (0.59 vs. 2.21 and 2.53, p < 0.001). VS measurement resulted in a higher percentage of polyps assigned to the same size category by all endoscopists compared to visual and snare measurements (experts: 69 % vs. 55 % and 59 %; trainees: 67 % vs. 51 % and 47 %) and reduced the maximum difference between individual endoscopists regarding the percentage of polyps assigned to the ≥ 10 mm size category (experts: 1.7 % vs. 10.0 % and 5.0 %; trainees: 2.5 % vs. 6.7 % and 11.7 %). Systematic differences between methods were < 0.5 mm. CONCLUSIONS : Use of the VS leads to lower polyp size measurement variability and more uniform polyp sizing by individual endoscopists compared to visual and snare measurements.

Trial registration: ClinicalTrials.gov NCT05499546.

Fig. 1

Schematic overview of the principle of distance estimation using the virtual scale endoscope. The reflection of the laser beam, as emitted from the endoscope, can be used to determine distance between the endoscope tip and the object (mucosal wall or polyp) the laser is positioned on based on the triangulation method. This information is used to continuously adapt the size of the virtual scale: the size of the virtual scale increases whenever the distance between the endoscope tip and the mucosal wall (or polyp) is shrinking, and decreases as the endoscope tips moves farther away. To enable real-time polyp size measurement, the virtual scale contains markings at 5, 10, and 20 mm.

Fig. 2

Endoscopic polyp size measurements methods. a Visual measurement (without the aid of a tool). b Measurement with the aid of the linear virtual scale. c Measurement with the aid of the circular virtual scale. d Measurement with the aid of a polypectomy snare of known size (maximum width of 9 mm).

Fig. 3

Bland–Altman plots illustrating the differences between polyp size measurements by expert endoscopists using different measurement methods. Within these plots, the polyp size according to two different measurement methods ( x axis) is plotted against the difference in polyp size according to these methods (y axis). Each plot comprises 960 observations, representing measurements of 120 polyps by eight different expert endoscopists. Count for each point within the plot is indicated by the legend on the top of each plot. Dotted blue lines represent mean differences. Dotted red lines represent upper and lower limits of the 95 %CIs. The plots represent the following methods ( y axis): a Visual and snare measurements. b Visual and virtual scale measurements. c Snare and virtual scale measurements.

Fig. 4

Bar plots. a The percentage of polyps assigned to the same category (≤ 5 mm, 6–9 mm, ≥ 10 mm) by all endoscopists using different polyp size measurements methods. b The maximum difference between individual endoscopists regarding the percentage of polyps assigned to the ≥ 10-mm size category.