TY - JOUR
T1 - Clinical validity of flexible spectral imaging color enhancement (FICE) in small-bowel capsule endoscopy: a systematic review and meta-analysis
AU - Yung, Diana
AU - Boal Carvalho, Pedro
AU - Giannakou, Andry
AU - Kopylov, Uri
AU - Rosa, Bruno
AU - Rondonotti, Emmanuele
AU - Toth, Ervin
AU - Plevris, John
AU - Koulaouzidis, Anastasios
PY - 2017/1/25
Y1 - 2017/1/25
N2 - Patients and methods A comprehensive literature search was conducted. We measured pooled rate of lesion visualization improvement and improvement in lesion detection comparing FICE settings 1 – 3 and WLE, for angioectasias and ulcers/erosions. Pooled results were derived using the random-effects model because of high heterogeneity as measured by I 2. Repeated-measures analysis of variance (ANOVA) was used to measure differences in lesion detection between WLE and the three FICE modes.
Results 13 studies were analyzed. All studies used the PillCam SB 1 and/or SB 2 devices. Most used experienced readers. Improvement in delineation had been investigated in 4 studies; in the 3 studies entered into the meta-analysis, using FICE setting 1, 89 % of angioectasias and 45 % of ulcer/erosions were considered to show improved delineation. For FICE settings 2 and 3, small proportions of images showed improved delineation. Heterogeneity of studies was high with I 2 > 90 % in 4/6 analyses. Lesion detection had been investigated in 10 studies; meta-analysis included 5 studies. Lesion detection did not differ significantly between any of the FICE modes and WLE.
Conclusions Overall, the use of the three FICE modes did not significantly improve delineation or detection rate in SBCE. In pigmented lesions, FICE setting 1 performed better in lesion delineation and detection.
AB - Patients and methods A comprehensive literature search was conducted. We measured pooled rate of lesion visualization improvement and improvement in lesion detection comparing FICE settings 1 – 3 and WLE, for angioectasias and ulcers/erosions. Pooled results were derived using the random-effects model because of high heterogeneity as measured by I 2. Repeated-measures analysis of variance (ANOVA) was used to measure differences in lesion detection between WLE and the three FICE modes.
Results 13 studies were analyzed. All studies used the PillCam SB 1 and/or SB 2 devices. Most used experienced readers. Improvement in delineation had been investigated in 4 studies; in the 3 studies entered into the meta-analysis, using FICE setting 1, 89 % of angioectasias and 45 % of ulcer/erosions were considered to show improved delineation. For FICE settings 2 and 3, small proportions of images showed improved delineation. Heterogeneity of studies was high with I 2 > 90 % in 4/6 analyses. Lesion detection had been investigated in 10 studies; meta-analysis included 5 studies. Lesion detection did not differ significantly between any of the FICE modes and WLE.
Conclusions Overall, the use of the three FICE modes did not significantly improve delineation or detection rate in SBCE. In pigmented lesions, FICE setting 1 performed better in lesion delineation and detection.
U2 - 10.1055/s-0042-122015
DO - 10.1055/s-0042-122015
M3 - Article
SN - 0013-726X
VL - 49
SP - 258
EP - 269
JO - Endoscopy
JF - Endoscopy
IS - 03
ER -