Evaluation of a new composite score combining SPICE and protrusion angle scores to distinguish submucosal lesions from innocent bulges

Rev Esp Enferm Dig. 2022 Mar;114(3):151-155. doi: 10.17235/reed.2021.8080/2021.

Abstract

Introduction and aim: in capsule endoscopy (CE), small bowel subepithelial lesions (SBSL) are difficult to distinguish from innocent mucosal protrusions. The SPICE score (smooth, protruding lesions index on CE) and a score that assesses the SBSL protrusion angle were developed. The aim of the study was to determine if a composite score is superior to the proposed models.

Methods: all CE between 01/2010 and 12/2020 were included in the study if a smooth, round protruding lesion was identified. Both scores and a composite score (SPICE > 2 and angle < 90°) were calculated after video review. Mucosal protrusions were defined as SBSL if they had a histological/imaging diagnosis and innocent protrusions if otherwise. All patients without at least one appointment and an additional diagnostic exam after CE were excluded.

Results: a total of 34 CE were included; 64.7 % were males, aged 65.4 ± 14.7 years. The most common indication for CE was anemia (52.9 %). SBSL was identified in 17 cases, with lipomas (14.7 %) being the most frequent diagnosis. Both the SPICE (AUROC 0.90, p < 0.001) and protrusion angle scores (AUROC 0.74, p = 0.019) accurately distinguished SBSL from innocent protrusions. Applying a 90° cut-off, the protrusion angle had a sensitivity of 52.9 % and specificity of 88.2 %. Applying a cut-off of > 2 points, the SPICE score has a sensitivity of 64.7 % and specificity of 94.2 %. The composite score had a sensitivity, specificity, positive and negative predictive value of 47.0 %, 100 %, 100 % and 65.4 %.

Conclusion: we propose that additional follow-up investigation should always be undertaken in cases where both a SPICE > 2 and angle of < 90° are obtained, as the likelihood of SBSL is high.

MeSH terms

  • Capsule Endoscopy* / methods
  • Female
  • Humans
  • Intestine, Small / pathology
  • Male
  • Predictive Value of Tests