Diagnostic performance of CT with Valsalva maneuver for the diagnosis and characterization of inguinal hernias

Abstract

Abstract Purpose Inguinal hernias are mainly diagnosed clinically, but imaging can aid in equivocal cases or for treatment planning. The purpose of this study was to evaluate the diagnostic performance of CT with Valsalva maneuver for the diagnosis and characterization of inguinal hernias. Methods This single-center retrospective study reviewed all consecutive Valsalva-CT studies between 2018 and 2019. A composite clinical reference standard including surgery was used. Three blinded, independent readers (readers 1–3) reviewed the CT images and scored the presence and type of inguinal hernia. A fourth reader measured hernia size. Interreader agreement was quantified with Krippendorff’s α coefficients. Sensitivity, specificity, and accuracy of Valsalva-CT for the detection of inguinal hernias was computed for each reader. Results The final study population included 351 patients (99 women) with median age 52.2 years (interquartile range (IQR), 47.2, 68.9). A total of 381 inguinal hernias were present in 221 patients. Sensitivity, specificity, and accuracy were 85.8%, 98.1%, and 91.5% for reader 1, 72.7%, 92.5%, and 81.8% for reader 2, and 68.2%, 96.3%, and 81.1% for reader 3. Hernia neck size was significantly larger in cases correctly detected by all three readers (19.0 mm, IQR 13, 25), compared to those missed by all readers (7.0 mm, IQR, 5, 9; p < 0.001). Interreader agreement was substantial (α = 0.723) for the diagnosis of hernia and moderate (α = 0.522) for the type of hernia. Conclusion Valsalva-CT shows very high specificity and high accuracy for the diagnosis of inguinal hernia. Sensitivity is only moderate which is associated with missed smaller hernias.