Daniel C Baumgart1, Hans P Müller, Ulrike Grittner, Diana Metzke, Andreas Fischer, Olaf Guckelberger, Andreas Pascher, Ingolf Sack, Michael Vieth, Birgit Rudolph. 1. From the Division of Gastroenterology and Hepatology, Department of Medicine (D.C.B., H.P.M., D.M., A.F.), Department of Biostatistics and Clinical Epidemiology (U.G.), Department of General, Visceral, and Transplant Surgery (O.G., A.P.), Department of Experimental Radiology (I.S.), and Institute of Pathology (B.R.), Charité Medical School, Virchow Hospital, Humboldt-University of Berlin, Augustenburger Platz 1, 13353 Berlin, Germany; and Institute of Pathology, Klinikum Bayreuth, Germany (M.V.).
Abstract
PURPOSE: To assess whether ultrasonography (US)-based real-time elastography (RTE) can be used to detect gut fibrosis. MATERIALS AND METHODS: In this institutional review board-approved, prospective, proof-of-concept study, unaffected and affected gut segments in 10 patients with Crohn disease (four women, six men; median age, 49 years) were examined pre-, intra-, and postoperatively with US, including RTE to assess strain. Disease activity was scored by using the Limberg index on the basis of (a) bowel wall thickness and (b) size and extent of Doppler signal. After surgical resection, strain of full gut wall segments was measured with direct tensiometry. Gut wall layers, fibrosis, and collagen content were quantified histologically. Aggregated data per patient, disease status, and available measurements were assessed with mixed-effects models. RESULTS: Unaffected versus affected gut segments yielded higher RTE (mean ± standard deviation, 169.0 ± 27.9 vs 43.0 ± 25.9, respectively) and tensiometry (mean, 77.1 ± 21.4 vs 13.3 ± 11.2, respectively) values used to assess strain (both P < .001). There was good correlation between pre-, intra-, and postoperative RTE values of unaffected (intraclass correlation coefficient, 0.572) and affected (intraclass correlation coefficient, 0.830) segments. RTE was not associated with pre- or intraoperative Limberg scores (median, 1 vs 2; P = .255 and .382, respectively). Affected internal (median, 2011 vs 1363 μm; P = .011) and external (median, 929 vs 632 μm; P = .013) muscularis propria, serosa (median, 245 vs 64 μm; P = .019), and muscularis mucosae (median, 451 vs 80 μm; P = .031) were wider than unaffected segments. Width differences of internal muscularis propria and mucularis mucosae were associated with RTE-assessed strain (P = .044 and .012, respectively) and tensiometry-assessed strain (P = .006 and .014, respectively). Masson trichrome (median, 4 vs 0; P < .001) and elastica-van Gieson (median, 805 346 μm(2) vs 410 649 μm(2); P < .001) stains and western blotting (median, 2.01 vs 0.87; P = .009) demonstrated a higher collagen content in affected versus unaffected segments and were associated with RTE-assessed strain (both P < .001) and tensiometry-assessed strain (P < .001 and 0.025, respectively). CONCLUSION: RTE can be used to detect fibrosis in human Crohn disease. Online supplemental material is available for this article. RSNA, 2015
PURPOSE: To assess whether ultrasonography (US)-based real-time elastography (RTE) can be used to detect gut fibrosis. MATERIALS AND METHODS: In this institutional review board-approved, prospective, proof-of-concept study, unaffected and affected gut segments in 10 patients with Crohn disease (four women, six men; median age, 49 years) were examined pre-, intra-, and postoperatively with US, including RTE to assess strain. Disease activity was scored by using the Limberg index on the basis of (a) bowel wall thickness and (b) size and extent of Doppler signal. After surgical resection, strain of full gut wall segments was measured with direct tensiometry. Gut wall layers, fibrosis, and collagen content were quantified histologically. Aggregated data per patient, disease status, and available measurements were assessed with mixed-effects models. RESULTS: Unaffected versus affected gut segments yielded higher RTE (mean ± standard deviation, 169.0 ± 27.9 vs 43.0 ± 25.9, respectively) and tensiometry (mean, 77.1 ± 21.4 vs 13.3 ± 11.2, respectively) values used to assess strain (both P < .001). There was good correlation between pre-, intra-, and postoperative RTE values of unaffected (intraclass correlation coefficient, 0.572) and affected (intraclass correlation coefficient, 0.830) segments. RTE was not associated with pre- or intraoperative Limberg scores (median, 1 vs 2; P = .255 and .382, respectively). Affected internal (median, 2011 vs 1363 μm; P = .011) and external (median, 929 vs 632 μm; P = .013) muscularis propria, serosa (median, 245 vs 64 μm; P = .019), and muscularis mucosae (median, 451 vs 80 μm; P = .031) were wider than unaffected segments. Width differences of internal muscularis propria and mucularis mucosae were associated with RTE-assessed strain (P = .044 and .012, respectively) and tensiometry-assessed strain (P = .006 and .014, respectively). Masson trichrome (median, 4 vs 0; P < .001) and elastica-van Gieson (median, 805 346 μm(2) vs 410 649 μm(2); P < .001) stains and western blotting (median, 2.01 vs 0.87; P = .009) demonstrated a higher collagen content in affected versus unaffected segments and were associated with RTE-assessed strain (both P < .001) and tensiometry-assessed strain (P < .001 and 0.025, respectively). CONCLUSION: RTE can be used to detect fibrosis in humanCrohn disease. Online supplemental material is available for this article. RSNA, 2015