OBJECTIVE: The purpose of this study was to compare the accuracy rates of unenhanced sonography, contrast-enhanced sonography, and MRI in displaying the anatomic features of cystic pancreatic masses larger than 1.5 mm in diameter. MATERIALS AND METHODS: Unenhanced and contrast-enhanced sonographic and MRI examinations of 33 patients who underwent resection of a cystic pancreatic mass were retrospectively reviewed. Two radiologists blinded to the final histologic diagnosis reviewed the images, specifically assessing the presence of intralesional mural nodules and septa. Sensitivity, specificity, positive and negative predictive values, and accuracy were calculated on the basis of correlation with surgical findings. Results of unenhanced sonography, contrast-enhanced sonography, and MRI were compared by McNemar test. Correlation of unenhanced and contrast-enhanced sonographic versus pathologic results was established with Spearman's test. Interobserver variability was determined. RESULTS: Contrast-enhanced sonography correctly depicted intralesional septa in 14 of 15 lesions (sensitivity, 93.3%; specificity, 88.8%; positive predictive value, 87.5%; negative predictive value, 94.1%; accuracy, 90.9%) and nodules in six of eight lesions (sensitivity, 75%; specificity, 96%; positive predictive value, 85.7%; negative predictive value, 92.3%; accuracy, 90.9%). MRI correctly depicted intralesional septa in 14 of 15 lesions (sensitivity, 93.3%; specificity, 61.1%; positive predictive value, 66.6%; negative predictive value, 91.6%; accuracy, 75.7%) and nodules in seven of eight lesions (sensitivity, 87.5%; specificity, 80%; positive predictive value, 58.3%; negative predictive value, 95.2%; accuracy, 81.8%). The difference between the diagnostic accuracy of contrast-enhanced sonography and that of MRI was not significant (p = 0.05, McNemar test) in the identification of septa and nodules. The correlation between contrast-enhanced sonographic findings and pathologic results (Rs = 0.93; p < 0.001) was significantly better than that between sonographic and pathologic results (Rs = 0.52; p < 0.0001). Interobserver agreement had a kappa value of 0.86-0.94. CONCLUSION: Contrast-enhanced sonography compares favorably with MRI in displaying the anatomic features of cystic pancreatic masses seen on transabdominal sonography.
Comparison of contrast-enhanced sonography and MRI in displaying anatomic features of cystic pancreatic masses
D'ONOFRIO, Mirko;FACCIOLI, Niccolò;MALAGO', Roberto;FALCONI, Massimo;POZZI MUCELLI, Roberto
2007-01-01
Abstract
OBJECTIVE: The purpose of this study was to compare the accuracy rates of unenhanced sonography, contrast-enhanced sonography, and MRI in displaying the anatomic features of cystic pancreatic masses larger than 1.5 mm in diameter. MATERIALS AND METHODS: Unenhanced and contrast-enhanced sonographic and MRI examinations of 33 patients who underwent resection of a cystic pancreatic mass were retrospectively reviewed. Two radiologists blinded to the final histologic diagnosis reviewed the images, specifically assessing the presence of intralesional mural nodules and septa. Sensitivity, specificity, positive and negative predictive values, and accuracy were calculated on the basis of correlation with surgical findings. Results of unenhanced sonography, contrast-enhanced sonography, and MRI were compared by McNemar test. Correlation of unenhanced and contrast-enhanced sonographic versus pathologic results was established with Spearman's test. Interobserver variability was determined. RESULTS: Contrast-enhanced sonography correctly depicted intralesional septa in 14 of 15 lesions (sensitivity, 93.3%; specificity, 88.8%; positive predictive value, 87.5%; negative predictive value, 94.1%; accuracy, 90.9%) and nodules in six of eight lesions (sensitivity, 75%; specificity, 96%; positive predictive value, 85.7%; negative predictive value, 92.3%; accuracy, 90.9%). MRI correctly depicted intralesional septa in 14 of 15 lesions (sensitivity, 93.3%; specificity, 61.1%; positive predictive value, 66.6%; negative predictive value, 91.6%; accuracy, 75.7%) and nodules in seven of eight lesions (sensitivity, 87.5%; specificity, 80%; positive predictive value, 58.3%; negative predictive value, 95.2%; accuracy, 81.8%). The difference between the diagnostic accuracy of contrast-enhanced sonography and that of MRI was not significant (p = 0.05, McNemar test) in the identification of septa and nodules. The correlation between contrast-enhanced sonographic findings and pathologic results (Rs = 0.93; p < 0.001) was significantly better than that between sonographic and pathologic results (Rs = 0.52; p < 0.0001). Interobserver agreement had a kappa value of 0.86-0.94. CONCLUSION: Contrast-enhanced sonography compares favorably with MRI in displaying the anatomic features of cystic pancreatic masses seen on transabdominal sonography.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
