Objective: Prospective comparison study of helical computed tomography (HCT) and magnetic resonance imaging (MRI, 1.5-tesla field strength) with regard to their role in planning surgical therapy in a defined, homogeneous patient collective with tumors of the pancreatic head.
Patients and methods: Sixty-one patients with clinical/ultrasound evidence of tumor were examined by HCT and MRI. Findings were compared with surgical anatomy and histology using a data documentation sheet. Factors evaluated included the malignancy of the tumors and the probability of correct diagnosis in relation to tumor size, as well as the occurrence of lymph node and distant metastases, arterial and portal-venous infiltration, extrapancreatic infiltration, determination of resectability and tumor staging (TNM classification). These factors were compared on the basis of their respective sensitivity, specificity, negative and positive predictive value and accuracy.
Results: Malignancy was correctly diagnosed by HCT and MRI in 82% and 77%, respectively, with a corresponding sensitivity for malignancy of 86% and 68%. Tumors with diameters > 2 cm were detected in 100% of cases by HCT and in 88% by MRI, while tumors with diameters < 2 cm were correctly diagnosed by HCT and MRI in 60% and 40% of cases. Respective rates of correct diagnosis by HCT and MRI were 80% and 66% for extrapancreatic infiltration; 76% and 79% for pancreatic cysts; 87% and 87% for calcifications; and 77% and 92% for atrophy. The diagnostic accuracy for vascular infiltration by HCT and MRI stood at 59% and 77% with sensitivities of 88% and 75%, respectively. Local resectability was correctly determined by HCT in 71%, by MRI in 62% of cases. Lymph node metastases were correctly identified by HCT and MRI with an accuracy of 76% and 72%, while distant metastases were correctly diagnosed by HCT and MRI with an accuracy of 93% and 93%. Correct T stage was identified by HCT in 66%, by MRI in 60%. HCT tended to assign patients to a higher T stage than later determined by surgery and/or histology.
Conclusion: Our data do not indicate a clear superiority for either HCT or MRI in the diagnosis of pancreatic malignancies.