[Diagnostic usefulness of PET/CT for pancreatic malignancy]

The purpose of this study was to evaluate the diagnostic usefulness of PET/CT for pancreatic malignancy.

We retrospectively analyzed medical records of 115 patients with pathologically diagnosed pancreatic cancer between January 2003 to August 2008 who underwent abdominal CT and PET/CT examination before histological confirmation. CT and PET/CT images were reviewed in single-blinded status and diagnostic ability on primary pancreatic lesion, regional lymph node metastasis, and distant metastasis was evaluated.

99 patients (86%) had malignant diseases including 91 cases of adenocarcinoma, and 16 patients (14%) benign diseases. Only CA 19-9 value and SUV were significantly different between PET/CT positive and negative groups (p=0.001, p<0.001). Sensitivity, specificity and positive predictive values (PPV) of both modality for pancreatic lesion were same (94%, 62%, and 95%, respectively), and negative predictive values (NPV) were 67% on CT and 57% on PET/CT. PET/CT correctly diagnosed 8 cases (6.9%) of falsely diagnosed pancreatic lesion on CT. Nine cases (15.7%) of misdiagnosed lymph node metastasis on CT were correctly diagnosed on PET/CT. But, there was no significant difference in the diagnosis of regional lymph node metastasis. 3 out of 29 cases of distant metastasis, except 2 cases of supraclavicular lymph node metastasis, were additionally diagnosed by PET/CT. But, overall sensitivity of distant metastasis was significantly higher in CT (83% vs 69%, p=0.045).

Although PET/CT provided additional correct diagnoses in many cases, it showed fair diagnostic power for primary pancreatic lesion and lymph node metastasis, and lower sensitivity for distant metastasis. Therefore, PET/CT should be used as an supplementary modality of CT in diagnosing pancreatic malignancy.