Format

Send to

Choose Destination
See comment in PubMed Commons below
Am Surg. 1988 Apr;54(4):221-6.

Efficacy of computerized tomography in the preoperative staging of pancreatic carcinoma.

Author information

1
Department of Surgery, Vanderbilt University Medical Center, Nashville, TN 37232.

Abstract

The increasing use of nonoperative methods for the diagnosis and palliative treatment of pancreatic cancer has placed greater emphasis on computerized tomography (CT) in staging of this malignancy. The present study was done to review our experience with CT in staging pancreatic cancer, specifically its efficacy in predicting lesions amenable to curative versus noncurative surgical treatment. Sixty six consecutive patients with pancreatic cancer who were considered potential candidates for surgical resection and who had preoperative CT scans over the 4-year period from 1982 through 1986 were studied. Preoperative CT scans were reviewed by a radiologist without knowledge of patients' surgical management. CT criteria for unresectable disease include hepatic and distant metastasis as well as evidence of locally advanced disease, including peripancreatic fascial extension, extension of tumor to locally contiguous structures, vascular encasement/invasion, and local lymphadenopathy. CT predicted resectability with a sensitivity of 75 per cent and a positive predictive value of 38 per cent. Unresectability was predicted with a sensitivity of 72 per cent and positive predictive value of 93 per cent. CT incorrectly predicted unresectable disease in three patients who had a curative resection. CT was most reliable when it predicted unresectability due to the presence of hepatic and/or distant metastasis. CT predicted unresectability with least sensitivity using criteria for locally advanced disease. Therapeutic decisions for nonoperative management of patients with pancreatic cancer based upon CT predictions of unresectable disease, especially predictions of unresectability solely on the basis of locally advanced disease, can not be recommended at this time.

PMID:
2451460
[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Loading ...
    Support Center