Diagnostic and therapeutic dilemmas in periampullary lesions.

ABSTRACT

BACKGROUND/AIMS: Periampullary lesions often present diagnostic and therapeutic dilemmas. This study is to clarify the justification of pancreaticoduodenectomy for the resectable periampullary lesion without histological confirmation of malignancy. METHODOLOGY: Clinical data for periampullary lesions with presumed malignancy were retrieved from our prospectively-collected computer database. The surgical risks and test performance characteristics in diagnosis were determined. RESULTS: There were 636 patients undergoing pancreaticoduodenectomy, including 572 with malignancy and 64 (10.1% false positive rate) with benign lesions. No resection was attempted for 32 patients, but 8 (25% false negative rate) eventually turned out to be malignant. Our data showed a sensitivity of 98.6% (572/580), a specificity of 27% (24/88) and an accuracy of 89.2% (596/668) in detecting periampullary malignancy. The surgical risks after pancreaticoduodenectomy were significantly lower in the benign group, with 28.1% morbidity (vs. 43.7% in the malignant group), no pancreatic leakage (vs. 11.5% in malignant group) and no surgical mortality (vs. 7.3% in the malignant group). CONCLUSIONS: Pancreaticoduodenectomy is justified for a periampullary lesion without histological confirmation whenever malignancy is suspected. Moreover, a nihilistic approach could be associated with a significant false negative rate (25%) if left unresected and might preclude a patient with periampullary malignancy from cure.