Background: EUS is becoming a standard procedure for preoperative staging of pancreatic cancer. Chronic pancreatitis can sometimes make the interpretation more difficult, and there is a perception that staging accuracy is decreased in the presence of chronic pancreatitis. Methods: The EUS database from 1993 to 1997 was reviewed and all patients who had surgery for suspected pancreatic carcinoma were identified. All patients with complete T and N staging were included. Two groups were analyzed: those with pancreatic cancer (Pea) alone and those with pancreatic cancer and chronic pancreatitis (Pea + chpan) on surgical pathology. Preoperative T stage accuracy, sensitivity and specificity were calculated for each stage and receiver operator characteristic (ROC) curves were generated for both groups. Results: A total of 276 patients were identified with pancreatic carcinoma. There were 84 patients with complete surgical T & N staging. There were 31 patients with Pea alone; 58% were men and the mean age was 64. There were 53 patients with Pea + chpan; 51% were men and the mean age was 62. Table 1 below stratifies EUS test characteristics by T stage. The ROC curve area for Pea patients was .66 while that for Pca + chpan is .64 (P-value = 0.4 for the diff. between curves). Pea Pca+chpan EUS T1 sensitivity 0.25 0.43 specificity 1.00 1.00 accuracy 0.90 0.85 EUS T2 sensitivity 0.89 0.73 specificity 0.77 0.68 accuracy 0.84 0.70 EUS T3 sensitivity 0.89 0.82 specificity 0.86 0.81 accuracy 0.87 0.81 Conclusions: 1.) EUS has fair discrimination in T staging of Pea which may be due to our selection criteria. 2.) The test characteristics in pts. with Pca + chpan are no different than those with Pca. 3.) Further study of EUS test characteristics in this setting requires a consecutive cohort of patients in whom surgical decisions are made independent of EUS.
|Original language||English (US)|
|State||Published - Dec 1 1998|
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging