All potential cures for pancreatic adenocarcinoma (PA) include surgical resection.  Endoscopic ultrasound (EUS) has been identified as potentially important due to its potential contribution to analyses of 1) resectability and 2) histology.  This study utilized a single-center prospective database to assess the utility of EUS.

Patients undergoing evaluation for potentially resectable pancreatic adenocarcinoma at a single center, 2002-2009, were identified.  Demographics and potentially important covariates were analyzed.  Outcome measures were identified a priori and included cytology from ERCP or EUS-FNA, as well as resectability by CT- or EUS evidence of vascular involvement.  Sensitivities and specificities were calculated for each diagnostic modality.

71 patients were identified.  22 had ERCP brushings, and 65 had EUS-FNA, with 12 and 55, respectively, identified as suspicious for malignancy or malignant.  All had EUS performed, and 65 also underwent CT, of whom 54 and 58, respectively, had resectability matching post-hoc determined resectability status.  Table 1 shows the calculated sensitivities/specificities.

Tissue diagnostic sensitivity is substantially higher by EUS-FNA than ERCP.  Combined use of both tests (either one positive) improves sensitivity.  CT and EUS are both sensitive tests for determining resectability (most patients appropriately considered for surgery).  Hence, either modality alone may be utilized.  Sensitivity/specificity are essentially unchanged by combining modalities.  EUS had substantially lower specificity than CT, overestimating resectability.  Therefore, for patients with considerable comorbidities, for whom the risks of an unsuccessful resection attempt might be unacceptably high, CT should be used for clinical staging.

Tissue Diagnosis

Resectability

ERCP

positive

EUS-FNA

positive

ERCP

OR

EUS-FNA

(either positive)

CT

positive

EUS

positive

CT

OR

EUS

(either positive)

CT

AND

EUS

(both required to be positive)

Sensitivity

54.5%

84.6%

88.2%

98.0%

92.6%

100.0%

92.0%

Specificity

---*

---*

---*

60.0%

23.5%

20.0%

60.0%

Table 1:

Sensitivity and specificity of EUS for tissue diagnosis and determination of resectability, compared to and in combination with ERCP and/or CT scan.

(*Note:  Because the cohort consists of confirmed pancreatic adenocarcinoma, by definition the number of true negatives and false positives cannot be determined for ERCP or EUS-FNA in this study; hence specificity was not calculated for the tissue diagnostic tests.)