Final Diagnosis: Pancreatic adenocarcinoma with involvement
of the portal vein on surgical resection.
1. What is the predictive value of EUS for resectability
of pancreatic adenocarcinoma?
Pancreatic adenocarcinoma is the 5th leading cause for cancer
related death in the U.S and has an extremely poor prognosis with <5% of
patients surviving 5 years after diagnosis. Surgical resection offers the
only chance for cure. Unfortunately, only 5-25% of patients sent for surgery
actually undergo the intended curative resection. Improved preoperative staging
should allow some unresectable patients to avoid unnecessary surgery. However,
the staging procedures must not "over stage" patients so that patients
with resectable disease are not classified as unresectable. Visual
1 describes the TNM classification for pancreatic cancer. The crucial findings
that limit resectability are vessel involvement (portal vein and mesentaric
artery), major organ invasion, and lymph node metastases. If preoperative
staging is T3N0 or less, then surgery is planned.
Initial studies showed that EUS was superior to CT and MR
in pancreatic cancer staging. The studies used surgery findings as the gold
standard. Rosch et al. evaluated 46 patients with pancreatic carcinoma
and 14 with ampullary cancer. EUS accuracy for portal vein involvement was
85%, as compared to nonhelical CT 75%, and transabdominal ultrasound 55%.
Gress et al. evaluated 81 patients with pancreatic cancer and compared
EUS with nonhelical CT. They found a T stage accuracy of 85% for EUS and 70%
for CT.
EUS measurements of the local tumor size is very good. Our
patient was very accurately assessed as having a 3.7 by 3.4 cm tumor (see
EUS photo). This makes the
tumor at least T2. A crucial assessment in EUS is the interface between the
portal vein and the tumor. In most initial studies, tumors that involved the
wall of the portal vein by EUS were thought to be unresectable (T4). However,
some of these tumors can be dissected away from the portal vein by the experienced
surgeon. Further experience has emphasized that EUS unresectability should
mean that the tumor actually extends into the lumen of the portal vein. The
implication is that for some patients "resectability" will be assessed
at surgery. Our patient is a good example. The EUS showed high intensity echos
at the interface between the tumor and the portal vein (see arrow in EUS
photo). However, during the surgery the tumor
could easily be removed from the vein. This is an example of how over staging
or under staging can happen.
Two recent studies have shown the limitations of EUS.
Ahmad et al. conducted a retrospective review of 89
patients evaluated preoperatively with EUS for pancreatic adenocarcinoma.
Patients with distant metastases, or vascular involvement on CT/MR were excluded.
EUS resectability was defined as tumor stage of T3 or less and either absence
of lymph node metastases or metastases confined to the peripancreatic lymph
nodes. Preoperative TNM classification was compared with surgical and histopathological
staging. The overall accuracy of EUS for T and N staging was 69% and 54% respectively.
The overall proportion of tumors deemed resectable by EUS that were surgically
resectable was 46%. The proportion of tumors staged as T4N1, T4N0, T3N1, and
T3N0 by EUS that were found to be resectable was 45%, 37%, 44% and 62% respectively.
So both over staging and under staging was common.
Rosch et al. evaluated 75 patients with pancreatic
adenocarcinoma by EUS to evaluate mesenteric venous invasion. A composite
gold standard was used for comparison including surgical resection and angiography.
EUS staging of T stage was wrong 20-30% of the time. That is, about 25% of
patients T3 by EUS were unresectable by surgery or angiography. Such understaging
is especially common when tumor size is greater than 3 cm. Also, about 25%
of patients T4 by EUS had a curable resection. About 20% of patients staged
N0 by EUS were N1 at surgery. About 50% of patients N1 by EUS were N0 at surgery.
It seems like about 10-15% of patients get a curative resection and EUS-based
T and N staging cannot reliably identify those patients.
2. What are the newer studies available for staging
pancreatic carcinoma and what role does EUS have?
Dual-phase (first scan within 20 seconds of a rapid contrast
injection, second scan after one minute of equilibration of contrast) helical
CT, in which imaging occurs during an arterial/pancreatic and then hepatic
contrast phase enhancement has recently been described as an accurate means
of determining resectability. Legmann et al. evaluated 22 patients
with pancreatic or ampullary cancer. They found that this technique was
just as sensitive as EUS for detection of the mass lesion and also just
as accurate (90%) for determining resectability. Sheridan et al
evaluated dynamic contrast enhanced MR imaging and dual phase contrast
CT in 33 patients preoperatively in assessment for suspected pancreatic
cancer. They found MR to be functioning around the 76% level in determining
resectability. Wiersema et al. suggest that given the high accuracy
of CT/MR in determining locoregional spread and distant metastases, the
role of EUS may be restricted to small tumors not detected by CT/MR. EUS
would also be useful when CT/MR are inconclusive regarding the
extent of locoregional spread and when pancreatic tumors and lymph nodes
are not accessible to percutaneous techniques.
