Comparison of Endoscopic Ultrasonography and Multidetector Computed Tomography for Detecting and Staging Pancreatic Cancer

To the Editor:

We congratulate Dewitt et al. for, "Comparison of Endoscopic Ultrasonography and Multidetector Computed Tomography for Detecting and Staging Pancreatic Cancer" (1). We highlight one inaccuracy and indicate limitations that temper acceptance of their conclusions into current patient care.

The authors incorrectly refer to the University of Michigan study (2) comparing EUS with helical CT as including patients with distant metastatic disease. Mertz et al. (3) did include 9 patients with metastatic disease in their study.

Our concerns relate to the authors' primary outcome: detection of unresectability. Tumors were deemed unresectable by distant metastatic disease or invasion of mesenteric vessels. Detection of distant metastases is best accomplished with CT, while both CT and EUS are useful for defining vascular invasion. This biases the analysis in favor of CT.

The Michigan study (2) specifically focused on vascular invasion. Dewitt et al. reported sensitivity for detecting T4 disease (EUS 88% vs. CT 71%). Presumably most had vascular invasion, but it is not specifically stated. Furthermore, it seems likely that at least 9 (6 liver and 3 distant nodal metastases detected intraoperatively) did not have complete dissection. This weakens the analysis regarding the ability of the tests to detect vascular invasion.

The protocol of performing EUS with fine needle aspiration (FNA) in all patients prior to CT is atypical. It leads to the unnecessary use of EUS in patients with distant disease and may lead to inflammatory changes in the pancreas confounding CT staging accuracy.

Most important is the problem of overstaging. Dewitt et al. reported that EUS and helical CT predictions of T4 disease were incorrect in 17% and 14%. This indicates patients with resectable disease would be denied potentially curative resection. The concept of deeming a patient unresectable by two imaging modalities has been suggested as the optimal approach, but is not perfect (4, 5). The authors' conclusions that CT alone should guide decision-making cannot be generalized into current practice. Future studies should not focus upon which single imaging modality is best, but rather what combination of tests, or staging algorithm, optimizes clinical outcomes.

1. DeWitt J, Devereaux B, Chriswell M, McGreevy K, Howard T, Imperiale TF, Ciaccia D, Lane KA, Maglinte D, Kopecky K, LeBlanc J, McHenry L, Madura J, Aisen A, Cramer H, Cummings O, Sherman S. Comparison of Endoscopic Ultrasonography and Multidetector Computed Tomography for Detecting and Staging Pancreatic Cancer. Annals Int Med. 2004;141 : 753- 763

2. Tierney WM, Francis IR, Eckhauser F, Elta G, Nostrant TT, Scheiman JM. The accuracy of EUS and helical CT in the assessment of vascular invasion by peripapillary malignancy. Gastrointest Endosc. 2001;53:182-8.

3. Mertz HR, Sechopoulos P, Delbeke D, Leach SD. EUS, PET, and CT scanning for evaluation of pancreatic adenocarcinoma. Gastrointest Endosc. 2000;52:367-71.

4. Tierney WM, Fendrick AM, Hirth RA, Scheiman JM. The Clinical and economic impact of alternative staging strategies for Adenocarcinoma of the Pancreas. Am J Gastro 2000; 95(7): 1708-13.

5. Kochman ML. EUS in pancreatic cancer. Gastrointest Endosc. 2002;56(4):S6-12.

Conflict of Interest:

None declared

We thank Drs. Tierney, Kochman and Scheiman for their interest in our article. We apologize for the incorrect reference cited.

We agree that CT and MRI are superior to EUS for detection of hepatic metastases from pancreatic cancer, since most of the right lobe of the liver cannot be seen by EUS. Therefore, EUS clearly cannot replace, but may supplement, other modalities for staging of the liver. However, EUS and EUS-FNA may detect and accurately sample small metastatic liver lesions missed by other imaging modalities (1). It also may be superior to CT for detection of celiac node metastases (2) and small quantities of peritoneal fluid (3). Therefore, the overall superiority of CT to detect distant metastases may bias the spectrum of disease for the study but does not bias the analysis. We do not believe that improved MDCT imaging of the right lobe of the liver, compared to EUS, creates a significant bias in favor of CT. Furthermore, limiting analysis of enrolled patients to only those patients with confirmed locoregional disease diminishes clinical application of our results.

Only four patients with pancreatic cancer in our study who underwent surgery did not have complete assessment of vascular invasion. Although not stated in our paper, all patients with T4 malignancy had invasion into vessels other than the splenic artery or splenic vein. Information concerning vascular invasion was omitted from our study principally due to the space limitation and to focus on the issues of detection, staging and resectability. We agree with Tierney et al. that this information is critical to determine preoperative staging and intend to publish this information separately.

All patients had either CT or MRI performed outside our institution prior to enrollment in the study. Furthermore, those with obvious metastatic disease were excluded. EUS-FNA was performed prior to MDCT used in our study. We believe that this practice is more the rule rather than the exception among tertiary care centers such as our institution. Although the risk of acute pancreatitis following EUS-FNA is 1-2%, there are no data to support the contention that this potential inflammation may alter accuracy of tumor staging by CT or MRI. In our study, most CT scans were performed the same day as the EUS potentially limiting this problem.

We agree that preoperative overstaging of pancreatic tumors would potentially preclude resectable tumors from proceeding to surgery. However, the protocol we employed generally utilized surgical resection only when one or both tests showed resectability. The more relevant question, however, is whether the use of two tests permits a clinically meaningful increase in resectability as compared to one study alone. Our study did not demonstrate this but may have been underpowered to demonstrate any difference. Despite a slightly increased positive predictive value of resectability when CT or MRI are used in combination EUS (4,5), this strategy remains debatable despite possible cost reductions (4). The use of EUS for pancreatic tumors, however, will likely remain dependent on availability, referral patterns and local expertise.

References

1. DeWitt J, LeBlanc J, McHenry L, Ciaccia D, Imperiale T, Chappo J, Cramer H, McGreevy K, Chriswell M, Sherman S. Endoscopic ultrasound- guided fine needle aspiration cytology of solid liver lesions: a large single-center experience. Am J Gastroenterol. 2003;98:1976-81.

2. Romagnuolo J, Scott J, Hawes RH, Hoffman BJ, Reed CE, Aithal GP, Breslin NP, Chen RY, Gumustop B, Hennessey W, Van Velse A, Wallace MB. Helical CT versus EUS with fine needle aspiration for celiac nodal assessment in patients with esophageal cancer. Gastrointest Endosc. 2002; 55:648-54.

3. Chang KJ, Albers CG, Nguyen P. Endoscopic ultrasound-guided fine needle aspiration of pleural and ascitic fluid. Am J Gastroenterol. 1995; 90:148-50.

4. Soriano A, Castells A, Ayuso C, Ayuso JR, de Caralt MT, Gines MA, Real MI, Gilabert R, Quinto L, Trilla A, Feu F, Montanya X, Fernandez-Cruz L, Navarro S. Preoperative staging and tumor resectability assessment of pancreatic cancer: prospective study comparing endoscopic ultrasonography, helical computed tomography, magnetic resonance imaging, and angiography. Am J Gastroenterol. 2004; 99:492-501.

5. Ahmad NA, Lewis JD, Siegelman ES, Rosato EF, Ginsberg GG, Kochman ML.Role of endoscopic ultrasound and magnetic resonance imaging in the preoperative staging of pancreatic adenocarcinoma. Am J Gastroenterol. 2000; 95:1926-31.

Conflict of Interest:

None declared