AJR:181, September 2003 775 Preoperative Detection of Pancreatic Insulinomas on Multiphasic Helical CT OBJECTIVE. The objective was to analyze enhancement characteristics of insulinomas and to determine the ability of multiphase CT to localize these tumors. MATERIALS AND METHODS. Prospective interpretations of multiphase helical CT scans were reviewed in 30 patients who had insulinomas resected over a 5-year period. CT scans were retrospectively reviewed to determine enhancement characteristics, tumor conspi- cuity in each phase of enhancement, and potential causes for false-negative findings. RESULTS. Sixty-three percent (19/30) of tumors were identified on CT prospectively. An additional six tumors were visualized in retrospect, allowing characterization of 25 (83%) of 30 tumors. Most tumors were hyperdense on at least one phase (n = 19), three tumors were hy- poattenuating, and three were isodense and pedunculated. Insulinomas were most conspicuous on the early phase in 15 patients and in the portal venous phase in three. All tumors that under- went pancreatic phase imaging were seen (13/13), whereas three of 18 arterial and six of 25 portal venous phase findings were inconclusive for tumor. In the six examinations with false- negative findings in which the tumor could be seen in retrospect, two tumors were isodense and pedunculated, three were in close proximity to vessels, and one had a cystic appearance. CONCLUSION. Multiphasic CT has a moderate sensitivity in the detection of insulino- mas. Most tumors are more conspicuous on the earlier phases of enhancement. The pancreatic phase may be more useful than the arterial phase. Potential sources of false-negative results in- clude tumors adjacent to vessels, pedunculated morphology, or nonhyperattenuating lesions. istorically, the preoperative eval- uation for patients with suspected insulinomas has been controver- sial. Some experts believe that after an in- sulinoma has been confirmed biochemically, preoperative imaging is not necessary be- cause nearly all insulinomas are located in the pancreas and can be detected with intra- operative inspection, palpation, and sonogra- phy. In addition, noninvasive imaging techniques are perceived as having relatively poor sensitivities [1–4]. Other experts believe that preoperative imaging can be valuable by providing infor- mation regarding the type of surgery re- quired. For example, insulinomas located near the surface of the gland may undergo simple enucleation; however, if the tumor is located deep in the gland, partial pancreatec- tomy may be required. Identification of a fo- cal mass can help exclude nesidioblastosis as a cause of the hypoglycemia, which might require an extensive resection. Nesidioblas- tosis, or islet cell hyperplasia, is character- ized by a proliferation of abnormal B cells throughout the pancreas. Finally, preopera- tive localization can reduce the operative pal- pation that is necessary [3, 5, 6]. Thus, if the sensitivity of noninvasive preoperative imag- ing improves, dedicated pancreatic imaging may be more widely used in preoperative evaluation algorithms. Technical advances have led to an improve- ment in the quality of CT, and recent reports have suggested improved performance for the detection of islet cell tumors [7–11]. However, further studies are necessary to validate these promising results. In addition, disagreement exists on the appropriate CT protocol. With the development of helical CT and, more recently, multidetector scanners, we are now able to rapidly image the pan- creas during multiple phases of contrast en- hancement. Some investigators have shown improved conspicuity of islet cell tumors on the arterial phase of enhancement, whereas J. L. Fidler 1 J. G. Fletcher 1 C. C. Reading 1 J. C. Andrews 1 G. B. Thompson 2 C. S. Grant 2 F. J. Service 3 Received February 10, 2003; accepted after revision April 1, 2003. Presented at the annual meeting of the American Roentgen Ray Society, San Diego, May 2003. 1 Department of Radiology, Mayo Clinic and Mayo Foundation, 200 First St. S.W., Rochester, MN 55905. Address correspondence to J. Fidler (fidler.jeff@mayo.edu). 2 Departmemt of Surgery, Mayo Clinic and Mayo Foundation, Rochester, MN 55905. 3 Department of Endocrinology, Mayo Clinic and Mayo Foundation, Rochester, MN 55905. AJR 2003;181:775–780 0361–803X/03/1813–775 © American Roentgen Ray Society H Downloaded from www.ajronline.org by 52.73.204.196 on 05/16/22 from IP address 52.73.204.196. Copyright ARRS. For personal use only; all rights reserved