Search

Background. Pancreatic cancer remains the fourth leading cause of cancer-related death. Surgery is the only option for cure. Multidetector computed tomography (CT) and advanced postprocessing techniques, can provide solutions for difficult problems in diagnosing and staging disease, and can aid radiologists in communicating findings to surgeons and oncologists. Multiplanar reformats can provide additional information on involvement of the portal vein, hepatic artery, and common bile duct. Maximum intensity projection (MIP) images and volume rendered images can aid in identification of important vascular variants. MIP images can show the relationship of tumor to the pancreatic duct or biliary tree. Curved reformatted images can show the relationship of tumor to the pancreatic duct or vascular structures.

A teaching file of Digital Imaging and Communications in Medicine (DICOM) images was developed. Selected computed tomographic, magnetic resonance, angiographic, nuclear medicine, and ultrasonographic images were transferred to a personal computer (PC) with a 75-GB hard drive. The database created from the DICOM header information was modified to include American College of Radiology (ACR) Learning File categories. Database modification allowed cases to be searched by diagnosis, by findings, and by patient age, sex, and demographics. Technologists inputted ACR codes into the DICOM headers before the images were acquired, thereby reducing database maintenance time. Images were manipulated and reviewed and were used for case presentations, board review, and conferences. Selected images were transferred via a hospital network to workstations for consultation, saving time for referring clinicians. The teaching file is reliable, cost-effective, and easy to install and takes advantage of recent technologic advances in computer hardware and software. Rapid cost reductions and increases in PC storage capacity allow maintenance of the teaching file in DICOM format without intermediate steps involving automation or manual conversion. This setup rivals any of the proprietary software currently available from other vendors.

Current commercial radiology information systems (RIS) are designed for scheduling, billing, charge collection, and report dissemination. Academic institutions have additional requirements for their missions for teaching, research and clinical care. The newest versions of commercial RIS offer greater flexibility than prior systems. We sent questionnaires to Cerner Corporation, ADAC Health Care Information Systems, IDX Systems, Per-Se’ Technologies, and Siemens Health Services regarding features of their products. All of the products we surveyed offer user customizable fields. However, most products did not allow the user to expand their product’s data table. The search capabilities of the products varied. All of the products supported the Health Level 7 (HL-7) interface and the use of structured query language (SQL). All of the products were offered with an SQL editor for creating customized queries and custom reports. All products included capabilities for collecting data for quality assurance and included capabilities for tracking “interesting cases,” though they varied in the functionality offered. No product offered dedicated functions for research. Alternatively, radiology departments can create their own client-server Windows-based database systems to supplement the capabilities of commercial systems. Such systems can be developed with “web-enabled” database products like Microsoft Access or Apple Filemaker Pro.

Computed tomography (CT) is the modality of choice in the evaluation of blunt renal injury. Intravenous urography is used primarily for gross assessment of renal function in hemodynamically unstable patients. Selective renal arteriography or venography can provide detailed information regarding vascular injury. Retrograde pyelography is valuable in assessing ureteral and renal pelvic integrity in suspected ureteropelvic junction injury. Ultrasonography is useful in detecting hemoperitoneum in patients with suspected intraperitoneal injury but has limited value in evaluating those with suspected extraperitoneal injury. Occasionally, radionuclide renal scintigraphy or magnetic resonance imaging may prove helpful. Renal injuries can be classified into four large categories based on imaging findings. Category I renal injuries include minor cortical contusion, subcapsular hematoma, minor laceration with limited perinephric hematoma, and small cortical infarct. Category II lesions include major renal lacerations extending to the medulla with or without involvement of the collecting system and segmental renal infarct. Category III lesions are catastrophic renal injuries and include multiple renal lacerations and vascular injury involving the renal pedicle. Category IV injuries are ureteropelvic junction injuries. CT is particularly useful in evaluating traumatic injuries to kidneys with preexisting abnormalities and can help assess the extent of penetrating injuries in selected patients with limited posterior stab wounds. Integration of the imaging findings in renal injury with clinical information is critical in developing a treatment plan.

Implementation of a second-generation radiology information system (RIS) requires attention to many issues, including work flow, system design, training, consideration for emerging technologies, and support. This presentation convers the issues, and solutions, involved in implementation and subsequent day-to-day usage of a client-server-based RIS tightly integrated with a hospital information system (HIS) using “thin client” software to limit hardware requirements for the client computers. The software and hardware implementation had to be designed for potential enterprise-wide scalability of the RIS for a system involving 11 hospitals. Issues arose regarding initial increased turn-around times, errors in importation of historical data, and problems with implementation of interfaces between the RIS and the billing system and between the RIS and the digital dictation system. Network errors and difficulties in using a thin client implementation had to be overcome. Hierarchical training was implemented, as support was switched from an “on-site” support group to a central Information System Division with responsibility to the entire enterprise.

Rationale and Objectives. The Internet offers many advantages for educating patients but has no standards for publication. This limitation could negatively affect patient care. The purpose of this study was to evaluate the quality of information on mammography that a patient could find on the Internet. Materials and Methods. Three search utilities were used to research the term “mammography.” For each utility, the first 50 addresses for Web pages were evaluated (or all the addresses, if fewer than 50 were returned). Web sites selected contained information that could guide an asymptomatic woman in deciding whether to undergo screening mammography. These sites were then evaluated for whether they indicated their sponsorship, authorship, the currency of information, and references. Sites were also noted if they advised women older than 50 years to undergo screening mammography at intervals of longer than 1 year or otherwise severely diminished the role of mammography. Results. Thirty-eight Web sites were identified. Ten indicated authorship, 29 indicated the currency of the information, and 27 provided references. All of the Web sites indicated sponsorship. Three sites recommended screening mammography at intervals of longer than 1 year for women 50 years of age or older. Two sites suggested that mammography is not substantially more sensitive than physical examination. Conclusion. Many Web sites do not meet the standards for disseminating information required in professional peer-reviewed journals. Some Web sites contain statements that might lead asymptomatic women over age 50 years to delay screening mammography or to undergo screening at intervals of longer than 1 year.

Picture archiving and communications systems (PACS) workstations are reported to improve workflow by making studies immediately available for review upon their completion. This study tested the hypothesis that a workstation would decrease the time from completion of a study to dictation of results (report time). A four-monitor, 2K x 2K workstation (Imation Cemax-Icon, Fremont, CA), was installed in a body imaging computed tomography (CT) reading room. Use of the workstation by the staff radiologists was voluntary. Images were also printed on film and films continued to be hung at the routine hanging times. To evaluate the workstation's maximum impact, data were collected for report times for studies completed during the routine day shift of the staff radiologist (Monday to Friday, 8 AM to 5 PM). Data were collected before workstation installation (August 1997 to November 1997) and for the subsequent 6 months. Histograms of the number of studies (743 v 103) versus report time (mean, 11.7 v 7.4 hours) showed a bimodal distribution, with peaks at approximately 6 and 24 hours, both before (8/97-11/97) and after (6/98) the workstation's installation. However, the number of studies dictated greater than 60 hours (25.2% v 20.4%) and the percentage of studies in the second peak (16 to 48 hours; 4.4% v 0%) both decreased. In conclusion, the workstation decreased the mean (11.7 v 7.4 hours) and standard deviation (19.8 v 9.1 hours) for report times. This was due to a decrease in both the number of cases dictated the day following their completion and the number of outliers (markedly delayed dictations). The decrease in outliers is probably due to a decrease in the number of "lost" film-based studies.

To fully take advantage of the widespread use of digital imaging systems and to update and eliminate redundant steps involved in medical radiographic publication, we present our experience of processing Digital Imaging and Communications in Medicine (DICOM)-3 digital images from the point of acquisition to the point of publisher-ready radiographic images without intervening hardcopies.

neoplastic resections. 272 (27%) of cases had preoperative evalution by EUS FNA. The overall accuracy of EUS-FNA in diagnosing malignancy was 84% (sensitivity of 82.4 %). 1.8% of resections were benign in the EUS group and 5.3% in the non-EUS group. There were significantly fewer benign lesions resected in the pre-operative EUS group than in the surgery only group (p 0.024). With regards to accuracy of staging, overall, EUS was able to differentiate T1/T2 from locally advanced disease (T3/T4) with an accuracy of 91.2%, sensitivity of 90.5% and a specificity of 91.5%. The accuracy of N-stage by EUS was 64.7%, sensitivity was 94.7%, and specificity was 53.1%. By univariate survival analysis there was no difference in overall survival among patients undergoing resection for pancreatic cancer with and without preoperative EUS FNA (Log rank p 0.1628). In a multivariate Cox proportional hazards model, undergoing preoperative EUS-FNA was not a significant predictor of time to recurrence (p 0.16). Conclusions: Our study suggests that preoperative EUS-FNA of suspicious but resectable appearing pancreatic masses may reduce the number of resections of benign pancreatic lesions. EUS had a high diagnostic and T staging accuracy but a relatively low N stage accuracy among patients with confirmed malignancy. We found no association between EUS-FNA and recurrence or survival in patients with resectable pancreatic cancer.

To evaluate the radiographic appearance of a pancreaticogastrostomy (PG) and its complications.Seventy-two patients underwent pancreaticoduodenectomy and PG or pancreaticojejunostomy. Those who underwent PG and were evaluated postoperatively with T-tube cholangiography and upper gastrointestinal (UGI) series constitute the study group (n = 22; 10 men, 12 women; age range, 33-88 years).Twenty-one of the patients had a gastric filling defect radiographically detected. Four patients had clinically apparent delayed gastric emptying and one patient had a clinically apparent pancreatic fistula not detected radiographically. Two patients outside the study group had complications: One had a pancreatic fistula seen only with sinography, and one had a PG leak seen only with repeat UGI series and computed tomography (CT).PG caused a gastric filling defect in most patients. Suspected pancreatic fistulas are best confirmed with sinography, and suspected PG leaks may require repeat evaluation and CT. Clinical findings of delayed gastric emptying do not correlate with findings of UGI.

4051 Background: Preop therapy for resectable pancreatic cancer (PC) maximizes access to multimodality therapy and increases the R0 resection rate. Based on our phase I chemoradiation (chemoRT) trial in PC patients, we hypothesized that the addition of bev to gem-based chemoRT in pts with resectable PC may allow more pts to undergo pancreaticoduodenectomy (PD), improve the rate of R0 resections, and prolong overall survival. Methods: Pts with biopsy proven, stage I/II adenocarcinoma of the pancreatic head or uncinate process received 6 weekly infusions of gem (400 mg/m2) and 3 infusions of bev (10 mg/kg, every 2 wks) with concomitant RT, 50.4 Gy at 1.8 Gy/fr. Pts were restaged 4-6 wks after the last dose of radiation. Those without disease progression and with good PS underwent surgery. Pts with adequate recovery received bev (10 mg/kg) every 2 weeks for 3 mo starting ~ 6 wks after surgery. Results: This study enrolled 11 of a planned 31 pts but was terminated early due to unforeseen postop complications. Median age is 64 yrs; all pts had ECOG-PS 0-1. Median CA19-9 was 52. Median follow-up from surgery was 41 mths. All completed chemoRT; 10 underwent restaging and 1 pt died from cardiac arrest before restaging. 9 pts (90 %) went to surgery and underwent a successful R0 PD. Pathologic PR rate (>50 % tumor kill) was 56 %. As we have previously reported, preop grade 3-4 toxicities were infrequent and major postop complications occurred in 5 of the 9 pts (56%) who underwent PD. Median overall survival (OS) was 30.1 mths (range: 2.6 - 63.9) for the entire cohort. Of the 9 resected pts, median OS was 45.5 mths with 5 of 11 pts (45%) achieving survival longer than best historical control (median 58.2 mths). Conclusions: The addition of bev to our prior backbone of gem-based preoperative chemoRT led to a higher resection rate (90%) than our previous protocols. Updated survival results of this study are promising. Taken together, these results may prompt further investigation of this regimen with modifications in the timing of bev delivery.

Renal oncocytomas are well-differentiated, benign tumors of the proximal tubules. They affect men more often than women by a ratio of 1.7~1, and are usually asymptomatic, the majority being noted incidentally [l]. They usually occur in the sixth or seventh decade, with size on presentation ranging from 3 to 26 cm. Overall, they comprise about 4% of all renal tumors [2]. We report a case of a renal oncocytoma presenting 33 years after the patient had undergone radiation therapy for an adrenal neuroblastoma.

4113 Background: There is no data on the value of surveillance CT scans in pts with resected pancreatic adenocarcinoma (PC). In practice, onset of symptoms and/or abnormal blood tests including ris...

Cystic lesions in a variety of organs are being increasingly recognized as an incidental finding on cross-sectional imaging. These lesions can be benign, premalignant or malignant. When these cystic lesions are small it can be difficult to characterize them radiologically. However, with appropriate clinical history and knowledge of typical imaging features of cystic pancreatic lesions this can enable accurate diagnosis and thus guide appropriate treatment. In this review, we provide an overview of the most common types of cystic lesions and their appearance on computer tomography, magnetic resonance imaging and ultrasound. We will also discuss the follow up and management strategies of these cystic lesions.

The pancreas is a retroperitoneal organ with a close anatomic relationship to the peritoneal reflections in the abdomen, including the transverse mesocolon and the small bowel mesentery, and is directly contiguous to peritoneal ligaments such as the hepatoduodenal ligament, gastrohepatic ligament, splenorenal ligament, gastrocolic ligament, and the greater omentum. Understanding of these anatomic relationships of the pancreas is aided by knowledge of its embryologic development. These reflections and ligaments are potential pathways for spread of disease processes such as pancreatitis and pancreatic carcinoma. One can recognize these ligaments and reflections by identifying the blood vessels that traverse them.

To improve radiology access to full uncompressed Digital Imaging and Communications in Medicine (DICOM) data sets, we evaluated satellite access to a DICOM server. Radiologists′ home computers were connected by satellite to a medweb DICOM server (Medweb, San Francisco, CA). A 10.2-kb data set containing a 19-image head computed tomography (CT) scan was transferred using DirecPC (Hughes Electronics Corp, Arlington, VA) at three different times of the day; 6 AM, 3 PM, and 8 PM. The average transfer time for all 19 images from the DICOM server was 4 minutes and 17 seconds (257 seconds). The slowest transfer rate of 670 seconds (121 kbps) was obtained at 8 PM. The best transfer rate of 2 minutes, 54 seconds (467 kbps) was obtained at 6 AM. The full 16-bit DICOM images were viewed with bone, brain, and soft tissue windows. The Medweb plug-in viewer loaded the first image within 30 seconds of selecting the case for satellite transfer. In conclusion, satellite internet transfer of radiology studies is suitable for timely review of full DICOM data sets and can expand the range of teleradiology consultation.

4037 Background: Over 80% of pts who undergo potentially curative pancreaticoduodenectomy (PD) for pancreatic adenocarcinoma develop local or distant recurrence. Chemotherapy and external-beam radiation therapy (EBRT) prior to surgery maximizes the number of pts who receive multimodality therapy and undergo a complete (R0) resection. Methods: Pts with biopsy proven, stage I/II adenocarcinoma of the pancreatic head or uncinate process received systemic therapy with 4 infusions of gem (750 mg/m2) + cisplatin (30 mg/m2) every 2 wks followed by 30 Gy of EBRT (3 Gy/Fx, M-F over 2 wks) and 4 weekly doses of gem (400 mg/m2). Pts underwent complete restaging 4–6 wks after the last dose of gem. Those without disease progression and with acceptable PS underwent surgery. Results: This study has currently enrolled 87 pts over 4 years. Median age is 65 yrs (38–80), most pts (96%) had ECOG-PS (0–1) at study entry. Median CA19–9 is 170. Six pts were excluded from analysis and 4 are currently on treatment. Of the 77 pts who finished chemoradiation, 72 completed restaging (3 await restaging, 2 drop outs). Of these 72 pts, 10 had disease progression, 1 had a decline in PS, and 61 (79 %) went to surgery. 49 of 61 pts (64 %) underwent a successful PD and metastatic or locally advanced disease was found in 12 (15 %). 44 (90 %) of 49 pts underwent R0 resection. Pathologic PR rate (>50 % tumor kill) was 61 %. The predominant toxicity prompting hospitalization during preoperative therapy was biliary stent occlusion (44 %). Gastrointestinal toxicity ≤ grade 3 was common but controllable. Grade 3 and 4 neutropenia was seen in 19 (25 %) and 3 pts (4 %) respectively. Grade 3–4 thrombocytopenia was rare. Median survival for resected pts was ∼ 21 months. Conclusions: This study, builds on our previous gem-based preop chemoradiation program ( ASCO 2002, Abs # 516) producing good local treatment effect and a high R0 resection rate. However, interim analysis suggests that the addition of systemic chemotherapy delivered prior to preoperative chemoradiation does not appear to improve survival. Optimal treatment sequencing of chemoradiation, systemic therapy and surgery needs further refinement. Supported in part by Eli Lilly and Co. [Table: see text]

4033 Background: Pancreatic cancers are resistant to RT. Vascular endothelial growth factor (VEGF) is overexpressed in many tumor types incuding PA. In vivo studies have shown that neutralizing VEG...

High quality images from clinical CT, MRI and angiography were available to remote desktop computers at virtually no cost. Film queue data files were relatively small and represented the findings of interest. No proprietary software was required for viewing. Residents were able to transfer images into the file server after only a few minutes of training.