High contrast flow rate best for pancreas CT

Nov 7, 2006

The conspicuity of pancreatic tumors improves markedly in multidetector-row CT (MDCT) when contrast flow rates are higher and scan delays are individualized for each patient, researchers from Austria report.

The stakes are always high for pancreatic imaging, and the scan delay becomes more critical as the number of detector rows increases, the researchers from the Medical University of Vienna in Austria wrote in Radiology (September 18, 2006). CT imaging is difficult, but high contrast flow rates can improve results.

The five-year survival rate for pancreatic cancer is just 4% in the U.S., and surgical resection is the only potential cure, though only about 20% of cases are resectable at the time of diagnosis, Dr. Gerd Schueller, and his group explained. Patient selection is of paramount importance because of the substantial morbidity associated with laparotomy (20% to 30%), and the Whipple procedure carries a mortality rate of up to 5%.

Contrast delivery and timing issues are key to answering the critical question of resectability. Most authors recommend dual-phase CT with image acquisition in the pancreatic and portal venous phases, the group wrote. In a previous study, Lu et al called for fixed scan delays of 40 seconds.

Although MDCT has improved z-axis resolution, lesion conspicuity is still problematic, with up to 11% of adenocarcinomas isoattenuating at CT, the team wrote. Researchers attempting to improve conspicuity have increased flow rates to as high as 5 mL/sec, and scan delays of 35-40 seconds have been used, but fixed scan delays still present problems in fast acquisitions of 3-4 seconds in 16- and 64-detector scanners.

"Such a short acquisition time is not likely to fall at the peak enhancement of the pancreas in all patients if a fixed scan delay is used," the authors wrote. "Moreover, an interval for the contrast material delivery of 30 seconds (given 150 mL contrast material at a flow rate of 5 mL/sec) produces an enhancement curve with a broad peak that does not conform to a scan duration of only 3-4 seconds."

A key question is whether scans would benefit from individualized scan delay times along with higher flow rates. To find out, the study sought to prospectively compare pancreatic lesion conspicuity with two different contrast flow rates: either a high flow rate (8 mL/sec) and individualized delay timing or a normal contrast flow rate of 4 mL/sec.

The study examined 40 patients with suspected pancreatic cancer (21 women and 19 men, mean age 67.1) who underwent biphasic CT imaging of the pancreas. They were randomized to receive 150 mL of nonionic contrast material at a flow rate of 4 mL/sec or alternatively at 8 mL/sec, followed by a 40-mL saline flush.

The contrast agent (iopromide, 300 mg of iodine/mL, Schering, Berlin) was administered with an Injection CT2 power injector (Medtronic, Minneapolis, MN). A 16-detector CT scanner (Somatom Sensation 16, Siemens Medical Solutions, Erlangen, Germany) was used to scan the patients at 1.5-mm collimation, 120 kVp, mean tube current of 75 mAs (range 61-95 mAs) for the dynamic phase, and 105 mAs (range 99-114) for portal venous phase imaging.

The images were reviewed on a workstation (Impax, Agfa HealthCare, Mortsel, Belgium) and were analyzed quantitatively for attenuation and time-to-peak attenuation, aortic transit time, and attenuation of tumors versus normal pancreatic tissue. Tumor conspicuity was graded on a five-point scale (0 not seen, 5 excellent), and resected tumors were analyzed histologically.

The results showed that peak contrast enhancement in the pancreas was seen significantly earlier (mean ± standard deviation, 28.7 seconds ± 3.5 versus 48.2 seconds ± 5.3; p < 0.05) and was significantly higher (129.0 HU ± 25.7 versus 106.2 HU ± 35.4, p < 0.05) using a flow rate of 8 mL/sec compared with 4 mL/sec.

The tumor-to-pancreas contrast greater than 40 HU lasted significantly longer at the high flow rate versus the low rate (26.4 seconds ± 11.9 versus 8.6 seconds ± 8.3, p < 0.05), according to the authors.

Using the high flow rate, an individualized scan delay of 19 seconds after aortic transit time showed a higher tumor-to-pancreas contrast and better lesion conspicuity than fixed scan delay, the group reported.

In the low-flow-rate cohort, the maximum attenuation of the pancreas was significantly higher (p < 0.05) with an individualized scan delay of 29 seconds rather than a fixed delay of 40 seconds (individualized aortic transit time to maximum tumor-to-pancreas contrast). In the high-flow-rate group, the maximum attenuation of the pancreas was significantly higher (p < 0.05) with an individualized delay of 19 seconds than with fixed delays of 34 and 40 seconds, the team reported.

Tumors were located in the pancreatic head in 14 patients (low flow: n = 8, high flow: n = 6), in the pancreatic body in five patients (low flow: n = 1, high flow: n = 4), and in the pancreatic tail of one patient (low-flow group). Follow-up showed that 20 of the 40 patients had a normal pancreas (low flow rate: n = 11, high flow rate: n = 9), and 20 had a histologically proven pancreatic adenocarcinoma (low flow rate: n = 10, high flow rate: n = 10).

"With 16-section CT, increased contrast material flow rate of 8 mL/sec and individualized scan delay were associated with improved pancreatic enhancement and tumor-to-pancreas contrast compared with the flow rate of 4 mL/sec and fixed scan delay," wrote Schueller and his colleagues.

To the authors' knowledge, the high flow rate had not been previously reported for pancreatic imaging; however, other contrast material applications into the forearm veins, such as perfusion liver imaging, have called for flow rates as high as 10 mLsec, they noted.

Limitations included the lack of a dedicated detection study for pancreatic cancer and the reporting of attenuation values only at 34 and 40 seconds after initiation of contrast injection. Larger studies "are necessary to confirm whether this protocol translates into better detection of small pancreatic tumors," the authors wrote.

By Eric Barnes
AuntMinnie.com staff writer
November 8, 2006

CT features diagnose autoimmune pancreatitis, February 2, 2006

MDCT yields reliable diagnosis of pancreas divisum, April 14, 2005