Applied Radiology

Dr. Knake received his MD in 1999 from Northwestern University Medical School, Chicago, IL. He currently is a third-year Diagnostic Radiology Resident at the University of Virginia, Charlottesville, VA.

Advances in magnetic resonance imaging (MRI) technology have solidified the role of abdominal MRI in the armamentarium of the body imager. This article will review the application of body MRI in clinical practice with special focus on the use of dynamic contrast-enhanced MRI techniques for the differential diagnosis of hepatic, renal, and pancreatic disease.

The major strengths of magnetic resonance imaging (MRI) are well established and include high intrinsic soft-tissue contrast, multiplanar imaging capability, and high sensitivity for the presence or absence of contrast enhancement. ¹ Another important hallmark of MRI is its safety; unlike ionizing radiation, the radiofrequency pulses and magnetic fields MRI employs are not known to cause harmful effects on biological systems. Moreover, gadolinium chelates, the primary MRI contrast agents, have a better safety profile than the iodine-based agents used for computed tomography (CT) and they are unlikely to cause renal damage.

Historically, MRI has been used mainly in neuroimaging and musculoskeletal imaging, playing only a minor role in abdominal imaging. Artifacts from respiratory motion, bowel peristalsis, cardiac pulsations, and patient motion combined with long scan times and difficulties with patient compliance (eg, long breath holds) significantly limited the role of MRI in the diagnosis of intra-abdominal disease. ² Recent developments in MRI techniques and contrast agents, however, have led to a broadened role for MRI in the evaluation of abdominal disease, including its use in the detection and characterization of intra-abdominal lesions. These advances have clarified body MRI's vital role in abdominal imaging and have shown that body MRI is no longer merely a last option. ³

Dynamic bolus MRI

Unenhanced T1- and T2-weighted MRI has the ability to detect intra-abdominal pathology primarily because of the high intrinsic soft-tissue contrast shown. However, contrast-enhanced imaging, specifically dynamic contrast-enhanced imaging, provides physiologic information about masses and organs that permits more specific differential diagnosis. Dynamic bolus MRI refers to the technique in which serial postcontrast images are acquired in the arterial, parenchymal, and delayed phases specific to the organ of primary analysis. The exact timing varies somewhat according to the organ of interest. This article will focus on the specific principles, techniques, and applications of dynamic contrast-enhanced parenchymal MRI, emphasizing the liver, kidneys, and pancreas. Three-dimensional contrast-enhanced MR angiography utilizes different dynamic techniques and is not addressed in this article.

Liver

The optimal imaging strategy for analysis of hepatic lesions has been a topic of extensive debate over the last decade. The fact that both malignant and benign lesions may occur concomitantly in the same patient emphasizes the importance of the detection and the characterization of individual lesions so that appropriate therapeutic procedures can be planned and executed. ⁴ Currently used modalities include triphasic CT, ultrasound, and MRI. Invasive computed tomographic arterial portography (CTAP) has fallen into disfavor due to the technological improvements of noninvasive CT and MRI. Streamlined MRI techniques, combined with improved gradients, hardware, and contrast agents, have created the capability of complete and comprehensive noninvasive hepatic imaging. ⁵ Research presented by leading academicians at the 2001 Annual Meeting of the Radiological Society of North America supports the general consensus that MRI, performed with one or more contrast agents, is the superior modality for liver lesion characterization and is equal to or better than the other modalities for lesion detection.

Although not the main focus of this article, a discussion of hepatic MRI is incomplete without an overview of examination technique. To date, no single universally accepted sequence for hepatic imaging has emerged. It is now feasible to obtain nearly artifact-free images with sufficient contrast-to-noise ratios across a broad range of techniques with rapid scanning. ⁶ Multiple possible sequences, each with different names or acronyms depending on the equipment manufacturer, tend to confuse the MR neophyte.

Disregarding vendor-specific terminology, comprehensive liver imaging should include pre-contrast axial T1-weighted gradient-echo images (usually to include "in-phase" and "opposed-phase" sequences), unenhanced T2-weighted images using fast sequences--either fast-spin echo or turbo-spin echo--and contrast-enhanced T1-weighted gradient-echo dynamic images obtained in the arterial, portal venous, and equilibrium phases of enhancement. Fat-saturation techniques may be applied to either the T1- or T2-weighted unenhanced images.

There are currently three classes of approved contrast agents for MRI of the liver: 1) T1-shortening nonspecific extracellular gadolinium-based compounds; 2) T2-shortening reticuloendothelial cell-specific iron-oxide­based compounds (ferumoxides); and 3) T1-shortening hepatocyte-selective manganese-based compounds (Mn-DPDP). ¹ Neither the reticuloendothelial cell-specific nor iron-oxide­specific agents are used in dynamic imaging, and although their availability has added a new dimension to liver MRI, it has not drastically altered the role of gadolinium chelates. ⁷

Why gadolinium?

The gadolinium chelates were the first contrast agents available for abdominal MRI and, consequently, there is a large body of clinical evidence for their use. These paramagnetic substances exert their primary effect by producing T1-shortening/relaxation and thereby increase signal intensity on T1-weighted scans within the targeted tissue. The extracellular characteristic of the agents is the feature that permits their use in dynamic imaging because equilibration between intravascular and extracellular spaces occurs rapidly after injection. ⁸ After entering the liver, intravascular gadolinium redistributes at different rates into both neoplastic lesions and normal parenchyma. The distribution into the vascular and interstitial compartments of both normal and pathologic tissues explains the need and rationale for rapid, dynamic, contrast-enhanced MRI so that maximum differential enhancement between tumors and liver parenchyma can be achieved. ⁷ An understanding of the enhancement patterns of various hepatic lesions then allows the radiologist to generate appropriate differential diagnoses.

How to use gadolinium chelates

On high-field strength (1.5 T) systems with dedicated coils for abdominal imaging, rapid gradient-echo T1-weighted sequences obtained during a short breath hold allow for effective use of gadolinium chelates with near complete elimination of respiratory motion, high signal-to-noise ratio, and subsequent optimization of contrast between tumor and adjacent parenchyma. A single dose of gadolinium (0.1 mmol/kg) is injected with a power injector as a bolus through a peripheral intravenous (IV) needle at a rate of at least 1 mL/sec, preferably 2 to 3 mL/sec, followed by a saline flush. The three gadolinium chelates approved for use in the United States are gadoteridol (Prohance; Bracco Diagnostics, Princeton, NJ), gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) and gadodiamide (Omniscan; Amersham Health, Wayne, PA).

Precise image acquisition timing after injection is essential in order to obtain optimal phase-specific images. Normal hepatic parenchyma in a patient with normal circulation receives approximately 20% of its blood supply from the common hepatic artery and 80% from the portal vein. In the patient with normal circulation, the hepatic arteries are contrasted about 25 to 30 seconds after bolus contrast injection, while the portal and hepatic veins are still unenhanced and remain hypointense to surrounding parenchyma. This is the arterial phase. After approximately 55 to 60 seconds, maximal liver parenchymal enhancement occurs, characterizing the portal venous (or nonequilibrium) phase of enhancement. After 3 to 5 minutes, the gadolinium has diffused across the capillary endothelium into the interstitial space of liver and tumors. Images acquired at this time are referred to as the delayed or equilibrium-phase images. It is important for the radiologist to be aware of variations of circulation times among patients and to have protocols in place for initiating appropriate time of scan delays so that optimal arterial, portal venous, and delayed images are obtained for every patient.

Arterial phase

Optimal images for this phase must be acquired before maximal sinusoidal enhancement of the liver, usually 20 to 30 seconds after the start of contrast injection. ⁶ The abdominal viscera have a characteristic appearance at this time. The hepatic arteries and often main portal vein branches are opacified (high signal on T1-weighted images), but the hepatic veins and liver parenchyma are not enhanced. At this time, pancreatic and renal cortical capillary enhancement will also be observed, the kidney demonstrating the characteristic renal "cortico-medullary differentiation." The spleen normally exhibits a serpiginous, arciform pattern of enhancement 25 to 30 seconds after contrast injection. ² Since some hypervascular metastases and hepatocellular carcinomas have only an hepatic arterial blood supply, this phase of imaging is crucial for their detection (figures 1 and 2). ⁶ One must take care not to mistake hypointense venous structures for pathology during this phase.

Portal venous phase

Image acquisition at approximately 1 minute after bolus contrast injection produces an enhanced liver parenchyma as well as enhanced hepatic and portal veins. Hypovascular metastases and tumors are well visualized during this phase, appearing as hypointense lesions relative to the enhancing liver (figure 3). ⁹ Additionally, because this phase of liver enhancement provides excellent visualization of the hepatic and portal vein branches. In this phase, segmental liver anatomy is depicted easily, allowing for precise description of lesion location, which is especially important to provide to the surgeon.

Equilibrium phase

Approximately 3 to 5 minutes after contrast injection, the gadolinium chelate has diffused out of the vascular compartment and into the interstitial space of both normal liver parenchyma and tumors. ⁸ The majority of tumors are less visible in this phase. However, because cholangiocarcinomas have a prominent interstitial space, they characteristically accumulate more contrast than adjacent normal liver and thereby increase their conspicuity on delayed images. ¹⁰ Hemangiomas will usually fill in and some extrahepatic disease, such as peritoneal implants and areas of inflammation, are also well visualized on delayed images because of their slow gadolinium accumulation.

Many comprehensive reviews in the radiology literature describe the detailed MRI characteristics and classifications of the various benign and malignant hepatic lesions. ^5,11 The following provides a short review of the classification of focal liver lesions by MRI, using vascularization patterns accentuated with dynamic imaging. Based on perfusion patterns, three primary groups of focal liver lesions are distinguishable: hypervascular in arterial phase, hypovascular in arterial phase, and lesions with delayed persistent enhancement. ²

Hypervascular lesions

Hypervascularized lesions are characterized by intense contrast enhancement in arterial phase images (25 to 30 sec). Regardless of specific lesion morphology, early enhancement reflects hepatic arterial blood supply. Additionally, the characteristic peripheral rim enhancement of hypervascular metastases is distinct from the nodular, discontinuous early peripheral enhancement characteristic of hemangiomas. The lesions most commonly hypervascular in the arterial phase include primary liver tumors and hypervascular metastases (Table 1). ^2,7,9,11,12

Hypovascular lesions

Characteristically hyposvascular liver lesions are identified by reduced or absent perfusion during all phases of dynamic scanning, which results in a hypointense appearance of the lesion relative to the normal surrounding liver parenchyma. The image appearance reflects the relative lack of blood supply to the tumor from the hepatic artery and portal vein. Hypovascular metastases may demonstrate minimal peripheral hypervascularity or irregular, heterogeneous, segmental internal enhancement--characteristics that are consistent with malignancy of the lesion--while the absence of irregular enhancement/perfusion is consistent with benignity, as with simple hepatic cysts or bilomas. ² The greatest difference in liver-to-lesion signal intensity occurs during the portal venous phase of dynamic scanning when greatest enhancement of the liver parenchyma occurs. ¹² Common hypovascular lesions are listed in Table 2. ^2,7,9,11,12

Lesions with delayed and persistent enhancement

Some hepatic lesions are characterized by delayed contrast media uptake, thereby appearing predominantly hypointense during the arterial and portal venous phases but iso- or hyperintense up to 12 to 15 minutes after contrast injection. The appearance reflects contrast washout from normal liver tissue with concomitant contrast pooling in the diseased liver. The hemangioma is the characteristic benign liver lesion with delayed enhancement, while the characteristic malignant lesion with delayed enhancement is the intrahepatic cholangiocarcinoma. Metastases from malignant gastrointestinal stromal tumors (formerly referred to as leiomyosarcomas) may also exhibit this enhancement pattern. ^2,7,9,11,12

Summary of hepatic MRI

Dynamic serial gadolinium-enhanced MRI of the liver is now a well-studied and understood technique that should be a routine component of a comprehensive MRI examination of the liver. When combined with unenhanced T1- and T2-weighted images, dynamic imaging helps increase the radiologist's accuracy and confidence in diagnosis, thereby helping to achieve the ultimate goal--optimal patient care. Similar utility is obtained with dynamic gadolinium-enhanced imaging of the kidney and pancreas.

Kidney

Low cost, widespread availability, and lack of ionizing radiation are the primary reasons why ultrasonography remains the initial screening modality of choice for imaging the kidneys. Renal masses and other abnormalities detected on ultrasound necessitate further radiologic workup which, for many years, has been performed solely with contrast-enhanced CT. However, as with liver MRI, new pulse sequences, improved technology, and the use of bolus dynamic imaging have changed the role of MRI in the workup of renal lesions. ²

Renal disease leads to both morphological change and functional disturbance of the kidney, ^13,14,19 and MRI has proven itself as the superior technique for both renal lesion detection and characterization. In head-to-head studies, CT and MRI have been shown to have similar sensitivity and specificity for detection of renal masses, but for evaluation of lesion extent and for analysis of peri-renal and vascular involvement, MRI is superior. ^12,15,16 MRI is also especially helpful for analyzing lesions that are equivocal on CT--a common example being differentiation between a small complicated cyst and cystic or hypovascular renal cell carcinoma. ² Additionally, gadolinium-enhanced renal MRI is an attractive option for the evaluation of native and transplanted kidneys because it can be used in patients with renal failure, ¹⁷ although this technique is not approved by the FDA.

Various protocols exist for renal MRI. Although specific sequences differ among institutions, imaging ought to include T1- and T2-weighted pre-contrast images (including T1-weighted images with fat saturation) followed by rapidly acquired breath-hold T1-weighted dynamic post-gadolinium images. With dynamic serial imaging after IV gadolinium injection, four distinct phases of contrast enhancement are possible to identify: capillary (cortical) phase, early tubular, ductal, and excretory. ¹⁸ These phases may be captured with imaging immediately after contrast injection (capillary phase, demonstrating corticomedullary differentiation), followed by repeat imaging at 1 to 2 minutes, 3 minutes, and 5 to 10 minutes after contrast injection. Immediate post-injection images capture the arterial phase of renal enhancement, the 1- to 2-minute and 3-minute images capture the "nephrogram" phase during which renal parenchyma is maximally enhanced, and the 5- to 10-minute images capture the excretory phase of renal physiology. Each phase may be acquired in a single breath-hold interval, thereby minimizing image degradation from respiratory motion.

Contrast-enhanced renal MRI requires a single bolus of gadolinium-based contrast administered at a standard dose of 0.1 mmol/kg at 2 to 3 mL/sec. The paramagnetic gadolinium chelates are filtered completely in the glomerulus, and the characteristic concentration-dependent change in tissue relaxation times that is induced provides insight into renal excretory kinetics. ^18,19 The T1-shortening effects of gadolinium produce a visibly increased T1-signal intensity. In the arterial phase, increased renal cortical signal intensity is observed as early as 10 to 20 seconds after contrast injection, with maximum intensity reached between 20 to 50 seconds post-injection. ²⁰ This imaging characteristic is explained by the heavy arterial perfusion of the renal cortex relative to the remaining renal parenchyma. With the onset of glomerular filtration, gadolinium is diluted in extravascular spaces, and renal cortical signal intensity decreases slowly and constantly.

Signal intensity changes in the renal medulla are different from those in the cortex, with increased signal intensity observed approximately 10 to 20 seconds later in the renal medulla than in the cortex. Maximum medullary signal intensity reaches values quantitatively equal to those in the cortex, and 30 to 40 seconds after the first visible signs of perfusion, medullary signal intensity sharply decreases to precontrast levels. ¹⁴ When gadolinium is dilute, T1 shortening occurs, creating high-signal-intensity urine. However, when its concentration is high (secondary to water resorption in the proximal convoluted tubules, loop of Henle, and collecting tubules), gadolinium induces signal loss, leading to low-signal-intensity urine.

These dynamic changes allow for study of the kidneys' concentrating ability with IV gadolinium chelates. ²¹ Baseline knowledge of functional contrast kinetics enhances the radiologist's understanding of MR image interpretation. Dynamic renal MRI can be useful for many nonvascular renal diagnoses, including those discussed below.

Renal cell carcinoma

MRI's ability to detect renal masses as small as 1 cm solidifies its role in early detection of renal cancer. Using a combination of breath-hold and fat-saturation techniques with IV gadolinium enables reliable differentiation between solid and cystic masses. ^4,16,22 Classic renal cell carcinoma appears as an irregular mass with ill-defined margination from normal renal parenchyma. Immediate arterial phase post-gadolinium images demonstrate heterogeneous enhancement reflecting the hypervascularity of the tumor (figure 4). Delayed images show diminished enhancement of the lesion. The low signal intensity of these neoplasms relative to parenchyma on delayed images reflects the physiology of the both the neoplasm and normal kidney, ie, contrast washout from the hypervascular tumor on a background of contrast-retaining renal tubules. ^18,22 Small, homogeneously enhancing tumors may be difficult to distinguish from normal cortex on immediate post-contrast images, a point that reinforces the necessity of routine serial image acquisition in the multiple postcontrast phases, as previously described. Although recent controlled studies have shown that MRI is slightly superior to CT in the detection, characterization, and staging of renal cell carcinoma, its higher cost, and still moderately time-consuming qualities (relative to CT) have prevented it from becoming the routine primary modality for diagnosis of renal masses. ¹⁶ The most widely accepted indications for MRI of renal masses continue to be renal failure, iodine allergy, and further analysis of masses considered indeterminate by CT.

Renal cysts

The renal cortical cyst is the most common renal mass in the adult patient. MRI characteristics of simple cysts are well established and include sharp margination from renal parenchyma, absent signal on T1-weighted images, homogeneously high signal on T2-weighted images, and no enhancement in any phase after gadolinium injection. More complex cysts with internal blood, calcification, or septations are also identified readily, although the pathognomonic lack of enhancement may be more difficult to discern.

Angiomyolipoma

Proper identification of these lesions is possible with both CT and MRI, although MRI may be able to detect and characterize somewhat smaller lesions than CT. Again, the role of dynamic MRI is usually reserved for the equivocal lesion. Angiomyolipomas are composed of blood vessels, smooth muscle, and fat, with the imaging characteristics dependent on the proportion of each component within the lesion. The typical angiomyolipoma has high fat relative to vascular content, creating a lesion with high noncontrast signal on T1-weighted images (that diminishes with fat saturation) and that is hypovascular on post-contrast imaging. ^2,12,22 Angiomyolipomas with low fat and high vascular content can be difficult to diagnose, as these appear hypervascular and have an enhancement pattern similar to that of renal cell carcinoma. To further complicate diagnosis, the rare renal cell carcinoma contains fat. These equivocal lesions are either removed surgically or followed closely with serial imaging, depending on the institution.

Renal cystic disease

Both CT and MRI readily identify the characteristic findings associated with both autosomal dominant polycystic kidney disease and acquired cystic disease of dialysis. The signal intensity of the cysts will vary on MRI depending on the presence of blood products of different ages and on the presence or absence of infection. The increased incidence of renal cell carcinoma in these patients is of extreme interest; but even with today's advanced techniques, differentiation between infected, hemorrhagic, and malignant cysts is often not possible with MRI, CT, or ultrasound because of the small size and high number of these cysts. ^2,12

Ischemia

Serial dynamic MRI can provide relative quantification of renal blood flow by comparing signal intensity increases between the kidneys following contrast injection. This provides the ability to diagnose both acute and chronic ischemia and highlights the utility of dynamic MRI for functional analysis of the kidney. ¹⁸ Both acute and chronic ischemia result in diminished contrast enhancement of the affected kidney relative to the normal one, though appropriate corticomedullary differentiation persists in both. The acutely ischemic kidney is enlarged, while the chronically ischemic kidney is small and shrunken.

Collecting system filling defects

Calculi are the most common filling defects in the renal collecting system and, regardless of the calcium content, appear as signal voids (black) on all MRI sequences. As discussed above, dilute gadolinium in the renal collecting system creates high-signal-intensity urine on T1-weighted images between 2 and 30 minutes after contrast injection, allowing for excellent contrast between signal-void stones and high-signal urine. Other entities such as fungus balls, neoplasia, and blood clots are similarly demonstrated. Complications of calculi, primarily renal obstruction, are also well depicted with dynamic imaging. The affected kidney is enlarged during acute obstruction, with resultant retained parenchymal contrast. The subsequent imaging findings are a prolonged nephrogram phase and diminished corticomedullary differentiation. ²³ The chronically obstructed kidney is small with decreased perfusion, leading to decreased cortical enhancement. The resultant increased transit time of the contrast agent leads to the delayed nephrogram. Renal cortical scarring and thinning that results from chronic reflux is best appreciated with MRI on immediate postcontrast images at which time corticomedullary differentiation is maximal.

Pancreas

Normal pancreas on unenhanced T1-weighted images has intermediate signal intensity--similar to liver--and is hypointense to surrounding fat. With T1-weighted fat-saturation, the pancreas becomes the highest-signal structure in the upper abdomen and is readily detectable. On nonenhanced T2-weighted images, the normal pancreas is iso- or hyperintense to liver. The pancreatic duct and common bile duct within the head of the pancreas exhibit very high signal on T2-weighted images. ^2,24 Regarding contrast dynamics, after bolus gadolinium-based contrast injection, the pancreas enhances before the liver. As in liver imaging, dynamic image acquisition is important for improved pancreatic lesion detection, characterization, and differential diagnosis. Unenhanced fat-suppressed T1-weighted and immediate (arterial) fat-suppressed T1-weighted post-contrast images are by consensus the most useful for diagnosis of pancreatic disease. ^25,26 In general, it is the aqueous content of the normal pancreas that creates high signal on T1-weighted images. Pancreatic tumors and chronic pancreatitis contain less aqueous tissue than normal pancreas, leading to their characteristically low signal intensity. The following is a brief description of MRI findings in pancreatic disease with an emphasis, where applicable, on the use of dynamic contrast-enhanced imaging.

Neoplastic disease

Pancreatic ductal adenocarcinoma accounts for 80% to 95% of pancreatic malignancies. These lesions are most common in the pancreatic head, followed by the body and tail. ^2,25 Typical MRI appearance is that of a low-signal-intensity mass within the high signal intensity of normal pancreatic tissue on T1-weighted images. Adenocarcinoma also usually has low signal intensity on T2-weighted sequences, a characteristic that can help with differentiation from islet cell tumors. The usual T1 tissue contrast of the adenocarcinoma can be decreased or even undetectable when coexisting carcinoma-induced pancreatitis is present. In such cases, there is an abnormally low signal intensity of the pancreas, with secondary atrophy, that often obscures the presence of the carcinoma. ^24,26 This combination reinforces the need for routine dynamic contrast-enhanced pancreatic MRI.

Almost all pancreatic adenocarcinomas are hypovascular and have scirrhous character secondary to the dense fibrotic tissue within the mass. As a result, these tumors are distinctly hypointense to pancreas during the arterial phase of dynamic enhancement, at which time the pancreas is maximally enhanced. The maximized contrast between tumor and pancreas during the arterial phase is what facilitates the detection and characterization of small pancreatic carcinomas that are often too small to alter the contour of the gland, and that may be missed on CT. ²⁶ On delayed images, contrast has leaked into the extracellular fluid compartment of the tumor, creating an iso- or hyperintense signal and a reduced lesion-to-parenchyma contrast differential on T1-weighted images. This reduced contrast leads to compromised tumor conspicuity and reinforces the utility of immediate post-contrast images.

Pancreatic neuroendocrine tumors are commonly referred to as islet cell tumors, the most common of the functional class being the insulinoma and gastrinoma, followed by the rarer glucagonomas, VIPomas, somatostatinomas, and GRFomas (growth hormone-releasing factor). ^2,24,26 Although these tumors have different clinical presentations, they share a common hypervascular nature that aids in their detection with imaging. These lesions have characteristically low signal on T1-weighted images and high signal on T2-weighted images relative to the pancreas. After IV bolus injection of gadolinium chelates, these lesions demonstrate early, intense, and usually homogeneous or ring enhancement (figure 5).

Cystic pancreatic neoplasms have a somewhat complex terminology and classification that has changed over the years. Current classification includes the mucinous cystic neoplasms (formerly macrocystic adenoma/adenocarcinoma), serous cystadenomas (microcystic adenomas), and intraductal mucin-producing pancreatic tumors composed of main duct and branch duct subtypes (ductectatic tumors). ²⁶ Aptly named, these tumors are primarily cystic though any soft-tissue or septated portions will often exhibit early arterial enhancement after gadolinium administration.

Inflammatory disease

Although early acute pancreatitis may be diagnosed by either CT or MRI, it remains primarily a clinical diagnosis, with CT and MRI used to evaluate for complications and to aid in difficult cases, such as ruling in or out concomitant malignancy. The MRI appearance of the uncomplicated and acutely inflamed pancreas is a normal signal intensity on T1-weighted fat-saturated images with normal arterial enhancement. ^25,26 Unenhanced T1-weighted fat-saturation images will show dusky peripancreatic fat (as on CT) consistent with inflammatory change. As acute pancreatitis progresses, the pancreas swells, unenhanced signal intensity falls on T1-weighted images, and the margins of the pancreas blur. ^18,25 In advanced cases, T2-weighted sequences identify peripancreatic fluid collections and pseudocysts easily. Pseudocysts appear as signal-void lesions on gadolinium-enhanced imaging. Contrast-enhanced imaging also aids in delineating focal areas of pancreatic necrosis that fail to enhance completely.

Chronic pancreatitis is defined as irreversible destruction of pancreatic parenchyma that is replaced by dense fibrosis and altered ductal architecture. ²⁶ On MRI, the pancreas exhibits decreased signal intensity, and there is decreased and heterogeneous enhancement on immediate postcontrast T1-weighted imaging. ²⁵ MRI findings include gland atrophy, irregular pancreatic ductal dilitation, pancreatic calcifications, chronic pseudocysts, focal pancreatic enlargement, and, occasionally, no abnormality. ^2,18,24 A particularly difficult differential diagnosis is between focal chronic pancreatitis and adenocarcinoma. Lack of a well-defined mass lesion on immediate post-gadolinium images combined with diffuse, lower-than-normal unenhanced pancreatic parenchymal signal on T1-weighted sequences favors chronic pancreatitis. ^18,24

Conclusion

The role of dynamic contrast-enhanced MRI of abdominal disease continues to expand as MR technology advances and as radiologists become more familiar with its utility in fields beyond musculoskeletal radiology and neuroradiology. As the speed and accessibility of MRI continue to increase, and as radiologists become more accustomed to abdominal MRI techniques, applications, and interpretation, abdominal MRI will no doubt complete the transformation that has already begun from niche modality to front-line tool of the abdominal imager.