Contrast-enhanced hepatic mri

CONTRAST-ENHANCED HEPATIC MRI: COMPARISON OF HALF-DOSE AND STANDARD-DOSE GADOLINIUM DTPA ADMINISTRATION IN LESION CHARACTERIZATION WITH T1-WEIGHTED GRADIENT ECHO SEQUENCES DOUGLAS R. DE CORATO, MD, GLENN A. KRINSKY MD, NEIL M. ROFSKY MD, JAMES P. EARLS MD, JONATHAN LEBOWITZ MD, JEFFERY C. WEINREB MD

The objective of this article was to compare halfdose (0.05 mm/kg) gadolinium-enhanced dynamic hepatic MR imaging to standard doses (0.10 mm/kg). Eighteen patients for follow-up hepatic MR received 0.05 mm/kg of gadolinium DTPA dynamically with gradient-echo imaging. Imaging parameters were identical to a 0.10-mm/kg study; patients were imaged during multiple phases of contrast enhancement. Two readers assessed for enhancement patterns and characterization. Quantitative signal-to-noise ratios (S/N) were obtained for abdominal viscera and contrast-to-noise ratios (C/N) were obtained on up to three lesions. No significant difference for the arterial dominant phase (P . 0.05) was found. Significant differences were found in all categories during the portal venous phase (except pancreas) and equilibrium phase (except liver). Lesion C/N ratios were not significant at any point (P . 0.05). Sixty-two out of 64 lesions (97%) were identically characterized. Therefore, half-dose dynamic gadolinium-enhanced MR may have diagnostic value.  Elsevier Science Inc., 2000

From the Department of Radiology, New York University Medical Center, New York, NY 10017. Address correspondence to: Dr. Douglas R. DeCorato, Department of Radiology, Memorial Sloan–Kettering Cancer Center, 1275 York Avenue, New York, NY 10021. Received March 1, 1999; accepted September 1, 1999. CLINICAL IMAGING 2000;23:302–310  Elsevier Science Inc., 2000. All rights reserved. 655 Avenue of the Americas, New York, NY 10010

KEY WORDS:

MRI; Comparison studies; Liver; Contrast dose; Gadolinium

INTRODUCTION Hepatic CT or MRI is commonly performed for the detection and characterization of focal hepatic masses. Extracellular contrast agents (i.e., gadolinium chelates or iodine-based compounds) are nonspecific and routinely used in this evaluation. Optimizing timing, dose, and rate of administration of iodinated contrast material has been extensively researched with both conventional and helical CT (1– 9). Studies have demonstrated that the dose of iodinated contrast material has the greatest effect on hepatic enhancement on CT, despite altered rates (10, 11). Gadolinium chelates, like the common iodinated contrast mediums, are extracellular agents, and are nonspecific in nature, yet are routinely used for lesion characterization. Gadolinium-enhanced hepatic MRI is usually performed with a “standard” dose of gadolinium chelates based on body weight (0.10 mmol/kg). This widely used dose has evolved primarily from the early literature in the evaluation of the central nervous system (CNS) with predominately spin-echo imaging (12, 13). Studies comparing different doses of gadolinium for hepatic MR imaging have been performed mainly in animal models (14–16). 0899-7071/00/$–see front matter PII S0899-7071(99)00150-3

SEPTEMBER/OCTOBER 1999

Technological advances in both hardware and software for MRI have significantly decreased imaging time allowing for complete coverage of the liver during a single breath hold, as with helical CT. Additionally, excellent T1-weighted contrast and signal-to-noise ratios (SNR) can be obtained with spoiled gradient-echo pulse sequences with large flip angles and phased array multicoils (17). Despite the improvements in technique, little interest has been paid to optimizing gadolinium chelates by attempting to lower the dose administered. Contrary to CT, larger doses of gadolinium chelates have been administered without additional diagnostic benefits (18, 19). Recently, attempts to lower doses of gadolinium (0.05 mm/kg) with magnetization transfer to evaluate intracranial neoplasms have been performed. This study concluded that a half-dose gadolinium-enhanced MR has limited utility in the evaluation of intracranial tumors (20). Hepatic hemodynamics are rather complex, with a variety of factors effecting blood flow and thus lesion appearance on both noncontrast and contrast-enhanced images (21). Clearly, it has been established for CT that the contrast dose is directly proportional to the hepatic enhancement (10, 11). In addition, decreased hepatic enhancement due to lower contrast doses may obscure lesions on the CT. MRI has superior tissue contrast and lesion conspicuity in comparison to CT. This fact led us to investigate whether a lower contrast dose, despite the expected decrease in hepatic enhancement, would allow for adequate lesion characterization based on enhancement patterns. MATERIALS AND METHODS Eighteen consecutive patients (12 men and 6 women), ranging in age from 44–80 years (mean 66 years), referred for follow-up hepatic MR over a 3-month period, comprise the patient population. All studies were performed at 1.5 T with 25 mTesla gradients and a 600-ms rise time (Seimens Vision Iselin NJ). Each patient had previously been studied with dynamic T1-weighted GRE breath-hold imaging using a standard dose (0.1 mm/kg) of gadopentetate dimeglumine (Magnevist Berlex Laboratories, Wayne, NJ). Subsequently, all patients underwent follow-up studies using low dose (0.05 mm/kg) contrast administration. The time interval between the two studies ranged from 42–504 days (mean 177.3 days 6 141.1 days, median 142.5 days). Routine hepatic imaging protocol includes a nonbreath-hold Turbo spin-echo (TSE) T2-weighted imaging sequence and either nonbreath-hold or breath-hold TURBO

HALF-DOSE CONTRAST-ENHANCED HEPATIC MRI

303

STIR imaging for lesion identification. T1-weighted imaging was obtained through the liver using an inphase spoiled gradient recalled echo sequence [FLASH (Fast Low Angle Shot)] acquired prior to, during, and after the dynamic administration of intravenous gadolinium in both studies. Imaging parameters for the FLASH sequences were held constant for each patient between the standard and lowdose studies. Parameters are as follows: TR/TE: 180– 220 ms/ 4.0–4.2 ms, flip angle: 80–908, rectangular field of view 300–400 mm, slice thickness: 7–10 mm, interslice gap: 0–3 mm and 128–196 phase encoding steps with 256 frequency encoding steps. Acquisition times ranged from 18 to 23 s. A quadrature phased array multicoil was used in 15 examinations, and the body coil was used in the remaining three to duplicate the original examination. Sixteen of 18 patients had documented primary malignancies: five had gastrointestinal malignancies, four had hepatocellular carcinoma (HCC), two had metastatic melanoma, one each had: metastatic sarcoma, breast carcinoma, renal cell carcinoma, germ cell tumor, and pancreatic carcinoma. Two patients had no known malignancy; one had multiple hemangiomas, the other pancreatitis. Biopsy proof was obtained in all of the HCCs. Five other patients had biopsies of at least one hepatic lesion, which were proven metastasis. In addition, two patients (not included in the above groups) demonstrated an increased number of lesions from the standard to the low-dose study. A third patient had a subsequent MRI, which revealed an increased number and size of lesions (the selected lesion for ROI analysis demonstrated interval growth). The remaining lesions were initially characterized by established criteria including the appearance on T2-weighted images as well as enhancement patterns. Contrast dose was based solely on the patients weight at the time of the examination and, for the reduced dose group, ranged from 5 to 10 cc (mean dose 8 6 1.4 cc) of gadopentetate dimeglumine. Contrast was hand injected through a 22-gauge catheter placed in the anticubital fossa in 15 patients, and through indwelling subclavian catheters in three patients. The contrast injection was immediately followed by a 20-cc saline flush. Dynamic imaging was performed as follows: contrast was injected as a bolus at approximately 2 cc/s followed by a 20-cc bolus of saline. The first postcontrast sequence was initiated 18 seconds after beginning the injection to approximate the hepatic arterial dominant phase. The second phase (portal venous) was acquired approximately 45 seconds after the

304

D.R. DE CORATO ET AL.

start of the injection. Additional images were taken at 120 and 180 seconds (equilibrium phase). Quantitative analysis of signal intensity (SI) was performed with user-defined regions of interest (ROI) drawn over: avascular portions of hepatic parenchyma, spleen, pancreas, renal cortex, aorta, and portal vein. Systemic noise was obtained by an ROI drawn in the upper right-hand corner of the image of the abdomen, away from the phase encoding artifact of the heart. The standard deviation (Std) for that value was utilized as the denominator, thus yielding signal-to-noise ratios (SNR). Signal intensity for lesions was obtained over a homogeneously enhancing portion of the lesion when present. Identical lesions were chosen from each study. When only one lesion was present on the initial study, the same single lesion was used for further analysis. Contrast-to-noise ratios (CNR) for up to three lesions per patient were determined in the following manor: the absolute value of [(Liver SI 2 Lesion SI)/Std noise]. Two readers blinded to dosage, patient history, demographics, and the results of the TSE T2 and TURBO STIR images performed the qualitative analysis in the following manner: The readers reviewed all studies on hard copy by consensus. The readers were given the pre contrast images followed by all four postcontrast images (standard dose and low dose studies were randomly dispersed). They were asked to identify the number of lesions and enhancement patterns, and to characterize the lesions based on enhancement patterns alone. Enhancement patterns were categorized as follows: none, peripheral nodular, ring, irregular (spiculated), or homogenous blush. The following enhancement patterns were considered malignant for the purpose of this study: ring enhancement, with or without peripheral washout, hypervascular homogeneous enhancement with rapid washout (no patients had FHN or adenomas that would demonstrate similar enhancement), and complete ring enhancement (either thick or irregular) during the arterial dominant phase. Benign patterns of enhancement included peripheral nodular enhancement with centripetal filling on delayed images, hypervascular homogenous enhancement with retained contrast paralleling the portal veins, and complete lack of contrast enhancement. Confidence in diagnosis was recorded as follows: (1 5 definitely benign, 2 5 probably benign, 3 5 uncertain, 4 5 probably malignant, and 5 5 definitely malignant). The number of lesions was identified; if more than 10 were present, a score of 10 was assigned to the patient. One patient with diffusely infiltrative disease was not assigned a lesion number. In addition, the readers were asked if

CLINICAL IMAGING VOL. 23, NO. 5

a successful arterial dominant phase was obtained according to published criteria (22).

Statistical Analysis A paired student’s t-test was used to determine statistical significance. The following variables were compared: SNR ratios for liver, spleen, renal cortex, pancreas, aorta and portal vein, and CNR ratios of focal hepatic lesions. Forty-two matched lesions were used for statistical analysis. RESULTS

Quantitative The quantitative data is summarized in Table 1. Evaluation of abdominal viscera and vascular structures yielded no significant differences for the precontrast and arterial dominant phase images despite administered dose. Significant difference for portal venous phase imaging was seen in the following abdominal viscera and vessels: liver, spleen, renal cortex, aorta, and portal vein. Only the pancreas failed to reach significance during the portal venous phase. The third postcontrast acquisition failed to demonstrated significant differences for the liver, renal cortex, pancreas, and portal vein. The only values of significance during this phase of enhancement are the spleen and aorta. Additionally, the equilibrium phase demonstrated significance for the following structures: spleen, renal cortex, pancreas, aorta, and portal vein. Liver was the only organ to fail to reach a significant difference in enhancement during the equilibrium phase. Mean lesion C/N ratios were not significantly different though out the study. Mean S/N ratios based on dose of the abdominal viscera and vascular structure are presented graphically in Figure 1.

Qualitative Analysis The consensus review demonstrated 64 distinct lesions on the standard dose study. Eighty-one lesions were identified on the low dose study, and as such, 17 interval lesions were noted (in no patient were less lesions present on the low-dose study). One patient had a single lesion identified on the standard dose study. The low-dose study, which was performed 8 months later, demonstrated greater than 10 lesions. The remaining eight interval lesions were noted in five of the remaining 17 patients. Lesion characterization based on enhancement patterns is summarized in Table 2 , and the enhance-

SEPTEMBER/OCTOBER 1999

HALF-DOSE CONTRAST-ENHANCED HEPATIC MRI

305

TABLE 1. Signal-to-Noise and Contrast-to-Noise Ratios for the Multiple Phases of Contrast Enhancement Liver Pre low Pre std P-value Art low Art std P-value PV low PV std P-value 3rd low 3rd std P-value 4th low 4th std P-value

P

P

P

P

P

39.9 47.7 5 0.10 40.5 45.3 . 0.15 42.2 55.0 , 0.02 43.8 52.3 . 0.15 42.2 54.3 . 0.05

Renal cortex

Spleen

P

P

P

P

P

31.4 34.6 . 0.25 49.3 58.9 . 0.25 46.8 66.6 , 0.01 45.5 60.8 5 0.05 43.1 63.2 , 0.01

P

P

P

P

P

33.2 38.2 5 0.10 62.3 72.4 . 0.30 58.3 74.9 , 0.03 59.1 52.8 . 0.20 56.9 73.7 , 0.01

Panc.

Aorta

41.3 45.7 P . 0.30 51.22 57.9 P . 0.30 46.5 56.0 P . 0.07 48.7 45.4 P . 0.45 47.3 56.7 P , 0.02

23.5 22.8 P 5 0.75 86.3 104.8 P 5 0.10 68.7 81.1 P , 0.01 64.6 73.3 P , 0.05 60.7 70.9 P , 0.03

P

P

P

P

P

Portal Vein

CNR Les. 1

CNR Les. 2

24.7 23.6 . 0.65 41.0 46.2 . 0.35 54.6 71.9 , 0.03 53.5 65.3 . 0.10 49.2 64.0 , 0.03

15.0 17.0 . 0.55 12.8 17.2 . 0.30 14.1 21.3 . 0.10 13.6 22.7 . 0.10 13.1 20.6 . 0.05

15.7 22.6 . 0.05 15.4 30.3 . 0.10 17.5 30.8 . 0.06 11.7 24.7 . 0.15 10.1 25.7 . 0.20

P

P

P

P

P

P

P

P

P

P

CNR Les. 3 20.6 18.8 P . 0.45 19.6 31.98 P . 0.30 21.8 29.4 P . 0.45 17.7 24.7 P . 0.45 14.2 20.4 P . 0.35

Pre: precontrast exam, Art: arterial dominant phase, PV: portal venous phase, 3rd: third post-gadolinium phase, 4th: fourth post-contrast phase; low: lowdose study (0.05 mm/kg). Std: standard dose study (0.1 mm/kg); Panc.: pancreas; les.: lesion.

ment patterns themselves are summarized in Table 3. Two lesions in one patient demonstrated classic peripheral nodular enhancement on both studies with filling of the lesion from the periphery, and were diagnosed as hemangiomas (Figure 2 ). Two out of 64 lesions (3.0%) were not classified identically with respect to patterns of enhancement or whether benign or malignant. One of these lesions demonstrated a homogenous arterial blush on the standard dose study. This initially was incorrectly thought to be a hemangioma with flash filling. On the subsequent low-dose study, the patient had the appearance of an interval lesion. The original lesion and the new lesion both demonstrated ring enhancement, and were considered definitely malignant (Figure 3). One patient had a single lesion on the standard dose study, and two lesions identified on the low-dose study. The original lesion was classified as having a homogenous blush on the standard dose study, but was classified as uncertain in nature. The same patient had two lesions on the low-dose study performed 2 months later. The previously seen lesion subsequently demonstrated ring enhancement, and the second (interval) lesion was described as having irregular enhancement; both lesions were again characterized as uncertain. All other lesions were matched as to pattern of enhancement and to their status. All hepatocellular carcinomas, which are commonly hypervascular lesions, were characterized identically (all HCCs were biopsy proven) as malignant (Figure 4). Evaluations of the arterial-dominant phase images were as follows: 4 of 18 studies (22%) using the standard dose had suboptimal arterial dominant phase

(two early and two late), while 3 of 18 (17%) lowdose studies were suboptimal (one early, two late). DISCUSSION The role of MRI in many practices is to characterize hepatic lesions detected on other imaging modalities. This characterization is based on a variety of factors including, but not limited to, lesion appearance: on T2weighted images, on heavily T2-weighted images, T1weighted images, and enhancement patterns. Lesion characterization with intravenous extracellular contrast agents is often used, and multiple studies have validated the efficacy of that approach (23–30). The dose of gadolinium chelates administered for hepatic MR imaging is 0.1 mm/kg, based primarily on the neurological imaging experience. Larger doses of contrast have been safely administered without apparent diagnostic benefit for hepatic imaging (18, 19). Clinically useful doses of contrast material must yield recognizable enhancement characteristics while maintaining adequate lesion visualization. Enhancement patterns alone are not always diagnostic, as definite overlap exists between benign and malignant lesions: hypervascular lesions with rapid washout can be seen with focal nodular hyperplasia, adenomas, and HCC. Despite this overlap, enhancement patterns or lack thereof, remain important for lesion evaluation. To modify the clinical dose of an agent, recognized enhancement patterns should not be altered. In our study, statistically significant differences were present in quantitative organ and vascular enhancement during several phases of enhancement

306

D.R. DE CORATO ET AL.

CLINICAL IMAGING VOL. 23, NO. 5

FIGURE 1. Mean signal-to-noise ratios based on dose and phase of enhancement (A) Aortic enhancement; (B) portal venous enhancement; (C) hepatic parenchymal enhancement; (D) pancreatic enhancement; (E) splenic enhancement; (F) renal cortical enhancement. (Note the scale for aortic enhancement is twice the values of the other scales.) Solid black bar standard dose (0.10mm/kg); Solid light gray bar low dose (0.05mm/kg). Pre: precontrast S/N ratios. Art: Arterial dominant phase. Portal: Portal venous phase images. 3rd and 4th are equilibrium phases.

between the standard and reduced doses. Specifically, during the portal venous phase, where a majority of hepatic lesions are identified on CT. This finding is supported by previous data demonstrating a dose-dependent enhancement for organs after the administration of intravenous gadolinium (31, 32). Despite significant differences in visceral enhancement, C/N ratios of the lesions were not significant.

Additionally, the reduced dose of gadolinium DTPA (0.05 mm/kg) did not adversely effect lesion enhancement patterns or characterization. Many routine hepatic CTs are performed as uniphasic examinations, thus requiring maximal differences during that single phase. MR uses different pulse sequences to maximize lesions conspicuity prior to the administration of contrast. Contrast enhancement patterns

TABLE 2. Summary of Lesion characterization Based on Enhancement Patterns by Dose

Standard dose Low dose

Definitely malignant

Probably malignant

Uncertain

Probably benign

Definitely benign

41 58

4 4

1 2

0 0

18 18

SEPTEMBER/OCTOBER 1999

HALF-DOSE CONTRAST-ENHANCED HEPATIC MRI

307

TABLE 3. Summary of the Lesion Enhancement Patterns Identified by Dose

Standard dose Low dose

Irregular

Ring

Arterial blush

Peripheral modular

None

13 26

26 31

8 6

2 2

15 16

are used to aid in characterization as opposed to aid in identification. Lower administered doses of gadolinium chelates have been performed in other areas of MRI. Recently, lower doses of contrast are being used in gadolinium-enhanced MR angiography, as the technique has evolved, especially with the shorter acquisition times now achievable (33, 34). Contrast-enhanced MRA and hepatic imaging utilize different hemodynamics; despite this, lowering doses in hepatic imaging yielded similar results. There are potential advantages to lowering the dose of contrast material, the most obvious being a decrease in the cost of the examination. Contrast cost for hepatic MR has decreased recently, thus making this advantage less important. Smaller doses, however, could theoretically allow for a tighter bolus being delivered within the central portion of the k space. This may allow better capture of the arterial dominant phase, before significant portal venous enhancement occurs. Additionally, as newer contrast agents become available, lowering the dose of gadolinium may allow for dual agent scanning. Semelka et al. (35) have recently shown that scanning with both T1 and T2 contrast agents is possible, and may be complementary, further increasing diagnostic

yield. Further work evaluating this technique is needed, as well as optimizing the doses required for each agent. Additionally, lower doses of gadolinium chelates may improve the economic feasibility of this approach. Finally, if gadolinium-enhanced MR angiography is to be performed during the same study, the lower dose used for dedicated hepatic MR will result in less background enhancement. There are several limitations of this study. Biopsy proof was available on all hepatocellular carcinomas, histological proof was also available for five other lesions, totaling 9 of 16 patients with malignant disease (56%). Other lesions were proven by either stability on follow-up, increased size and number on subsequent examinations, or imaging characteristics. The lack of histologic proof of all lesions hinders assessment of the accuracy of characterization based on enhancement patterns; however, this does not alter the fact that characterization was almost identical. Atypical appearances of lesions based on enhancement have been documented (36, 37), and this study was not designed to yield a definitive answer based solely on one criterion. We set out to determine if characterization would be similar with lower doses of intravenous contrast material. In addition, diagnostic confidence of enhancement pat-

FIGURE 2. Hepatic hemangioma (initial images not shown demonstrated peripheral nodular enhancement) equilibrium phase images show peripheral discontinuous enhancement (arrow). A hepatic cyst is also present (arrowhead). (A) Standard dose (0.10 mm/kg); (B) low-dose study (0.05 mm/kg) 7.0 cc of contrast material at a similar level.

308

D.R. DE CORATO ET AL.

CLINICAL IMAGING VOL. 23, NO. 5

FIGURE 3. Biopsy-proven metastasis from renal cell carcinoma during the arterial dominant phase. (A) Standard dose (0.10 mmol/kg) study demonstrates a lesion with homogenous enhancement characterized as uncertain. (B) Low-dose study (0.05 mmol/kg) after 10.0 cc of gadolinium DTPA demonstrates the lesion to have slightly increased in size. This lesion was characterized as definitely malignant with ring enhancement.

terns was not altered by a lower dose. This was important for the following reason: if lesion enhancement patterns were identical but reader confidence was severely affected, then lowering the dose of administered contrast could have grave consequences. A second limitation to this study was the time interval between patients. Ideally, a much shorter time interval between studies should be obtained. Larger spans of time may allow slow growing masses to reach a “critical size,” permitting improved detection and characterization. This point may account for the altered characterization of one of the discrepancies (Figure 3). A related point is that all low-dose

studies were performed after the standard dose studies. This lack of randomization can lead to bias. A large prospective randomized trial based on several doses including low doses to quantitate the optimal dose should be performed. At this time no such study exists; however, our preliminary work would suggest that lower doses of contrast might be feasible, despite significant differences in hepatic enhancement. Another limitation is that the patient population in our study was biased towards malignant disease. Forty-one definitely malignant lesions were identified on the standard dose study and 57 on the low-dose study, with 18 lesions characterized as

FIGURE 4. Recurrent hepatocellular carcinoma in a patient with prior embolization. Arterial-dominant phase images demonstrate enhancing tumor nodules (arrow). (A) Standard dose study (0.10 mmol/kg) (B) Low-dose study (0.05 mmol/kg) after 8.5 cc of gadolinium.

SEPTEMBER/OCTOBER 1999

definitely benign on both studies. As is typically the case, once diagnosed, benign lesions are less likely to be followed, and if followed, may be followed with more cost efficient studies such as ultrasonography. This small number of patients with benign disease could clearly bias the observers toward a definitely malignant characterization without adequate evaluation. However, our study was designed to require both the pattern of enhancement and characterization based on that pattern to attempt to minimize this bias. It should also be noted that none of the patients with cirrhosis had demonstrable arterial portal shunting (transient hepatic attenuation defects); in this situation, reducing the dose of contrast could decreased the diagnostic yield of hypervascular lesions. The relatively small sample size is also a limitation; although lesion C/N ratios approached but did not reach significance, a larger sample size may alter these results, and as such requires further study. In conclusion, preliminary results demonstrate no significant difference in enhancement patterns or characterization of lesions based on a reduced dose of gadolinium. A significant difference was noted in hepatic enhancement, especially in the portal venous phase. A dose between our chosen 0.05 mmol/ kg and the standard dose would likely improve hepatic enhancement with similar results in lesion enhancement. Further investigation is warranted in determining the optimal dose, concentration, and injection rates for dynamic gadolinium-enhanced hepatic MR imaging.

REFERENCES 1. Baron RL. Understanding and optimizing use of contrast material for CT of the liver. AJR 1994;163:323–331. 2. Mitsuzaki K, Yamashita Y, Ogata I, Nishiharu T, Urata, J, Takahashi M. Multiple-Phase helical CT of the liver for detecting small hepatomas in patients with liver cirrhosis: Contrast-injection protocol and optimal timing. AJR 1996; 167:753–757. 3. Garcia PA, Bonaldi VM, Bret PM, Liang L, Reinhold C, Atri M. Effect of rate of contrast medium injection on hepatic enhancement at CT. Radiology 1996;199:185–189. 4. Bluemke DA, Soyer P, Fishman EK. Helical (spiral) CT of the liver. Radiol Clin North Am 1995;33:863–886. 5. Silverman PM, Roberts SC, Ducic I, et al. Assessment of a technology that permits individualized scan delays on helical hepatic CT: A technique to improve efficiency in use of contrast material. AJR 1996;167:79–84. 6. Fredrick MG, McElaney BL, Singer A, et al. Timing of parenchymal enhancement on dual-phase dynamic helical CT of the liver: How long does the arterial phase predominate. AJR 1996;166:1305–1310. 7. Silverman PM, Cooper C, Trock B, Garra B, Davaros WJ, Zeman RK. The optimal temporal window for CT of the liver us-

HALF-DOSE CONTRAST-ENHANCED HEPATIC MRI

309

ing a time-density analysis: Implications for helical (spiral) CT. J Comput Assist Tomogr 1995;19:73–79. 8. Bluemke DA, Fishman EK, Anderson JH. Dose Requirements for a nonionic contrast agent for spiral computed tomography of the liver in rabbits. Invest Radiol 1994;29:195–200. 9. Berland LL, Lee JY. Comparison of contrast media injection rates and volumes for hepatic dynamic incremental computed tomography. Invest Radiol 1988;23:918–922. 10. Chambers TP, Baron RL, Lush RM. Hepatic CT enhancement Part I. Alterations in the volume of contrast material within the same patients. Radiology 1994;193:513–517. 11. Chambers TP, Baron RL, Lush RM. Hepatic CT enhancement Part II. Alterations in the volume of contrast material and rate of injection within the same patients. Radiology 1994; 193:518–522. 12. Neindorf HP, Lanaido M, Semmler W, Schorner W, Flex R. Dose administration of gadolinium DTPA in MR imaging of intracranial tumors. AJNR 1987;8:803–815. 13. Laniado M, Weinmann HJ, Schorner W, Felix R, Speck U. First use of gadolinium DTPA in man. Physiol Chem Phys Med NMR 1984;16:157–165. 14. Brasch RC, Weinmann H, Wesber GE. Contrast-enhanced NMR imaging: Animal studies using gadolinium DTPA complex. AJR 1984;142:625–630. 15. Runge VM, Kirsch JE, Woolfolk C, Brack MA, Garneau RA. Gadoteridol dose dependence in MR imaging of a liver abscess model. J Magn Reson Imaging 1994;4:343–350. 16. Reimer P, Saini S, Kwong KK, Cohen MS, Weissleder R, Brady T. Dynamic gadolinium-enhanced echo-planar MR imaging of the liver: Effect of pulse sequence and dose on enhancement. J Magn Reson Imaging 1994;4:1–5. 17. Campeau NG, Johnson CD, Felmlee JP, et al. MR imaging of the abdomen with a phased-array multicoil: Prospective clinical evaluation. Radiology 1995;195:769–776. 18. Petersein J, Saini S, Mitchell DG, et al. Gadoteridol-enhanced MR imaging of malignant hepatic tumors: Effect of triple dose versus standard doses on lesion liver contrast. AJR 1995; 165:1157–1161. 19. Mitchell DG, Saini S, Weinreb JC, et al. Hepatic metastasis and cavernous hemangiomas: Distinction with standard- and triple-dose gadoteridol-enhanced MR imaging. Radiology 1994;193:49–57. 20. Han D, Chang K-H, Han MH, et al. Half-dose gadoliniumenhanced MR imaging with magnetization transfer technique: Comparison with conventional contrast-enhanced MR imaging. AJR 1998;170:189–193. 21. Itai Y, Matsui O. Blood flow and liver imaging. Radiology 1997;202:306–14. 22. Mitchell DG, Liver I: Currently available gadolinium chelates. MRI Clin North Am 1996;4:37–51. 23. Hamm B, Mahfouz A, Taupitz M, et al. Liver metastases: Improved detection with dynamic gadolinium-enhanced MR imaging. Radiology 1997;202:677–682. 24. Schmiedel U, Kolbel G, Hess CF, Klose U, Kurtz B. Dynamic sequential MR imaging of focal liver lesions: Initial experience in 22 patients at 1.5 T. J Comput Assist Tomogr 1990; 14:600–607. 25. Hamm B, Fischer E, Taupitz M. Differentiation of hepatic hemangiomas from metastases by dynamic contrast-enhanced MR imaging. J Comput Assist Tomogr 1990;14:205–216. 26. Ohtomo K, Itai Y, Yoshikawa K, et al. Hepatic tumors: Dynamic MR imaging. Radiology 1987;163:27–31. 27. Semelka RC, Shoenut JP, Kroeker MA, et al. Focal liver diseases: Comparison of dynamic contrast enhanced CT and T2-

310

D.R. DE CORATO ET AL.

CLINICAL IMAGING VOL. 23, NO. 5

weighted fat-suppressed, FLASH, and dynamic gadoliniumenhanced MR imaging at 1.5 T. Radiology 1992;184:687–694. 28. Semelka RC, Shoenut JP, Ascher SM, et al. Solitary hepatic metastasis: Comparison of dynamic contrast-enhanced CT and MR imaging with fat-suppressed T2-weighted, breathhold T1-weighted FLASH, and dynamic gadolinium-enhanced FLASH sequences. J Magn Reson Imaging 1994;4:319–323. 29. Mahfouz A, Hamm B, Taupitz M, Wolf K. Hypervascular liver lesions: Differentiation of focal nodular hyperplasia from malignant tumors with dynamic gadolinium-enhanced MR imaging. Radiology 1993;186:133–138. 30. Hamm B, Thoeni RF, Gould RG, et al. Focal liver lesion: Characterization with nonenhanced and dynamic contrast material-enhanced MR imaging. Radiology 1994;190:417–423. 31. Tweedle MF, Wedeking P, Telser J, et al. Dependence of MR signal intensity on GD tissue concentration over a broad dose range. Magn Reson Med 1991;22:191–194. 32. Wedeking P, Sotak CH, Telser J, Kumar K, Chang CA, Tweedle MF. Quantitative dependence of MR signal intensity on tissue

33.

34.

35.

36.

37.

concentration of Gd(HP-DO3A) in the nephrectomized rat. Magn Reson Imaging 1992;10:97–108. Earls JP, Rofsky NM, DeCorato DR, Krinsky GA, Weinreb JC. Optimization of the hepatic arterial phase in dynamic gadolinium-enhanced MR using a test dose and power injector. Radiology 1997;202:268–273. Rofsky NM, Johnson G, Adelman MA, Rosen RJ, Krinsky GA, Weinreb JC. Peripheral vascular disease evaluated with reduced dose gadolinium-enhanced MR angiography. Radiology 1997;205:163–169. Semelka RC, Lee JKT, Worawattankul S, Noone TC, Patt RH, Ascher SM. Sequential use of ferumoxide particles and gadolinium chelate for the evaluation of focal liver lesions. J Magn Reson Imaging 1998;8:670–674. Hanafusa K, Ohashi I, Himeno Y, Suzuki S, Shibuya H. Hepatic hemangioma: Findings with two-phase CT. Radiology 1995;196:465–469. Choi CS, Freeny PC. Triphasic helical CT of hepatic focal nodular hyperplasia: Incidence of atypical findings. AJR 1998: 170:391–395.