Russell N. Low, MD Richard C. Semelka, MD Suvipapun Worawattanakul, MD Gregg D. Alzate, MD Joel S. Sigeti, MD

Index terms: Abdomen, neoplasms, **.32,2 **.332 Magnetic resonance (MR), comparative studies Neoplasms, CT, **.121152 Neoplasms, MR, **.12142,**.1214152, 2 **.12143 Radiology 1999; 210:625–632 Abbreviations: FMPSPGR ⫽ fast multiplanar spoiled gradient-recalled GRE ⫽ gradient echo RARE ⫽ rapid acquisition with relaxation enhancement TE ⫽ echo time TR ⫽ repetition time 1

From the Department of Radiology, Sharp and Children’s MRI Center, Sharp Memorial Hospital, 7901 Frost St, San Diego, CA 92123 (R.N.L., G.D.A., J.S.S.), and the Department of Radiology, University of North Carolina at Chapel Hill (R.C.S., S.W.). Received January 12, 1998; revision requested April 2; final revision received July 29; accepted October 7. Address reprint requests to R.N.L.

2

**. Multiple body systems

r RSNA, 1999

Author contributions: Guarantors of integrity of entire study, R.N.L., R.C.S.; study concepts, R.N.L., R.C.S.; study design, R.N.L.; definition of intellectual content, R.N.L., R.C.S.; literature research, R.N.L., R.C.S.; clinical studies, R.N.L., R.C.S., S.W., G.D.A., J.S.S.; data acquisition, R.N.L., S.W., G.D.A., J.S.S.; data analysis, R.N.L., S.W.; statistical analysis, R.N.L.; manuscript preparation, R.N.L.; manuscript editing, R.N.L., R.C.S.; manuscript review, R.N.L., R.C.S., G.D.A.

Extrahepatic Abdominal Imaging in Patients with Malignancy: Comparison of MR Imaging and Helical CT, with Subsequent Surgical Correlation1 PURPOSE: To compare state-of-the-art magnetic resonance (MR) imaging with single-phase helical computed tomography (CT) in abdominal screening for extrahepatic disease in patients with proved malignancy. MATERIALS AND METHODS: Fifty-seven patients with known malignancy underwent abdominal contrast material–enhanced helical CT and MR imaging from 1994 through 1997. Prospective interpretations of CT scans and MR images were used to assess each modality’s sensitivity in depicting malignant extrahepatic tumor at 17 anatomic sites. Imaging findings were compared with surgical results in all patients. RESULTS: Helical CT depicted 101 (66%) of 154 surgically confirmed extrahepatic tumor sites; MR imaging depicted 139 (90%) (P ⬍ .001). MR imaging depicted tumor in more patients at 11 of the 17 anatomic sites; at six sites, MR imaging and helical CT were equivalent. MR imaging showed significantly greater depiction of extrahepatic tumor for the peritoneum (P ⬍ .05), bowel (P ⬍ .01), and mesentery (P ⬍ .05). False-negative interpretations would have altered patient care had the extrahepatic tumor remained undetected in 13 patients for helical CT and in six patients for MR imaging. CONCLUSION: State-of-the-art MR imaging can be used for effective abdominal screening for extrahepatic tumor in patients with malignancy. Compared with single-phase helical CT, MR imaging depicted more sites of extrahepatic tumor and was particularly advantageous for the peritoneum, mesentery, and bowel.

For the evaluation of patients with malignancy, computed tomographic (CT) scanning has long been established as the predominant imaging modality for depiction of abdominal tumor (1–9). In this clinical setting, oncologists and surgeons depend on imaging tests to accurately depict both hepatic and extrahepatic tumor. Although magnetic resonance (MR) imaging has been successful in depicting and characterizing hepatic malignancy (10–14), the widespread use of MR imaging as a screening test has been limited by its perceived inability to depict extrahepatic tumor. Because of this limitation, abdominal MR imaging is most often used as a problem-solving examination to address questions left unanswered by helical CT scanning. At most institutions, abdominal MR imaging is rarely used as the initial imaging test in patients with malignancy. Advances in MR imaging, including faster pulse sequences, breath-hold imaging, use of intravenously administered contrast agents, and surface coils, have improved the quality of images and shortened examination times (15–19). The use of MR imaging for depiction of extrahepatic tumor involving the peritoneum, gastrointestinal tract, and visceral organs has been reported in separate studies (20–26). In this study, we investigated how newer MR imaging sequences perform, compared with helical CT scanning, in depicting all forms of 625

extrahepatic abdominal tumor in 57 patients with malignancy who subsequently underwent surgical exploration.

MATERIALS AND METHODS Patients At two separate institutions (Sharp Memorial Hospital/Sharp and Children’s MRI Center, San Diego, Calif; and University of North Carolina at Chapel Hill), 57 patients with surgical proof of abdominal malignancy underwent helical CT scanning and MR imaging of the abdomen from 1994 through 1997. The 57 patients included 30 men and 27 women, with a mean age of 58 years (range, 28–85 years). All 57 patients had confirmed malignancy, with a history of colon carcinoma (14 patients), lung cancer (one patient), ovarian carcinoma (11 patients), renal cell carcinoma (three patients), pancreatic carcinoma (six patients), gastric adenocarcinoma (five patients), gastric leiomyoblastoma (one patient), esophageal carcinoma (one patient), hepatocellular carcinoma (two patients), cholangiocarcinoma (one patient), ampullary carcinoma (two patients), islet cell tumor (one patient), pheochromocytoma (one patient), endometrial carcinoma (one patient), leiomyosarcoma (one patient), pseudomyxoma peritonei (two patients), adenocarcinoma of unknown primary site (one patient), lymphoma (one patient), retroperitoneal sarcoma (one patient), or bladder carcinoma (one patient).

a.

Imaging Examination and Image Interpretation All patients were examined with singlephase helical CT and MR imaging. The two examinations were performed within 4 weeks, with a mean interval between the two examinations of 7 days (range, 0–27 days). In 36 patients, the CT scan was obtained before the MR examination, and in the remaining 21 patients, the MR examination was the initial study performed. Thirty-six patients were examined at institution 1, and 21 patients were examined at institution 2. At institution 1, a total of 15 of the MR and helical CT examinations included the abdomen, and the remaining 21 examinations included the abdomen and pelvis. At institution 2, a total of 12 MR and helical CT examinations included the abdomen, and the remaining nine examinations included the abdomen and pelvis. Indications for the imaging investigation were as follows: primary staging of 626 • Radiology • March 1999

b. Figure 1. Images of the abdomen of a 70-year-old man with metastatic carcinoma of the colon. (a) Helical CT scans at two contiguous levels, with 5-mm collimation, show metastasis in the right hepatic lobe (curved arrow) and a thin rim of perihepatic ascites (long straight arrows). Isoattenuating perihepatic soft tissue (short straight arrow) is poorly defined. (b) Gadolinium chelate–enhanced FMPSPGR MR images (140/2.6, 70° flip angle) with fat suppression. Left: Portal venous phase image shows hypointense metastasis in the right hepatic lobe. Note the abnormal enhancement of the right subphrenic peritoneum (arrow). Right: Delayed equilibrium-phase FMPSPGR image is from a section 1 cm lower than the image at left. Note the thick rind of enhancing right subphrenic and perihepatic peritoneal tumor (arrows). Surgical findings of extensive peritoneal metastatic tumor correlated with the tumor extent in b.

intraabdominal malignancy (41 patients), restaging of abdominal malignancy after treatment or tumor recurrence or both (five patients), second-look laparotomy for ovarian cancer (seven patients), and surgical exploration and restaging before intraperitoneal chemotherapy (four pa-

tients). In 14 patients, sequential helical CT and MR imaging were performed before surgery as part of a study to determine the presence and extent of peritoneal tumor in patients with intestinal or gynecologic malignancy (20). At institution 1, helical CT scans were Low et al

Depiction of Extrahepatic Disease in 57 Patients with Malignancy: Helical CT versus MR Imaging Site and Imaging Technique Gallbladder and biliary system Helical CT MR imaging Pancreas Helical CT MR imaging Spleen Helical CT MR imaging Kidneys Helical CT MR imaging Adrenal glands Helical CT MR imaging Peritoneum Helical CT MR imaging Omentum Helical CT MR imaging Mesentery Helical CT MR imaging Bowel Helical CT MR imaging Retroperitoneum Helical CT MR imaging Abdominal wall Helical CT MR imaging Lymph nodes Helical CT MR imaging Ascites Helical CT MR imaging Lung bases Helical CT MR imaging Bones Helical CT MR imaging Vascular structures Helical CT MR imaging Pelvis Helical CT MR imaging Total Helical CT MR imaging

TP

FN

FP

TN

Sensitivity

Specificity

Accuracy

10 13

3 0

0 0

44 44

0.77 1.0

1.0 1.0

0.95 1.0

11 12

3 2

0 0

43 43

0.79 0.86

1.0 1.0

0.95 0.96

4 6

2 0

1 0

50 51

0.67 1.0

0.98 1.0

0.95 1.0

3 3

0 0

0 0

54 54

1.0 1.0

1.0 1.0

1.0 1.0

3 3

0 0

0 0

54 54

1.0 1.0

1.0 1.0

1.0 1.0

11 19

8 0

4 2

34 36

0.58 1.0*

0.89 0.95

0.79 0.96†

8 11

3 0

0 0

46 46

0.73 1.0

1.0 1.0

0.95 1.0

2 8

9 3

1 1

45 45

0.18 0.73*

0.98 0.98

0.82 0.93

15 24

14 5

1 3

27 25

0.52 0.83†

0.96 0.89

0.74 0.88*

3 3

0 0

0 0

54 54

1.0 1.0

1.0 1.0

1.0 1.0

2 3

1 0

0 1

54 53

0.67 1.0

1.0 0.98

0.98 0.98

9 13

4 0

1 0

43 44

0.69 1.0

0.98 1.0

0.91 1.0

7 8

3 2

0 0

47 47

0.70 0.80

1.0 1.0

0.95 0.96

2 2

1 1

0 0

54 54

0.67 0.67

1.0 1.0

0.98 0.98

0 0

0 0

0 1

57 56

NA‡ NA‡

1.0 0.98

1.0 0.98

3 3

1 1

0 0

53 53

0.75 0.75

1.0 1.0

0.98 0.98

8 8

1 1

1 1

47 47

0.89 0.89

0.98 0.98

0.96 0.96

101 139

53 15

9 9

806 806

0.66 0.90§

0.99 0.99

0.94 0.98

Note.—Fifty-seven patients with malignancy and surgical proof of extrahepatic tumor were evaluated with helical CT and MR imaging. Numbers represent the number of true-positive (TP), false-negative (FN), false-positive (FP), and true-negative (TN) prospective interpretations of malignant extrahepatic disease. *P ⬍ .05. †P ⬍ .01. ‡NA ⫽ not applicable, no true-positive bone metastases. §P ⬍ .001.

interpreted by one of 12 radiologists, all of whom had more than 8 years of experience in interpretation of abdominal CT scans. The abdominal MR images were Volume 210 • Number 3

interpreted by one of three radiologists with more than 6 years of experience in the interpretation of body MR images. All MR images and CT scans were interpreted

in a separate blinded fashion by different radiologists. At institution 2, helical CT scans were interpreted by one of eight radiologists, all of whom had more than 8 years of experience in the interpretation of abdominal CT scans. The abdominal MR images were interpreted by one of three radiologists, each with more than 2 years of experience in the interpretation of body MR images. All images were interpreted in a separate blinded fashion by independent radiologists.

MR Imaging Institution 1.—MR imaging was performed with a 1.5-T Signa imager (GE Medical Systems, Milwaukee, Wis) and included the abdomen in all patients and the abdomen and pelvis in 18 patients. Dilute barium sulfate (1,450 mL of ReadiCat 2; E-Z-Em, Westbury, NY) was used as an orally administered contrast material in 20 patients. Body-coil imaging was used in 25 patients, and an abdominal flex coil (Medical Advances, Milwaukee, Wis) was used in 11 patients. Nonenhanced axial T1-weighted images were obtained with either a conventional spinecho (SE) T1-weighted sequence or a spoiled gradient-echo (GRE) T1-weighted sequence. Conventional SE imaging used respiratory-ordered phase encoding, with a repetition time (TR) of 300 msec and an echo time (TE) of 11 msec (300/11), a 256 ⫻ 192 matrix, two signals acquired, 7-mm section thickness, and a 3-mm intersection gap. Time of acquisition was 8 minutes 17 seconds for each set of 24 sections. The spoiled GRE imaging used breath-hold fast multiplanar spoiled gradient-recalled (FMPSPGR) T1-weighted images (111/4.2), a 256 ⫻ 192 matrix, and one signal acquired. Contiguous 10mm-thick sections were acquired. The time of acquisition was 28 seconds for 12 sections. For the T2-weighted fast SE sequence, parameters were 4,600/91–96, a 256 ⫻ 256 matrix, two signals acquired, an echo train length of 8, and receiver bandwidth of ⫾32 kHz. Sections were acquired with a 7-mm section thickness and a 3-mm intersection gap. Flow compensation and fat saturation were used for artifact reduction. The time of acquisition for this sequence was 5 minutes 14 seconds for each set of 24 sections. Dynamic gadolinium chelate–enhanced imaging used a fat-suppressed FMPSPGR acquisition during suspended respiration after the rapid intravenous injection of a bolus of 0.1 mmol of gado-

Extrahepatic Abdominal Imaging in Patients with Malignancy • 627

pentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) per kilogram of body weight. Axial FMPSPGR images were obtained immediately after the intravenous injection of the gadolinium chelate, with additional axial images obtained at 1 minute and at 5–10 minutes. Coronal fat-suppressed FMPSPGR images were obtained before the delayed axial images in all patients. Imaging parameters included 141/2.6, a 256 ⫻ 192 matrix with a three-quarter rectangular field of view, one signal acquired, 10-mmthick sections with no intersection gap, a receiver bandwidth of ⫾16 kHz, and a 70° flip angle. Noninterleaved sets of 8–12 sections were obtained during each 24second breath hold. In patients in whom the abdomen and pelvis were imaged, the 1-minute postcontrast acquisition was omitted. Institution 2.—MR examinations were performed at 1.5 T on SP4000 or Vision MR imagers (Siemens Medical Systems, Iselin, NJ). The T1-weighted breath-hold spoiled GRE sequence was performed with the following parameters: 140/4.5, matrix of 256 ⫻ 128, flip angle of 80°, 8-mm section thickness with a 20% gap, and one signal acquired, with 14–21 sections obtained in a 20-second breath hold. For evaluation of fatty infiltration of the liver, T1-weighted breath-hold out-of-phase spoiled GRE images were acquired with the following parameters: 140/2.2, 6.7; matrix of 256 ⫻ 128; flip angle of 80°; 8-mm section thickness with a 20% intersection gap; and one signal acquired. T2-weighted fat-suppressed SE or turbo SE images were also obtained, with the following parameters: 2,400–4,000/90– 100, matrix of 256 ⫻ 144, 8-mm section thickness with a 20% intersection gap, two signals acquired, and 18 sections obtained in 3.5–10.8 minutes. A phasedarray surface coil was used in 40 patients on the Vision system, and the remaining patients were imaged with the body coil. In 38 patients, T2-weighted half-Fourier, rapid acquisition with relaxation enhancement (RARE) SE images (TR/effective TE, ⬁/90) were also obtained in the coronal plane. The gadolinium chelate used was either gadopentetate dimeglumine (Magnevist; Berlex Laboratories) or gadoversetamide (Optimark; Mallinckrodt Medical, St Louis, Mo). The gadolinium chelate was administered in a dose of 0.1 mmol/kg as a rapid bolus injection, followed by a flush with physiologic saline. T1-weighted spoiled GRE image acquisition was initiated after the flush and subsequently at 45 seconds, 90 seconds, and 5–10 min628 • Radiology • March 1999

a.

c.

utes. Spoiled GRE imaging parameters were as stated previously in this section. Fat suppression was used for the 90second spoiled GRE acquisition. At both institutions, MR imaging in patients suspected of having pancreatic cancer included fat-suppressed T1-weighted SE sequences. In patients with adrenal masses, MR imaging included in-phase and opposed-phase spoiled GRE sequences through the adrenal glands.

CT Scanning Helical CT scanning was performed with a Somatom Plus or Somatom Plus S CT scanner (Siemens) for all patients at both institutions. The iodinated contrast material used was either iohexol (Omnipaque 300 [320 mg of iodine per milliliter]; Nycomed Amersham, Princeton, NJ) or diatrizoate meglumine (Hypaque 60%; Nycomed Amersham). The iodinated con-

b. Figure 2. Images of the abdomen of a 48-yearold man with renal cell carcinoma. (a) Helical CT scan, (b) spoiled GRE MR image (140/4, 80° flip angle), and (c) coronal T2-weighted halfFourier RARE SE image (TR/effective TE, ⬁/90). In a, a large heterogeneous enhancing mass (arrow) with central necrosis arises from the lower pole of the right kidney. Venous tumor involvement is poorly shown in a and was not detected prospectively. In b, the large heterogeneous enhancing renal cancer (long arrow) with central low signal intensity is shown. Well-defined low-signal-intensity material (short arrow) is clearly seen in the vena cava, consistent with venous tumor involvement, which was prospectively described. In c, the renal mass (white arrow) and the extent of venous tumor involvement, which extends to the level of the intrahepatic vena cava (black arrow), are seen. The MR demonstration of venous tumor invasion affected treatment planning in this patient. Renal cell carcinoma with venous tumor extension was confirmed surgically.

trast material was injected intravenously with a power injector (LFII; LiebelFlarsheim, Cincinnati, Ohio [at institution 1]; or Medrad, Pittsburgh, Pa [at institution 2]) as a uniphasic bolus of 125–150 mL at a rate of 2–3 mL/sec. At 60-70 seconds after initiation of contrast material injection, image acquisition commenced as a spiral acquisition with a pitch of 1.0. The table incrementation speed was 8–10 mm/sec, with 8–10-mm collimation at 210 mAs and 120 kVp. The scanning duration was 24 seconds, during which time the patient was instructed to suspend respiration. For those examinations that included the pelvis, the remainder of the pelvis from the lowest helical scan was imaged dynamically with contiguous 10-mm sections or as a second spiral acquisition. Thin 5-mm sections were obtained in patients with pancreatic carcinoma and at the discretion of the radiologist. Low et al

a.

b.

Figure 3. Images of the abdomen of a 79-year-old woman with primary ovarian cancer. (a) Helical CT scan shows perihepatic ascites (arrows). Fluid or soft tissue (arrowheads) is also present adjacent to the spleen. Distinguishing between ascites and peritoneal implants can be difficult at CT. (b) Fat-suppressed FMPSPGR MR image (140/2.6, 70° flip angle) obtained 10 minutes after intravenous injection of gadopentetate dimeglumine shows a rim of enhancing peritoneal tumor (short solid straight arrows) in the right subphrenic space. Note that the tumor extends into the left intersegmental fissure (open arrow), into the superior recess of the lesser sac (curved arrow), and along the portal vein (arrowheads). Enhancing perisplenic tumor (long solid straight arrow) is also present. Extensive peritoneal tumor was confirmed surgically.

Patients fasted for 3 hours before CT scanning. At institution 1, dilute 2% barium sulfate (Readi-Cat 2; E-Z-Em) was administered orally as a contrast agent. Doses of 450 mL were administered every 30 minutes starting 90–120 minutes before the CT examination, for a total oral dose of 1,450–1,800 mL. Rectal contrast material for pelvic imaging consisted of a water-soluble contrast material (Hypaque; Nycomed Amersham) at a volume of 500–1,000 mL. At institution 2, a dose of 800 mL of a combination of diatrizoate meglumine and diatrizoate sodium (Gastrografin; Bristol-Myers Squibb, Princeton, NJ) was given orally 45–60 minutes before an abdominal examination and 1–2 hours before an abdominal and pelvic examination. This dose was followed by oral administration of 300 mL of diluted 2.2% barium sulfate (Cheetah; Cheetah Pharmaceuticals, Lafayette, Ind) immediately before the examination. A volume of 500–1,000 mL of dilute Gastrografin was used for rectal contrast.

Proof of Disease For the 57 patients who underwent surgical exploration, surgical reports and histopathologic findings were reviewed to confirm or exclude the presence of malignant extrahepatic disease. Each site of extrahepatic tumor was proved by surgical reports, interviews of the surgeon immediately following laparotomy, or histopathologic findings. The results of other Volume 210 • Number 3

concurrent imaging tests or follow-up imaging examinations were also reviewed for each patient. The time of clinical follow-up ranged from 1 to 36 months (mean, 7 months). Follow-up CT scans and MR images were available in four and 20 patients, respectively, and four patients had both follow-up CT scans and MR images.

Tabulation of Data The prospective interpretations, as indicated in the original written CT and MR examination reports, were reviewed. For 14 patients at institution 1 who were included in a previous study (20), the blinded, prospective interpretations of the CT scans and MR images by different radiologists were used to indicate the presence or absence of extrahepatic tumor. For each test, the presence or absence of extrahepatic disease at 17 anatomic sites was recorded. These 17 sites were the gallbladder and biliary system, pancreas, spleen, kidneys, adrenal glands, peritoneum, mesentery, omentum, bowel, retroperitoneum, abdominal wall, lymph nodes, ascites, lung bases, vascular structures, bones, and pelvis. Malignant extrahepatic disease was recorded for each site. The findings of the prospective interpretations were compared with the sites of surgically proved tumor. Errors in interpretation that would have affected patient care if the findings had remained undetected with other tests or at surgery

were recorded for helical CT and MR imaging. The effect of imaging findings on patient care was evaluated in consultation with oncologists and surgical oncologists.

Statistical Analysis The sensitivity and accuracy of helical CT scanning and of MR imaging for depicting extrahepatic disease were compared by using the McNemar test of correlated proportions. In the application of the McNemar test, data were paired for the helical CT scan and MR image for each patient. Two-tailed P values are reported, with the null hypothesis rejected for a P value greater than .05.

RESULTS In the 57 patients with a history of malignancy, 51 patients had extrahepatic tumor confirmed surgically, and the remaining six patients had no evidence of extrahepatic tumor. Surgical exploration confirmed 154 sites of extrahepatic tumor or findings directly related to the presence of tumor. MR imaging depicted 139 sites of tumor but did not depict 15 sites (sensitivity, 0.90), compared with helical CT, which depicted 101 sites of extrahepatic tumor but did not depict 53 sites (sensitivity, 0.66) (P ⬍ .001) (Fig 1). The Table compares the results of MR imaging and helical CT for depicting ex-

Extrahepatic Abdominal Imaging in Patients with Malignancy • 629

trahepatic tumor at the 17 anatomic sites. Compared with helical CT, MR imaging depicted tumor in more patients at 11 of the 17 anatomic sites, and MR imaging and helical CT results were equivalent at the remaining six sites (Fig 2). This difference between MR imaging and helical CT achieved statistical significance for evaluation of the peritoneum (Fig 3) (P ⬍ .05), mesentery (P ⬍ .05), and bowel (P ⬍ .01). With helical CT, there were nine falsepositive interpretations of extrahepatic tumor; these false-positive results involved the following sites: bowel, one; peritoneum, four; mesentery, one; spleen, one; lymph nodes, one; and pelvis, one. With MR imaging, there were nine falsepositive interpretations of extrahepatic tumor; these false-positive results involved the following sites: peritoneum, two; bowel, three; abdominal wall, one; mesentery, one; pelvis, one; and spine, one. In these patients with false-positive results, surgery showed no evidence of extrahepatic tumor at these sites. In some cases, the distinction between benign and malignant findings was incomplete, leading to false-positive interpretations. In one patient, interpretations of helical CT scans and MR images predicted an ovarian malignancy in a patient with bilateral pelvic masses. Histopathologic evaluation revealed benign ovarian neoplasms. In another patient with primary gastric carcinoma, helical CT scan interpretation predicted nodal metastases, and MR imaging predicted extraserosal tumor extension. Histopathologic evaluation confirmed tumor confined to the stomach, with adjacent inflammation and normal perigastric lymph nodes. In the 57 patients, helical CT and MR imaging were equivalent for depicting extrahepatic tumor in 36 patients, and in the remaining 21 patients, MR imaging depicted more sites of tumor (Fig 4) than did helical CT scanning (P ⬍ .001). For MR imaging, the failure to depict extrahepatic tumor could have affected patient care in six patients, including one patient with pulmonary metastases, four patients with bowel invasion by an adjacent mass or bowel metastases, and one patient with an undetected primary gastrointestinal malignancy. With helical CT, the failure to depict findings of extrahepatic tumor could have affected patient care in 13 patients, including four patients in whom undetected peritoneal tumor was the only evidence of extrahepatic tumor, one patient with direct extrahepatic tumor extension into the peritoneum and lung base, one patient with a small pancre630 • Radiology • March 1999

a.

b. Figure 4. Images of the abdomen of a 52-yearold woman who presented with jaundice. (a) Helical CT scan, (b) spoiled GRE MR image (140/4, 80° flip angle) obtained immediately after injection of gadopentetate dimeglumine, and (c) coronal T2-weighted half-Fourier RARE SE MR image (TR/effective TE, ⬁/90). In a, dilated distal common bile duct (arrow) is seen at the level of the pancreatic head. No tumor is identified in a. In b, MR image obtained with a phased-array surface coil shows a small 1-cm, low-signal-intensity mass (arrows) surrounding the distal common bile duct at the level of the inferior pancreatic head; this mass represents a small ampullary carcinoma. In c, the level of obstruction is clearly demonstrated. An abnormal low-signal-intensity mass (arrow) at the ampulla of Vater represents an ampullary carcinoma. The ampullary carcinoma depicted on the MR images was confirmed surgically.

c.

atic carcinoma, three patients with malignant bowel invasion or metastases, three patients with undetected primary gastrointestinal malignancy (Fig 5), and one patient with ovarian cancer in whom undetected nodal metastases were the only evidence of tumor recurrence.

DISCUSSION The accurate depiction of extrahepatic abdominal tumor in patients with malignancy is critical to determine initial tumor staging, to monitor response to therapy, and to detect sites of tumor recurrence or progression. Clinical decisions regarding initial therapeutic options, surgical planning, and the reinitiation of treatment for recurrent tumor require accurate knowledge of the presence and extent of all abdominal tumor sites. Incorrect or incomplete informa-

tion could adversely affect patient care and outcome. Helical CT scanning has assumed the role of the primary cross-sectional imaging test for evaluation of the abdomen in patients with cancer. The accessibility of helical CT scanners, the short examination times, and the familiarity of radiologists and clinicians with CT scan interpretation make helical CT the natural choice for abdominal imaging at most institutions (6,7,27–29). Recent advances in MR imaging hardware and software have shortened acquisition times and improved the quality of MR images (15–18). The routine use of breath-hold MR imaging and the use of intravenously administered contrast agents have improved MR depiction of hepatic and extrahepatic tumor sites. For MR imaging to be used as a primary imaging test for patients with cancer, the modality must consistently and accurately depict extrahepatic abdominal tumor. The results of our study show that the routine use of combined T1-weighted, RARE T2Low et al

a.

b.

c.

d.

Figure 5. Images of the abdomen of a 60-year-old woman with metastatic adenocarcinoma discovered at laparoscopic cholecystectomy. (a) Helical CT scan shows colon distended with stool. Mild thickening (⬎2 mm) of the descending colon (arrow) was not noted prospectively. (b) Helical CT scan 6 cm caudal to a shows rectal contrast material with air and fluid in the rectosigmoid colon (arrow) but no evidence of an obstructing mass. (c) Gadolinium chelate–enhanced FMPSPGR MR image (141/2.6, 70° flip angle) with fat suppression shows marked mural thickening and enhancement (arrows) of the descending colon, which represent hyperemia and ischemic colitis proximal to a distal colonic obstruction. Enhancing right and left paracolic tumor was depicted on other MR images (not shown). (d) Gadolinium chelate–enhanced fat-suppressed FMPSPGR MR image (141/2.6, 70° flip angle) obtained 7 cm below c shows an enhancing mass (arrows) in the rectosigmoid colon. The position of d is lower than that of b because of the presence of rectal contrast material that distended and elevated the sigmoid colon. Primary colon carcinoma with peritoneal metastases was confirmed surgically.

weighted, and fat-suppressed gadolinium chelate–enhanced spoiled GRE T1-weighted MR imaging allows one to accurately depict malignant extrahepatic disease. With the use of these MR sequences at our institutions, MR imaging was equal or superior to helical CT in depicting extrahepatic abdominal disease at 17 different anatomic sites. At our institutions, these findings have allowed MR imaging to evolve from a problem-solving examination to a primary imaging modality for patients with malignancy. In this setting, MR imaging is used to screen the abdomen and pelvis and is the primary modalVolume 210 • Number 3

ity for accurately detecting all abdominal tumor sites. An inherent advantage of MR imaging compared with helical CT is the ability to acquire multiple types of images, each of which may be useful for depicting different forms of abdominal disease. In our experience, fat-suppressed T2-weighted images are most useful for depicting nodal metastases, abdominal wall and retroperitoneal tumor sites, ascites, and gallbladder and biliary abnormalities. Liver metastases are also well shown on fat-suppressed T2weighted images (10,12,14), although this was not included in our current study.

Fat-suppressed gadolinium-enhanced spoiled GRE T1-weighted MR images, such as FMPSPGR or fast low-angle shot images, are most useful for depicting peritoneal and omental tumor sites; bowel tumor sites; osseous metastases; biliary, renal, pancreatic, and splenic tumors; and vascular thrombosis or tumor encasement of abdominal vessels (20–22,24,30– 34). In our experience, osseous metastases in the spine and pelvis are often depicted on the fat-suppressed gadolinium-enhanced images. Our oncologists use these images to screen the spine in patients with malignancy. Because our study group was limited to patients with surgical proof, patients with osseous metastases were excluded because they were not surgical candidates. Gadolinium-enhanced spoiled GRE imaging performed immediately after the injection of gadolinium-based contrast material is important for depicting pancreatic and splenic tumors. The sequence is also very useful for evaluating liver disease (10–14), although this was not included in the current study. The breathhold spoiled GRE sequence with gadolinium enhancement and fat suppression is, in our opinion, the single most useful MR acquisition for rapid screening of the extrahepatic abdomen. T1-weighted images are useful for confirming abnormalities and depicting subtle mesenteric or omental infiltration. Individual studies (20–22,30,31,33–35) have compared gadolinium-enhanced MR imaging and CT for the depiction of extrahepatic tumor involving the pancreas, kidneys, biliary system, peritoneum, and intestines. Characterization of adrenal masses with in-phase and opposed-phase MR imaging has been well described, and comparisons with CT have also been performed (36–38). To our knowledge, however, this report is the first to compare MR imaging and helical CT for the depiction of all forms of extrahepatic tumor. Other factors, including the availability of an MR imager, cost, and the expertise of the radiologist, will markedly affect the decision to use helical CT or MR imaging in patients with malignancy. At institution 1, the charges for an abdominal MR examination and for helical CT of the abdomen are identical; however, at most institutions, the cost of an MR examination is more than that for helical CT of the abdomen. This difference in cost would have to be factored into any cost-benefit analysis, which is beyond the scope of this article. Our results indicate that duplication of imaging tests is not

Extrahepatic Abdominal Imaging in Patients with Malignancy • 631

necessary. MR imaging can be used effectively to evaluate hepatic and extrahepatic malignancy. At the institutions involved in this study, MR imaging serves as the primary imaging examination for some patients with malignancy. A savings in cost can be achieved by performing one examination, instead of a series of tests, which often includes helical CT and an MR examination. Although not evaluated in this study, the addition of other MR tests, such as MR angiography or MR cholangiopancreatography, would further increase the value of a comprehensive MR examination. Limitations of this study should be acknowledged. The helical CT modality could have been improved by acquiring data during multiple phases after intravenous injection of the contrast material; however, the multiphase helical CT technique is more important for hepatic imaging and is not typically used for extrahepatic tumor. Collimation of 8–10 mm was used for helical CT, compared with a section thickness of 7–10 mm for the MR imaging. The use of narrower collimation for helical CT would have been preferable and might have improved the depiction of tumor sites with helical CT scans; however, even in cases in which narrower collimation was used for CT scanning, the superior soft-tissue contrast of MR imaging was the predominant factor in determining tumor depiction (Fig 1). We also did not evaluate separately the use of breathing-independent T2-weighted MR imaging with single-shot fast SE sequences, which show promise for evaluating extrahepatic abdominal structures. In conclusion, MR imaging with T1weighted, fat-suppressed T2-weighted, and fat-suppressed gadolinium chelate– enhanced spoiled GRE sequences is effective in depicting malignant extrahepatic disease. With the use of these sequences at our two institutions, MR imaging was superior to single-phase helical CT in screening the abdomen for extrahepatic tumor in patients with known malignancy.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

References 1.

Chezmar JL, Rumancik WM, Megibow AJ, Hulnick DH, Nelson RC, Bernadino ME. Liver and abdominal screening in patients with cancer: CT versus MR imaging. Radiology 1988; 168:43–47. 2. Heiken JP, Weyman PJ, Lee JKT, et al. Detection of focal hepatic masses: prospective evaluation with CT, delayed CT, CT during arterial portography, and MR imaging. Radiology 1989; 171:47–51. 3. Nelson RC, Chezmar JL, Sugarbaker PH, Bernardino ME. Hepatic tumors: comparison of CT during arterial portography, delayed CT, and MR imaging for preoperative evaluation. Radiology 1989; 172:27–34. 632 • Radiology • March 1999

19. 20.

21.

Megibow AJ, Zhou XH, Rotterdam H, et al. Pancreatic adenocarcinoma: CT versus MR imaging in the evaluation of resectability— report of the Radiology Diagnostic Oncology Group. Radiology 1995; 195:327–332. Muller MF, Meyenberger C, Bertschinger P, Schaer R, Marincek B. Pancreatic tumors: evaluation with endoscopic US, CT, and MR imaging. Radiology 1994; 190:745–751. Cohan RH, Sherman LS, Korobkin M, Bass JC, Francis IR. Renal masses: assessment of corticomedullary-phase and nephrographic-phase CT scans. Radiology 1995; 196:445–451. Jamis-Dow CA, Choyke PL, Jennings SB, Marston Linehan W, Thakore KN, Walther MM. Small (ⱕ3 cm) renal masses: detection with CT versus US and pathologic correlation. Radiology 1996; 198:785–788. Korobkin M, Brodeur FJ, Yutzy GG, et al. Differentiation of adrenal adenomas from nonadenomas using CT attenuation values. AJR 1966; 166:531–536. Zerhouni EA, Rutter C, Hamilton SR, et al. CT and MR imaging in the staging of colorectal carcinoma: report of the Radiology Diagnostic Oncology Group II. Radiology 1996; 200: 443–451. Larson RE, Semelka RC, Bagley AS, Molina PL, Brown ED, Lee JKT. Hypervascular malignant liver lesions: comparison of various MR imaging pulse sequences and dynamic CT. Radiology 1994; 192:393–399. Hamm B, Thoeni RF, Gould RG, et al. Focal liver lesions: characterization with nonenhanced and dynamic contrast material-enhanced MR imaging. Radiology 1994; 190: 417–423. Semelka RC, Worawattanakul S, Kelekis NL, et al. Liver lesion detection, characterization, and effect on patient management: comparison of single-phase spiral CT and current MR techniques. JMRI 1997; 7:1040–1047. Yamashita Y, Mitsuzaki K, Yi T, et al. Small hepatocellular carcinoma in patients with chronic liver damage: prospective comparison of detection with dynamic MR imaging and helical CT of the whole liver. Radiology 1996; 200:79–84. Semelka RC, Schlund JF, Molina RL, et al. Malignant liver lesions: comparison of spiral CT arterial portography and MR imaging for diagnostic accuracy, cost, and effect on patient management. JMRI 1996; 6:39–43. Semelka RC, Willms AB, Brown MA, Brown ED, Finn JP. Comparison of breath-hold T1weighted MR sequences for imaging of the liver. JMRI 1994; 4:759–765. Low RN, Francis IR, Sigeti JS, Foo TKF. Abdominal MR imaging: comparison of T2weighted fast and conventional spin-echo and contrast-enhanced fast multiplanar spoiled gradient-recalled imaging. Radiology 1993; 186:803–811. Outwater EK, Mitchell DG, Vinitski S. Abdominal MR imaging: evaluation of a fast spin-echo sequence. Radiology 1994; 190:425–429. Rydberg JN, Lomas DJ, Coakley KJ, et al. Comparison of breath-hold fast spin-echo and conventional spin-echo pulse sequences for T2-weighted MR imaging of liver lesions. Radiology 1995; 194:431–437. Low RN. Contrast agents for MR imaging of the liver. JMRI 1997; 7:56–67. Low RN, Barone RM, Lacey C, Sigeti JS, Alzate GD, Sebrechts CP. Peritoneal tumor: MR imaging with dilute oral barium and intravenous gadolinium-containing contrast agents compared with unenhanced MR imaging and CT. Radiology 1997; 204:513–520. Low RN, Carter WD, Saleh F, Sigeti JS. Ovarian cancer: comparison of findings with perfluorocarbon-enhanced MR imaging, In-111– CYT-103 immunoscintigraphy, and CT. Radiology 1995; 195:391–400.

22.

23.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Low RN, Francis IR. MR imaging of the gastrointestinal tract with IV gadolinium and diluted barium oral contrast media compared with unenhanced MR imaging and CT. AJR 1997; 169:1051–1059. Matsushita M, Oi H, Murakami T, et al. Extraserosal invasion in advanced gastric cancer: evaluation with MR imaging. Radiology 1994; 192:87–91. Rofsky NM, Weinreb JC, Bosniak MA, Libes RB, Birnbaum BA. Renal lesion characterization with gadolinium-enhanced MR imaging: efficacy and safety in patients with renal insufficiency. Radiology 1991; 180:85–89. Semelka RC, Cumming MJ, Shoenut JP, et al. Islet cell tumors: comparison of dynamic contrast-enhanced CT and MR imaging with dynamic gadolinium enhancement and fat suppression. Radiology 1993; 186:799–802. Semelka RC, Kroeker MA, Shoenut JP, Kroeker R, Yaffe CS, Micflikier AB. Pancreatic disease: prospective comparison of CT, ERCP, and 1.5-T MR imaging with dynamic gadolinium enhancement and fat suppression. Radiology 1991; 181:785–791. Zeman RK, Fox SH, Silverman PM, et al. Helical (spiral) CT of the abdomen. AJR 1993; 160:719–725. Hollett MD, Jeffrey RB, Nino-Murcia M, Jorgensen MJ, Harris DP. Dual-phase helical CT of the liver: value of arterial phase scans in the detection of small (⬍1.5 cm) malignant hepatic neoplasms. AJR 1995; 164:879–884. Bluemke DA, Cameron JL, Hruban RH, et al. Potentially resectable pancreatic adenocarcinoma: spiral CT assessment with surgical and pathologic correlation. Radiology 1995; 197: 381–385. Low RN, Sigeti JS, Francis IR, et al. Evaluation of malignant biliary obstruction: efficacy of fast multiplanar spoiled gradient-recalled MR imaging vs spin-echo MR, CT, and cholangiography. AJR 1994; 162:315–323. Semelka RC, Shoenut JP, Kroeker MA, et al. Bile duct disease: prospective comparison of ERCP, CT and fat-suppressed MRI. Gastrointest Radiol 1992; 17:347–352. Semelka RC, Shoenut JP, Magro CM, Kroeker MA, MacMahon R, Greenberg HM. Renal cancer staging: comparison of contrast-enhanced CT and gadolinium-enhanced fatsuppressed spin-echo and gradient-echo MR imaging. JMRI 1993; 3:597–602. Ichikawa T, Hardome H, Hachiya J, et al. Pancreatic ductal adenocarcinoma: preoperative assessment with helical CT versus dynamic MR imaging. Radiology 1997; 202:655– 662. Semelka RC, Kelekis NL, Molina PL, Sharp TJ, Calvo B. Pancreatic masses with inconclusive findings on spiral CT: is there a role for MRI? JMRI 1996; 6:585–588. Semelka RC, Shoenut JP, Silverman R, Kroeker MA, Yaffe CS, Micflikier AB. Bowel disease: prospective comparison of CT and 1.5-T preand postcontrast MR imaging with T1weighted fat-suppressed and breath-hold FLASH sequences. JMRI 1991; 1:625–632. Mayo-Smith WW, Lee MJ, McNicholas MMJ, Hahn PF, Boland GW, Saini S. Characterization of adrenal masses (⬍5 cm) by use of chemical shift MR imaging: observer performance versus quantitative measures. AJR 1995; 165:91–95. Mitchell DG, Crovello M, Matteucci T, et al. Benign adrenocortical masses: diagnosis with chemical shift MR imaging. Radiology 1992; 185:345–351. Outwater EK, Siegelman ES, Tadecki PD, Piccoli CW, Mitchell DG. Distinction between benign and malignant adrenal masses: value of T1-weighted chemical-shift MR imaging. AJR 1995; 165:579–583.

Low et al