- Email: [email protected]
Intraoperative ultrasound does not improve detection of liver metastases in resectable pancreatic cancer
r
Intraoperative Ultrasound Does Not Improve Detection of Liver Metastases in Resectable Pancreatic Cancer Christina Finlayson, Marlane Guttmann,
MD, John Hoffman, MD, Raymond Yeung, MD, Howard MD, Andrew Shaer, MD, Michael Clair, MD, Philadelphia,
BACKGROUND: In colorectal cancer, intraoperative ultrasound (IOUS) is superior to other imaging studies in characterizing hepatic metastases. The value of IOUS in detecting liver metastases from pancreatic cancer has not been evaluated previously. METHODS: Between 1990 and 1995, IOUS was prospectively employed to evaluate the liver for metastases in 32 patients with resectable pancreatic adenocarcinoma. Preoperatively, all patients had computed tomography (CT) and 22 patients had CT pot-tography. RESULTS: At exploration, 5 of the 32 patients (15%) had extrapancreatic disease, 3 (9%) with liver implants. IOUS did not identify any additional hepatic metastases. Four preoperative studies were suspicious for metastatic disease in the liver. In these 4 patients, no hepatic metastases were identified by exploration or intraoperative ultrasound. CONCLUSIONS: We no longer routinely perform hepatic IOUS when evaluating patients with pancreatic adenocarcinoma for pancreaticoduodenectomy. When a preoperative study indicates possible hepatic involvement, IOUS can confirm the presence or absence of liver metastases. Am J Sorg. 1998;175:99-101.0 1993 by Excerpta Medica, Inc.
T
he surgical treatment of pancreatic adenocarcinoma requires a careful search for extrapancreatic spread of malignancy prior to embarking on pancreaticoduodenectomy. The presence of hepatic metastases may be a contraindication to resection. In colorectal cancer, intraoperative ultrasound (IOUS) has been shown to he superior to other radiologic tests in identifying the presence of hepatic metastases. Several studies have demonstrated that IOUS can increase the detection of hepatic metastases from colorectal malignancy by as much ;as 15% to 47%
From the Departments of Surgical Oncology (CF, JH, RY) and Department of Radiology (HK, MG, AS, MC), Fox Chase Cancer ;enter, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Christina Finlayson, MD, Department of Surgery, UCHSC. Box C-311, 4200 E. \linth Ave., Denver, Colorado 80262. Manuscript submitted March 11,1997 and accepted in revised orm September 11, 1997.
0 1998 by Excerpta All rights reserved.
Medica,
Inc.
Kessler,
MD,
Pennsylvania
when used in addition to preoperative radiologic evaluation and manual exploration.1-6 Preoperative staging of pancreatic cancer usually includes one or more imaging studies including abdominal computed tomography (CT), CT portography, transabdominal ultrasound, endoscopic retrograde pancreatography (ERCP), angiography, or magnetic resonance imaging (MRI). Although a positive study can prevent an unnecessary exploration, each of these modalities has a known incidence of false negative results. ~‘3~ Intraoperatively, manual exploration can identify surface deposits of malignancy on the liver and other peritoneal surfaces. Several studies have evaluated the use of intraoperative ultrasound to assess resectability in pancreatic adenocarcinoma. However, these published reports have either not fully addressed the utility of IOUS for evaluating the liver or utilized laparoscopic ultrasound to replace the manual examination of the liver for hepatic metastases.“” We prospectively performed intraoperative hepatic ultrasound as part of the assessment of resectability for pancreatic neoplasm to determine if this increased our sensitivity in identifying those patients with hepatic metastases.
METHODS Between 1990 and 1995, 32 patients underwent exploration for pancreatic adenocarcinoma with planned pancreaticoduodenectomy, and intraoperative ultrasound of the liver was performed searching for evidence of hepatic metastasis. All of these patients had preoperative abdominal CT and 22 patients had preoperative CT portography. In 2 patients, an abnormality suspicious for metastatic disease was identified, 1 by CT and 1 hy CT portography. Two other CT portograms had flow ahnormalities that could not exclude a mass lesion. There were 14 men and 18 women in this study. The average age was 63 years with a range from 41 to 79 years. The tumors were located in the head of the pancreas in 25 patients, the head and hody in 4 patients, and the head and uncinate process in 3 patients. Twenty-two patients received preoperative chemoradiation therapy prior to surgical exploration. All patients who received preoperative chemoradiation were restaged hy radiographic imaging prior to surgery. In the operating room, all patients underwent a thorough tnanual exploration of the abdomen that included examination of all peritoneal surfaces, the stomach, small howel, omenturn, pelvic organs, and the liver surface as well as palpation of the celiac axis and periaortic region for lymphadenopathy. Intraoperative ultrasound was then performed on all patients, including 5 patients who were identified on 0002-961 O/98/$1 PII SOOO2-9610(97)00276-6
9.00
99
LTRASOUND TABLE
FOR
DETECTING
LIVER
I
TABLE Results
of Staaina
Study
Studies Positive
Computed tomography Computed tomography Exploration lntraoperative ultrasound
l/32 3/22 5/32 O/32
portography
(3%) (14%) (16%) (0%)
Negative 31/32 19/22 27/32 32/32
(97%) (86%) (84%) (100%)
II False-Positive Computed Tomography/Computed Tomography Portogram Outcome
Preoperative Study
Time to Recurrence
CT CT portography CT portography CT pottography
44 21 11 9
CT = computed
TABLE
months months months months
Site of First Recurrence
Overall Survival
Liver Lung Local Pleura
51 30 14 17
tomography.
Ill Site
of First
Recurrence Number
Number of patients with Site of first recurrence Liver Local Peritoneal Lung/pleura
Figure. Arrows astatic
False-positive computed tomography indicate hypodense areas that could lesion.
portogram not exclude
scan. a met-
manual exploration to have extrapancreatic disease. The ultrasound was performed by the radiologist with the assistance of the ultrasound technologist. The surgical team was present throughout the ultrasound examination. Abnormalities identified on physical examination were biopsied and sent for frozen section analysis. No abnormalities were identified on ultrasound examination. Postoperative follow-up to death or last clinic visit ranged from 1 to 51 months with median follow-up of 11.5 months. Six months have elapsed since the last patient enrollment on the stuily. No patients were lost to follow-up.
RESULTS The results of each imaging study and exploration are detailed in Table I. At operation, 27 of the 32 patients had no evidence of extrapancreatic disease and IOUS did not identify any parenchymal hepatic disease in this group of patients. Five of the 32 patients (15%) had extrapancreatic disease identified by manual exploration, including 3 (9%) with surface liver implants. Intraoperative ultrasound did not identify any further evidence of parenchymal hepatic metastases in these patients. In the 4 patients with a suspicious preoperative CT or CT portogram, no evidence of hepatic disease was identified by either manual exploration or intraoperative ultrasound. An example of one of the CT portograms interpreted as being positive is shown in the Figure. Hypodense areas in the left lobe of the liver were suspicious for metastatic disease. However, at intraoperative ultrasound no mass lesion could DO
THE
AMERICAN
JOURNAL
OF
SURGERY@
VOLUME
175
months months months months
of Patients
recurrence
21 6 2 5 8
(29%) (9%) (24%) (38%)
be demonstrated. Long-term follow-up of the 4 patients with a false-positive CT or CT portogram is shown in Table II. Only 1 patient had a first site of recurrence in the liver, and this occurred at greater than 3 years after initial operation. Therefore, we do not believe that gross metastatic disease was present in the liver of these patients at the time of preoperative or intraoperative evaluation. Twenty-one patients had recurrence of their pancreatic tumor. The site of first recurrence is detailed in Table 111. For those patients who had the site of first recurrence in the liver, the median time to recurrence was 10 months with a range of 4 to 44 months.
COMMENTS Pancreatic cancer is characterized by late presentation with a high incidence of concomitant extrapancreatic metastases. Although the published operative mortality and morbidity for pancreaticoduodenectomy is decreasing,’ the risk-benefit ratio for palliative resection is only beginning to be elucidated’-” Therefore, careful preresection staging is necessary to identify those patients most likely to benefit from a potentially curative operation. The radiologic identification of hepatic metastases from other gastrointestinal primaries has been evaluated extensively. Intraoperative ultrasound has been shown to be superior to standard CT, CT portography, and manual exploration in identifying these lesionsW6 Pancreatic cancer is well known for a propensity toward intraperitoneal spread as well as hematogenous and lymphatic metastases. Metastatic involvement of the liver is a common finding in the natural history of this disease. Therefore, we proposed that IOUS might provide a similar benefit in the staging of this disease. In this series, intraoperative ultrasound provided no additional benefit over preoperative imaging and manual exploration of the liver in identifying hepatic metastases in patients with adenocarcinoma of the pancreas. Even in the FEBRUARY
1998
1 ULTRASOUND
3 patients who were found to have surface metastases at the time of exploration, IOUS did not identify any additional metastases in the deeper parenchyma. In another report using open IOUS, similar findings were observed. Plainfosse and colleagues7 performed ultrasonography of the liver during the surgical exploration of 22 patients. Four patients were found to have liver metastases on physical examination. In the remaining 18 patients, ultrasound did not identify any additional patients with hepatic metastases although 2 patients who had surface deposits of malignancy also had deep parenchymal metastases. Although IOUS was not helpful in identifying occult hepatic metastases, it was valuable in further evaluating suspicious liver lesions that were seen on other preoperative imaging studies. The false-positive rate for CT portography in evaluating liver metastases from a variety of primaries has been reported to be as high as 42%.” In this experience, 4 patients had preoperative studies that demonstrated indeterminate abnormalities and, therefore, metastatic disease could not be excluded. In these patients, IOUS demonstrated no corresponding mass lesions. The lack of early recurrence and long survival prior to late recurrence in the liver indicate that this represented a false-positive result of the CT or CT portography rather than a false-negative result of IOUS.
CONCLUSIONS Based on these findings, intraoperative hepatic ultrasound is most useful when evaluating patients with pancreatic adenocarcinoma who have an equivocal preoperative imaging study. In this situation, ultrasound imaging of the liver may exclude a metastatic lesion prior to pancreaticoduodenectomy. We no longer routinely perform intraoperative hepatic ultrasound when evaluating patients with pancreatic adenocarcinoma.
THE
AMERICAN
FOR
DETECTING
LIVER
METASTASEWFINLAYSON
ET AL
REFERENCES 1. Machi 1, Isomoto H, Kurohiii T, et al. Accuracy of intraouerative ultrasonography in diagnosing liver metastasis from colorectal cancer: evaluation with postoperative follow-up results. World I strrg. 1991;15:551. 2. Machi J, Sigel B, Zaren HA, et al. Operative ultrasonography during hepatobiliary and pancreatic surgery. World J Surg. 1993;17: 640. 3. Hagspiel KD, Neidi KF’W, Eichenberger AC, et al. Detection of liver metastases: comparison of superparamagnetic iron oxide-enhanced and unenhanced MR imaging at 1.5 T with dynamtc CT, intraoperative US, and percutaneous US. Radtology. 1995;196:471. 4. Parker GA, Lawrence W, Horsley JS, et al. Intraoperative ultrasound of the liver affects operative decision making. Ann Surg. 1989;209:569. 5. Rifkin MD, Rosato FE, Branch HM, et al. Intraoperative ultrasound of the liver: an Important adjunctive tool for decision making in the operating room. .Ann Surg. 1987;205:466. 6. Fortunato L, Clair M, Hoffman J, et al. is CT portography (CTAP) really useful in patients with liver tumors who undergo intraoperattve ultrasonography (IOUS)? Am Surg. 1995;61:560. 7. Plainfosse MC, Bouillot MD, Rivaton F, et al. The use of operative sonography in carcinoma of the pancreas. World .I Surg. 1987;11:654. 8. Bemelman WA, DeWit LT, Van Delden CM, et al. Diagnostic laparoscopy combined with laparoscopic ultrasonography in staging of cancer of the pancreatic head region. Br J Surg. 1995;82:820. 9. Pitt HA. Curative treatment of pancreatic neoplasms: standard resection. Surg Clin iV Am. 1996;75:891. 10. Lillemoe KD, Cameron JL, Yeo CJ, et al. Pancreaticoduodenectomy. Does it have a role m the palliatton of pancreatic cancer? Ann Surg. 1996;223:718. 11. Lillemoe KD, Barnes SA. Surgical palliation of unresectable pancreatic carcinoma. Surg Clm N Am. 1995;75:953. 12. Haiken JP, Weyman PJ, Lee JKT, et al. Detection of focal hepatic masses: prospective evaluation wtth CT, delayed CT, CT during arterial portography, and MR imaging. Radiology. 1989;171: 47.
JOURNAL
OF SURGERY@
VOLUME
175
FEBRUARY
1998
101
Intraoperative Ultrasound Does Not Improve Detection of Liver Metastases in Resectable Pancreatic Cancer Christina Finlayson, Marlane Guttmann,
MD, John Hoffman, MD, Raymond Yeung, MD, Howard MD, Andrew Shaer, MD, Michael Clair, MD, Philadelphia,
BACKGROUND: In colorectal cancer, intraoperative ultrasound (IOUS) is superior to other imaging studies in characterizing hepatic metastases. The value of IOUS in detecting liver metastases from pancreatic cancer has not been evaluated previously. METHODS: Between 1990 and 1995, IOUS was prospectively employed to evaluate the liver for metastases in 32 patients with resectable pancreatic adenocarcinoma. Preoperatively, all patients had computed tomography (CT) and 22 patients had CT pot-tography. RESULTS: At exploration, 5 of the 32 patients (15%) had extrapancreatic disease, 3 (9%) with liver implants. IOUS did not identify any additional hepatic metastases. Four preoperative studies were suspicious for metastatic disease in the liver. In these 4 patients, no hepatic metastases were identified by exploration or intraoperative ultrasound. CONCLUSIONS: We no longer routinely perform hepatic IOUS when evaluating patients with pancreatic adenocarcinoma for pancreaticoduodenectomy. When a preoperative study indicates possible hepatic involvement, IOUS can confirm the presence or absence of liver metastases. Am J Sorg. 1998;175:99-101.0 1993 by Excerpta Medica, Inc.
T
he surgical treatment of pancreatic adenocarcinoma requires a careful search for extrapancreatic spread of malignancy prior to embarking on pancreaticoduodenectomy. The presence of hepatic metastases may be a contraindication to resection. In colorectal cancer, intraoperative ultrasound (IOUS) has been shown to he superior to other radiologic tests in identifying the presence of hepatic metastases. Several studies have demonstrated that IOUS can increase the detection of hepatic metastases from colorectal malignancy by as much ;as 15% to 47%
From the Departments of Surgical Oncology (CF, JH, RY) and Department of Radiology (HK, MG, AS, MC), Fox Chase Cancer ;enter, Philadelphia, Pennsylvania. Requests for reprints should be addressed to Christina Finlayson, MD, Department of Surgery, UCHSC. Box C-311, 4200 E. \linth Ave., Denver, Colorado 80262. Manuscript submitted March 11,1997 and accepted in revised orm September 11, 1997.
0 1998 by Excerpta All rights reserved.
Medica,
Inc.
Kessler,
MD,
Pennsylvania
when used in addition to preoperative radiologic evaluation and manual exploration.1-6 Preoperative staging of pancreatic cancer usually includes one or more imaging studies including abdominal computed tomography (CT), CT portography, transabdominal ultrasound, endoscopic retrograde pancreatography (ERCP), angiography, or magnetic resonance imaging (MRI). Although a positive study can prevent an unnecessary exploration, each of these modalities has a known incidence of false negative results. ~‘3~ Intraoperatively, manual exploration can identify surface deposits of malignancy on the liver and other peritoneal surfaces. Several studies have evaluated the use of intraoperative ultrasound to assess resectability in pancreatic adenocarcinoma. However, these published reports have either not fully addressed the utility of IOUS for evaluating the liver or utilized laparoscopic ultrasound to replace the manual examination of the liver for hepatic metastases.“” We prospectively performed intraoperative hepatic ultrasound as part of the assessment of resectability for pancreatic neoplasm to determine if this increased our sensitivity in identifying those patients with hepatic metastases.
METHODS Between 1990 and 1995, 32 patients underwent exploration for pancreatic adenocarcinoma with planned pancreaticoduodenectomy, and intraoperative ultrasound of the liver was performed searching for evidence of hepatic metastasis. All of these patients had preoperative abdominal CT and 22 patients had preoperative CT portography. In 2 patients, an abnormality suspicious for metastatic disease was identified, 1 by CT and 1 hy CT portography. Two other CT portograms had flow ahnormalities that could not exclude a mass lesion. There were 14 men and 18 women in this study. The average age was 63 years with a range from 41 to 79 years. The tumors were located in the head of the pancreas in 25 patients, the head and hody in 4 patients, and the head and uncinate process in 3 patients. Twenty-two patients received preoperative chemoradiation therapy prior to surgical exploration. All patients who received preoperative chemoradiation were restaged hy radiographic imaging prior to surgery. In the operating room, all patients underwent a thorough tnanual exploration of the abdomen that included examination of all peritoneal surfaces, the stomach, small howel, omenturn, pelvic organs, and the liver surface as well as palpation of the celiac axis and periaortic region for lymphadenopathy. Intraoperative ultrasound was then performed on all patients, including 5 patients who were identified on 0002-961 O/98/$1 PII SOOO2-9610(97)00276-6
9.00
99
LTRASOUND TABLE
FOR
DETECTING
LIVER
I
TABLE Results
of Staaina
Study
Studies Positive
Computed tomography Computed tomography Exploration lntraoperative ultrasound
l/32 3/22 5/32 O/32
portography
(3%) (14%) (16%) (0%)
Negative 31/32 19/22 27/32 32/32
(97%) (86%) (84%) (100%)
II False-Positive Computed Tomography/Computed Tomography Portogram Outcome
Preoperative Study
Time to Recurrence
CT CT portography CT portography CT pottography
44 21 11 9
CT = computed
TABLE
months months months months
Site of First Recurrence
Overall Survival
Liver Lung Local Pleura
51 30 14 17
tomography.
Ill Site
of First
Recurrence Number
Number of patients with Site of first recurrence Liver Local Peritoneal Lung/pleura
Figure. Arrows astatic
False-positive computed tomography indicate hypodense areas that could lesion.
portogram not exclude
scan. a met-
manual exploration to have extrapancreatic disease. The ultrasound was performed by the radiologist with the assistance of the ultrasound technologist. The surgical team was present throughout the ultrasound examination. Abnormalities identified on physical examination were biopsied and sent for frozen section analysis. No abnormalities were identified on ultrasound examination. Postoperative follow-up to death or last clinic visit ranged from 1 to 51 months with median follow-up of 11.5 months. Six months have elapsed since the last patient enrollment on the stuily. No patients were lost to follow-up.
RESULTS The results of each imaging study and exploration are detailed in Table I. At operation, 27 of the 32 patients had no evidence of extrapancreatic disease and IOUS did not identify any parenchymal hepatic disease in this group of patients. Five of the 32 patients (15%) had extrapancreatic disease identified by manual exploration, including 3 (9%) with surface liver implants. Intraoperative ultrasound did not identify any further evidence of parenchymal hepatic metastases in these patients. In the 4 patients with a suspicious preoperative CT or CT portogram, no evidence of hepatic disease was identified by either manual exploration or intraoperative ultrasound. An example of one of the CT portograms interpreted as being positive is shown in the Figure. Hypodense areas in the left lobe of the liver were suspicious for metastatic disease. However, at intraoperative ultrasound no mass lesion could DO
THE
AMERICAN
JOURNAL
OF
SURGERY@
VOLUME
175
months months months months
of Patients
recurrence
21 6 2 5 8
(29%) (9%) (24%) (38%)
be demonstrated. Long-term follow-up of the 4 patients with a false-positive CT or CT portogram is shown in Table II. Only 1 patient had a first site of recurrence in the liver, and this occurred at greater than 3 years after initial operation. Therefore, we do not believe that gross metastatic disease was present in the liver of these patients at the time of preoperative or intraoperative evaluation. Twenty-one patients had recurrence of their pancreatic tumor. The site of first recurrence is detailed in Table 111. For those patients who had the site of first recurrence in the liver, the median time to recurrence was 10 months with a range of 4 to 44 months.
COMMENTS Pancreatic cancer is characterized by late presentation with a high incidence of concomitant extrapancreatic metastases. Although the published operative mortality and morbidity for pancreaticoduodenectomy is decreasing,’ the risk-benefit ratio for palliative resection is only beginning to be elucidated’-” Therefore, careful preresection staging is necessary to identify those patients most likely to benefit from a potentially curative operation. The radiologic identification of hepatic metastases from other gastrointestinal primaries has been evaluated extensively. Intraoperative ultrasound has been shown to be superior to standard CT, CT portography, and manual exploration in identifying these lesionsW6 Pancreatic cancer is well known for a propensity toward intraperitoneal spread as well as hematogenous and lymphatic metastases. Metastatic involvement of the liver is a common finding in the natural history of this disease. Therefore, we proposed that IOUS might provide a similar benefit in the staging of this disease. In this series, intraoperative ultrasound provided no additional benefit over preoperative imaging and manual exploration of the liver in identifying hepatic metastases in patients with adenocarcinoma of the pancreas. Even in the FEBRUARY
1998
1 ULTRASOUND
3 patients who were found to have surface metastases at the time of exploration, IOUS did not identify any additional metastases in the deeper parenchyma. In another report using open IOUS, similar findings were observed. Plainfosse and colleagues7 performed ultrasonography of the liver during the surgical exploration of 22 patients. Four patients were found to have liver metastases on physical examination. In the remaining 18 patients, ultrasound did not identify any additional patients with hepatic metastases although 2 patients who had surface deposits of malignancy also had deep parenchymal metastases. Although IOUS was not helpful in identifying occult hepatic metastases, it was valuable in further evaluating suspicious liver lesions that were seen on other preoperative imaging studies. The false-positive rate for CT portography in evaluating liver metastases from a variety of primaries has been reported to be as high as 42%.” In this experience, 4 patients had preoperative studies that demonstrated indeterminate abnormalities and, therefore, metastatic disease could not be excluded. In these patients, IOUS demonstrated no corresponding mass lesions. The lack of early recurrence and long survival prior to late recurrence in the liver indicate that this represented a false-positive result of the CT or CT portography rather than a false-negative result of IOUS.
CONCLUSIONS Based on these findings, intraoperative hepatic ultrasound is most useful when evaluating patients with pancreatic adenocarcinoma who have an equivocal preoperative imaging study. In this situation, ultrasound imaging of the liver may exclude a metastatic lesion prior to pancreaticoduodenectomy. We no longer routinely perform intraoperative hepatic ultrasound when evaluating patients with pancreatic adenocarcinoma.
THE
AMERICAN
FOR
DETECTING
LIVER
METASTASEWFINLAYSON
ET AL
REFERENCES 1. Machi 1, Isomoto H, Kurohiii T, et al. Accuracy of intraouerative ultrasonography in diagnosing liver metastasis from colorectal cancer: evaluation with postoperative follow-up results. World I strrg. 1991;15:551. 2. Machi J, Sigel B, Zaren HA, et al. Operative ultrasonography during hepatobiliary and pancreatic surgery. World J Surg. 1993;17: 640. 3. Hagspiel KD, Neidi KF’W, Eichenberger AC, et al. Detection of liver metastases: comparison of superparamagnetic iron oxide-enhanced and unenhanced MR imaging at 1.5 T with dynamtc CT, intraoperative US, and percutaneous US. Radtology. 1995;196:471. 4. Parker GA, Lawrence W, Horsley JS, et al. Intraoperative ultrasound of the liver affects operative decision making. Ann Surg. 1989;209:569. 5. Rifkin MD, Rosato FE, Branch HM, et al. Intraoperative ultrasound of the liver: an Important adjunctive tool for decision making in the operating room. .Ann Surg. 1987;205:466. 6. Fortunato L, Clair M, Hoffman J, et al. is CT portography (CTAP) really useful in patients with liver tumors who undergo intraoperattve ultrasonography (IOUS)? Am Surg. 1995;61:560. 7. Plainfosse MC, Bouillot MD, Rivaton F, et al. The use of operative sonography in carcinoma of the pancreas. World .I Surg. 1987;11:654. 8. Bemelman WA, DeWit LT, Van Delden CM, et al. Diagnostic laparoscopy combined with laparoscopic ultrasonography in staging of cancer of the pancreatic head region. Br J Surg. 1995;82:820. 9. Pitt HA. Curative treatment of pancreatic neoplasms: standard resection. Surg Clin iV Am. 1996;75:891. 10. Lillemoe KD, Cameron JL, Yeo CJ, et al. Pancreaticoduodenectomy. Does it have a role m the palliatton of pancreatic cancer? Ann Surg. 1996;223:718. 11. Lillemoe KD, Barnes SA. Surgical palliation of unresectable pancreatic carcinoma. Surg Clm N Am. 1995;75:953. 12. Haiken JP, Weyman PJ, Lee JKT, et al. Detection of focal hepatic masses: prospective evaluation wtth CT, delayed CT, CT during arterial portography, and MR imaging. Radiology. 1989;171: 47.
JOURNAL
OF SURGERY@
VOLUME
175
FEBRUARY
1998
101