| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Gastrointestinal Imaging |
1 From the Department of Radiology, Duke University Medical Center, Box 3808, Erwin Rd, Durham, NC 27710 (D.H.S., J.S.K., E.K.P., D.M.D., R.C.N.) and Windsong Radiology, Buffalo, NY (A.M.F.). Received April 15, 1999; revision requested June 2; revision received June 30; accepted August 11. Address reprint requests to D.H.S. (e-mail: sheaf001@mc.duke.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Triple-phase helical computed tomography (CT) was performed in 60 patients with known breast cancer without visible hepatic metastases. Peak hepatic attenuation and enhancement, and attenuation and enhancement at 25 and 30 seconds were obtained. Ratios of hepatic attenuation or enhancement at 25 and 30 seconds to peak hepatic attenuation or enhancement were calculated. A Wilcoxon rank sum test was used to compare patients with and those without subsequent hepatic metastases.
RESULTS: During a mean 18-month follow-up, 18 patients (30%) developed hepatic metastases. Decreases in peak hepatic attenuation and enhancement and increases in hepatic attenuation and enhancement ratios at 25 and 30 seconds were seen in patients who developed metastases compared with those who did not (P < .05). When corrected for chemotherapy interval, these differences were not statistically significant. Using a threshold value of 0.40 or more for the enhancement ratio at 30 seconds resulted in sensitivity of 28%, specificity of 92%, and accuracy of 55%.
CONCLUSION: Patients with breast cancer who develop subsequent hepatic metastases have higher relative hepatic arterial perfusion during triple-phase CT; however, after correction for chemotherapy interval, this difference was not statistically significant. Threshold values cannot be used reliably to identify patients who will develop metastases.
Index terms: Breast neoplasms, metastases, 00.79 • Computed tomography (CT), helical, 761.12114, 761.12115, 761.12119 • Liver, blood supply, 761.91 • Liver, CT, 761.12114, 761.12115, 761.12119 • Liver neoplasms, secondary, 761.332
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The addition of imaging without the administration of contrast material or during the hepatic arterial–dominant phase can improve the sensitivity and specificity of hepatic CT for hypervascular primary tumors (1–3). Even with scan optimization, however, standard CT techniques have a reported sensitivity of 38%–81% for the detection of hepatic metastases (4–6). Furthermore, a subset of patients with no metastases detectable by means of conventional imaging modalities (CT, magnetic resonance imaging, CT portography, intraoperative ultrasonography [US]) will have micrometastases at the time of their initial screening. Detection of hepatic micrometastases could have a profound effect on the treatment of these patients and allow earlier treatment with chemotherapy, bone marrow transplantation, or surgical resection.
Patients with known metastatic disease to the liver have altered hepatic blood flow, with overall increases in hepatic arterial perfusion (7–9). Therefore, it has been hypothesized that patients with subclinical metastatic disease may also have increased hepatic arterial flow. Increases in hepatic arterial flow as measured by means of the Doppler perfusion index have been described in patients with hepatic metastatic disease (10–12). Platt et al (13) recently suggested that hepatic CT enhancement characteristics can be used to identify patients at higher risk for the subsequent development of radiographically identifiable hepatic metastases. Conversely, however, Miles et al (14) showed that increased hepatic arterial perfusion correlated with increased survival in patients with metastatic colorectal carcinoma. The purpose of this study was to determine whether patients with breast cancer who develop hepatic metastases have altered hepatic perfusion detectable by means of functional CT techniques.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
None of the remaining 60 patients (age range, 29–70 years) had detectable metastatic disease on initial scans, and they represent the study group. Absence of metastatic disease was determined independently and unanimously by three radiologists (D.H.S., E.K.P., R.C.N.) experienced in abdominal CT, following blinded review of all three phases of the CT examination. Patient demographic information including age, weight, and time interval between most recent chemotherapy and initial CT scan were recorded prospectively.
The CT protocol was as follows: All images were obtained with a helical CT scanner (HiSpeed Advantage; GE Medical Systems, Milwaukee, Wis) following the oral administration of 800–1,000 mL of 3% diatrizoate meglumine (Gastrografin; Bracco Diagnostics, Princeton, NJ) or 2.1% weight-to-volume ratio of barium sulfate suspension (READI-CAT; E-Z-Em, Westbury, NY). After a nonenhanced study was performed through the entire liver, 175 mL of iopamidol (Isovue 300; Bracco Diagnostics) was injected at 5 mL/sec into an antecubital vein by using a mechanical power injector (Medrad, Pittsburgh, Pa). The delay between administration of contrast material and the onset of scanning was 20 seconds for the arterial-dominant phase and 65 seconds for the portal venous–dominant phase.
We used a 7-mm section thickness and reconstruction interval with a pitch of 1.5 to 1.0 for the scan obtained during the hepatic arterial–dominant phase. We used identical section thickness and reconstruction intervals with a pitch of 1 to 1 for the scans obtained during the nonenhanced and portal venous–dominant phases. We used 140 kVp and 170–220 mA.
For each patient, absolute densitometric measurements of the liver in Hounsfield units were obtained 25 and 30 seconds after initiation of injection of contrast material by using a standardized square region of interest containing 296 voxels. Measurements were performed by observers (J.S.K., n = 10; A.M.F., n = 50) who had no knowledge of the subjects' clinical outcome. The two images obtained at 25 and 30 seconds after administration of contrast material were chosen to allow comparison with enhancement parameters previously reported by Platt et al (13).
For each sampling time, three region-of-interest attenuation measurements were recorded from distinct regions of the liver, which included the medial segment of the left hepatic lobe and the anterior and posterior segments of the right hepatic lobe. Care was taken to not place a cursor over a vascular structure or artifact, when present, and to avoid overlapping regions of interest. The three measurements for each time (ie, image) were then averaged. Mean hepatic attenuation was measured throughout the portal venous phase to determine peak hepatic attenuation and time to peak hepatic attenuation. The initial nonenhanced attenuation value for the liver was also obtained, which allowed calculation of peak hepatic attenuation and enhancement values. In addition, hepatic enhancement and attenuation values were obtained at 25 and 30 seconds following the onset of contrast material injection.
Attenuation and enhancement ratios were calculated subsequently. Specifically, the ratio of hepatic attenuation at 25 seconds to peak hepatic attenuation, the ratio of hepatic attenuation at 30 seconds to peak hepatic attenuation, the ratio of hepatic enhancement at 25 seconds to peak hepatic enhancement, and the ratio of hepatic enhancement at 30 seconds to peak hepatic enhancement were determined.
Clinical and radiologic follow-up were available in all 60 patients, with a mean follow-up of 18 months ± 9 (SD). Forty-two patients (70%) remained free of detectable hepatic metastases, with a mean follow-up of 21 months ± 9 (range, 9–37 months). The remaining 18 patients (30%) developed radiologically documented hepatic metastases within a mean of 10 months ± 5 (range, 2–12 months) from the initial CT examination.
Histopathologic proof was not available in all patients, as hepatic resection or hepatic biopsies are not universally used in patients with metastatic breast cancer. Of patients with metastatic disease, histopathologic proof was available in four patients (22%). In the remaining 14 patients (78%), hepatic metastatic disease was considered present if subsequent CT examinations demonstrated new and/or progressive hepatic metastatic lesions.
A Wilcoxon rank sum test was used to compare the data sets between the two groups of patients. Owing to differences in patient demographics within the study group, an analysis of covariance in which a general linear models procedure was used was also performed. A P value less than .05 was selected to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Figure 1 is a graph of peak hepatic attenuation and peak hepatic enhancement values for patients with and those without the subsequent development of hepatic metastases. Attenuation and enhancement ratios at 25 seconds are presented in Figure 2, whereas attenuation and enhancement ratios at 30 seconds are presented in Figure 3. Although the mean values for peak attenuation and enhancement, as well as ratios at 25 and 30 seconds, were significantly different between groups (all P < .05), a large overlap of values was present.
|
|
|
|
Analysis of patient demographics revealed the following: The mean patient age was 47 years ± 8 with a mean patient weight of 66.8 kg ± 14.1. Forty-eight (80%) of 60 patients underwent chemotherapy prior to their initial CT examination, with a mean interval of 24.3 months ± 26.3 (range, 0.5–31.5 months) from completion of chemotherapy to initial CT examination. The mean weight of patients developing future hepatic metastasis was 71 kg ± 14, compared with 66 kg ± 15 in patients without future metastasis, but this difference was not statistically significant (P > .07).
However, there was a significantly shorter interval between the completion of the last course of chemotherapy and the initial CT scan for patients developing metastases compared with those without future metastases (Fig 4). The mean time following chemotherapy was 380 days ± 630 in patients developing metastases, compared with 910 days ± 840 in patients without future metastasis (P = .002).
|
The adjustment for chemotherapy intervals eliminated any statistically significant differences in hepatic perfusion between groups (see corrected P values, Table 1). Specifically, the ratio of hepatic attenuation at 25 seconds to peak hepatic attenuation, the ratio of hepatic enhancement at 25 seconds to peak hepatic enhancement, the ratio of hepatic attenuation at 30 seconds to peak hepatic attenuation, and the ratio of hepatic enhancement at 30 seconds to peak hepatic enhancement were no longer significantly different between groups (all P 
.1). Whereas there was still a trend toward decreased peak hepatic attenuation and enhancement in patients developing metastasis, these differences were no longer statistically significant (P = .056, .051).
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The normal liver has a dual blood supply; however, hepatic metastases are supplied almost exclusively by the hepatic arteries. This difference in vascular supply has been used to an imaging advantage for macroscopic disease, with improved sensitivities achieved with CT during arterial portography and CT during hepatic arteriography (15). CT also allows quantitation of vascular flow to the liver, which theoretically allows detection of alterations in hepatic hemodynamics caused by metastatic disease.
Results of studies (8,11,12,16) with dynamic scintigraphy and duplex Doppler US have confirmed increased hepatic arterial perfusion relative to portal perfusion in patients with overt hepatic metastases. Functional CT has also shown increased hepatic arterial flow relative to portal venous flow in patients with documented hepatic metastasis (7,17).
Using simpler hepatic densitometry techniques, Platt et al (10) documented significantly increased hepatic attenuation ratios in patients with documented metastatic disease. Specifically, hepatic arterial-phase attenuation at 25 and 40 seconds compared with peak hepatic attenuation was significantly higher in patients with metastatic disease, particularly in those with primary breast tumors (10).
Whereas results of initial studies demonstrated increased hepatic arterial perfusion in patients with documented metastatic disease, changes in hepatic hemodynamics have also been linked to the presence of radiologically occult metastases (11). In additional work by Platt et al (13), increased hepatic arterial perfusion was noted in patients who had a variety of nonhepatic primary neoplasms and subsequently developed hepatic metastases. Conversely, Miles et al (14) reported improved survival and no development of metastases in patients with increased hepatic arterial perfusion.
Our study results suggest that, if we ignore differences in chemotherapy intervals, patients who develop metastases have higher mean arterial phase attenuation and enhancement ratios at 25 and 30 seconds, as well as lower peak hepatic attenuation and enhancement. The effects of chemotherapy interval may be twofold, however. First, the use of chemotherapy may alter hepatic blood flow, with these effects potentially changing over time. Second, a more remote history of chemotherapy use may indicate a population at lower risk for hepatic metastasis compared with patients with more recent chemotherapy.
With the exception of a trend toward increases in peak attenuation and enhancement, correction for chemotherapy interval removed the significance of differences between the populations in our study. Moreover, whereas the mean values were significantly different between groups, there was a large overlap of values in individual patients (Figs 1–3). By using the threshold values for attenuation and enhancement ratios suggested by Platt et al (13), a high number of false negatives and false positives resulted (Table 2). When threshold values were recalculated to fit our data, sensitivities improved, but specificities became unacceptably low, as seen for peak hepatic attenuation and enhancement (sensitivity, 100%; specificity, 14%).
There are limitations to this study that need to be discussed. First, there was a large difference between the chemotherapy interval reported in patients who developed metastases and those who did not. The use of chemotherapy certainly can influence hepatic hemodynamics and potentially reduce the effects of micrometastases on measured attenuation and enhancement values. If patients with radiologically negative results in the study by Platt et al (13) underwent imaging prior to receiving chemotherapy, this could explain differences from our study results; however, the prevalence of chemotherapy use in the patients with occult disease in their study was not provided.
The effects of chemotherapy could not be removed entirely from our analysis. However, when logistic regression analysis was performed to account for this interval, the differences between our study groups were eliminated. Future attempts at predicting metastases may need to rely on initial staging scans obtained prior to the administration of chemotherapy.
Second, additional variables may affect hepatic enhancement characteristics, including patient weight, prandial status, hepatic disease, and hepatic steatosis (17–19). However, we were careful to exclude patients with hepatic steatosis, and there were no significant differences in patient weight between groups. A standardized volume of contrast material was used in our CT examinations, as is routine practice at our institution (Department of Radiology, Duke University Medical Center, Durham, NC).
Whereas use of a higher volume of contrast material compared with that used in prior studies could have influenced optimal threshold values, it does not account for the larger overlap between patient groups. In addition, as we usually performed arterial-phase imaging in less than 20 seconds, enhancement data at 40 seconds was not available for most patients. Thus, our enhancement data at 35 seconds is not entirely equivalent to data in the study by Platt et al (13).
Third, patients who developed overt metastases did not have histopathologic proof in the majority of cases. However, most patients with breast cancer do not undergo hepatic biopsy or resection as a part of management or treatment at our institution. Histopathologic proof of micrometastasis in patients developing subsequent overt metastasis was also not available. Metastases were considered present at follow-up if CT findings were unequivocal or progressed with time.
Whereas recent work in a mouse model suggests that documented micrometastases can alter hepatic hemodynamics, we cannot conclusively assert that patients with altered hemodynamics had micrometastases at the time of that study (20). Without this pathologic proof, we cannot exclude baseline increases in hepatic arterial perfusion that theoretically could increase the risk of subsequent hematogenous metastatic seeding.
Finally, when compared with patients in the study by Platt et al (13), some of the patients without metastasis in our study had a shorter follow-up. If all patients had been followed up for 18 months or longer, theoretically more metastases might have developed. The doubling time of breast cancer is relatively short (as little as 80 days) in premenopausal patient populations (21). Further, doubling times are shortest for small lesions (ie, <12 mm in diameter) (22). Therefore, in the relatively young patient population in our study owing to the referral patterns at our tertiary center, one might expect changes to occur within a year of a patient's initial CT scan, unless the patient was receiving treatment. Of patients developing metastasis in this study, CT documented these lesions within a mean of 11 months. Since pathologic proof of micrometastasis was not available, it is not clear whether patients developing metastases at a longer follow-up would have had normal or abnormal initial CT enhancement characteristics.
In conclusion, the results of our study tend to confirm those of earlier reports that patients developing subsequent hepatic metastases have altered hepatic hemodynamics at triple-phase helical CT. Unfortunately, likely owing to the large number of variables that can affect hepatic hemodynamics, this technique was not sufficiently robust to identify patients at increased risk of hepatic metastases.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. S. Koh, C. H. Thng, P. S. Lee, S. Hartono, H. Rumpel, B. C. Goh, and S. Bisdas Hepatic Metastases: In Vivo Assessment of Perfusion Parameters at Dynamic Contrast-enhanced MR Imaging with Dual-Input Two-Compartment Tracer Kinetics Model Radiology, October 1, 2008; 249(1): 307 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Nakashige, J Horiguchi, A Tamura, T Asahara, F Shimamoto, and K Ito Quantitative measurement of hepatic portal perfusion by multidetector row CT with compensation for respiratory misregistration Br. J. Radiol., September 1, 2004; 77(921): 728 - 734. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. B. Kruskal, P. Thomas, R. A. Kane, and S. N. Goldberg Hepatic Perfusion Changes in Mice Livers with Developing Colorectal Cancer Metastases Radiology, May 1, 2004; 231(2): 482 - 490. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Braga, R. C. Semelka, R. Pietrobon, D. Martin, N. de Barros, and U. Guller Does Hypervascularity of Liver Metastases as Detected on MRI Predict Disease Progression in Breast Cancer Patients? Am. J. Roentgenol., May 1, 2004; 182(5): 1207 - 1213. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K A Miles Perfusion CT for the assessment of tumour vascularity: which protocol? Br. J. Radiol., December 1, 2003; 76(suppl_1): S36 - S42. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K A Miles and M R Griffiths Perfusion CT: a worthwhile enhancement? Br. J. Radiol., April 1, 2003; 76(904): 220 - 231. [Full Text] [PDF]
|
|
|
|
|
|
M A Fuentes, C J Keith, M Griffiths, G Durbridge, and K A Miles Hepatic haemodynamics: interrelationships between contrast enhancement and perfusion on CT and Doppler perfusion indices Br. J. Radiol., January 1, 2002; 75(889): 17 - 23. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
