Gastrointest Radiol 13:145-151 (1988)
Gastrointestinal
Radiology 9 Springer-VerlagNew York Inc. 1988
Progressive CT Appearance of Hepatic Metastases from Colorectal Carcinoma Janis G. Letourneau, 1 William M. Thompson, 1 Marvin E. Goldberg, 1 Dale C. Snover, 2 Theodor B. Grage, 3 and Mathis P. Frick 1'4 Departments of 1Diagnostic Radiology, 2Surgical Pathology, and 3Surgery, University of Minnesota, Minneapolis, Minnesota; and 4Department of Radiology, Creighton University, Omaha, Nebraska, U S A
Abstract. An 8-year retrospective review of 106 serial computed tomographic (CT) examinations performed on 32 patients with colorectal carcinoma metastatic to the liver was done to determine if the CT appearance of such metastases changed with a favorable response to chemotherapy or with progression of disease. Of these 32 patients, 15 underwent placement of an infusion pump for delivery of chemotherapy directly into the hepatic artery, 3 underwent partial hepatectomy, 1 underwent both procedures, and 13 underwent neither. Regression of hepatic metastases (7 patients), only seen following infusion pump placement, was associated with a decrease in size and an increase in margination of lesions. In two of these patients regression of metastases was seen in one area of the liver with subsequent progression or development of metastases in another region, presumably due to preferential delivery of chemotherapeutic agent. Progression of disease (23 patients) was associated with an increase in both size and number of lesions that became progressively less well marginated. Development of poorly marginated or infiltrative characteristics at the periphery of the lesion was associated with a poor prognosis. Thus, the CT characteristics of hepatic metastases from colorectal carcinoma differ with a favorable response to chemotherapy and with progression of disease. Key words: Liver, secondary neoplasms - Colon, carcinoma - Liver metastasis, CT diagnosis.
carcinoma is well established [1-9]. The major problems with CT staging of colon and rectal carcinoma relate to difficulty in determining local infiltration and regional lymph node involvement [2]. Sensitivity and specificity of detection of hepatic metastases with CT are less problematic [2-4, 10-13]. In our institution CT is often used to determine the mode of treatment in patients with metastatic disease. Subtotal hepatectomy is performed when a small number of hepatic metastases are present in the absence of extrahepatic disease and when surgery is not contraindicated for some other reason. In unresectable cases an infusion pump is placed to allow delivery of chemotherapeutic agents directly into the hepatic artery. Occasionally, and more commonly early in this series, systemic chemotherapy is instituted. The CT appearance of hepatic metastases has been evaluated with serial CT examinations in a large number of patients at our hospital. The character of the hepatic metastases usually changed substantially during the course of follow-up. This study was designed to determine if the changes in the CT appearance of such hepatic metastases are different with a favorable response to chemotherapy and with progression of disease. To our knowledge there has been no report in the radiographic literature correlating the CT findings with these different clinical courses.
Subjects and Methods
The role of computed tomography (CT) in the staging and follow-up of patients with colorectal Address reprint requests to: Janis G. Letourneau, M.D., Box 292 U M H C , Harvard Street at East River Road, Minneapolis, M N 55455, U S A
Between August, 1978, and February, 1986, 32 patients with adenocarcinoma of the colon and rectum metastatic to the liver underwent 106 CT examinations. Each patient had at least two studies over a minimum scan interval of two months and over a maximum scan interval of 25 months. Of these 32 patients 18 were men and 14 women. Patient ages ranged from 30 to 84 years, with an average of 54.9 years. Patients with more than 1 known malignancy were excluded from this series. Of the
146
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases
patients included in the series, 15 underwent placement of a hepatic artery infusion pump for chemotherapy, 3 underwent partial hepatectomy, i underwent both procedures, and 13 underwent neither procedure, but received systemic chemotherapy or limited hepatic artery infusion of chemotherapy through a percutaneous transfemoral catheter. The CT scans were reviewed retrospectively with specific attention to certain features: number of metastases, size of metastases, degree of lesion margination, lesion inhomogeneity, and degree and nature of lesion calcification. Examinations were performed on a state-of-the-art CT scanner. Generally, CT scans were done following administration of intravenous contrast as a bolus followed by a drip infusion. Medical records of these patients were reviewed, noting major therapeutic interventions, such as placement of an infusion pump for chemotherapy and partial hepatectomy. The clinical status of the patient was determined in large part from the clinic note assessments, which synthesized all available data, including functional capacity of the patient, liver size, CT findings, and results of laboratory studies, such as level of liver enzymes and carcinoembryonic antigen. Twenty-one patients have died, all with evidence of worsening hepatic metastases. Five, including one who has undergone partial hepatectomy and one who has undergone both partial hepatectomy and infusion pump placement, have evidence of progressive metastatic disease within the liver. At the conclusion of the study metastatic disease within the liver is either stable or regressed in the remaining six pati, ents; five of these patients have hepatic infusion pumps in place and one has undergone partial hepatectomy. Two of these latter patients, both with hepatic artery infusion pumps, have CT evidence of progressive metastatic disease outside of the liver. Overall, seven patients have had laboratory and definitive CT evidence of temporary or sustained regression of hepatic metastases. All of these patients were treated with intrahepatie chemotherapy delivered by an infusion pump. Three of these patients are now dead, all with extensive extrahepatic disease, one has evidence of progression of hepatic disease, and three continue to have stable but diminished hepatic disease.
Results
The CT appearance of the hepatic metastases at the time of detection was heterogeneous and is summarized in Table 1. Most lesions were greater than 3 cm in diameter at the time of the first scan (88% of patients). The number of identifiable metastases was highly variable, but commonly only one or two lesions were seen (53% of patients). The degree of margination of metastases was also variable, with most lesions being well or moderately well circumscribed (75% of patients) and others being irregular and infiltrative at their borders (25% of patients). Upon initial detection the lesions were frequently inhomogeneous centrally following intravenous contrast enhancement (59% of patients). Both fine and coarse calcification was also frequently demonstrated (50% of patients). Pathologic examination of partial hepatectomy specimens was available in two patients following CT detection of liver metastases. In one patient
Table 1. Initial CT appearance of hepatic lesions (32 patients)
Number of patients Size Small (1-2.9 cm) Medium (3 4.9 cm) Large ( >__5 cm)
4 13 15
Number 1-2 3-5 >5
17 9 6
Margination Well circumscribed Moderately well circumscribed Poorly circumscribed Irregularly marginated
9 15 4 4
Homogeneity Homogeneous Inhomogeneous
13 19
Calcification Present Absent
16 16
Table 2. CT appearance of lesions with temporary or sustained
regression of metastases (7 patients) Number of patients Size Decreased Increased
7 0
Number Decreased Increased Unchanged
0 0 (2) a 7
Margination Decreased Increased Unchanged
1 5 1
Homogeneity Decreased Increased Unchanged
1 5 1
Calcification Increased Not present
5 2
a New lesions were ultimately detected elsewhere within liver while old lesions showed evidence of regression.
well-marginated lesion on CT was characterized grossly and microscopically by a sharp, regular interface between the tumor and the adjacent hepatic parenchyma (Fig. 1 A, B). In the other patient a less well-marginated lesion showed numerous irregular extensions of tumor into the adjacent nora
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases
147 Fig. 1. A CT scan reveals a wellmarginated focus (arrow) of low attenuation in the inferior aspect of the right hepatic lobe. B Photograph of the gross pathologic specimen obtained at partial hepatectomy from the patient in A corroborates the well-circumscribed margin of the lesion. C CT scan of another patient demonstrates a central area of enhancement with somewhat irregular boundaries (arrows). D The appearance of the margin of the gross pathologic specimen also obtained at partial hepatectomy correlates with the CT findings. The microscopic pathologic study of this specimen showed multiple small metastatic foci surrounding a dominant large lesion with an irregular border.
mal hepatic parenchyma. In addition, m a n y satellite microscopic foci of t u m o r were seen adjacent to the main t u m o r although separated from it by normal hepatocytes (Fig. 1 C, D). With regression o f metastatic disease within the liver, the size of the lesions always decreased to a variable percentage o f the original size (Table 2, Fig. 2). All lesions remained detectable. Margination, homogeneity, and calcification of lesions each increased in five patients (71%). In one patient, however, margination of the hepatic lesions was irregular initially and appeared even slightly more
infiltrative with treatment, despite a reduction in size of the metastases. Response to treatment was not sustained in this patient; the hepatic lesions subsequently increased in size and the patient died. Two patients d e m o n s t r a t e d a decrease in size and increase in margination o f lesions situated predominantly within one lobe, but showed progression of metastatic involvement within the other lobe (Fig. 3). The serial CT appearance o f hepatic lesions of the 23 patients who have died or have evidence of progressive hepatic metastases and who have
148
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases Table 3. CT appearance of lesions with progression of metastases (23 patients)" Number of patients Size Decreased Increased
1 22
Number Decreased Increased Unchanged
0 16 7
Margination Decreased 19 Well circumscribed 2 Increased 1 Moderately-well circumscribed 1 Unchanged 3 Poorly circumscribed 7 Irregularly marginated 13 Homogeneity Decreased Increased Unchanged
12 8 3
Calcification Increased Unchanged Not present
15 3 5
" Three patients with progressive metastatic disease who underwent partial hepatectomy are not included in this table because the initial lesions detected by CT were surgically removed.
not undergone partial hepatectomy is summarized in Table 3. The tendency toward loss of lesion margination was accompanied by a change in distribution of the character of metastases to poorly marginated or infiltrative in appearance (87% of patients) (Fig. 4). Inhomogeneity of the lesions commonly increased (52% of patients), as did degree of lesion calcification (65% of patients). Discussion
Because of its ability to detect subtle differences in density within soft tissue, CT is well suited to the characterization of hepatic metastases. The CT appearance of hepatic metastases from colorectal carcinoma at the time of detection is heteroge-
149
neous. Upon diagnosis most patients have a small number of lesions and most lesions are relatively large and well circumscribed. Relatively large lesion size in colorectal carcinoma metastatic to the liver has been seen by other investigators [14]. Pathologic material available in two of the study patients correlated with the CT appearance of the hepatic lesions. Thus, it seems that there may be some pathologic basis for the variable degree of lesion margination seen on CT. Frequently the soft tissue component of the lesion is inhomogeneous, and this likely relates to central necrosis and fibrous stroma as well as malignant tissue [14, 15]. At the time of diagnosis calcification is often present within the lesion. Because a small number of lesions are commonly seen, a relatively large proportion of these patients would be potential candidates for partial hepatectomy as a means of controlling metastatic disease [16]. There is some evidence in clinical series that patients with this disease who undergo partial hepatectomy have improved survival rates [16-18]. This was the case in our series, in which average survival following partial hepatectomy was 16 months and that following placement of hepatic artery infusion pump was 10 months. Obviously, determination of appropriate treatment in these patients is very complicated because of the multiple influencing factors and is beyond the scope of this paper. Regression of metastatic disease within the liver was seen in this series only in patients treated with hepatic artery chemotherapy by infusion pump. Efficacy of hepatic infusion chemotherapy with an implanted pump has been described in the oncology literature [19], although prospective trials of this mode of treatment have been limited [20]. Regression of metastases in our series was usually accompanied by reduction in size of lesions that became better defined at the periphery. All lesions remained detectable. Complete disappearance of lesions by CT has, however, been reported by other authors [19]. Regression of metastases was seen within one area of the liver despite appearance or progression of other lesions elsewhere within the
Fig. 2, A Metastatic lesions (arrows) are identified within the medial segment of the left lobe and the posterior segment of the right lobe. B With hepatic artery infusion chemotherapy, these lesions shrink and densely calcify (arrows), Fig. 3. A An ill-defined focus of decreased attenuation (solid arrow) is seen within the left hepatic lobe, There are also 3 ill-defined foci: 1 near the interlobar fissure and 2 more posteriorly in the right lobe (open arrows). B Postchemotherapy scan reveals some calcification and regression of the 3 latter lesions despite progression of the metastasis in the left lobe. Fig. 4. A A lobular area of decreased attenuation (arrow) is seen in the inferior aspect of the right lobe. B Clinical worsening was seen in conjunction with massive enlargement of this lesion. The margin of the lesion is now infiltrative in character.
150
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases
liver and is presumed, but not proven angiographically, to be due to differential delivery of chemotherapeutic agents to one area. Such differential delivery could be attributed to variable flow phenomena within normal vessels or to impaired flow through partially thrombosed vessels. Further investigation of such presumed flow differentials could be done with scintigraphic [21] or angiographic techniques and may be valuable in those patients in whom extrahepatic disease is well controlled. Progression of metastatic disease within the liver in patients who remain alive and in those that have died was associated with strong tendencies for lesions to increase in both size and number. The periphery of these lesions tended to become less well circumscribed and the development of poorly defined or infiltrative appearing margins was associated with a poor prognosis. Once these characteristics were identified mean patient survival was 4.5 months. It is interesting that poorly defined or irregularly marginated lesions were seen initially in eight patients, seven of whom were treated with chemotherapy and one with partial hepatectomy. Mean patient survival was 10 months in this group; three patients had temporary regression of hepatic metastases and one had no recurrence following surgical resection of disease. Such infiltrative features identified with CT may be related to the pathologic findings of infiltration of tumor along the liver sinusoids seen in advanced metastases of colorectal carcinoma. Small deposits of metastases around the periphery of larger lesions may also result in the CT appearance of lesion irregularity and inhomogeneity. Increased inhomogeneity of hepatic lesions may be due to development of central necrosis [15], central fibrous stroma [14], or a mixture of malignant elements, central necrosis, and fibrous tissue [14]. In summary, the CT appearance of hepatic metastases from colorectal carcinoma was heterogeneous at the time of initial detection. Variability in appearance was corroborated in two cases by the pathologic material available from partial hepatectomy. Although lesions in this series remained detectable by CT following a favorable response to hepatic artery infusion of chemotherapy, their appearance changed, decreasing in size and becoming better marginated. Regression of metastases within one region of the liver and progression within another was likely to be caused by differential delivery of chemotherapeutic agents. Deterioration of the clinical status of the patient was associated with an increase in both the size and number of lesions that were generally less well circum-
scribed. Preterminal scans frequently demonstrated poorly marginated or infiltrative appearing lesions and suggested a poor prognosis. Acknowledgments. We thank Fran Leppanen, R.N., Department of Surgery, University of Minnesota, for her help in accumulating patient data on this series and Wilfrido R. Castaneda, M.D., Department of Diagnostic Radiology, University of Minnesota, for his critical review of the manuscript.
References 1. Clark J, Bankoff M, Carter B, Smith TJ : The use of computerized tomography scan in the staging and follow-up study of carcinoma of the rectum. Surg Gynecol Obstet 159:335-342, 1984 2. Freeny PC, Marks WM, Ryan JA, Bolen JW: Colorectal carcinoma evaluation with CT: preoperative staging and detection of postoperative recurrence. Radiology 158: 347-353, 1986 3. Mayes GB, Zornoza J: Computed tomography of colon carcinoma. A JR 135:43-46, 1980 4. McCarthy SM, Barnes D, Deveney K, Moss AA, Goldberg HI: Detection of recurrent rectosigmoid carcinoma: prospective evaluation of CT and clinical factors. A JR 144: 577-579, 1985 5. Moss AA, Thoeni RF, Schnyder P, Margulis AR: Value of computed tomography in the detection and staging of recurrent rectal carcinomas. Y Comput Assist Tomogr 5:870-874, 1981 6. Thoeni RF, Moss AA, Schnyder P, Margulis AR: Detection and staging of primary rectal and rectosigmoid cancer by computed tomography. Radiology 141:135-138, 1981 7. vanWaes PFGM, Koehler PR, Feldberg MAM: Management of rectal carcinoma: impact of computed tomography. A JR 140:1137-1142, 1983 8. vanWaes PFGM, Koehler PR, Feldberg MAM : CT of rectal cancer: its accuracy and effect on patient management. Radiographics 4: 801-819, 1984 9. Zaunbauer W, Haertel M, Fuchs WA: Computed tomography in carcinoma of the rectum. Gastrointest Radiol 6:79-84, 1981 10. Alderson PO, Adams DF, McNeil BJ, Sanders R, Siegelman SS, Finberg HJ, Hessel SJ, Abrams HL: Computed tomography, ultrasound, and seintigraphy of the liver in patients with colon or breast carcinoma: a prospective comparison. Radiology 149: 225-230, 1983 11. Bernardino ME, Erwin BC, Steinberg HV, Baumgartner BR, Torres WE, Gedgaudas-McClees RK : Delayed hepatic CT scanning: increased confidence and improved detection of hepatic metastases. Radiology 159: 71-74, 1986 12. Finlay IG, Meek DR, Gray HW, Duncan JG, McArdle CS: Incidence and detection of occult hepatic metastases in colorectal carcinoma. Br Med J 284: 803-805, 1982 13. Finlay IG, McArdle CS: Effect of occult hepatic metastases on survival after curative resection for colorectal carcinoma. Gastroenterology 85: 596~599, 1983 14. Muramatsu Y, Takayasu K, Moriyama N, Shima Y, Goto H, Ushio K, Yamada T, Hasegawa H, Koyama Y, Hirohashi S : Peripheral low-density area of hepatic tumors : CTpathologic correlation. Radiology 160:49-52, 1986 15. Takayasu K, Moriyama N, Muramatsu Y, Suzuki M, Yamada T, Kishi K, Hasagawa H, Okazaki N: Hepatic arterial embolization for hepatocellular carcinoma: comparison of CT scans and resected specimens. Radiology 150:661-665, 1984
J.G. Letourneau et al. : CT of Hepatic Colorectal Carcinoma Metastases 16. Adson MA: Diagnosis and surgical treatment of primary and secondary solid hepatic tumors in the adult. Surg Clin North Am 61:181-196, 1981 17. Little JM: Hepatic secondaries: minimal tumor and resectable tumor. Worm J Surg 8:753-756, 1984 18. Morrow CE, Grage TB, Sutherland DER, Najarian JS: Hepatic resection for secondary neoplasms. Surgery 92:610-614, 1982 19. Shepard KV, Levin B, Karl RC, Faintuch J, DuBrow RA, Hagle M, Cooper RM, Beschorner J, Stablein D: Therapy for metastatic colorectal cancer with hepatic artery infusion chemotherapy using a subcutaneous implanted pump. J Clin Oncol 3:161-169, 1985
151
20. Kemeny N, Daly J: Preliminary results of a randomized study of intrahepatic infusion versus systemic infusion of 5-fluor-2-deoxyuridine for metastatic colorectal carcinoma. Recent Results Ca Res 100:171-178, 1986 21. Kaplan WD, D'Orsi CJ, Ensminger WD, Smith EH, Levin DC: Intra-arterial radionuclide infusion: a new technique to assess chemotherapy perfusion patterns. Cancer Treat Rep 62:69%703, 1978
Received: June 12, 1987; accepted: July 20, 1987