HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nazarian, L. N. Articles by Wechsler, R. J. Search for Related Content PubMed PubMed Citation Articles by Nazarian, L. N. Articles by Wechsler, R. J.

Size of Colorectal Liver Metastases at Abdominal CT: Comparison of Precontrast and Postcontrast Studies¹

¹ From the Department of Radiology, Thomas Jefferson University Hospital, 7th Fl, Main Bldg, 132 S 10th St, Philadelphia, PA 19107-5244. From the 1998 RSNA scientific assembly. Received December 17, 1998; revision requested February 18, 1999; revision received April 7; accepted June 17. Address reprint requests to L.N.N.

	Abstract

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
PURPOSE: To investigate whether measurements of hepatic metastases from colorectal carcinoma before contrast material administration are significantly different statistically from measurements after contrast material administration.

MATERIALS AND METHODS: Twenty-four patients with hepatic metastases from colorectal carcinoma underwent spiral computed tomography (CT) with 7-mm collimation. The liver was imaged before and in the portal-dominant phase after intravenous contrast material administration. For each scan, one to three discrete liver lesions were selected for measurement (n = 49). Three experienced radiologists performed independent measurements of the selected lesions on both pre- and postcontrast images at a computer workstation. A three-way analysis of variance (ANOVA) was performed: subjects by raters (the three independent radiologists) by pre- or postcontrast status. The dependent variable was the product of bidimensional measurements.

RESULTS: Sixty-seven percent (33 of 49) of the lesions were measured as larger on precontrast images; 33% (16 of 49), as smaller. There was high interrater reliability, with an intraclass correlation coefficient greater than 0.9. ANOVA showed significant subject, rater, and contrast material effects (P < .001) for the largest lesions in each liver. Contrast material status was a significant factor for all lesion sizes (P < .003).

CONCLUSION: On average, hepatic metastases from colorectal carcinoma are significantly smaller after contrast material administration.

Index terms: Colon, neoplasms, 75.321 • Computed tomography (CT), contrast enhancement, 761.12114 • Liver neoplasms, CT, 761.12111, 761.12114, 761.12115, 761.332 • Liver neoplasms, secondary, 761.331, 761.332

	Introduction

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
Over the past 2 decades, there has been extensive research on computed tomographic (CT) detection of liver metastases. It is generally agreed that contrast material should be intravenously administered with a bolus injection and scanning timing should be chosen to optimize attenuation differences between lesions and liver. Spiral CT has made it possible to image the entire liver in a single breath hold. The arterial-dominant phase, portal venous–dominant phase, or both can be demonstrated with the same bolus of contrast material. Hepatic metastases from colorectal carcinoma are typically hypovascular and are best imaged in the portal-dominant phase (1,2). CT detection of hepatic metastases from colorectal carcinoma optimizes treatment planning and can lead to increased survival in patients who meet defined criteria for surgical resectability (3).

There has been relatively little attention given to the optimal method of measuring individual hepatic metastases. These measurements are important, since many patients with unresectable hepatic metastatic disease are being enrolled in clinical trials for treatment of advanced cancer. To assess interval change in size of known hepatic metastases, oncologists rely on serial CT scans obtained at intervals of 4 weeks to several months (4).

Serial follow-up of hepatic metastases with scans obtained after the administration of contrast material, however, may not be the optimal approach. Multiple factors such as cardiac output, hydration status, body weight, renal function, time since last meal, adequacy of venous access, and reactions to contrast material affect both conspicuity and the apparent size of hepatic metastases (5–7). Furthermore, some have postulated that the degree of arterial enhancement of hepatic lesions may decrease after treatment (8). Metastatic lesions may be better followed with CT scanning performed without the administration of contrast material, which may provide more reproducible measurements from examination to examination (1).

This study was performed to investigate whether measurements of hepatic metastases from colorectal carcinoma before the administration of contrast material are significantly different statistically from postcontrast measurements.

	MATERIALS AND METHODS

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
Between November 1997 and February 1998, spiral CT (HiSpeed Advantage; GE Medical Systems, Milwaukee, Wis) was performed on 24 patients (11 men, 13 women; mean age, 63 years; age range, 28–79 years) with biopsy-proved hepatic metastases from colorectal carcinoma.

Oral contrast material (barium sulfate suspension [Readi-Cat 2; E-Z-Em, Westbury, NY] or diatrizoate meglumine and diatrizoate sodium solution [MD-Gastroview; Mallinckrodt Medical, St Louis, Mo]) was administered in all patients. Precontrast images were obtained through the entire liver in one breath hold at 7-mm collimation, 7-mm spacing, and a pitch of 1.5:1. Subsequently, 150 mL of iodinated contrast material, either 60% iothalamate meglumine or 68% ioversol (Conray or Optiray; Mallinckrodt Medical, St Louis, Mo), was administered intravenously at 3–5 mL/sec with a power injector (Medrad, Pittsburgh, Pa) through an 18–22-gauge angiographic catheter (Insyte; Becton Dickinson, Sandy, Utah) inserted into an antecubital vein or, less often, a hand or forearm vein.

Scanning delay was determined by SMARTPREP (GE Medical Systems). A region of interest was placed over liver parenchyma, with care taken to avoid obvious focal lesions or blood vessels. The system was set to initiate scanning when a threshold of 45 HU above baseline was reached. If this threshold was not reached (as occurred in one patient), a default scanning delay of 75 seconds was used to image in a portal venous–dominant phase. Postcontrast scans were identical to precontrast scans in collimation, spacing, and pitch.

All scans were transferred to a SPARC teleradiology workstation (Sun Microsystems, Mountain View, Calif). The precontrast and postcontrast images were viewed separately with RATIONAL IMAGING software (Intuitive Software Technology, West Hills, Calif). Initially, two radiologists (L.N.N., J.H.P.) reviewed the precontrast scans of the entire liver and chose, by consensus, up to a maximum of three of the largest and most well-defined liver lesions per patient. On the image where they appeared most conspicuous (not necessarily largest), lesions were numbered electronically on-screen on the precontrast images, and these numbers were entered onto data sheets. The same lesion was subsequently identified and numbered on the postcontrast images (Fig 1a).

View larger version (124K): [in this window] [in a new window]   Figure 1a. Contrast-enhanced transverse CT scans of the liver in a 59-year-old woman with metastatic colonic carcinoma. (a) Scan demonstrates how the largest, most discrete lesions were numbered at the computer workstation for later analysis. (b) Scan shows bidimensional measurements (the longest axis and its perpendicular).

View larger version (120K): [in this window] [in a new window]   Figure 1b. Contrast-enhanced transverse CT scans of the liver in a 59-year-old woman with metastatic colonic carcinoma. (a) Scan demonstrates how the largest, most discrete lesions were numbered at the computer workstation for later analysis. (b) Scan shows bidimensional measurements (the longest axis and its perpendicular).

The product of bidimensional measurements, which is proportional to lesion area, was used for statistical analysis (10) and will hereafter be referred to as "lesion area." The mean and SD for lesion area were computed for precontrast and postcontrast images as a mean for all three raters. The number of lesions measured as larger or smaller on postcontrast images was tabulated. Because lesions in each patient were undoubtedly correlated and did not constitute independent occurrences, separate analyses of variance (ANOVAs) were performed on each of the three lesions to determine whether there was a significant change in measured lesion size on the basis of rater and precontrast versus postcontrast status. Since the variance of the measured lesion area increased with lesion size, a natural logarithm transformation was applied to the data prior to ANOVA.

Reliability was assessed by using the intraclass correlation coefficient (ICC) for raters considered representative of a large population of similar raters. Portnoy and Watkins' formula "ICC" (11, eq [2,1]) was calculated separately for the six lesion measurements (pre- and postcontrast for lesions 1–3). The ICC is a ratio of between-subject, within-subject, and error variance derived from the ANOVAs. ICC is functionally equivalent to certain forms of the more familiar coefficient.

	RESULTS

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
A total of 49 lesions were measured in the 24 patients. At least one lesion was measured in each liver. By using the mean measurements of the three readers, 33 of 49 (67%) lesions measured larger on the precontrast images (Figs 2, 3) and 16 of 49 (33%) measured smaller (Fig 4). There was extremely high interrater reliability. The three pre- and postcontrast lesion measurements all had ICCs greater than 0.9. Table 1 shows the mean sizes of the liver lesions both before and after administration of contrast material. ANOVA showed significant subject, rater, and contrast material enhancement effects (P < .001) for the largest lesions in each liver. The rater factor was less significant for smaller lesions, but the presence of intravenous contrast material was a significant factor at all lesion sizes (P < .003). Thus, the analysis demonstrates that measurements of lesion area in the portal-dominant phase of postcontrast imaging are significantly different from those obtained at precontrast imaging. The differences (in percent) in lesion measurements, defined as [(CM_Pre – CM_Post) x 100%] / CM_Pre, where CM_Pre is precontrast measurement and CM_Post is postcontrast measurement, are presented in Table 2 and are shown in graphic form in Figure 5.

View larger version (117K): [in this window] [in a new window]   Figure 2a. Transverse CT images of the liver in a 59-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows a hepatic lesion (arrow) in the anterior segment of the right lobe. (b) Postcontrast image shows the lesion (arrow) to appear smaller than on the precontrast image.

View larger version (123K): [in this window] [in a new window]   Figure 2b. Transverse CT images of the liver in a 59-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows a hepatic lesion (arrow) in the anterior segment of the right lobe. (b) Postcontrast image shows the lesion (arrow) to appear smaller than on the precontrast image.

View larger version (118K): [in this window] [in a new window]   Figure 3a. Transverse CT images of the liver in a 68-year-old woman with metastatic colonic carcinoma. (a) Precontrast image shows a large calcified hepatic metastasis. (b) Postcontrast image shows the metastasis as smaller.

View larger version (112K): [in this window] [in a new window]   Figure 3b. Transverse CT images of the liver in a 68-year-old woman with metastatic colonic carcinoma. (a) Precontrast image shows a large calcified hepatic metastasis. (b) Postcontrast image shows the metastasis as smaller.

View larger version (156K): [in this window] [in a new window]   Figure 4a. Transverse CT images of the liver in a 65-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows an ill-defined lesion (arrow) in the posterior segment of the right lobe of the liver. (b) Postcontrast CT image shows the lesion as larger and more conspicuous.

View larger version (160K): [in this window] [in a new window]   Figure 4b. Transverse CT images of the liver in a 65-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows an ill-defined lesion (arrow) in the posterior segment of the right lobe of the liver. (b) Postcontrast CT image shows the lesion as larger and more conspicuous.

View larger version (28K): [in this window] [in a new window]   Figure 5. Graph shows the distribution of differences (in percent) between precontrast and postcontrast lesion measurements among the three radiologists. The two groupings of bars to the right depict the observations in which precontrast measurements were at least 10% larger than postcontrast measurements. The two groupings of bars to the left show observations in which precontrast measurements were at least 10% smaller than postcontrast measurements.

	DISCUSSION

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

Oncologists generally ask the radiologist to report bidimensional measurements of all indicator lesions (the largest, most well-defined lesions); oncologists multiply these measurements to obtain a number proportional to the area of the lesion, and it is this number that is compared from scan to scan (10). Decisions on whether to continue, suspend, or alter chemotherapy are based on whether the patient's condition is stable, progressing (more lesions or greater than 25% enlargement of lesions), or regressing (disappearance of lesions or greater than 50% shrinkage of lesions) (10,12). These percentages have been chosen to ensure that observed changes in lesion size are not simply due to intra- and interobserver variation, which can range from 5% to 15% (13,14).

Other authors (15) have advocated using a three-dimensional volume method to observe tumor bulk with CT and have shown that this method differs significantly from the two-dimensional method. Although volumetric methods can theoretically provide more accurate assessment of tumor bulk and compensate for irregularities in lesion shape (4), the bidimensional method remains the standard.

Previous studies (16) have shown both pre- and postcontrast scans to be sensitive and specific in detecting hepatic masses. However, postcontrast scans are preferred due to approximately 10%–15% increased sensitivity (17) and increased diagnostic confidence (16). In fact, for routine evaluation of the liver for hypovascular metastases such as colonic carcinoma, most authors have advocated use of scans obtained solely in the portal venous phase (2,18).

However, relying on postcontrast images alone can be hazardous, since the conspicuity and apparent size of liver lesions is so dependent on scanning timing (1,8). Even in hypovascular metastases such as colonic carcinoma, contrast material diffusion into the extravascular space of the tumor decreases liver-to-lesion contrast and may eventually lead to partial or complete disappearance of the tumor (6,19,20). This phenomenon typically occurs in highly cellular, nonnecrotic metastases imaged during the equilibrium phase of contrast material enhancement when the iodine concentration in the liver is declining (19,21). Therefore, hepatic metastases imaged at two times in the same patient may appear different in size even when no real change has occurred.

Silverman et al (6) stressed the importance of performing CT scanning in patients with cancer during the same phase of contrast material enhancement. Otherwise, an artifactual decrease or increase in lesion size can occur, which gives a false impression of either lesion regression or progression. Several authors (1,20) have suggested that nonenhanced images provide tumor measurements that are more accurate and more consistent from study to study. Before deciding whether precontrast imaging is a better way to follow the size of hepatic metastases, we set out to determine whether there was a difference between measured lesion sizes on pre- and postcontrast images.

Our findings show that measurements of hepatic lesions on precontrast images are, on average, significantly larger than measurements on postcontrast images. This is true even though we used an automated computer technique (SMARTPREP) to attempt to obtain a more consistent level of hepatic enhancement from patient to patient than we would with a fixed scanning delay time (6). These data may have important ramifications for current radiologic practice.

If two interval studies are being compared, one with contrast material enhancement and one without, assessment of interval change in lesion size is problematic. Unfortunately, this is not a rare occurrence, since intravenous contrast material on any given day may have to be withheld due to renal failure, fluid status, poor venous access, patient refusal, or contrast material allergy at a prior examination. We contend that obtaining a nonenhanced scan through the liver at each examination could serve as a reproducible method of comparison. Ideally, postcontrast scanning would still be performed to optimize the detection of new lesions (Fig 6).

View larger version (120K): [in this window] [in a new window]   Figure 6a. Transverse CT images of the liver in a 71-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows a single lesion (arrow) in the right lobe of the liver. (b) Postcontrast image shows a smaller right lobe lesion (black arrow) and an additional lesion (white arrow) in the medial segment of the left hepatic lobe.

View larger version (121K): [in this window] [in a new window]   Figure 6b. Transverse CT images of the liver in a 71-year-old man with metastatic colonic carcinoma. (a) Precontrast image shows a single lesion (arrow) in the right lobe of the liver. (b) Postcontrast image shows a smaller right lobe lesion (black arrow) and an additional lesion (white arrow) in the medial segment of the left hepatic lobe.

Another potential limitation is that the sizes of metastases on pre- and postcontrast images may be difficult to compare directly if misregistration occurs from patient breathing, motion, and so on. In theory, spiral CT that obtains a volumetric data set helps limit misregistration by enabling retrospective reconstruction at overlapping intervals. However, such manipulation must be performed before raw data are deleted (18), and, because of time and computer memory constraints, we were unable to alter scan spacing in our study.

Other limitations included lack of an independent standard to define lesion size. For example, lesions may have been measured as larger on precontrast images simply because they were more poorly defined than after contrast material administration. Magnetic resonance imaging or intraoperative ultrasonographic correlation may have been useful as an independent correlation of true lesion size.

Finally, the patients in the study had been treated with various methods. It is possible that in patients treated with chemotherapy, tumor vascularity could have differed from that in nontreated patients, thus affecting the appearance of lesions after contrast material administration.

Although the ICC, a measure of agreement, was extremely high (>0.9) in our study, there was still a mean difference between raters in lesion size measurement, which amounted to about 10% of the mean lesion size. Therefore, we recommend that the same radiologist perform measurements on both current and comparison CT studies (14). Intrarater differences were not measured in this study but were shown by Hopper and associates (14) to be about 6% for individual tumor foci. However, in the study by Hopper et al, the radiologists were permitted to select different indicator lesions for measurement. We removed this variable by telling the radiologists which lesions to measure. In doing so, we could have introduced selection bias. We tried to minimize this potential bias by committing ourselves to choosing the lesions from the precontrast images before knowing what happened to the lesions after the administration of contrast material.

In conclusion, the results of our study suggest that colorectal carcinoma metastases appear different in size on pre- and postcontrast CT scans. We are currently investigating whether the overall impression of disease progression, stability, or regression in comparing of precontrast scans with each other would differ from that of the current practice of comparing postcontrast scans with each other.

	Footnotes

 
Abbreviations: ANOVA = analysis of variance ICC = intraclass correlation coefficient

Author contributions: Guarantor of integrity of entire study, L.N.N.; study concepts, all authors; study design, L.N.N., J.H.P., L.P., R.J.W.; definition of intellectual content, L.N.N.; literature research, L.N.N.; clinical studies, L.N.N., J.H.P., E.J.H., A.S.L.T., P.T.J.; data acquisition, J.H.P.; data analysis, L.N.N., J.H.P., L.P., E.J.H., R.J.W.; statistical analysis, L.P., E.J.H.; manuscript preparation, L.N.N.; manuscript editing and review, all authors.

	References

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 

1. Baron RL. Understanding and optimizing use of contrast material for CT of the liver. AJR 1994; 163:323-331.[Abstract/Free Full Text]
2. Zeman RK, Baron RL, Jeffrey RB, et al. Helical body CT: evolution of scanning protocols. AJR 1998; 170:1427-1438.[Free Full Text]
3. Sugarbaker PH. Surgical decision making for large bowel cancer metastatic to the liver. Radiology 1990; 174:621-626.[Free Full Text]
4. Van Hoe L, Van Cutsem E, Vergote I, et al. Size quantification of liver metastases in patients undergoing cancer treatment: reproducibility of one-, two-, and three-dimensional measurements determined with spiral CT. Radiology 1997; 202:671-675.[Abstract/Free Full Text]
5. Dodd GD, Baron RL. Investigation of contrast enhancement in CT of the liver: the need for improved methods. AJR 1993; 160:643-646.[Free Full Text]
6. Silverman PM, Roberts S, Tefft MC, et al. Helical CT of the liver: clinical application of an automated computer technique, SmartPrep, for obtaining images with optimal contrast enhancement. AJR 1995; 165:73-78.[Abstract/Free Full Text]
7. Walkey MM. Dynamic hepatic CT: how many years will it take 'til we learn?. Radiology 1991; 181:17-24.[Free Full Text]
8. Paulson EK, McDermott VG, Keogan MT, DeLong DM, Frederick MG, Nelson RC. Carcinoid metastases to the liver: role of triple-phase helical CT. Radiology 1998; 206:143-150.[Abstract/Free Full Text]
9. Fornage BD. Measuring masses on cross-sectional images (letter). Radiology 1993; 187:289.[Free Full Text]
10. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981; 47:207-214.[Medline]
11. Portnoy LG, Watkins MP. Statistical measures of reliability. In: Portnoy LG, Watkins MP, eds. Foundations of clinical research: applications to practice. Norwalk, Conn: Appleton & Lange, 1993; 512.
12. Hayward JL, Carbone PP, Heuson JC, Kumaoka S, Segaloff A, Rubens RD. Assessment of response to therapy in advanced breast cancer. Cancer 1977; 39:1289-1294.[Medline]
13. Moertel CG, Hanley JA. The effect of measuring error on the results of therapeutic trials in advanced cancer. Cancer 1976; 38:388-394.[Medline]
14. Hopper KD, Kasales CJ, Van Slyke MA, Schwartz TA, TenHave TR, Jozefiak JA. Analysis of interobserver and intraobserver variability in CT tumor measurements. AJR 1996; 167:851-854.[Abstract/Free Full Text]
15. Hopper KD, Kasales CJ, Eggli KD, et al. The impact of 2D versus 3D quantitation of tumor bulk determination on current methods of assessing response to treatment. J Comput Assist Tomogr 1996; 20:930-937.[Medline]
16. Berland LL, Lawson TL, Foley WD, Melrose BL, Chintapalli KN, Taylor AJ. Comparison of pre- and postcontrast CT in hepatic masses. AJR 1982; 138:853-858.[Abstract/Free Full Text]
17. Nelson RC. Techniques for computed tomography of the liver. Radiol Clin North Am 1991; 29:1199-1212.[Medline]
18. Berland LL. Slip-ring and conventional dynamic hepatic CT: contrast material and timing considerations. Radiology 1995; 195:1-8.[Abstract/Free Full Text]
19. Cox IH, Foley WD, Hoffmann RG. Right window for dynamic hepatic CT. Radiology 1991; 181:18-21.[Free Full Text]
20. Oliver JH, Baron RL, Federle MP, Jones BC, Sheng RS. Hypervascular liver metastases: do unenhanced and hepatic arterial-phase CT images affect tumor detection?. Radiology 1997; 205:709-715.[Abstract/Free Full Text]
21. Foley WD. Dynamic hepatic CT. Radiology 1989; 170:617-622.[Abstract/Free Full Text]

This article has been cited by other articles:

				S. C. Eberhardt, P. H. Choi, A. M. Bach, S. A. Funt, H. E. Felderman, and L. E. Hann Utility of Sonography for Small Hepatic Lesions Found on Computed Tomography in Patients With Cancer J. Ultrasound Med., April 1, 2003; 22(4): 335 - 343. [Abstract] [Full Text] [PDF]

				E. Neri, P. Giusti, L. Battolla, P. Vagli, P. Boraschi, R. Lencioni, D. Caramella, and C. Bartolozzi Colorectal Cancer: Role of CT Colonography in Preoperative Evaluation after Incomplete Colonoscopy Radiology, June 1, 2002; 223(3): 615 - 619. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nazarian, L. N. Articles by Wechsler, R. J. Search for Related Content PubMed PubMed Citation Articles by Nazarian, L. N. Articles by Wechsler, R. J.