Abstract
Objectives
Reports on the sensitivity and accuracy of contrast-enhanced helical computed tomography (HCT) in the preoperative evaluation of colorectal liver metastases (CLM) have been conflicting. Few studies have controlled for and reported on the time interval between HCT and eventual surgery.
Methods
A multi-institution, retrospective review of consecutive patients who underwent hepatic resection for CLM from January 1999 to September 2004 was conducted. Data regarding lesion characteristics and resectability were extracted from radiology reports, operative findings and histopathological records. Findings in HCT were evaluated according to their sensitivity for detecting CLM and ability to predict resectability.
Results
A total of 217 consecutive patients who underwent hepatic resection for CLM were identified. The overall sensitivity of HCT for detection of CLM was 83.2%. Prolonged time between imaging and surgery was a negative predictor for HCT sensitivity in univariate and multivariate analysis (P < 0.001). In predicting resectability, preoperative HCT was accurate 77.0% of the time. The time interval to surgery was negatively correlated with HCT prediction accuracy in univariate and multivariate analyses (P < 0.001).
Conclusions
The utility of HCT as a preoperative tool to evaluate CLM is inversely proportional to the time interval between imaging and surgery. This may explain conflicting reports of the accuracy of HCT in the current literature.
Keywords: colorectal metastases, liver, resection
Introduction
Colorectal liver metastasis (CLM) is a significant source of morbidity and mortality in patients with colorectal cancer (CRC). Up to 25% of patients present with CLM at the time of CRC diagnosis; a further 40% develop CLM following primary CRC resection.1,2 Hepatic resection remains the only potentially curative option for such patients, with reported 5-year survival rates of around 23–58%.3–9 However, only 10–30% of patients with CLM have surgically resectable hepatic disease.5 Thus, surgical resectability, defined as the removal of all macroscopic disease with clear margins and preservation of sufficient functioning liver, must be determined preoperatively. In order to avoid unnecessary surgery, the surgeon must have accurate information on the metastases, including the number of lesions, their distribution within the parenchyma and their encroachment on key vessels or bile ducts, the projected size of the liver remnant, and any vascular anomalies.
Contrast-enhanced helical computed tomography (HCT) remains the most commonly used imaging modality for the preoperative evaluation of CLM. Unlike magnetic resonance imaging (MRI), HCT is widely available at most institutions and allows for the evaluation of both intra- and extrahepatic disease. The development of multi-detector HCT scanners has further improved the sensitivity of HCT for the characterization of lesions, as well as evaluation of the hepatic vasculature.10,11 However, the true sensitivity of HCT for the detection of CLM remains unclear secondary to ongoing variability in the literature.12 In addition, with respect to the utility of HCT for predicting resectability, it should be noted that a significant proportion of patients are deemed to have unresectable disease at the time of surgery despite preoperative HCT indicating otherwise (Table 1). As such, these patients are subjected to the risk for morbidity and mortality associated with unnecessary exploratory surgery.
Table 1.
Sensitivity of helical computed tomography (HCT) and resectability rates in previously published studies
| Study | Year | Patients, n | HCT sensitivity | Resectable | Time interval |
|---|---|---|---|---|---|
| Schmidt et al.15 | 2000 | 33 | 94% | 79% | NR |
| Wallace et al.16 | 2001 | 179 | – | 81% | NR |
| Valls et al.17 | 2001 | 157 | 85% | 94% | NR |
| Zacherl et al.18 | 2002 | 61 | 83% | 87% | <28 days |
| Bhattacharjya et al.19 | 2004 | 120 | 73% | 83% | 30 days |
| Soyer et al.20 | 2004 | 60 | 92% | 100% | 6 days |
NR, not reported; Time interval, average time between HCT scan and hepatic resection
It is proposed that the discordance within the current literature regarding HCT sensitivity and the limited utility of HCT for predicting resectability are in part the results of a prolonged time interval between preoperative imaging and surgery. This study represents the first multi-institution review to examine the relationship between the time interval between imaging and surgery and the utility of such imaging.
Materials and methods
All patients who underwent surgery for proposed hepatic resection of hepatic metastases from CRC at the University Health Network, Mount Sinai Hospital and Sunnybrook Health Sciences Centre (Toronto, ON, Canada) from January 1999 to September 2004 were included in the study. All three institutions employ dedicated multidisciplinary hepatobiliary teams with appropriate radiological expertise. Patients who received preoperative chemotherapy or portal vein embolization within 6 months prior to surgery, underwent planned staged resections or did not undergo preoperative HCT were not included in the study. Demographic information, primary colorectal tumour location and stage, number and dates of HCTs, HCT results, date of hepatic resection, intraoperative findings and postoperative pathology were extracted by reviewing the charts of all patients.
Definitions
Time interval to operation is defined as the number of days between a preoperative HCT study and the date of hepatic resection. Sensitivity was defined as the ratio of the number of metastatic lesions reported on preoperative HCT to the actual number of lesions. Information from intraoperative findings and postoperative pathology was used to determine the actual number of lesions present. Intraoperative ultrasonography was employed to map the true location and number of lesions. Resected specimens were examined by experienced pathologists at each institution using standardized techniques involving slices of 3–5 mm in thickness. Among other features, pathology reports contained information on tumour numbers and sizes. Resectability is defined as the ability to resect all lesions in a single operation. For the purposes of this study, any patient who underwent staged operations to remove all lesions was defined as having unresectable disease.
Preoperative HCT reports were thoroughly reviewed to determine resectability defined by generally accepted criteria.13 Evidence in HCT of invasion of major vasculature (including the hepatic artery proper, bilateral hepatic arteries, main portal vein, bilateral portal veins, inferior vena cava, and any hepatic veins within 1 cm of the inferior vena cava), bilateral bile ducts, the common bile duct or adjacent organs (excluding the gallbladder and diaphragm) was considered to indicate unresectable disease. Imaging evidence of extensive liver metastasis with a predicted hepatic remnant of less than two segments or 25% after resection was likewise considered to indicate unresectable disease. True resectability was determined from the operative report. Accuracy of prediction was defined as agreement between the HCT prediction and operative findings.
The correlation between the time interval to operation and the sensitivity of preoperative HCT was determined using univariate and multivariate linear regression analysis. The normality assumption was tested via a normal probability plot on the residuals. The correlation between the time interval to operation and the accuracy of prediction was calculated using univariate and multivariate logistic regression. In this case, accuracy of prediction was defined as a binary outcome in which HCT predictions either agreed or disagreed with operative or histopathological findings. As a variable in multivariate regression, the Fong score was used to account for oncologic factors leading to CLM unresectability. This score was developed by Fong et al. to predict longterm survival in patients receiving hepatic resection for CLM.3 Subsequently, the score has been shown to predict resectability of CLM.14 Data analysis was performed using sas 9.1.3 (SAS Institute, Inc., Cary, NC, USA). An alpha level of 0.05 was defined as statistically significant.
Results
Patient and HCT characteristics
A total of 217 patients who underwent hepatic resection at the University Health Network, Mount Sinai Hospital or Sunnybrook Health Sciences Centre between January 1999 and September 2004 were included in this study (Table 2). Of these, 174 (80.2%) patients underwent resection. Of the 43 patients with lesions that were defined as unresectable, 38 required further operations to achieve full resection and five had unresectable disease.
Table 2.
Characteristics of patients and helical computed tomography (HCT) scans included in the current study
| Patient or HCT characteristic | Value |
|---|---|
| Number of patients | 217 |
| Number of HCT studies | 293 |
| Number of operations | 223 |
| Mean age, years (range) | 65 (37–83) |
| Gender | |
| Male, n | 138 |
| Female, n | 79 |
| Site of primary tumour | |
| Colon, n | 135 |
| Rectum, n | 75 |
| Recto-sigmoid, n | 7 |
| CLMs per HCT, n | 1 (0–8) |
| Size of largest CLM, cm (range) | 2.9 (0.8–15.7) |
| Time interval to resection, days (range) | 48 (1–179) |
| Sensitivity of CLM detection, % | 83.2 ± 30.0 |
| HCT prediction of resectability | |
| Resectable | 279 |
| Unresectable | 14 |
CLM, colorectal liver metastasis
A total of 293 HCT studies were performed preoperatively in the study group (Table 2). Of these, 115 (39.2%) studies reported the presence of multiple CLMs. The median diameter of the largest lesion was 2.9 cm (range 0.8–15.7 cm). The average sensitivity of HCT for detecting CLMs was 83.2%. Fourteen imaging studies concluded that lesions were unresectable according to previously defined criteria. The median time between HCT imaging and hepatic resection was 48 days (range 1–179 days).
Interval to surgery and HCT sensitivity
On univariate linear regression analysis, the time interval to surgery was inversely correlated with the sensitivity of HCT studies (P < 0.001) (Fig. 1). This model predicts that in order to achieve test sensitivity of over 90%, HCT must be performed no more than 26 days before surgery. Using age, gender, Fong score and time interval to surgery as variables, analysis by multivariate linear regression was performed. Time interval to surgery was the lone significant factor in predicting HCT sensitivity for correctly detecting CLMs (P < 0.001) (Table 3). The assumption of normality was supported by normal probability plots on residuals, which indicated straight lines.
Figure 1.
Relationship between sensitivity of helical computed tomography (HCT) and time interval to surgery in 293 HCT studies. The line represents the linear regression model. CLM, colorectal liver metastasis
Table 3.
Multivariate regression analyses of helical computed tomography sensitivity and prediction of resectability
| Variable | P-value | |
|---|---|---|
| Sensitivity | Prediction of resectability | |
| Age | 0.521 | 0.060 |
| Gender | 0.745 | 0.720 |
| Fong score | 0.129 | 0.116 |
| Time to surgery | <0.001 | <0.001 |
Interval to surgery and prediction of resectability
Of the 279 HCT studies that predicted resectable disease, 216 represented truly resectable disease and 63 represented either unresectable disease or disease that required multiple operations. Of 14 studies that predicted unresectable disease, 11 referred to disease that was truly unresectable or that required multiple operations and three referred to disease that was found to be resectable at the time of surgery. Overall, HCT provided an accurate prediction of resectability in 77.5% (227/293) of scans. Univariate logistic regression analysis suggested that time interval to surgery was inversely correlated with accuracy of prediction. Using this model, HCT must be performed no more than 43 days before surgery to achieve a prediction accuracy of 80%. Multivariate logistic regression with age, gender, Fong score and time interval to surgery showed that time interval to surgery was the lone significant factor in determining prediction accuracy (Table 3).
Discussion
Multiple non-invasive modalities have been applied in the preoperative evaluation of metastatic lesions in the liver. The choice of imaging modality used at each institution largely depends on local availability, technical expertise and cost. As the most commonly used modality, HCT has been evaluated by several previous studies for its sensitivity in detecting CLM, as well as its utility for predicting resectability (Table 1).15–20 All such studies have compared imaging results with intraoperative evaluation. Despite their similarities in experimental design, these studies have reported an unexpectedly wide range of values for both sensitivity and prediction accuracy. One explanation for these discrepancies concerns the varying lengths of time patients wait between preoperative imaging and hepatic resection. These time intervals were not reported in some studies and varied from 6 days to 2 months in others. The current study is the first multi-institution study to evaluate the sensitivity and prediction accuracy of HCT in relation to the length of time between imaging and surgery.
The current results for the overall sensitivity and prediction accuracy of HCT are comparable with values reported by previous studies.15–20 Time interval to surgery was found to be a negative predictor of both HCT sensitivity and prediction accuracy. These correlations remained significant in multivariate regression analyses, in which other variables, such as age, gender and Fong score, were not significant predictors of either outcome measures. In comparison with previous studies, the median time of 48 days between imaging and surgery in our cohort is relatively prolonged. According to the current analysis, this may result in lower sensitivity and prediction accuracy values. Soyer et al. compared HCT with intraoperative findings in 32 patients with CLM and found HCT to have a sensitivity of 92% and prediction accuracy of 100%.20 However, we must consider that the short average time interval between imaging and surgery of 6 days (range 3–17 days) may account for the extraordinarily positive results. By contrast, in studies by Zacherl et al. and Bhattacharjya et al., the mean time intervals to surgery were 28 and 30 days, respectively.18,19 These studies yielded slightly lower but comparable sensitivity and prediction accuracy results.
Although low sensitivity and accuracy may be inherent to HCT technology and radiology expertise, their relationship with the time interval between imaging and surgery shown in the current study suggests the physical growth of metastatic lesions during the time that elapses between HCT and surgery. Several studies have examined growth rates of untreated CLM. Havelaar et al. studied four patients with intra-abdominal CRC metastases with serial CT scans and determined that the mean tumour doubling time for CLM was 70 days.21 A more recent study in 74 patients with CLM suggested a mean tumour doubling time of 60.1 days.22 The latter study included only patients with surgically resectable disease and is thus comparable with data for our present cohort. In light of these tumour doubling times, it is conceivable that, in the present study, lesions that were previously resectable became unresectable over time. Similarly, lesions that were previously too small to be detected by HCT may have grown to detectable sizes, thus falsely lowering the reported sensitivity values.
This study has several important limitations. Firstly, resectable lesions were defined as those that were resectable in a single operation. Thus, patients who required further operations to resect all disease were considered unresectable in the current analysis. However, the definition used in this study was chosen to ensure a fair comparison because strict HCT resectability criteria that generally define resectability in a single operation were applied. Secondly, a proportion of patients in the cohort received multiple HCT scans prior to surgery. It was not possible to determine the reasons for repeated tests, but some patients may have represented for re-scanning as a result of the poor quality of the initial scan. Although such patients may have biased our sensitivity estimates, they should not affect the correlation between time interval and HCT sensitivity. Finally, all data regarding HCT scans were extracted from radiology reports without secondary review of HCT images. Although this may have slightly impacted the accuracy of HCT data, the imaging reports used had been produced at institutions with dedicated multidisciplinary teams with excellent radiological expertise.
The median time interval to surgery of 48 days observed in the current study is relatively long compared with those reported in other studies. In general, time between HCT and surgery is spent on extensive preoperative evaluation and the limiting of operative risks by treating co-morbid conditions. In jurisdictions where health care is publically financed, such extensive medical workups can often cause delays as a result of the volume of demand on the system. Although institutions in other jurisdictions may not experience significant waiting times, the correlation between time interval and HCT utility is nonetheless important.
The results of this study have some significant implications. Firstly, it has identified that the time interval between imaging and surgery is a significant factor that contributes to the discrepancy in HCT sensitivity values reported in the current literature. Any future studies of imaging sensitivity in CLM evaluation in which intraoperative findings are used as a reference should control for and report the time interval to surgery. Secondly, this study demonstrates that, regardless of the initial quality of imaging studies, their utility will decline unless surgery is performed in a timely manner. There is currently no standard that defines just how recent imaging should be prior to proceeding to hepatic resection surgery. It is proposed that HCT investigations should be performed no more than 30 days prior to the date of surgery in order to minimize any unnecessary surgery. According to the current models, a 30-day time interval can provide 89% sensitivity for lesion detection and 83% accuracy for resectability prediction.
Conclusions
The current study establishes that the utility of preoperative HCT in evaluating CLM declines over time until hepatic resection is performed. Any study comparing imaging modalities for the detection of liver metastases should be reported within the context of time interval to surgery. Ultimately, delays between imaging and surgery lead to increased patient morbidity caused by unnecessary surgery and greater costs to health care systems.
Conflicts of interest
None declared.
References
- 1.Bozzetti F, Cozzaglio L, Boracchi P, Marubini E, Doci R, Bignami P, et al. Comparing surgical resection of limited hepatic metastases from colorectal cancer to non-operative treatment. Eur J Surg Oncol. 1993;19:162–167. [PubMed] [Google Scholar]
- 2.Cummings LC, Payes JD, Cooper GS. Survival after hepatic resection in metastatic colorectal cancer: a population-based study. Cancer. 2007;109:718–726. doi: 10.1002/cncr.22448. [DOI] [PubMed] [Google Scholar]
- 3.Fong Y, Fortner J, Sun RL, Brennan MF, Blumgart LH. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases. Ann Surg. 1999;230:309–318. doi: 10.1097/00000658-199909000-00004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Pawlik TM, Schulick RD, Choti MA. Expanding criteria for resectability of colorectal liver metastases. Oncologist. 2008;13:51–64. doi: 10.1634/theoncologist.2007-0142. [DOI] [PubMed] [Google Scholar]
- 5.Scheele J, Altendorf-Hofmann A. Resection of colorectal liver metastases. Langenbecks Arch Surg. 1999;384:313–327. doi: 10.1007/s004230050209. [DOI] [PubMed] [Google Scholar]
- 6.Abdalla EK, Vauthey JN, Ellis LM, Ellis V, Pollock R, Broglio KR, et al. Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined resection/ablation for colorectal liver metastases. Ann Surg. 2004;239:818–825. doi: 10.1097/01.sla.0000128305.90650.71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wang X, Hershman DL, Abrams JA, Feingold D, Grann VR, Jacobson JS, et al. Predictors of survival after hepatic resection among patients with colorectal liver metastasis. Br J Cancer. 2007;97:1606–1612. doi: 10.1038/sj.bjc.6604093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sugawara Y, Yamamoto J, Yamasaki S, Shimada K, Kosuge T, Makuuchi M. Estimating the prognosis of hepatic resection in patients with metastatic liver tumours from colorectal cancer with special concern for the timing of hepatectomy. Surgery. 2001;129:408–413. doi: 10.1067/msy.2001.112001. [DOI] [PubMed] [Google Scholar]
- 9.Wei AC, Greig PD, Grant D, Taylor B, Langer B, Gallinger S. Survival after hepatic resection for colorectal metastases: a 10-year experience. Ann Surg Oncol. 2006;13:668–676. doi: 10.1245/ASO.2006.05.039. [DOI] [PubMed] [Google Scholar]
- 10.Kamel IR, Choti MA, Horton KM, Braga HJ, Birnbaum BA, Fishman EK, et al. Surgically staged focal liver lesions: accuracy and reproducibility of dual-phase helical CT for detection and characterization. Radiology. 2003;227:752–757. doi: 10.1148/radiol.2273011768. [DOI] [PubMed] [Google Scholar]
- 11.Sahani DV, Kalva SP. Imaging the liver. Oncologist. 2004;9:385–397. doi: 10.1634/theoncologist.9-4-385. [DOI] [PubMed] [Google Scholar]
- 12.Bipat S, van Leeuwen MS, Comans EF, Pijl ME, Bossuyt PM, Zwinderman AH, et al. Colorectal liver metastases: CT, MR imaging, and PET for diagnosis – meta-analysis. Radiology. 2005;237:123–131. doi: 10.1148/radiol.2371042060. [DOI] [PubMed] [Google Scholar]
- 13.Adam R, de Haas RJ, Wicherts DA, Aloia TA, Delvart V, Azoulay D, et al. Is hepatic resection justified after chemotherapy in patients with colorectal liver metastases and lymph node involvement? J Clin Oncol. 2008;26:3672–3680. doi: 10.1200/JCO.2007.15.7297. [DOI] [PubMed] [Google Scholar]
- 14.Li Destri G, Di Benedetto F, Torrisi B, Portale TR, Mosca F, Vecchio R, et al. Metachronous liver metastases and resectability: Fong's score and laparoscopic evaluation. HPB (Oxford) 2008;10:13–17. doi: 10.1080/13651820701851384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Schmidt J, Strotzer M, Fraunhofer S, Boedeker H, Zirngibl H. Intraoperative ultrasonography versus helical computed tomography and computed tomography with arterioportography in diagnosing colorectal liver metastases: lesion-by-lesion analysis. World J Surg. 2000;24:43–47. doi: 10.1007/s002689910009. [DOI] [PubMed] [Google Scholar]
- 16.Wallace JR, Christians KK, Quiroz FA, Foley WD, Pitt HA, Quebbeman EJ. Ablation of liver metastasis: is preoperative imaging sufficiently accurate? J Gastrointest Surg. 2001;5:98–107. doi: 10.1016/s1091-255x(01)80019-6. [DOI] [PubMed] [Google Scholar]
- 17.Valls C, Andia E, Sanchez A, Guma A, Figueras J, Torras J, et al. Hepatic metastases from colorectal cancer: preoperative detection and assessment of resectability with helical CT. Radiology. 2001;218:55–60. doi: 10.1148/radiology.218.1.r01dc1155. [DOI] [PubMed] [Google Scholar]
- 18.Zacherl J, Scheuba C, Imhof M, Zacherl M, Langle F, Pokieser P, et al. Current value of intraoperative sonography during surgery for hepatic neoplasms. World J Surg. 2002;26:550–554. doi: 10.1007/s00268-001-0266-2. [DOI] [PubMed] [Google Scholar]
- 19.Bhattacharjya S, Bhattacharjya T, Baber S, Tibballs JM, Watkinson AF, Davidson BR. Prospective study of contrast-enhanced computed tomography, computed tomography during arterioportography, and magnetic resonance imaging for staging colorectal liver metastases for liver resection. Br J Surg. 2004;91:1361–1369. doi: 10.1002/bjs.4699. [DOI] [PubMed] [Google Scholar]
- 20.Soyer P, Poccard M, Boudiaf M, Abitbol M, Hamzi L, Panis Y, et al. Detection of hypovascular hepatic metastases at triple-phase helical CT: sensitivity of phases and comparison with surgical and histopathologic findings. Radiology. 2004;231:413–420. doi: 10.1148/radiol.2312021639. [DOI] [PubMed] [Google Scholar]
- 21.Havelaar IJ, Sugarbaker PH, Vermess M, Miller DL. Rate of growth of intraabdominal metastases from colorectal cancer. Cancer. 1984;54:163–171. doi: 10.1002/1097-0142(19840701)54:1<163::aid-cncr2820540131>3.0.co;2-v. [DOI] [PubMed] [Google Scholar]
- 22.Tanaka K, Shimada H, Miura M, Fujii Y, Yamaguchi S, Endo I, et al. Metastatic tumour doubling time: most important prehepatectomy predictor of survival and non-recurrence of hepatic colorectal cancer metastasis. World J Surg. 2004;28:263–270. doi: 10.1007/s00268-003-7088-3. [DOI] [PubMed] [Google Scholar]