10.1245/ASO.2006.07.009
Annals of Surgical Oncology 13:542-546 (2006)
© 2006 Society of Surgical Oncology
Accuracy of Preoperative Imaging of Hepatic Tumors With Helical Computed Tomography
Courtney L. Scaife, MD1,
Chaan S. Ng, MD2,
Lee M. Ellis, MD1,
J. Nicolas Vauthey, MD1,
Chuslip Charnsangavej, MD2 and
Steven A. Curley, MD1
1 Department of Surgical Oncology, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 444, Houston, Texas 77030
2 Department of Radiology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 57, Houston, Texas 77030
Correspondence: Address correspondence and reprint requests to: Steven A. Curley, MD; E-mail: scurley{at}mdanderson.org.
 |
ABSTRACT
|
|---|
Background: The accuracy of preoperative computed tomography (CT) scans in the era of modern imaging techniques with helical, high-resolution CT has not been adequately assessed. We reviewed the data from our departmental prospective database with the hypothesis that intraoperative ultrasonography (IOUS) still detects more hepatic tumors than are evident on preoperative helical CT scans.
Methods: All patients who underwent surgical resection and/or radiofrequency ablation of primary or metastatic hepatic tumors between January 2001 and July 2002 were included in the review. All patients had preoperative helical CT imaging followed by hepatic IOUS. The number of malignant lesions and evidence of local disease identified by the preoperative CT scan versus IOUS and surgical exploration were compared.
Results: In this time period, 250 patients underwent surgical resection and/or radiofrequency ablation of hepatic tumors. In 67 (27%) of these patients, IOUS identified more hepatic tumors than were seen on preoperative helical CT scan. In eight patients (3%), CT underestimated local extension of the disease into the diaphragm. The incidence of inaccurate pre-operative prediction of the extent of disease increased significantly with a greater number of hepatic tumors.
Conclusions: IOUS identified additional hepatic tumors in 27% of patients who underwent hepatic resection after state-of-the-art preoperative CT imaging. This study provides evidence that IOUS remains an essential part of the complete assessment of hepatic malignancies in patients who receive surgical treatment.
Key Words: Liver imaging • Intraoperative ultrasonography • Preoperative CT scan • Liver resection
|
INTRODUCTION
|
|---|
Surgical resection remains the optimal treatment for primary and metastatic hepatic malignancies. This treatment approach is inherently dependent on accurate hepatic imaging for preoperative planning and intraoperative complete resection of the neoplastic disease. The primary imaging modalities for the liver remain preoperative computed tomography (CT) and intraoperative ultrasonography (IOUS).
Historically, preoperative CT scans have failed to identify 20% to 40% of hepatic malignant lesions demonstrated by hepatic IOUS.1–5 In 1994, Kane et al.4 compared CT hepatic imaging with intraoperative exploration and IOUS and found that the sensitivity of CT was 67%, compared with 78% sensitivity of exploration and 97% sensitivity of IOUS. Fortunato et al.5 also compared preoperative hepatic CT scanning with hepatic IOUS in the same era and found that the sensitivity of CT scanning was 67%, versus 75% sensitivity for exploration and 100% sensitivity for hepatic IOUS.
CT scanning techniques have changed significantly in the past decade to include high-speed helical scanners. In addition, scanning techniques have also improved to include multiphasic contrast studies that scan the liver parenchyma during a noncontrast, an arterial, and a portal venous flow phase. The accuracy of preoperative CT scans in this era of modern imaging techniques with helical, high-resolution CT has not been adequately assessed. The purpose of this study was therefore to review the data from our departmental prospective database with the hypothesis that IOUS still detects more hepatic tumors than are evident on state-of-the-art preoperative helical CT scans.
|
METHODS
|
|---|
We reviewed our institutional liver tumor database to identify all patients who had undergone surgical resection and/or intraoperative radiofrequency ablation of malignant primary or metastatic hepatic tumors. We queried the database for patients between January 2001 and July 2002, because this would include only patients whose preoperative imaging was completed with modern high-speed helical CT scanners within 3 months of the operative exploration. We excluded patients with neuroendocrine hepatic metastases, who typically have a miliary pattern of small hepatic metastases such that all imaging techniques are routinely inaccurate in the assessment of the number of tumors, and those with gallbladder tumors, which tend to involve the liver by direct tumor extension. All other primary and metastatic malignant histological diagnoses were included.
All patients underwent preoperative helical CT imaging based on one of two institutional protocols. The imaging protocol was arbitrarily assigned by the ordering primary physician. The two CT imaging protocols included a standard—uniphasic abdominal CT scan or a liver protocol—triphasic abdominal CT scan. All studies were obtained with 150 mL of nonionic intravenous contrast medium (320–350 mg I/ml) by using single-slice helical or multislice CT scanners (Hi-Speed or Light Speed GE Scanners; GE Medical Systems, Milwaukee, WI). Oral contrast medium was used to opacify the bowel before the study. When the uniphasic protocol was used, images of the liver were obtained 60 to 65 seconds after the commencement of pump injection of intravenous contrast medium at 3 mL/s (portal venous phase), and helical or multislice scanners were used (7.5-mm collimation, pitch of 3, and reconstruction interval of 7.5 mm). When the triphasic protocol was used, images of the liver were obtained at 30, 50, and 70 seconds after the commencement of pump injection of intravenous contrast medium at 5 mL/s (hepatic arterial, liver parenchymal, and portal venous phase, respectively), and multislice scanners were used (5-mm collimation, pitch of 6, and reconstruction interval of 2.5 mm). The radiologists’ reports were reviewed to record the pre-operatively identified number of hepatic lesions.
During surgery, all patients underwent hepatic IOUS by the operating surgeon. The intraoperative findings were scored on a data sheet at the time of operation. Data recorded included the number of the hepatic malignant lesions and evidence of local extension of disease.
We compared the reported preoperative findings with the intraoperative findings regarding the number and extension of the hepatic tumors. A 
2 test was performed to determine the statistical significance of variance in outcomes.
|
RESULTS
|
|---|
From January 2001 to July 2002, 271 patients underwent surgical resection and/or intraoperative radiofrequency ablation of hepatic tumors with a preoperative CT scan obtained within 3 months of operation. Twenty-one patients with a diagnosis of a neuroendocrine or gallbladder primary disease were excluded from this review (Fig. 1
). In 67 (27%) of the 250 patients reviewed, IOUS identified more hepatic tumors than were seen on preoperative helical CT scan. An example is demonstrated in Fig. 2.
View larger version (67K):
[in this window]
[in a new window]
|
FIG. 2. Lesion in segment 7 not seen on preoperative triphasic computed tomographic (CT) scan. (A) Preoperative CT scan. (B) Intraoperative ultrasonography: the solid arrow marks a segment 7 neoplasm.
|
|
The accuracy of the preoperative imaging decreased as the number of hepatic lesions increased. The overall sensitivity of CT imaging was 73%. The sensitivity of the preoperative CT scan decreased significantly as the number of lesions increased; sensitivity was 89% when one lesion was present and was only 52% when more than four lesions were present (Table 1). Thirty percent of patients who had six or more lesions were found to have missed lesions on preoperative scanning (Fig. 3).
View this table:
[in this window]
[in a new window]
|
TABLE 1. The sensitivity of preoperative CT scan compared with IOUS relative to the number of liver tumors identified
|
|
The sensitivity of the triphasic liver protocol was significantly higher (P = .02) than the sensitivity of the uniphasic protocol in predicting the number of lesions (Table 2), according to 238 assessable images. In this series, 12 patients had CT scans performed at an outside institution. However, all of these films were reviewed by our radiologists. These films were no longer available to assess which imaging contrast technique was used.
For eight patients (3%), the CT underestimated the local extension of the disease. Specifically, CT failed to detect diaphragmatic invasion in four patients, failed to identify the extent of segmental or caudate lobe involvement in three patients, and did not predict carcinomatosis in one patient.
The primary histological diagnosis of the neoplasm did influence the accuracy of preoperative imaging, as shown in Table 3. The sensitivity of CT imaging for hepatocellular carcinoma was less than the sensitivity for other histological categories, but this decrease was slight. We found no tumor or patient characteristic that was a predictor of a decrease in CT sensitivity.
|
DISCUSSION
|
|---|
The accuracy of preoperative CT scanning has improved with modern helical CT scanning techniques from the previously reported 67% sensitivity to our finding of a 73% sensitivity. However, the fact remains that CT scanning is not definitive in the assessment of hepatic neoplasms, because there continues to be a 27% rate of failure to detect all hepatic tumors.
Other authors have previously reported that CT arterial portography increases the sensitivity and specificity of preoperative hepatic imaging over standard CT imaging, with sensitivities of 65% to 93% and specificities of 75% to 90%.5–8 In these same studies, IOUS remained the superior imaging technique, with sensitivities of 98% to 100% and specificities of 95% to 100%. It requires no additional invasive procedure for IOUS, as opposed to CT arterial portography, which does require intraarterial catheterization and contrast injection. Other authors have reviewed the accuracy of magnetic resonance imaging for hepatic lesions, but, again, the sensitivity (66%–90%) and specificity (80%) remained inferior to those of IOUS.7,9,10 No patient in this series underwent CT arterial portography, and very few had preoperative magnetic resonance imaging scans, so we where not able to compare these imaging modalities with our findings.
Early studies have attempted to identify factors that influence the ability to detect lesions with preoperative imaging, including the size and segmental location of the intraparenchymal neoplasm. Our study is a retrospective review based on the information provided by the radiologist’s transcribed interpretation of the preoperative CT scan. This, unfortunately, did not allow us to compare the size or the location of the missed lesions, because this information was not consistently reported by the radiologist; rather, only the location of the dominant lesions and total number of lesions present were reported. Previous evidence indicates that lesions <1 cm are more commonly missed by preoperative CT scans and CT arterioportography.8,11 In addition, there is some evidence, reported by Foroutani et al.,11 that the segmental location did not significantly affect the accuracy of lesion identification; however, proximity to the falciform ligament did negatively affect the accuracy of CT imaging.
In conclusion, IOUS identified additional hepatic tumors in 27% of our patients undergoing hepatic resection and/or radiofrequency ablation after state-of-the-art preoperative CT imaging. Although these data support improved accuracy of preoperative imaging compared with historical findings, this study provides evidence that IOUS remains an essential part of the complete assessment of hepatic malignancies in patients undergoing surgical treatment for hepatic malignant disease.
Received for publication July 21, 2003.
Accepted for publication October 5, 2005.
|
REFERENCES
|
|---|
- Clarke MP, Steele G, Hamilton ES, Ravikumar TS, Onik G, Clouse ME. Prospective comparison of preoperative imaging in the detection of liver tumors. Surgery 1989; 106:849–55.[Medline]
- Soyer P, Levesque M, Elias D, Zeitoun G, Roche A. Preoperative assessment of respectability of hepatic metastases from colonic carcinoma: CT portography vs sonography and dynamic CT. AJR Am J Roentgenol 1992; 159:741–4.[Abstract/Free Full Text]
- Knol JA, Marn CS, Francis IR, Rubin JM, Bromberg J, Chang AE. Comparisons of dynamic infusion and delayed computed tomography, intraoperative ultrasound, and palpation in the diagnosis of liver metastases. Am J Surg 1993; 165:81–8.[CrossRef][Medline]
- Kane RA, Hughes LA, Cua EJ, Steele GD, Jenkins RL, Cady B. The impact of intraoperative ultrasonography on surgery for liver neoplasms. J Ultrasound Med 1994; 13:1–6.[Abstract]
- Fortunato L, Clair M, Hoffman J, et al. Is CT portography (CTAP) really useful in patients with liver tumors who undergo intraoperative ultrasonography (IOUS)? Am Surg 1995; 61:560–5.[Medline]
- Matsuo M, Kanematusu M, Inaba Y, et al. Preoperative detection of malignant hepatic tumors: value of combined helical CT during arterial portography and biphasic CT during hepatic arteriography. Clin Radiol 2001; 56:138–45.[Medline]
- Kehagias D, Metafa A, Hatziioannou A, et al. Comparison of CT, MRI, and CT during arterial portography in the detection of malignant hepatic lesions. Hepatogastroenterology 2000; 47:1399–403.[Medline]
- 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: lesions-by-lesion analysis. World J Surg 2000; 24:43–8.[CrossRef][Medline]
- Conlon R, Jacobs M, Dasgupta D, Lodge JPA. The value of intraoperative ultrasound during hepatic resection compared with improved preoperative magnetic resonance imaging. Eur J Ultrasound 2003; 16:211–6.[CrossRef][Medline]
- Hagspeil KD, Neidl KF, Eichenberger AC, Weder W, Marincek B. Detection of liver metastases: comparison of super-paramagnetic iron oxide-enhanced and unenhanced MR imaging at 1.5 T with dynamic CT, intraoperative US, and percutaneous US. Radiology 1995; 1969:471–8.
- Foroutani A, Garland A, Berber E, et al. Laparoscopic ultrasound vs triphasic computed tomography for detecting liver tumors. Arch Surg 2000; 135:933–8.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
Y.-s. Kim, H. Rhim, H. K. Lim, D. Choi, W. J. Lee, T. Y. Jeon, J. W. Joh, and S. J. Kim
Intraoperative Radiofrequency Ablation for Hepatocellular Carcinoma: Long-Term Results in a Large Series
Ann. Surg. Oncol.,
July 1, 2008;
15(7):
1862 - 1870.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
