- Email: [email protected]
Early Detection of Local RFA Site Recurrence Using Total Liver Volume Perfusion CT
Early Detection of Local RFA Site Recurrence Using Total Liver Volume Perfusion CT: Initial Experience1 Martijn R. Meijerink, MD, Jan Hein T. M. van Waesberghe, MD, PhD, Lineke van der Weide, BSc, Petrousjka van den Tol, MD, PhD, Sybren Meijer, MD, PhD, Emile F. Comans, MD, PhD, Richard P. Golding, MD, Cornelis van Kuijk, MD, PhD
Rationale and Objectives. The aim of this study was to prospectively evaluate the feasibility of a novel total liver volume perfusion computed tomographic technique in demonstrating treatment-site recurrence of liver metastases after radiofrequency ablation (RFA). Materials and Methods. Eleven patients considered to be at increased risk for local RFA-site tumor recurrence underwent both positron emission tomography (PET) and perfusion computed tomography (CTP): a 12-phase scan of the entire liver acquired before and 11 times after contrast injection. After coregistration, blood flow maps were created using the maximum slope method. Results. In all cases, the CTP-derived blood flow maps fully paralleled the PET images in showing either the absence (nine of 13 lesions) or presence (four of 13 lesions) of local RFA-site recurrence. Marginal lesions with high hepatic arterial perfusion (>50 mL/min/100 g) and low portal venous perfusion (<10 mL/min/100 g) represented recurring vital tumor tissue (P < .05). Conclusion. Total liver volume CTP seems feasible for the detection and localization of treatment-site recurrence after RFA. Key Words. Liver metastases; radiofrequency ablation (RFA); tumor recurrence; perfusion computed tomography (CTP); positron emission tomography (PET). ª AUR, 2009
Currently, the most widely used tumor ablative technique for the treatment of colorectal liver metastases (CRLMs) is radiofrequency ablation (RFA). In patients with unresectable hepatic tumors, RFA has proved to be safe and feasible (1,2). Furthermore, there are indications that RFA can improve both short-term and long-term survival (3). Unfortunately, up to 40% of treated patients have recurring disease, and 12% are found to have recurrence at a treatment site only 1 year after RFA (4–6). If recurrence is limited to the treatment site, Acad Radiol 2009; 16:1215–1222 1
From the Department of Radiology (M.R.M., J.H.T.M.v.W., R.P.G., C.V.K.), the Master of Oncology Program (L.v.d.W.), the Department of Surgical Oncology (P.v.d.T., S.M.), and the Department of Nuclear Medicine & PET Research (E.F.C.), VU University Medical Center, De Boelelaan 1117, Postbus 7057 Amsterdam, The Netherlands. Received October 29, 2008; accepted March 30, 2009. Address correspondence to: M.R.M. e-mail: mr.meijerink@ vumc.nl
ª AUR, 2009 doi:10.1016/j.acra.2009.03.023
most often, a second RFA procedure or surgical resection can be performed. At present, positron emission tomography (PET) is the best follow-up imaging modality for the early detection of recurrence, although PET cannot be used as an image-guiding technique for local ablative therapies such as RFA (7). However, PET combined with computed tomography (CT) can provide exact localization of active tumor tissue (8–12), although this equipment is expensive and as yet not widely available. Using conventional CT, one study showed that of 38 ablated hepatocellular carcinomas that initially showed no enhancement, eight actually did recur. The authors concluded that short-term follow-up CT performed <3 months after treatment is not a reliable method to determine remission (13). Differentiation between vital tumor tissue and postRFA necrosis with perilesional inflammation can be difficult if not impossible using CT or ultrasound. We evaluated the feasibility of a technique using total volume perfusion CT (CTP), combining morphologic and dynamic enhancement
1215
MEIJERINK ET AL
information from a multiphase computed tomographic scan with hemodynamic blood flow maps, for the detection of local treatment-site recurrence after RFA. MATERIALS AND METHODS Patients Between November 2006 and August 2008, 11 consecutive patients with maximum 1-year histories of RFA for either CRLMs (10 patients) or adenocarcinoma of unknown primary (ACUP; 1 patient) were prospectively included in this study. The patients were selected on the basis of their PET referrals (routinely performed approximately 6 and 12 months after the RFA procedures). All patients were considered to be at increased risk for the presence of tumor recurrence on the basis of either a history of RFA for lesions >4 cm, the proximity of lesions to large vessels, and/or increases in the carcinoembryonic antigen tumor marker (Table 1). All patients underwent both 2- [18F]fluoro-2-deoxyglucose PET and dynamic total liver volume contrast-enhanced CTP within a period ranging from 3 to 12 months after their RFA procedures. If no recurrence was found on CT and PET, patients returned to regular follow-up plans that included additional PET and CT performed 6 months later. When the results of CT and PET were positive for tumor recurrence, patients underwent second intraoperative procedures for tumor resection and/or ablation. Patients were treated with systemic chemotherapy if extensive hepatic or extrahepatic disease was present. The results of 6-month follow-up PET and CT were used as the gold standard if no subsequent intraoperative procedures were undertaken or if no histology was obtained during this procedure. The study was approved by the institutional ethical and scientific review board, and all patients gave written informed consent. The procedures carried out were in accordance with the ethical standards of the World Medical Association (Declaration of Helsinki). Scanning Protocol and Image Postprocessing Dynamic perfusion computed tomographic measurements were obtained using a 64-slice multidetector computed tomographic system (Somatom Sensation; Siemens Medical Systems, Erlangen, Germany). Twelve-phase helical computed tomographic examinations of the total liver volume (collimation, 0.6 mm; rotation time, 0.33 seconds; pitch, 0.75) were performed before and 11 times after the rapid intravenous injection (6 mL/s) of 100 mL low-osmolar nonionic contrast agent with an iodine concentration of 300 mg/ mL (Ultravist-300 Iopromide; Schering AG, Berlin, Germany), followed by a 20-mL saline chasing bolus into the left antecubital vein, using an injection pump through an 18-gauge needle. Bolus tracking (threshold +100 Hounsfield units) was used to start the computed tomographic data
1216
Academic Radiology, Vol 16, No 10, October 2009
acquisition, placing a region of interest over the right atrium. With a minimal interscan delay of 3 seconds between two sets of series, the first four series were obtained in a single breath hold at maximum inspiration. After these first series, one ‘‘breathe out and breathe in’’ command (lasting 6 seconds) was given before adding two series in a single breath hold at maximum inspiration. This was repeated until a total of 11 series were acquired. Assuming a scan time of 3 seconds (depending on the scan range) and a contrast arrival time of 6 seconds (time to reach the threshold of +100 Hounsfield units in the right atrium), the postcontrast acquisition times were 9, 15, 21, 27, 36, 42, 51, 57, 66, 72, and 300 seconds after the start of injection. The 10th series (considered to represent the portal venous phase series) was obtained with conventional tube voltage and current (120 kV and a maximum of 180 mAs with dose modulation). In all other series, a fixed lower tube current was used to reduce radiation exposure (120 kV and 80 mAs). Phase coregistration was performed using a commercially available three-dimensional image fusion program (Vinci 2.36.0; Max-Planck Institute for Neurologic Research, Cologne, Germany). For the creation of blood flow maps, the software program Basama Perfusion 3.0.4.8 (Kanazawa, Ishikawa, Japan) was used (9). This program estimates tissue perfusion as the maximum slope of the tumor time-density curve divided by peak arterial enhancement. Because of the dual blood supply of the liver, hepatic tissue perfusion is divided into hepatic artery and portal vein perfusion as the maximum slope of the tumor time-density curve before versus after the splenic peak enhancement divided by, respectively, peak aortic and portal enhancement (10). The scanning protocol, the mathematic technique, and the effective radiation dose (approximately 24.0 mSv) have recently been described in more detail (11). Statistical Analysis To retrospectively analyze differences in perfusion values between vital and nonvital rim regions surrounding the postRFA lesions and to obtain P values, we used an independentsamples t test. RESULTS Results of PET In seven of 11 patients (nine of 13 post-RFA lesions), no pathologic intrahepatic or extrahepatic tracer uptake was found (Table 1). In three of these nine lesions, a minimal increase of tracer uptake was visible throughout the rim region surrounding the post-RFA lesion; this was characterized as an inflammatory response. In one patient, a solitary avid lesion was found not at the RFA site but elsewhere in the liver, which after resection was histologically confirmed to represent a CRLM. In four patients (four of 13 post-RFA
CTP PET
Subject 1 2 3 4 5
Time After RFA (mo)
Lesion Size (cm)
1 2 3 4 5 6 7 8 9 10 11 12 13
CRLM CRLM CRLM CRLM CRLM CRLM CRLM CRLM CRLM ACUP CRLM CRLM CRLM
6 12 12 13 6 6 6 8 7 3 9 6 6
4.0 4.0 4.2 4.9 3.5 5.5 4.0 4.8 5.4 8.0 5.0 3.5 5.7
Center
Rim
=/+ =/+ =/+ + + + +
Arterial
Portal Venous
Rim Arterial (mL/min/ 100 g) = = = = = =/+ =/+ =/+ =/+ + + + +
18 5 14 21 9 31 62 29 24 60 140 96 105
Portal Venous (mL/min/100 g) = = = = = = = = =
69 34 102 55 60 44 28 12 82 <5 <5 <5 <5
Histology No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy + + +
6-mo Follow-Up PET and CT
Final Diagnosis
+ Repeat RFA Repeat RFA Repeat RFA
No recurrence No recurrence No recurrence No recurrence No local recurrence* Marginal inflammation Marginal inflammation Marginal inflammation Reactive marginal hyperemia RFA-site recurrence RFA-site recurrence RFA-site recurrence RFA-site recurrence
ACUP, adenocarcinoma of unknown primary; CRLM, colorectal liver metastasis; CT, computed tomography; CTP, perfusion computed tomography; PET, positron emission tomography; RFA, radiofrequency ablation; , decreased or absent tracer uptake or tissue blood flow compared to normal liver parenchyma; = , comparable tracer uptake or tissue blood flow compared to normal liver parenchyma; = /+, vaguely increased tissue blood flow compared to normal liver parenchyma; +, increased tracer uptake or tissue blood flow in at least focal spot. * Intrahepatic recurrence but not at the RFA site.
1217
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
6 7 8 9 10 11
Lesion
Primary Tumor
Center
Academic Radiology, Vol 16, No 10, October 2009
Table 1 Results of PET and CTP for 13 Post-RFA Liver Lesions in 11 Patients for the Detection of Local RFA-Site Tumor Recurrence
MEIJERINK ET AL
lesions), tracer uptake was characterized as suspect for tumor recurrence at the margin of the RFA site (cases 1 to 4). Results of CTP On the basis of the conventional computed tomographic images in two patients, local RFA-site recurrence was suspected (subjects 8 and 11); in one patient, the reviewing radiologist was unsure (subject 10); and in a fourth patient (subject 9), no signs of local recurrence were reported. The CTP-derived blood flow maps fully paralleled the positron emission tomographic images in showing either the absence (nine of 13 lesions) or presence (four of 13 lesions) of local RFA-site recurrence (Table 1). No enhancement and therefore absent hepatic artery and portal venous blood flow was found in the center of all 13 post-RFA lesions, fully corresponding to the positron emission tomographic images, which showed no tracer uptake in these central areas. Compared to normal-appearing liver tissue, the margins of the nine PET-negative post-RFA lesions showed either comparable hepatic artery and portal venous blood flow (five of nine lesions) or slight increases in hepatic artery perfusion together with decreases in portal venous perfusion (four of nine lesions). This vague increase in perilesional hepatic artery perfusion corresponded to a minimally increased perilesional uptake of tracer on PET in three of four cases. Sustained negative results on PET and the absence of growth on CT 6 months after initial PET and CTP were taken as confirmation of the absence of recurrence at the RFA site in all seven patients (nine lesions). The four patients with positive results on PET for local RFA-site recurrence are discussed in more detail below. In these patients, the arterial blood flow maps were clearly increased in focal marginal spots, corresponding to the results of PET. Case 1 A 64-year-old man (subject 8) had undergone RFA for a very large (8 cm), unresectable, and rapidly growing ACUP in the right liver lobe, originating from segment 8. Prior to surgical exploration and RFA, the lesion had progressed during both first-line and second-line chemotherapy. At surgery, the tumor was considered unresectable because of its extent into segment 4 of the liver and what was judged to be an insufficient future liver remnant to perform extended right hemihepatectomy. Despite the size of the lesion, RFA was performed, mainly for downsizing purposes to enable a possible future right hemihepatectomy or repeat RFA. After 3 months, PET revealed local RFA-site recurrence at three marginal spots. The arterial blood flow maps fully paralleled the positron emission tomographic images, and the portal venous blood flow map inversely paralleled the positron emission tomographic images (Fig 1). Because of the severe extent of the multifocal RFA-site recurrence and the presence of new liver metastases elsewhere in the liver, no core biopsy
1218
Academic Radiology, Vol 16, No 10, October 2009
or secondary RFA or resection was performed. Follow-up CT 6 months later showed significant progression of all of these lesions, confirming the presence of tumor recurrence and progressive disease. Case 2 A 56-year-old man (subject 9, Fig 2) had undergone RFA for a large, solitary CRLM (5 cm) in segment 5 of the liver. Nine months after RFA, PET revealed a small caudal marginal spot with pathologic tracer uptake considered to be suspect for tumor recurrence. The arterial blood flow maps showed both slightly increased perilesional hepatic artery blood flow, most likely representing reactive inflammation with hyperemia, and a focal spot with very high hepatic artery and absent portal venous blood flow that paralleled the PETpositive focus. At surgery, a core biopsy of the suspected marginal lesion proved the presence of RFA-site tumor recurrence. The lesion was retreated with RFA. One year after this additional procedure, the patient is alive without detectable disease. Case 3 A 74-year-old woman (subject 10, Fig 3) had undergone RFA 6 months earlier for a CRLM (3.5 cm) in segment 8. She was considered to be at high risk for local recurrence because of the proximity of the inferior caval vein. PET and blood flow maps clearly showed an avid lesion with high arterial and low portal venous flow at the lateral margin of the postRFA lesion. Intraoperative core biopsy of the lesion confirmed the presence of RFA-site tumor recurrence. The lesion was retreated with RFA. The results of follow-up PET and CT 6 months later showed no disease. Case 4 A 59-year-old man (subject 11, Fig 4) had undergone RFA 6 months earlier for a large CRLM (5.7 cm) in segment 8 of the liver. PET and blood flow maps clearly showed a small avid lesion with high arterial and low portal venous flow at the margin of the post-RFA lesion. Intraoperative core biopsy of the lesion confirmed the presence of RFA-site tumor recurrence. The lesion was recently retreated with RFA. Quantification The center of all post-RFA lesions showed no enhancement, indicating a complete absence of both hepatic artery and portal venous perfusion. The average hepatic artery perfusion of rim regions surrounding this center was 23.7 mL/min/100 g (range, 5–62 mL/min/100 g) in lesions in which no recurrence was present and 100.3 mL/min/100 g (range, 60–140 mL/min/100 g) in marginal focal spots representing RFA-site recurrence (Fig 5); this difference was statistically significant (P = .0001). The average portal venous perfusion of rim regions without recurrence (54.0 mL/ min/100 g; range, 12–102 mL/min/100 g) was also
Academic Radiology, Vol 16, No 10, October 2009
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
Figure 1. Multifocal radiofrequency ablation (RFA)–site recurrence 3 months after the ablation of a very large unresectable adenocarcinoma of unknown primary. The conventional four-phase images on computed tomography (CT) depict a large, nonenhancing, strongly hypodense necrotic area (a–d), representing the post-RFA lesion, with three adjacent areas showing vague hyperdense blushes on the arterial (art) phase image (b). All three areas (white boxes) are clearly visible on the blood flow (BF) maps as heterogeneous regions of high arterial perfusion (e) and low or absent portal venous (port ven) perfusion (f). Positron emission tomography (PET) suggests the presence of local RFA-site recurrence in all three areas (g).
significantly different from rim regions with local recurrence (1.5 mL/min/100 g; range, 0–4 mL/min/100 g) (P = .004).
DISCUSSION We found blood flow maps derived from dynamic CTP to correlate fully with 2- [18F]fluoro-2-deoxyglucose positron emission tomographic images of patients with and without local RFA-site tumor recurrence. Areas with significantly increased hepatic artery and decreased or absent portal venous perfusion, compared to normal-appearing liver tissue, would appear to represent recurring tumor tissue. The quantification of marginal spots surrounding postRFA lesions confirmed the presence of vital recurring tumor tissue in all cases using the following cutoff values: hepatic arterial perfusion > 50 mL/min/100 g and portal venous perfusion < 10 mL/min/100 g (P < .05). This is in agreement with other studies, in which higher hepatic artery and
lower portal venous perfusion was found within liver metastases compared to normal liver parenchyma (9–15). The slight increase in hepatic artery blood flow often present within rim regions of post-RFA lesions corresponded to the more well known perilesional areas of increased tracer uptake on PET most likely representing a temporary inflammatory response of surrounding normal liver parenchyma to RFA. Nevertheless, as is the case in PET, differentiation between perilesional post-RFA inflammation and tumor recurrence remains a matter of interpretation without quantification. Contrary to PET, in which tracer uptake is based on the metabolic activity of the tumor tissue, CTP is based on the differences in blood supply between the tumor tissue (mainly supplied by the hepatic artery with an early peak increase of enhancement) and normal liver tissue (mainly supplied by the portal vein with a later peak increase of enhancement). CTP offers both morphologic and hemodynamic information about post-RFA lesions.
1219
MEIJERINK ET AL
Academic Radiology, Vol 16, No 10, October 2009
Figure 2. Focal radiofrequency ablation (RFA)–site recurrence 9 months after the ablation of a large colorectal liver metastasis. Arterial (art) (a) and portal venous (port ven) (b) phase images on computed tomography (CT) with typical perilesional slightly increased hepatic artery blood flow (BF) (c), normal portal venous blood flow (d), and somewhat increased 2- [18F]fluoro-2-deoxyglucose uptake on positron emission tomography (PET) (e). This rim region probably represents reactive inflammatory tissue and should not be mistaken for tumor recurrence. In the same patient, a few centimeters caudally, a superficial focal spot that is hyperdense during the arterial phase (f) and isodense during the portal venous phase (g) is shown. This spot has a very high hepatic artery blood flow (h) and absent portal venous blood flow (i), which clearly parallels the PET-positive focus (j). A core biopsy proved the presence of tumor recurrence.
Figure 3. Focal radiofrequency ablation (RFA)–site recurrence 6 months after the ablation of a colorectal liver metastasis. The patient was considered to be at increased risk for local tumor recurrence because of the proximity of the inferior caval vein. A lesion with hyperdensity during the arterial (art) phase (b) and hypodensity during the unenhanced (a), portal venous (port ven) (c), and equilibrium (d) phases is visible on the lateral margin of a small avascular post-RFA lesion abutting several hemostasis clips. The blood flow (BF) maps (e,f) and positron emission tomography (PET) (g) clearly show an avid lesion with high arterial and low portal venous flow at the lateral margin of the post-RFA lesion. The lesion clearly shows contrast enhancement on the microbubble contrast-enhanced ultrasound (US) image (h). This is suggestive of tumor recurrence, which was confirmed by core biopsy.
1220
Academic Radiology, Vol 16, No 10, October 2009
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
Figure 4. Focal radiofrequency ablation (RFA)–site recurrence 6 months after the ablation of a large colorectal liver metastasis. Unenhanced, early and late arterial (art), portal venous (port ven), and equilibrium phase images (a–e) show a post-RFA lesion with heterogeneous enhancement at the dorsal margin. The blood flow (BF) maps (f,g) clearly display a marginal lesion with high arterial and low portal venous flow, which parallels a marginal enhancing lesion on contrast-enhanced ultrasound (US) (h) and a somewhat smaller avid lesion on positron emission tomography (PET), acquired 8 weeks earlier (i). Intraoperative core biopsy confirmed the presence of RFA-site tumor recurrence.
Figure 5. Data comparison graph showing box-and-whisker plots of hepatic artery and portal venous perfusion values of rim regions surrounding post–radiofrequency ablation (RFA) lesions. For both hepatic artery and portal venous perfusion values, the difference between marginal tumor recurrence and perilesional liver tissue without recurrence was statistically significant (P = .0001 and P = .004, respectively).
CTP was well tolerated by all patients and could, without much difficulty, be implemented in follow-up computed tomographic schemes after RFA. However, at present, image fusion and the creation of blood flow maps is very time consuming; the present semiautomatic method takes about 1.5 hours per scan, as has been previously described (11). One important drawback of this study was the limited number of patients included, especially the small number of patients with local RFA-site recurrence. The number of local recurrences has dropped significantly during the past year in our hospital, possibly because of the introduction of a novel bipolar radiofrequency system (InCircle; RFA Medical, Inc, Fremont, CA) for very large lesions in which we have not as yet seen any local recurrences for lesions up to 7 cm, as well as the introduction of transarterial chemoembolization with irinotecan 1 day after ablation to increase the ablation zone. Patients treated with these novel systems and schemes were excluded from the present study. The method of total liver volume CTP entails a high radiation dose, which is another important limitation, especially considering that all patients are treated with curative intent. In all cases, the presence or absence of tumor recurrence could be predicted on the hepatic artery blood flow maps alone; both nonrecurring postRFA lesions and recurring tumor tissue had absent or very low portal venous blood flow. By excluding portal venous blood flow information, we could significantly reduce scan time and therefore radiation dose. Nevertheless, at least one portal venous series will remain necessary for the detection of
1221
MEIJERINK ET AL
new lesions elsewhere in the liver. Because differences in respiratory volume or motion-induced artifacts may potentially lead to false interpretations of the blood flow maps, we acquired all series at maximum inspiration and strongly recommend using three-dimensional volumetric image fusion to minimize the degree of misregistration. In conclusion, these preliminary results from a small-scale study suggest that total liver volume CTP may be feasible for the detection and localization of recurrent tumor after RFA.
Academic Radiology, Vol 16, No 10, October 2009
7.
8.
9.
10. REFERENCES 1. Curley SA, Izzo F, Delrio P, et al. Radiofrequency ablation of unresectable primary and metastatic hepatic malignancies: results in 123 patients. Ann Surg 1999; 230:1–8. 2. de Baere T. Radiofrequency ablation of the liver. AJR Am J Roentgenol 2001; 177:1213–1215. 3. Leen E, Horgan PG. Radiofrequency ablation of colorectal liver metastases. Surg Oncol 2007; 16:47–51. 4. Solbiati L, Livraghi T, Goldberg SN, et al. Percutaneous radio-frequency ablation of hepatic metastases from colorectal cancer: long-term results in 117 patients. Radiology 2001; 221:159–166. 5. Scaife CL, Curley SA. Complication, local recurrence, and survival rates after radiofrequency ablation for hepatic malignancies. Surg Oncol Clin N Am 2003; 12:243–255. 6. Abdalla EK, Vauthey JN, Ellis LM, et al. Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined
1222
11.
12.
13.
14.
15.
resection/ablation for colorectal liver metastases. Ann Surg 2004; 239: 818–825. Donckier V, Van Laethem JL, Goldman S, et al. [F-18] fluorodeoxyglucose positron emission tomography as a tool for early recognition of incomplete tumor destruction after radiofrequency ablation for liver metastases. J Surg Oncol 2003; 84:215–223. Veit P, Antoch G, Stergar H, Bockisch A, Forsting M, Kuehl H. Detection of residual tumor after radiofrequency ablation of liver metastasis with dual-modality PET/CT: initial results. Eur Radiol 2006; 16:80–87. Tsushima Y, Funabasama S, Aoki J, Sanada S, Endo K. Quantitative perfusion map of malignant liver tumors, created from dynamic computed tomography data. Acad Radiol 2004; 11:215–223. Miles KA, Hayball M, Dixon AK. Colour perfusion imaging: a new application of computed tomography. Lancet 1991; 337:643–645. Meijerink MR, van Waesberghe JH, van der Weide L, van den Tol P, Meijer S, van Kuijk C. Total-liver-volume perfusion CT using 3-D image fusion to improve detection and characterization of liver metastases. Eur Radiol 2008; 18:2345–2354. Tsushima Y, Funabasama S, Sanada S, Aoki J, Endo K. Development of perfusion CT software for personal computers. Acad Radiol 2002; 9: 922–926. Leggett DA, Kelley BB, Bunce IH, Miles KA. Colorectal cancer: diagnostic potential of CT measurements of hepatic perfusion and implications for contrast enhancement protocols. Radiology 1997; 205:716–720. Miles KA, Leggett DA, Kelley BB, Hayball MP, Sinnatamby R, Bunce I. In vivo assessment of neovascularization of liver metastases using perfusion CT. Br J Radiol 1998; 71:276–281. Meijerink MR, van Cruijsen H, Hoekman K, et al. The use of perfusion CT for the evaluation of therapy combining AZD2171 with gefitinib in cancer patients. Eur Radiol 2007; 17:1700–1713.
Rationale and Objectives. The aim of this study was to prospectively evaluate the feasibility of a novel total liver volume perfusion computed tomographic technique in demonstrating treatment-site recurrence of liver metastases after radiofrequency ablation (RFA). Materials and Methods. Eleven patients considered to be at increased risk for local RFA-site tumor recurrence underwent both positron emission tomography (PET) and perfusion computed tomography (CTP): a 12-phase scan of the entire liver acquired before and 11 times after contrast injection. After coregistration, blood flow maps were created using the maximum slope method. Results. In all cases, the CTP-derived blood flow maps fully paralleled the PET images in showing either the absence (nine of 13 lesions) or presence (four of 13 lesions) of local RFA-site recurrence. Marginal lesions with high hepatic arterial perfusion (>50 mL/min/100 g) and low portal venous perfusion (<10 mL/min/100 g) represented recurring vital tumor tissue (P < .05). Conclusion. Total liver volume CTP seems feasible for the detection and localization of treatment-site recurrence after RFA. Key Words. Liver metastases; radiofrequency ablation (RFA); tumor recurrence; perfusion computed tomography (CTP); positron emission tomography (PET). ª AUR, 2009
Currently, the most widely used tumor ablative technique for the treatment of colorectal liver metastases (CRLMs) is radiofrequency ablation (RFA). In patients with unresectable hepatic tumors, RFA has proved to be safe and feasible (1,2). Furthermore, there are indications that RFA can improve both short-term and long-term survival (3). Unfortunately, up to 40% of treated patients have recurring disease, and 12% are found to have recurrence at a treatment site only 1 year after RFA (4–6). If recurrence is limited to the treatment site, Acad Radiol 2009; 16:1215–1222 1
From the Department of Radiology (M.R.M., J.H.T.M.v.W., R.P.G., C.V.K.), the Master of Oncology Program (L.v.d.W.), the Department of Surgical Oncology (P.v.d.T., S.M.), and the Department of Nuclear Medicine & PET Research (E.F.C.), VU University Medical Center, De Boelelaan 1117, Postbus 7057 Amsterdam, The Netherlands. Received October 29, 2008; accepted March 30, 2009. Address correspondence to: M.R.M. e-mail: mr.meijerink@ vumc.nl
ª AUR, 2009 doi:10.1016/j.acra.2009.03.023
most often, a second RFA procedure or surgical resection can be performed. At present, positron emission tomography (PET) is the best follow-up imaging modality for the early detection of recurrence, although PET cannot be used as an image-guiding technique for local ablative therapies such as RFA (7). However, PET combined with computed tomography (CT) can provide exact localization of active tumor tissue (8–12), although this equipment is expensive and as yet not widely available. Using conventional CT, one study showed that of 38 ablated hepatocellular carcinomas that initially showed no enhancement, eight actually did recur. The authors concluded that short-term follow-up CT performed <3 months after treatment is not a reliable method to determine remission (13). Differentiation between vital tumor tissue and postRFA necrosis with perilesional inflammation can be difficult if not impossible using CT or ultrasound. We evaluated the feasibility of a technique using total volume perfusion CT (CTP), combining morphologic and dynamic enhancement
1215
MEIJERINK ET AL
information from a multiphase computed tomographic scan with hemodynamic blood flow maps, for the detection of local treatment-site recurrence after RFA. MATERIALS AND METHODS Patients Between November 2006 and August 2008, 11 consecutive patients with maximum 1-year histories of RFA for either CRLMs (10 patients) or adenocarcinoma of unknown primary (ACUP; 1 patient) were prospectively included in this study. The patients were selected on the basis of their PET referrals (routinely performed approximately 6 and 12 months after the RFA procedures). All patients were considered to be at increased risk for the presence of tumor recurrence on the basis of either a history of RFA for lesions >4 cm, the proximity of lesions to large vessels, and/or increases in the carcinoembryonic antigen tumor marker (Table 1). All patients underwent both 2- [18F]fluoro-2-deoxyglucose PET and dynamic total liver volume contrast-enhanced CTP within a period ranging from 3 to 12 months after their RFA procedures. If no recurrence was found on CT and PET, patients returned to regular follow-up plans that included additional PET and CT performed 6 months later. When the results of CT and PET were positive for tumor recurrence, patients underwent second intraoperative procedures for tumor resection and/or ablation. Patients were treated with systemic chemotherapy if extensive hepatic or extrahepatic disease was present. The results of 6-month follow-up PET and CT were used as the gold standard if no subsequent intraoperative procedures were undertaken or if no histology was obtained during this procedure. The study was approved by the institutional ethical and scientific review board, and all patients gave written informed consent. The procedures carried out were in accordance with the ethical standards of the World Medical Association (Declaration of Helsinki). Scanning Protocol and Image Postprocessing Dynamic perfusion computed tomographic measurements were obtained using a 64-slice multidetector computed tomographic system (Somatom Sensation; Siemens Medical Systems, Erlangen, Germany). Twelve-phase helical computed tomographic examinations of the total liver volume (collimation, 0.6 mm; rotation time, 0.33 seconds; pitch, 0.75) were performed before and 11 times after the rapid intravenous injection (6 mL/s) of 100 mL low-osmolar nonionic contrast agent with an iodine concentration of 300 mg/ mL (Ultravist-300 Iopromide; Schering AG, Berlin, Germany), followed by a 20-mL saline chasing bolus into the left antecubital vein, using an injection pump through an 18-gauge needle. Bolus tracking (threshold +100 Hounsfield units) was used to start the computed tomographic data
1216
Academic Radiology, Vol 16, No 10, October 2009
acquisition, placing a region of interest over the right atrium. With a minimal interscan delay of 3 seconds between two sets of series, the first four series were obtained in a single breath hold at maximum inspiration. After these first series, one ‘‘breathe out and breathe in’’ command (lasting 6 seconds) was given before adding two series in a single breath hold at maximum inspiration. This was repeated until a total of 11 series were acquired. Assuming a scan time of 3 seconds (depending on the scan range) and a contrast arrival time of 6 seconds (time to reach the threshold of +100 Hounsfield units in the right atrium), the postcontrast acquisition times were 9, 15, 21, 27, 36, 42, 51, 57, 66, 72, and 300 seconds after the start of injection. The 10th series (considered to represent the portal venous phase series) was obtained with conventional tube voltage and current (120 kV and a maximum of 180 mAs with dose modulation). In all other series, a fixed lower tube current was used to reduce radiation exposure (120 kV and 80 mAs). Phase coregistration was performed using a commercially available three-dimensional image fusion program (Vinci 2.36.0; Max-Planck Institute for Neurologic Research, Cologne, Germany). For the creation of blood flow maps, the software program Basama Perfusion 3.0.4.8 (Kanazawa, Ishikawa, Japan) was used (9). This program estimates tissue perfusion as the maximum slope of the tumor time-density curve divided by peak arterial enhancement. Because of the dual blood supply of the liver, hepatic tissue perfusion is divided into hepatic artery and portal vein perfusion as the maximum slope of the tumor time-density curve before versus after the splenic peak enhancement divided by, respectively, peak aortic and portal enhancement (10). The scanning protocol, the mathematic technique, and the effective radiation dose (approximately 24.0 mSv) have recently been described in more detail (11). Statistical Analysis To retrospectively analyze differences in perfusion values between vital and nonvital rim regions surrounding the postRFA lesions and to obtain P values, we used an independentsamples t test. RESULTS Results of PET In seven of 11 patients (nine of 13 post-RFA lesions), no pathologic intrahepatic or extrahepatic tracer uptake was found (Table 1). In three of these nine lesions, a minimal increase of tracer uptake was visible throughout the rim region surrounding the post-RFA lesion; this was characterized as an inflammatory response. In one patient, a solitary avid lesion was found not at the RFA site but elsewhere in the liver, which after resection was histologically confirmed to represent a CRLM. In four patients (four of 13 post-RFA
CTP PET
Subject 1 2 3 4 5
Time After RFA (mo)
Lesion Size (cm)
1 2 3 4 5 6 7 8 9 10 11 12 13
CRLM CRLM CRLM CRLM CRLM CRLM CRLM CRLM CRLM ACUP CRLM CRLM CRLM
6 12 12 13 6 6 6 8 7 3 9 6 6
4.0 4.0 4.2 4.9 3.5 5.5 4.0 4.8 5.4 8.0 5.0 3.5 5.7
Center
Rim
=/+ =/+ =/+ + + + +
Arterial
Portal Venous
Rim Arterial (mL/min/ 100 g) = = = = = =/+ =/+ =/+ =/+ + + + +
18 5 14 21 9 31 62 29 24 60 140 96 105
Portal Venous (mL/min/100 g) = = = = = = = = =
69 34 102 55 60 44 28 12 82 <5 <5 <5 <5
Histology No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy No biopsy + + +
6-mo Follow-Up PET and CT
Final Diagnosis
+ Repeat RFA Repeat RFA Repeat RFA
No recurrence No recurrence No recurrence No recurrence No local recurrence* Marginal inflammation Marginal inflammation Marginal inflammation Reactive marginal hyperemia RFA-site recurrence RFA-site recurrence RFA-site recurrence RFA-site recurrence
ACUP, adenocarcinoma of unknown primary; CRLM, colorectal liver metastasis; CT, computed tomography; CTP, perfusion computed tomography; PET, positron emission tomography; RFA, radiofrequency ablation; , decreased or absent tracer uptake or tissue blood flow compared to normal liver parenchyma; = , comparable tracer uptake or tissue blood flow compared to normal liver parenchyma; = /+, vaguely increased tissue blood flow compared to normal liver parenchyma; +, increased tracer uptake or tissue blood flow in at least focal spot. * Intrahepatic recurrence but not at the RFA site.
1217
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
6 7 8 9 10 11
Lesion
Primary Tumor
Center
Academic Radiology, Vol 16, No 10, October 2009
Table 1 Results of PET and CTP for 13 Post-RFA Liver Lesions in 11 Patients for the Detection of Local RFA-Site Tumor Recurrence
MEIJERINK ET AL
lesions), tracer uptake was characterized as suspect for tumor recurrence at the margin of the RFA site (cases 1 to 4). Results of CTP On the basis of the conventional computed tomographic images in two patients, local RFA-site recurrence was suspected (subjects 8 and 11); in one patient, the reviewing radiologist was unsure (subject 10); and in a fourth patient (subject 9), no signs of local recurrence were reported. The CTP-derived blood flow maps fully paralleled the positron emission tomographic images in showing either the absence (nine of 13 lesions) or presence (four of 13 lesions) of local RFA-site recurrence (Table 1). No enhancement and therefore absent hepatic artery and portal venous blood flow was found in the center of all 13 post-RFA lesions, fully corresponding to the positron emission tomographic images, which showed no tracer uptake in these central areas. Compared to normal-appearing liver tissue, the margins of the nine PET-negative post-RFA lesions showed either comparable hepatic artery and portal venous blood flow (five of nine lesions) or slight increases in hepatic artery perfusion together with decreases in portal venous perfusion (four of nine lesions). This vague increase in perilesional hepatic artery perfusion corresponded to a minimally increased perilesional uptake of tracer on PET in three of four cases. Sustained negative results on PET and the absence of growth on CT 6 months after initial PET and CTP were taken as confirmation of the absence of recurrence at the RFA site in all seven patients (nine lesions). The four patients with positive results on PET for local RFA-site recurrence are discussed in more detail below. In these patients, the arterial blood flow maps were clearly increased in focal marginal spots, corresponding to the results of PET. Case 1 A 64-year-old man (subject 8) had undergone RFA for a very large (8 cm), unresectable, and rapidly growing ACUP in the right liver lobe, originating from segment 8. Prior to surgical exploration and RFA, the lesion had progressed during both first-line and second-line chemotherapy. At surgery, the tumor was considered unresectable because of its extent into segment 4 of the liver and what was judged to be an insufficient future liver remnant to perform extended right hemihepatectomy. Despite the size of the lesion, RFA was performed, mainly for downsizing purposes to enable a possible future right hemihepatectomy or repeat RFA. After 3 months, PET revealed local RFA-site recurrence at three marginal spots. The arterial blood flow maps fully paralleled the positron emission tomographic images, and the portal venous blood flow map inversely paralleled the positron emission tomographic images (Fig 1). Because of the severe extent of the multifocal RFA-site recurrence and the presence of new liver metastases elsewhere in the liver, no core biopsy
1218
Academic Radiology, Vol 16, No 10, October 2009
or secondary RFA or resection was performed. Follow-up CT 6 months later showed significant progression of all of these lesions, confirming the presence of tumor recurrence and progressive disease. Case 2 A 56-year-old man (subject 9, Fig 2) had undergone RFA for a large, solitary CRLM (5 cm) in segment 5 of the liver. Nine months after RFA, PET revealed a small caudal marginal spot with pathologic tracer uptake considered to be suspect for tumor recurrence. The arterial blood flow maps showed both slightly increased perilesional hepatic artery blood flow, most likely representing reactive inflammation with hyperemia, and a focal spot with very high hepatic artery and absent portal venous blood flow that paralleled the PETpositive focus. At surgery, a core biopsy of the suspected marginal lesion proved the presence of RFA-site tumor recurrence. The lesion was retreated with RFA. One year after this additional procedure, the patient is alive without detectable disease. Case 3 A 74-year-old woman (subject 10, Fig 3) had undergone RFA 6 months earlier for a CRLM (3.5 cm) in segment 8. She was considered to be at high risk for local recurrence because of the proximity of the inferior caval vein. PET and blood flow maps clearly showed an avid lesion with high arterial and low portal venous flow at the lateral margin of the postRFA lesion. Intraoperative core biopsy of the lesion confirmed the presence of RFA-site tumor recurrence. The lesion was retreated with RFA. The results of follow-up PET and CT 6 months later showed no disease. Case 4 A 59-year-old man (subject 11, Fig 4) had undergone RFA 6 months earlier for a large CRLM (5.7 cm) in segment 8 of the liver. PET and blood flow maps clearly showed a small avid lesion with high arterial and low portal venous flow at the margin of the post-RFA lesion. Intraoperative core biopsy of the lesion confirmed the presence of RFA-site tumor recurrence. The lesion was recently retreated with RFA. Quantification The center of all post-RFA lesions showed no enhancement, indicating a complete absence of both hepatic artery and portal venous perfusion. The average hepatic artery perfusion of rim regions surrounding this center was 23.7 mL/min/100 g (range, 5–62 mL/min/100 g) in lesions in which no recurrence was present and 100.3 mL/min/100 g (range, 60–140 mL/min/100 g) in marginal focal spots representing RFA-site recurrence (Fig 5); this difference was statistically significant (P = .0001). The average portal venous perfusion of rim regions without recurrence (54.0 mL/ min/100 g; range, 12–102 mL/min/100 g) was also
Academic Radiology, Vol 16, No 10, October 2009
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
Figure 1. Multifocal radiofrequency ablation (RFA)–site recurrence 3 months after the ablation of a very large unresectable adenocarcinoma of unknown primary. The conventional four-phase images on computed tomography (CT) depict a large, nonenhancing, strongly hypodense necrotic area (a–d), representing the post-RFA lesion, with three adjacent areas showing vague hyperdense blushes on the arterial (art) phase image (b). All three areas (white boxes) are clearly visible on the blood flow (BF) maps as heterogeneous regions of high arterial perfusion (e) and low or absent portal venous (port ven) perfusion (f). Positron emission tomography (PET) suggests the presence of local RFA-site recurrence in all three areas (g).
significantly different from rim regions with local recurrence (1.5 mL/min/100 g; range, 0–4 mL/min/100 g) (P = .004).
DISCUSSION We found blood flow maps derived from dynamic CTP to correlate fully with 2- [18F]fluoro-2-deoxyglucose positron emission tomographic images of patients with and without local RFA-site tumor recurrence. Areas with significantly increased hepatic artery and decreased or absent portal venous perfusion, compared to normal-appearing liver tissue, would appear to represent recurring tumor tissue. The quantification of marginal spots surrounding postRFA lesions confirmed the presence of vital recurring tumor tissue in all cases using the following cutoff values: hepatic arterial perfusion > 50 mL/min/100 g and portal venous perfusion < 10 mL/min/100 g (P < .05). This is in agreement with other studies, in which higher hepatic artery and
lower portal venous perfusion was found within liver metastases compared to normal liver parenchyma (9–15). The slight increase in hepatic artery blood flow often present within rim regions of post-RFA lesions corresponded to the more well known perilesional areas of increased tracer uptake on PET most likely representing a temporary inflammatory response of surrounding normal liver parenchyma to RFA. Nevertheless, as is the case in PET, differentiation between perilesional post-RFA inflammation and tumor recurrence remains a matter of interpretation without quantification. Contrary to PET, in which tracer uptake is based on the metabolic activity of the tumor tissue, CTP is based on the differences in blood supply between the tumor tissue (mainly supplied by the hepatic artery with an early peak increase of enhancement) and normal liver tissue (mainly supplied by the portal vein with a later peak increase of enhancement). CTP offers both morphologic and hemodynamic information about post-RFA lesions.
1219
MEIJERINK ET AL
Academic Radiology, Vol 16, No 10, October 2009
Figure 2. Focal radiofrequency ablation (RFA)–site recurrence 9 months after the ablation of a large colorectal liver metastasis. Arterial (art) (a) and portal venous (port ven) (b) phase images on computed tomography (CT) with typical perilesional slightly increased hepatic artery blood flow (BF) (c), normal portal venous blood flow (d), and somewhat increased 2- [18F]fluoro-2-deoxyglucose uptake on positron emission tomography (PET) (e). This rim region probably represents reactive inflammatory tissue and should not be mistaken for tumor recurrence. In the same patient, a few centimeters caudally, a superficial focal spot that is hyperdense during the arterial phase (f) and isodense during the portal venous phase (g) is shown. This spot has a very high hepatic artery blood flow (h) and absent portal venous blood flow (i), which clearly parallels the PET-positive focus (j). A core biopsy proved the presence of tumor recurrence.
Figure 3. Focal radiofrequency ablation (RFA)–site recurrence 6 months after the ablation of a colorectal liver metastasis. The patient was considered to be at increased risk for local tumor recurrence because of the proximity of the inferior caval vein. A lesion with hyperdensity during the arterial (art) phase (b) and hypodensity during the unenhanced (a), portal venous (port ven) (c), and equilibrium (d) phases is visible on the lateral margin of a small avascular post-RFA lesion abutting several hemostasis clips. The blood flow (BF) maps (e,f) and positron emission tomography (PET) (g) clearly show an avid lesion with high arterial and low portal venous flow at the lateral margin of the post-RFA lesion. The lesion clearly shows contrast enhancement on the microbubble contrast-enhanced ultrasound (US) image (h). This is suggestive of tumor recurrence, which was confirmed by core biopsy.
1220
Academic Radiology, Vol 16, No 10, October 2009
PERFUSION CT TO DETECT RFA-SITE RECURRENCE
Figure 4. Focal radiofrequency ablation (RFA)–site recurrence 6 months after the ablation of a large colorectal liver metastasis. Unenhanced, early and late arterial (art), portal venous (port ven), and equilibrium phase images (a–e) show a post-RFA lesion with heterogeneous enhancement at the dorsal margin. The blood flow (BF) maps (f,g) clearly display a marginal lesion with high arterial and low portal venous flow, which parallels a marginal enhancing lesion on contrast-enhanced ultrasound (US) (h) and a somewhat smaller avid lesion on positron emission tomography (PET), acquired 8 weeks earlier (i). Intraoperative core biopsy confirmed the presence of RFA-site tumor recurrence.
Figure 5. Data comparison graph showing box-and-whisker plots of hepatic artery and portal venous perfusion values of rim regions surrounding post–radiofrequency ablation (RFA) lesions. For both hepatic artery and portal venous perfusion values, the difference between marginal tumor recurrence and perilesional liver tissue without recurrence was statistically significant (P = .0001 and P = .004, respectively).
CTP was well tolerated by all patients and could, without much difficulty, be implemented in follow-up computed tomographic schemes after RFA. However, at present, image fusion and the creation of blood flow maps is very time consuming; the present semiautomatic method takes about 1.5 hours per scan, as has been previously described (11). One important drawback of this study was the limited number of patients included, especially the small number of patients with local RFA-site recurrence. The number of local recurrences has dropped significantly during the past year in our hospital, possibly because of the introduction of a novel bipolar radiofrequency system (InCircle; RFA Medical, Inc, Fremont, CA) for very large lesions in which we have not as yet seen any local recurrences for lesions up to 7 cm, as well as the introduction of transarterial chemoembolization with irinotecan 1 day after ablation to increase the ablation zone. Patients treated with these novel systems and schemes were excluded from the present study. The method of total liver volume CTP entails a high radiation dose, which is another important limitation, especially considering that all patients are treated with curative intent. In all cases, the presence or absence of tumor recurrence could be predicted on the hepatic artery blood flow maps alone; both nonrecurring postRFA lesions and recurring tumor tissue had absent or very low portal venous blood flow. By excluding portal venous blood flow information, we could significantly reduce scan time and therefore radiation dose. Nevertheless, at least one portal venous series will remain necessary for the detection of
1221
MEIJERINK ET AL
new lesions elsewhere in the liver. Because differences in respiratory volume or motion-induced artifacts may potentially lead to false interpretations of the blood flow maps, we acquired all series at maximum inspiration and strongly recommend using three-dimensional volumetric image fusion to minimize the degree of misregistration. In conclusion, these preliminary results from a small-scale study suggest that total liver volume CTP may be feasible for the detection and localization of recurrent tumor after RFA.
Academic Radiology, Vol 16, No 10, October 2009
7.
8.
9.
10. REFERENCES 1. Curley SA, Izzo F, Delrio P, et al. Radiofrequency ablation of unresectable primary and metastatic hepatic malignancies: results in 123 patients. Ann Surg 1999; 230:1–8. 2. de Baere T. Radiofrequency ablation of the liver. AJR Am J Roentgenol 2001; 177:1213–1215. 3. Leen E, Horgan PG. Radiofrequency ablation of colorectal liver metastases. Surg Oncol 2007; 16:47–51. 4. Solbiati L, Livraghi T, Goldberg SN, et al. Percutaneous radio-frequency ablation of hepatic metastases from colorectal cancer: long-term results in 117 patients. Radiology 2001; 221:159–166. 5. Scaife CL, Curley SA. Complication, local recurrence, and survival rates after radiofrequency ablation for hepatic malignancies. Surg Oncol Clin N Am 2003; 12:243–255. 6. Abdalla EK, Vauthey JN, Ellis LM, et al. Recurrence and outcomes following hepatic resection, radiofrequency ablation, and combined
1222
11.
12.
13.
14.
15.
resection/ablation for colorectal liver metastases. Ann Surg 2004; 239: 818–825. Donckier V, Van Laethem JL, Goldman S, et al. [F-18] fluorodeoxyglucose positron emission tomography as a tool for early recognition of incomplete tumor destruction after radiofrequency ablation for liver metastases. J Surg Oncol 2003; 84:215–223. Veit P, Antoch G, Stergar H, Bockisch A, Forsting M, Kuehl H. Detection of residual tumor after radiofrequency ablation of liver metastasis with dual-modality PET/CT: initial results. Eur Radiol 2006; 16:80–87. Tsushima Y, Funabasama S, Aoki J, Sanada S, Endo K. Quantitative perfusion map of malignant liver tumors, created from dynamic computed tomography data. Acad Radiol 2004; 11:215–223. Miles KA, Hayball M, Dixon AK. Colour perfusion imaging: a new application of computed tomography. Lancet 1991; 337:643–645. Meijerink MR, van Waesberghe JH, van der Weide L, van den Tol P, Meijer S, van Kuijk C. Total-liver-volume perfusion CT using 3-D image fusion to improve detection and characterization of liver metastases. Eur Radiol 2008; 18:2345–2354. Tsushima Y, Funabasama S, Sanada S, Aoki J, Endo K. Development of perfusion CT software for personal computers. Acad Radiol 2002; 9: 922–926. Leggett DA, Kelley BB, Bunce IH, Miles KA. Colorectal cancer: diagnostic potential of CT measurements of hepatic perfusion and implications for contrast enhancement protocols. Radiology 1997; 205:716–720. Miles KA, Leggett DA, Kelley BB, Hayball MP, Sinnatamby R, Bunce I. In vivo assessment of neovascularization of liver metastases using perfusion CT. Br J Radiol 1998; 71:276–281. Meijerink MR, van Cruijsen H, Hoekman K, et al. The use of perfusion CT for the evaluation of therapy combining AZD2171 with gefitinib in cancer patients. Eur Radiol 2007; 17:1700–1713.