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

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2473070414v1 247/3/726    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Goh, V. Articles by Bartram, C. I. Search for Related Content PubMed PubMed Citation Articles by Goh, V. Articles by Bartram, C. I.

Quantitative Assessment of Colorectal Cancer Tumor Vascular Parameters by Using Perfusion CT: Influence of Tumor Region of Interest¹

¹ From the Paul Strickland Scanner Centre, Mount Vernon Hospital, Northwood, Middlesex, England (V.G., A.G.); Department of Academic Radiology, University College Hospital, 235 Euston Rd, Level 2 Podium, London NW1 2BU, England (S.H.); Intestinal Imaging Centre, St Mark's Hospital, Harrow, England (A.G., C.I.B.); and Health Research and Development Support Unit, University of Hertfordshire, Hatfield, England (D.W., J.S.). From the 2006 RSNA Annual Meeting. Received March 2, 2007; revision requested May 9; revision received July 19; accepted August 17; final version accepted November 13. Supported in part by a pump priming grant from the Royal College of Radiologists, London, England. Address correspondence to S.H. (e-mail: s.halligan{at}ucl.ac.uk).

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Purpose: To prospectively determine whether position and size of tumor region of interest (ROI) influence estimates of colorectal cancer vascular parameters at computed tomography (CT).

Materials and Methods: After institutional review board approval and informed consent, 25 men and 22 women (mean age, 65.8 years) with colorectal adenocarcinoma underwent 65-second CT perfusion study. Blood volume, blood flow, and permeability–surface area product were determined for 40- or 120-mm² circular ROIs placed at the tumor edge and center and around (outlining) visible tumor. ROI analysis was repeated by two observers in different subsets of patients to assess intra- and interobserver variation. Measurements were compared by using analysis of variance; a difference with P = .002 was significant.

Results: Blood volume, blood flow, and permeability–surface area product measurements were substantially higher at the edge than at the center for both 40- and 120-mm² ROIs. For 40-mm² ROI, means of the three measurements were 6.9 mL/100 g (standard deviation [SD], 1.4), 108.7 mL/100 g per minute (SD, 39.2), and 16.9 mL/100 g per minute (SD, 4.2), respectively, at the edge versus 5.1 mL/100 g (SD, 1.5), 56.3 mL/100 g per minute (SD, 33.1), and 13.9 mL/100 g per minute (SD, 4.6), respectively, at the center. For 120-mm² ROI, means of the three measurements were 6.6 mL/100 g (SD, 1.3), 96.7 mL/100 g per minute (SD, 42.5), and 16.3 mL/100 g per minute (SD, 5.6), respectively, at the edge versus 5.1 mL/100 g (SD, 1.4), 58.3 mL/100 g per minute (SD, 32.5), and 13.4 mL/100 g per minute (SD, 4.3) at the center (P < .0001). Measurements varied substantially depending on the ROI size; values for the ROI for outlined tumor were intermediate between those at the tumor edge and center. Inter- and intraobserver agreement was poor for both 40- and 120-mm² ROIs.

Conclusion: Position and size of tumor ROI and observer variation substantially influence ultimate perfusion values. ROI for outlined entire tumor is more reliable for perfusion measurements and more appropriate clinically than use of arbitrarily determined smaller ROIs.

© RSNA, 2008

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Techniques that are used to image tumor vascular parameters, such as perfusion computed tomography (CT), have been promoted as useful tools for the assessment of antivascular chemotherapy, as these techniques combine robust quantitative in vivo assessment of tumor vascularity with excellent anatomic detail (1–5). The availability of commercial perfusion software has enabled perfusion CT to be incorporated easily into clinical practice. Vascular parameters obtained through imaging are regarded as an appropriate surrogate for angiogenesis because they have correlated positively with histologic measures, such as microvessel density, in a variety of tumors, including lung and renal cancer (6,7).

Although commercial software has enabled evaluation of tumor vascular parameters, such as blood flow, blood volume, and permeability–surface area product, following drug therapy, how the measurements should be made is a topic of debate. In particular, it is unclear whether analysis should include the entire tumor volume or a representative tumor section and whether the representative tumor section is sufficient to reflect therapeutic effect accurately. Integral to quantitative assessment is the definition of a tumor region of interest (ROI) from which vascular parameters are derived. It is well recognized that tumor perfusion is spatially heterogeneous, yet, to date, there has been no systematic evaluation as to what extent the size and position of the tumor ROI influence ultimate values. This is particularly pertinent to colorectal cancer, as the first antiangiogenic drug to be licensed for clinical use, bevacizumab (Avastin; Genentech, San Francisco, Calif), is now available. Many other compounds also are being developed and evaluated. Thus, the aim of our study was to prospectively determine whether the size and position of the tumor ROI influence estimates of colorectal cancer vascular parameters at CT.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
GE Healthcare Technologies, Waukesha, Wis, provided software for analysis. The authors retained full control of all data collected and information submitted for publication.

Patients
Institutional review board approval was obtained for this prospective study. Each patient received information that detailed the study, including information on radiation exposure, and written informed consent was obtained after the study was explained to each patient personally. Adult patients presenting for pretreatment staging of a biopsy-proved colorectal cancer were eligible for this study. Forty-seven adult patients (25 men, 22 women; mean age, 65.8 years; range, 28.1–88.7 years) were examined. Nineteen tumors were located in the rectum. The remaining tumors were in the colon: cecum (n = 9), ascending colon (n = 3), transverse colon (n = 1), descending colon (n = 1), and sigmoid colon (n = 14). The mean tumor length was 5.4 cm (range, 2–13.3 cm) and the mean tumor depth was 4.6 cm (range, 2–8 cm) at contrast material–enhanced CT. Patients were excluded if they had renal impairment, if they had a known allergy to intravenous contrast material, if the physician was unable to place an intravenous cannula of at least 18 gauge in size, if the tumor could not be identified on the initial nonenhanced scan, or if the tumor was less than 2 cm in depth. Twenty-two additional patients were excluded as a consequence. Of these 22 patients, 10 were excluded because the tumor could not be identified (n = 5) or the tumor was less than 2 cm in depth (n = 5).

CT Technique
Following a 4-hour fast, 1000 mL of water-soluble contrast material with a concentration of 2%–4% meglumine and sodium diatrizoate (Gastrografin; Bracco, Milan, Italy) was ingested 30 minutes prior to scanning to opacify the small bowel as per usual practice in our institution. Twenty milligrams of the spasmolytic agent hyoscine butylbromide (Buscopan; Boehringer Ingelheim, Ingelheim am Rhein, Germany) was administered intravenously to all patients immediately prior to data acquisition by means of an 18-gauge cannula sited in the antecubital fossa to minimize bowel peristalsis during the CT study. An abdominal band restraint was placed to minimize abdominal wall motion. All patients were examined by using a four–detector row CT scanner (Lightspeed Plus; GE Healthcare Technologies). An abdominal-pelvic study was performed initially without intravenous contrast material to identify the CT spatial coordinates of the known colorectal tumor by using the following parameters: 120 kV, 180 mA, 0.6-second rotation speed, 10-mm section collimation, 50-cm field of view, and 512 x 512-mm matrix. The tumor margins were identified by one of three supervising radiologists (including V.G.) with a range of 4–10 years of experience in body CT. The scan coordinates were noted, and these coordinates were used to plan the subsequent dynamic study.

For the dynamic study, a pump injector (Percupump Touchscreen; E-Z-Em, Westbury, NY) was used to inject intravenously 100 mL of iopamidol 340 (Niopam 340; Bracco) at a rate of 5 mL/sec. Four contiguous sections, each collimated to 5 mm (providing z-axis tumor coverage of 2 cm), were obtained at 1-second intervals by using a cine-mode acquisition (120 kV, 60 mA, 1-second rotation speed, 50-cm field of view, 512 x 512-mm matrix, 10-mSV effective dose). This acquisition commenced 5 seconds following the start of intravenous injection to allow acquisition of baseline nonenhanced images and continued for a total duration of 65 seconds. Respiration was not suspended during this study, but patients were coached prior to the examination to minimize respiratory excursion. The dynamic study was followed by a diagnostic portal venous phase abdominal-pelvic study that started 75 seconds following the commencement of intravenous injection by using the following parameters: 120 kV, 280 mA, 0.6-second rotation speed, 5-mm collimation, 50-cm field of view, and 512 x 512-mm matrix. This diagnostic study was used to determine the local and distant stage of the tumor, and a radiology report was issued following interpretation of these images as per day-to-day clinical practice.

Image Analysis
All scan data were transferred to a stand-alone workstation (Advantage Windows 4.2; GE Healthcare Technologies). Viewing and analysis were performed initially by one radiologist (V.G.) with 6 years of experience in perfusion CT by using perfusion software that was based on distributed parameter analysis (Body protocol, Perfusion 3.0; GE Healthcare Technologies). The first of the four 5-mm contiguous sections acquired was analyzed by using the software package as follows. A processing threshold of 0–120 HU was chosen to optimize soft-tissue visualization. An arterial input was defined by using the mouse to place a circular ROI, 10 mm² in area, in the best-visualized artery (aorta, iliac artery, or femoral artery) on the selected image. The arterial time enhancement curve was derived automatically by using the software, and resulting parametric maps were produced, with each pixel representing a parameter value. Tumor blood volume, blood flow, and permeability–surface area product measurements were obtained by using different tumor ROIs: (a) a circular ROI of 40 mm² in area selected from the available perfusion software ROI templates, (b) a circular ROI of 120 mm² in area selected from the available perfusion software ROI templates, and (c) an ROI defined by tracing a line freehand around the perceived tumor margins by using the cursor and mouse (mean ROI area, 1017 mm²; range, 390–3338 mm²) (Fig 1). The circular ROI sizes were chosen for pragmatic reasons, given the size range of the tumors evaluated, to enable nonoverlapping ROI placement at the tumor edge and center, whereas the ROI outlining the tumor was chosen because this method is commonly performed in clinical practice.

View larger version (149K): [in this window] [in a new window] [Download PPT slide]   Figure 1a: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

View larger version (41K): [in this window] [in a new window] [Download PPT slide]   Figure 1b: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

View larger version (40K): [in this window] [in a new window] [Download PPT slide]   Figure 1c: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

View larger version (40K): [in this window] [in a new window] [Download PPT slide]   Figure 1d: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

View larger version (42K): [in this window] [in a new window] [Download PPT slide]   Figure 1e: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

View larger version (42K): [in this window] [in a new window] [Download PPT slide]   Figure 1f: (a) Morphologic image shows cecal cancer (arrow). (b–f) Corresponding blood volume parametric maps show 40-mm2 ROI at (b) tumor edge and (c) center, 120-mm2 ROI at (d) tumor edge and (e) center, and (f) ROI outlining visible tumor. 2 = ROI identifier.

To provide an assessment of the intraobserver variability, analysis for each of the circular ROIs was repeated five times by the same observer, with 3 weeks between each analysis. Analysis for the ROI outlining the tumor was performed in a more conventional manner and was repeated twice only because variation was expected to be lower, as noted in previous published data (8–11). Thus, five repeated estimates for each parameter value for each of the randomly placed circular ROIs were obtained for each section, and the process resulted in 20 estimates for each parameter value for each of the circular ROIs per patient. Two repeated estimates for each parameter value for the ROI outlining the tumor were obtained for each section, and the process resulted in eight estimates for each parameter value per patient. A subset of five patients, chosen arbitrarily as every eighth patient from an alphabetical list of the study data set, was assessed in the same manner by a second radiologist (A.G.), with 6 years of experience in body CT and 1 year of experience in perfusion CT, to assess interobserver variability.

Statistical Analysis
Statistical analysis was performed by a statistician (J.S. or D.W.) with a software package (Stata 7.0; Stata, College Station, Tex). The mean and standard deviation (SD) were determined for each ROI defined for each of the perfusion parameters. The t tests for paired data were performed to examine differences in the perfusion parameters between the circular ROIs. Bonferroni correction was used for multiple comparisons, and a difference with P .002 was considered significant.

Intra- and interobserver variability were examined for each of the perfusion parameters for each ROI with univariate analysis of variance. Intraobserver variability was assessed with one-way analysis of variance, with the patient as factor for the total sample of 47 patients. In the subsample of five patients with two ratings, both intra- and interobserver variability were assessed with two-way random-effects analysis of variance, with patient and rater as factors. Variance components were estimated on the basis of the mean squares from each model.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 
Tumor ROI Position and Size
Blood volume, blood flow, and permeability–surface area product measurements were substantially higher at the tumor edge than at the tumor center for both the 40- and 120-mm² ROIs (P < .0001) (Tables 1, 2; Fig 2). At the tumor edge, there were small but significant differences between the 40-mm² and the 120-mm² circular ROIs for blood volume and blood flow (P .002) but not for permeability–surface area product. No significant differences were found between the 40- and 120-mm² circular ROIs for either measurement at the tumor center.

View larger version (14K): [in this window] [in a new window] [Download PPT slide]   Figure 2a: Range of values for (a) blood volume in milliliters per 100 g of tissue and (b) blood flow and (c) permeability–surface area product in milliliters per 100 g of tissue per minute for five tumor ROIs. Blood volume, blood flow, and permeability–surface area product were substantially different at tumor edge and tumor center. Highest measurements were at tumor edge; measurements obtained by outlining tumor were intermediate (between those at tumor edge and tumor center) in all cases.

View larger version (15K): [in this window] [in a new window] [Download PPT slide]   Figure 2b: Range of values for (a) blood volume in milliliters per 100 g of tissue and (b) blood flow and (c) permeability–surface area product in milliliters per 100 g of tissue per minute for five tumor ROIs. Blood volume, blood flow, and permeability–surface area product were substantially different at tumor edge and tumor center. Highest measurements were at tumor edge; measurements obtained by outlining tumor were intermediate (between those at tumor edge and tumor center) in all cases.

View larger version (15K): [in this window] [in a new window] [Download PPT slide]   Figure 2c: Range of values for (a) blood volume in milliliters per 100 g of tissue and (b) blood flow and (c) permeability–surface area product in milliliters per 100 g of tissue per minute for five tumor ROIs. Blood volume, blood flow, and permeability–surface area product were substantially different at tumor edge and tumor center. Highest measurements were at tumor edge; measurements obtained by outlining tumor were intermediate (between those at tumor edge and tumor center) in all cases.

Observer Agreement
The intra- and interobserver variances are expressed as absolute values for the 95% CIs, which together with the corresponding mean values show the variability in scores caused by intra- and interobserver agreement (Tables 3, 4). Both intraobserver and interobserver agreement were poor for placement of the 40- and 120-mm² ROIs, with variability ranging from 20% (1.3/6.6) to 66% (11.1/16.9) for intraobserver agreement and between 50% (2.2/4.4) and 80% (9.2/11.5) for interobserver agreement. Interobserver agreement was poorer than intraobserver agreement and was poorer for the smaller ROI (Table 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

We found that the position of the tumor ROI significantly influenced derived estimates of tumor perfusion. In terms of location, when we used an ROI of the same size, we found that estimates of blood volume, blood flow, and permeability–surface area product were highest at the tumor edge and lowest at the tumor center. These findings are concordant with morphologic data that have considered the distribution of vessels in colorectal cancers. For example, findings in a study of the three-dimensional vascular structure of colorectal tumors assessed by using microvessel corrosion casting and electron scanning micrography indicated elegantly that tumor vessels demonstrate a zonal distribution irrespective of tumor size (12). Vascularity decreased from the tumor edge to the center, with the tumor center poorly vascularized. Vessels in the tumor center also appeared compressed and elongated when they were compared with vessels (which were dilated) in areas of greater vessel density (12). The difference between rim and tumor center perfusion has been noted also in animal tumor models at other sites. For example, investigators in a previous study of rabbit VX2 tumor in the liver found that hepatic blood flow and volume were significantly higher at the tumor periphery than at the tumor center (13).

The size of the ROI also influenced final measurements, particularly at the tumor edge. Mean parameter values were substantially higher for the 40-mm² ROI than they were for the 120-mm² ROI at the tumor edge but not at the tumor center. Both intra- and interobserver agreement were poor but were worse for the 40-mm² ROI. In clinical practice where multiple repeated measurements may be performed to assess therapeutic effect in one patient, this level of measurement variation would be unacceptable. Human studies in which the drug effects have been investigated typically have been performed with the use of freehand ROIs traced around the visible tumor boundary. For example, when perfusion CT demonstrated that bevacizumab, an antiangiogenic drug, produced a direct tumor antivascular effect in rectal cancer, the ROI was drawn freehand around the rectal tumor (3). Likewise, when CT demonstrated the antivascular effect of N-nitro-L-arginine, a nitric oxide synthase inhibitor, in lung and pelvic cancer, the ROI was drawn freehand around the tumor for multiple contiguous 10-mm-thick sections to encompass the entire tumor (4). We found that measurements obtained by using freehand ROI outlining of the tumor lay intermediate to measurements obtained from the center and edge irrespective of the size (40- or 120-mm² ROI), thus reflecting averaging by the software. Although our findings suggest that evaluation by using one ROI may not accurately reflect the differences in tumor perfusion that are present at the zonal level, levels of observer agreement for this type of ROI analysis are more acceptable, as shown in our study and in the studies of others (8–11), and this may yet remain an appropriate form of analysis in the context of therapeutic assessment.

To provide a more encompassing evaluation, our data suggest that tumor measurements should be obtained from the edge, center, and entire tumor volume, but performance of these measurements adds additional complexity and is not without its problems. First, as our data show, ROI placement is not objective and is subject to considerable variation; the smaller ROI produced the greatest variation in values. This finding has been demonstrated also with other tumors, such as breast cancer enhancement at contrast-enhanced magnetic resonance imaging (14,15): Small ROIs are difficult to standardize within and between observers. In our study, both intra- and interobserver agreement were worse for the smaller ROIs. Furthermore, it can be difficult to define the edge and center consistently in some tumors.

A study of rabbit VX2 tumor defined central tumor through use of an ROI that was located where hepatic blood flow was less than 75% of normal (13). The rim was then defined as the difference between a whole-tumor ROI from which the central ROI was subtracted (13). This type of approach may not be possible at other tumor sites, including the colon and rectum. Although CT software that automatically segments tumors is available, for example, for the lung (16) and colon (17), automatic segmentation is presently unfeasible for segmenting different areas within a tumor. Other methods of displaying processed data may provide a better alternative. Such methods include histogram analysis of all of the pixel data, and histogram analysis is currently an option with some commercially available software platforms. Methods of assessing tumor heterogeneity, such as fractal analysis, also may become available commercially. To date, there has been no formal assessment by using these methods.

One of the limitations of our study is that we do not have correlative histopathologic data to substantiate our vascular findings. Although most of the patients eventually underwent surgical resection, and therefore were staged pathologically, assessment of vessel density by using microvessel counting is not performed routinely in our institution. Thus, we were unable to confirm our findings with histologic vessel distribution. Second, we have not explored the effect of the arterial input. Researchers in previous studies have addressed this issue within the cranial circulation of patients with cerebral artery territory infarction (18,19). Variations in arterial ROI size and placement have been found to have no substantial effect on ultimate perfusion values, although venous ROI placement influenced final measurements by as much as threefold in one study (18). Of course, this is irrelevant for body perfusion, as a venous ROI is not defined during analysis. We also have not assessed the reproducibility of the technique on this occasion by performing two separate studies on two separate occasions; however, published data that are based on the same technique suggest that the reproducibility of the technique is sufficiently robust for tumor evaluation (20,21).

In conclusion, we showed that the position and size of the tumor ROI influence estimates of perfusion obtained, and the degree of observer variation for placement of small ROIs found in our study would be unacceptable for clinical practice. Although an ROI outlining the tumor may not best reflect the spatial distribution of the tumor vasculature, outlining of the tumor may remain the most appropriate method for analysis. Until recommendations for standardized practice are made, we believe the type of tumor ROI analysis should be specified during therapeutic assessment to ensure consistency.

	ADVANCES IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

• The position and size of the tumor region of interest (ROI) influence vascular parameter values: Tumor vascular parameter values are highest at the tumor edge and lowest at the tumor center; values for an ROI outlining visible tumor lie between those at the tumor edge and tumor center, reflecting the averaging effect.
  
• Inter- and intraobserver agreement is poor for smaller ROIs; use of an ROI outlining visible tumor may be more reliable for therapeutic assessment.
  

	IMPLICATION FOR PATIENT CARE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

• How tumor ROI analysis is performed influences ultimate tumor vascular parameters; a consensus should be reached for such ROI analysis for therapeutic assessment.
  

	FOOTNOTES

 

Abbreviations: CI = confidence interval • ROI = region of interest • SD = standard deviation

See Materials and Methods for pertinent disclosures.

Author contributions: Guarantors of integrity of entire study, V.G., S.H.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, V.G.; clinical studies, V.G., A.G.; statistical analysis, V.G., D.W., J.S.; and manuscript editing, V.G., S.H., C.I.B.

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE IMPLICATION FOR PATIENT CARE References

 

1. Miles KA, Griffiths MR. Perfusion CT: a worthwhile enhancement? Br J Radiol 2003;76:220–231.[Free Full Text]
2. Goh V, Padhani AR, Rasheed S. Functional imaging of colorectal cancer angiogenesis. Lancet Oncol 2007;8:245–255.[CrossRef][Medline]
3. Willett CG, Boucher Y, Di Tomaso E, et al. Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 2004;10:145–147.[CrossRef][Medline]
4. Ng QS, Goh V, Milner J, et al. Effect of nitric oxide synthesis on tumour blood volume and vascular activity: a phase I study. Lancet Oncol 2007;8:111–118.[CrossRef][Medline]
5. Sahani DV, Kalva SP, Hamberg LM, et al. Assessing tumor perfusion and treatment response in rectal cancer with multisection CT: initial observations. Radiology 2005;234:785–792.[Abstract/Free Full Text]
6. Yi CA, Lee KS, Kim EA, et al. Solitary pulmonary nodules: dynamic enhanced multi–detector row CT study and comparison with vascular endothelial growth factor and microvessel density. Radiology 2004;233:191–199.[Abstract/Free Full Text]
7. Wang JH, Min PQ, Wang PJ, et al. Dynamic CT evaluation of tumor vascularity in renal cell carcinoma. AJR Am J Roentgenol 2006;186:1423–1430.[Abstract/Free Full Text]
8. Goh V, Halligan S, Hugill JA, Bassett P, Bartram CI. Quantitative assessment of colorectal cancer perfusion using MDCT: inter- and intraobserver agreement. AJR Am J Roentgenol 2005;185:225–231.[Abstract/Free Full Text]
9. Ng QS, Goh V, Fichte H, et al. Lung cancer perfusion at multi–detector row CT: reproducibility of whole tumor quantitative measurements. Radiology 2006;239:547–553.[Abstract/Free Full Text]
10. Fiorella D, Heiserman J, Prenger E, Partovi S. Assessment of the reproducibility of postprocessing dynamic CT perfusion data. AJNR Am J Neuroradiol 2004;25:97–107.[Abstract/Free Full Text]
11. Sanelli PC, Nicola G, Tsiouris AJ, et al. Reproducibility of postprocessing of quantitative CT perfusion maps. AJR Am J Roentgenol 2007;188:213–218.[Abstract/Free Full Text]
12. Konerding MA, Fait E, Gaumann A. 3D microvascular architecture of pre-cancerous lesions and invasive carcinomas of the colon. Br J Cancer 2001;84:1354–1362.[CrossRef][Medline]
13. Stewart EE, Chen X, Hadway J, Lee TY. Correlation between hepatic tumor blood flow and glucose utilization in a rabbit tumor model. Radiology 2006;239:740–750.[Abstract/Free Full Text]
14. Mussurakis S, Buckley DL, Horsman A. Dynamic MRI of invasive breast cancer: assessment of three region-of-interest analysis methods. J Comput Assist Tomogr 1997;21:431–438.[CrossRef][Medline]
15. Liney GP, Gibbs P, Hayes C, Leach MO, Turnbull LW. Dynamic contrast enhanced MRI in the differentiation of breast tumors: user-defined versus semi-automated region-of-interest analysis. J Magn Reson Imaging 1999;10:945–949.[CrossRef][Medline]
16. Li F, Arimura H, Suzuki K, et al. Computer-aided detection of peripheral lung cancers missed at CT: ROC analyses without and with localization. Radiology 2005;237:684–690.[Abstract/Free Full Text]
17. Burling D, Halligan S, Roddie ME, et al. Computed tomography colonography: automated diameter and volume measurement of colonic polyps compared with a manual technique—in vitro study. J Comput Assist Tomogr 2005;29:387–393.[CrossRef][Medline]
18. Sanelli PC, Lev MH, Eastwood JD, Gonzalez RG, Lee TY. The effect of varying user-selected input parameters on quantitative values in CT perfusion maps. Acad Radiol 2004;11:1085–1092.[CrossRef][Medline]
19. Kealey SM, Loving VA, Delong DM, Eastwood JD. User-defined vascular input function curves: influence on mean perfusion parameter values and signal-to-noise ratio. Radiology 2004;231:587–593.[Abstract/Free Full Text]
20. Purdie TG, Henderson E, Lee TY. Functional CT imaging of angiogenesis in rabbit VX2 soft-tissue tumor. Phys Med Biol 2001;46:3161–3175.[CrossRef][Medline]
21. Goh V, Halligan S, Hugill JA, Bartram CI. Quantitative assessment of tissue perfusion using MDCT: comparison of colorectal cancer and skeletal muscle measurement reproducibility. AJR Am J Roentgenol 2006;187:164–169.[Abstract/Free Full Text]

This article has been cited by other articles:

				L.-J. Zhang, Y.-E Zhao, S.-Y. Wu, B. M. Yeh, C.-S. Zhou, X.-B. Hu, Q.-J. Hu, and G.-M. Lu Pulmonary Embolism Detection with Dual-Energy CT: Experimental Study of Dual-Source CT in Rabbits Radiology, July 1, 2009; 252(1): 61 - 70. [Abstract] [Full Text] [PDF]

				C. M. Park, J. M. Goo, H. J. Lee, M. A Kim, H.-C. Kim, K. G. Kim, C. H. Lee, and J.-G. Im FN13762 Murine Breast Cancer: Region-by-Region Correlation of First-Pass Perfusion CT Indexes with Histologic Vascular Parameters Radiology, June 1, 2009; 251(3): 721 - 730. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) All Versions of this Article: 2473070414v1 247/3/726    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Goh, V. Articles by Bartram, C. I. Search for Related Content PubMed PubMed Citation Articles by Goh, V. Articles by Bartram, C. I.