- Email: [email protected]
Whole tumour perfusion of peripheral lung carcinoma: evaluation with first-pass CT perfusion imaging at 64-detector row CT
Clinical Radiology (2008) 63, 629e635
Whole tumour perfusion of peripheral lung carcinoma: evaluation with first-pass CT perfusion imaging at 64-detector row CT Y. Lia, Z.-g. Yanga,b,*, T.-w. Chena, Y.-p. Denga, J.-q. Yua, Z.-l. Lia a
Department of Radiology, and bNational Key Laboratory of Biotherapy Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China
Received 21 September 2007; received in revised form 21 December 2007; accepted 23 December 2007
AIM: To prospectively assess the feasibility of a whole-tumour perfusion technique using 64-detector row computed tomography (CT) and to analyse the variation of CT perfusion parameters in different histological types, sizes, and metastases in patients with peripheral lung carcinoma. METHODS AND MATERIALS: Ninety-seven pathologically proved peripheral lung carcinomas (less than 5 cm in largest diameter) underwent dynamic contrast-enhanced CT using a 64-detector row CT machine. Small amounts of iodinated contrast medium with a sharp bolus profile (50 ml, 6e7 ml/s), and 12 repeated fast acquisitions encompassing the entire tumour lesion were adopted to quantify perfusion of the whole-tumour during first-pass of contrast medium. Four kinetic parameters, including perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV), were measured and statistically compared among different histological types, sizes, and metastases. RESULTS: Mean values for perfusion, PEI, TTP, and BV of the 97 lung carcinomas were 57.5 45.4 ml/min/ml (range 5.9e243 ml/min/ml), 53.4 40.6 HU (range 10.3e234.4 HU), 34 11 s (range 11e60 s), and 30.1 21.7 ml/100 g (range 3.9e113.4 ml/100 g), respectively. No statistical differences were found between the histological types regarding the perfusion parameters (p > 0.05). Perfusion, PEI, and BV of stage T2 tumours were significantly lower than those of stage T1 tumours (all p < 0.05), whereas no statistically significant differences was found between other stages of tumours (all p > 0.05). Perfusion of the tumours with distant metastasis was significantly higher than that of the tumours without distant metastasis (p < 0.05), but there was no statistically significant difference between nodal metastasis positive and negative groups (p > 0.05). CONCLUSION: The present study of first-pass perfusion imaging using 64-detector row CT could provide a feasible method for assessment of whole-tumour perfusion. CT perfusion parameters of peripheral lung carcinoma may be associated with tumour size and distant metastasis. ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Lung cancer is the leading cause of cancer death in both men and women worldwide. Efforts have been made to identify contrast-enhanced computed tomography (CT) findings that can evaluate tumour vascularity based on the analyses of functional parameters, such as perfusion, permeability, and
* Guarantor and correspondent: Z.-g. Yang, Department of Radiology, West China Hospital, Sichuan University, 37# Guo Xue Xiang, Chengdu, Sichuan 610041, China. Tel./fax: þ86 28 85423817. E-mail address: [email protected] (Z.-g. Yang).
blood volume since the 1990s.1e6 Incremental acquisitions at long time intervals were used in these measurements, and the sample volume was often restricted to a single-section level. It is necessary to develop techniques that can be used to detect minor changes of tumour perfusion more accurately, rapidly, and comprehensively. The advent of new-generation helical CT machines offering high spatial and temporal resolution, versatile imaging sequences, and extensive availability, may allow the adoption of quantitative measures of whole-tumour perfusion within the first-pass of contrast medium. To the best of the authors’ knowledge, there has been no reported series of first-pass whole-tumour perfusion measurements
0009-9260/$ - see front matter ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2007.12.012
630
(i.e., dynamic CT imaging encompassing the entire tumour lesion during the first circulation of an injected contrast medium) of peripheral lung carcinomas using 64-detector row CT.1e7 The purpose of the present study was to prospectively evaluate, in patients with peripheral lung carcinoma, the feasibility of obtaining reliable data for whole-tumour perfusion by using 64-detector row CT and to analyse the variances of perfusion parameters in different histological types, sizes, and metastases.
Materials and methods Patients Between September 2006 and April 2007, 110 consecutive patients (79 men, 31 women; age range 23e81 years; mean age 57.9 years) suspected of peripheral lung carcinoma (less than 5 cm in largest diameter) underwent contrastenhanced dynamic CT perfusion imaging. Our institutional review board approved this prospective study, and informed written consent was obtained from all participants before the study. Ninety-seven primary lung carcinomas were finally diagnosed from the pathological specimens obtained at surgical resection (n ¼ 46), videoassisted thoracoscopic surgery (n ¼ 5), bronchoscopy (n ¼ 16) or CT-guided transthoracic needle aspiration (n ¼ 30). The patients ranged in age from 26e79 years (mean age 58.2 years), including 72 men and 25 women. All patients underwent diagnostic procedures, including brain CT or magnetic resonance imaging (MRI), whole-body CT, and bone scintigraphy. Each patient had one peripheral pulmonary carcinoma; 55 had adenocarcinoma, 30 had squamous cell carcinoma, four had large cell carcinoma, six had small cell carcinoma, and two had adenosquamous carcinoma. The 97 pulmonary carcinomas measured 1.8e5 cm in largest diameter (mean, 3.3 0.85 cm). For tumours that were surgically resected, the final staging was determined on the basis of surgical and pathological findings, according to the tumour-node-metastasis (TNM) classification system.8 For tumours that were not resected surgically, tumour stage was defined from the clinical data and reports of the diagnostic procedures. Thirty-two tumours were T1, 34 were T2, 13 were T3, and 18 were T4. Lymph node involvement was positive in 57 patients and negative in 40 patients. Eighteen patients had distant metastasis, such as brain, adrenal or liver metastasis.
Y. Li et al.
Dynamic CT imaging Imaging was performed using a 64-detector row CT machine (Philips Brilliance 64; Philips Medical System, Best, the Netherlands). A 19 G cannula (B. Braun, Melsungen, AG, Germany) was placed into cubital vein while the patient lay supine on the CT table. First, breath-hold helical CT (120 kV, 200 mAs) was undertaken, without intravenous contrast medium, in order to locate the tumour and plan acquisition of the CT perfusion imaging. Dynamic CT perfusion images were subsequently acquired by using 12 repeated spiral scans that encompassed the entire tumour lesion at 5 s intervals. In all patients, 50 ml iodinated contrast medium (Ultravist 300, iopamidol, Schering, Germany) was administrated as a bolus using a pump injector (Meorao-Stellant, Medrad, Inc., Pittsburgh, Pa., USA) at rate of 6e7 ml/s. The dynamic acquisition commenced 5 s after the start of bolus injection, and the total duration time varied from patient to patient, but was approximately 55 s. The following parameters were used: 120 kV; 100 mAs; rotation time 0.40 s; table speed 110 mm/s; collimation 64 0.625 mm; field of view 350 mm; matrix 512 512. In this protocol, the current used was half that used for conventional diagnostic imaging (i.e., 200 mAs) to minimize radiation dose but offer low noise. Patients were reminded to breathe gently during the dynamic scan acquisition to minimize movement. Dynamic images were reconstructed with 5 mm section thickness using a standard reconstruction algorithm without edge enhancement at a display window width of 350 HU and a window level of 40 HU. The total imaging time varied from patient to patient, but was approximately 55 s. An experienced radiologist (Z.G.Y., who had 22 years of experience in thoracic radiology) supervised the acquisition of the perfusion studies.
Data analysis On completion of the CT examination, data were transferred to the Extended Brilliance Workstation (Philips Medical System) and analysed by another radiologist (Y.L., with 3 years of experience interpreting perfusion CT images) by using commercial perfusion software (Brilliance perfusion 2.1.1). The artery input was determined by placing a circular region of interest (ROI) over the aorta or the left subclavian artery if the aorta was not included in the section. A circular or oval ROI for calculation of perfusion parameters was established around the peripheral margin of the tumour which was larger than 70% of diameter of the
Whole tumour perfusion of peripheral lung carcinoma
tumour, while avoiding atelectatic lung tissue, intratumoural necrosis, and cavitation.1,3 The analytical method used in this study was based on the maximum slope model and dedicated to perfusion measurement of lung carcinoma, yielding four major kinetic parameters: (1) perfusion (measured in ml/min/ml); (2) peak enhancement intensity (PEI, measured in HU); (3) time to peak (TTP, measured in s); and (4) blood volume (BV, measured in ml/ 100 g). Moreover, time attenuation curves (TACs) for the input artery and tumour were generated, respectively, by the software along with colour maps of the four kinetic parameters. We used a section-by-section averaging technique to evaluate the whole-tumour perfusion. First, we repeated drawing the ROI for each contiguous transverse level of the entire tumour lesion. Then, a global value representing the perfusion of the entire tumour was calculated by taking the mean value of all individual sections involved. To minimize operator-dependent bias, each perfusion study was analysed with the observer blinded to the patients’ clinical or histopathological data. To test intraobserver agreement, each case was reanalysed 1 month later. Results of the two sets of measurements were compared, and when good agreement was achieved between the replicated measurements, values of the first set of measurement were chosen to represent the perfusion parameters for lung carcinomas.
Statistical analysis All statistical analyses were performed by using commercially available software (SPSS, release 11.5; SPSS, Chicago, IL, USA). Repeatability between the two sets of measurements was assessed by methods described by Bland and Altman.9 The mean differences between measurements, SD of the differences, 95% confidence interval, and 95% limits of agreement for each of the four perfusion parameters were determined. We also used intraclass correlation coefficients (95% confidence intervals) to assess the level of agreement as described by Fleiss and Cohen.10 The perfusion parameters, including perfusion, PEI, TTP, and BV, were compared in different histological types, sizes, and metastases. Normality of distribution was firstly evaluated with the ShapiroeWilk test. Because none of the continuous variables was normally distributed, all statistical analyses were performed with nonparametric methods (KruskaleWallis test). Data were summarized with median and range from 27e75th percentile of interquartile range (IQR). A p-value of 0.05 or less was considered to indicate a statistically significant difference for all statistical tests.
631
Results Repeatability of CT perfusion parameters of peripheral lung carcinoma between the replicated measurements The mean values of first-pass perfusion, PEI, TTP, and BV were 57.5 45.4 ml/min/ml (range 5.9e243 ml/min/ml), 53.4 40.6 HU (range 10.3e 234.4 HU), 34 11 s (range 11e60 s), and 30.1 21.7 ml/100 g (range 3.9e113.4 ml/100 g), respectively, for the first set of measurements, and 56.8 43.8 ml/min/ml (range 6.3e239.4 ml/min/ ml), 53.1 38.2 HU (range 10.9e226.4 HU), 34.1 10.9 s (range 11e60 s), and 29.5 19.7 ml/ 100 g (range 3.7e87.9 ml/100 g), respectively, for the second set of measurements. Good agreements were obtained between the replicated measurements in terms of measuring four first-pass perfusion parameters (Table 1).
CT perfusion parameters of peripheral lung carcinoma The mean CT perfusion parameters of peripheral lung carcinoma according to pathological cell types are summarized in Table 2. Functional images of perfusion, PEI, TTP, and BV for each contiguous transverse level of an adenocarcinoma are shown in Fig. 1. Perfusion, PEI, and BV were highest for small cell carcinoma, followed by adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, but these differences were not statistically significant (all p > 0.05). Detailed analyses of perfusion parameters with regard to each T stage are summarized in Table 3. Perfusion, PEI, and BV of stage T2 tumours were significantly lower than those of stage T1 tumours (all p < 0.05), whereas no statistically significant differences were found between other tumours with different T stages (all p > 0.05). As for tumour metastases, perfusion of the distant metastasis group manifested significantly higher than those of the tumours without distant metastasis (Table 4, p < 0.05), whereas no significant differences were found between nodal metastasis positive and negative groups (Table 4, all p > 0.05).
Discussion A growing body of evidence indicates that contrastenhanced dynamic CT can be applied for quantitative, non-invasive evaluation of blood-flow patterns and
632
Y. Li et al.
Table 1 Repeatability for first-pass perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements Perfusion parameters
Differences between measurements (mean SD)
95% CI
Perfusion (ml/min/ml) PEI (HU) TTP (s) BV (ml/100 g)
0.66 3.56 0.3 3.82 0.08 0.34 0.55 3.41
0.06 0.47 0.15 0.13
to to to to
1.38 1.07 0.01 1.24
95% Limits of agreement
ICC (95% CI)
6.32 7.19 0.75 6.13
0.9973 0.9971 0.9995 0.9913
to to to to
7.63 7.78 0.59 7.23
(0.9962 (0.9957 (0.9992 (0.9870
to to to to
0.9983) 0.9981) 0.9996) 0.9942)
95% CI, 95% confidence interval; ICC, intraclass correlation coefficient.
angiogenesis of lung carcinomas1e7; however, the technique of previous measurements was rather limited. The sample volume was restricted to a single transverse level, which might be problematic for quantitative assessment of perfusion in lung lesions, especially in large tumours. With the development of multisection CT, the capability to cover a larger tissue volume, even whole-organ acquisition, is possible. Ng et al.7 assessed wholetumour perfusion using a helical acquisition technique and found that perfusion measurement derived from a greater tumour volume was more sensitive to variation. In the present study, spiral CT systems with acquisition capabilities of up to 64 sections per gantry rotation were introduced. The measurement of whole-tumour perfusion could be obtained using fast, repeated, spiral acquisitions and the section-by-section averaging technique. Furthermore, owing to lower noise and higher spatial resolution, perfusion images obtained using 64-detector row CT machines might be of better quality than other techniques. As the early phase of contrast enhancement seems to be very important to the characterization
of tumour blood flow pattern, a faster imaging technique that focused on the first-pass of contrast medium was attempted in the present study. According to Miles,11,12 the first-pass of contrast material through the vascular system typically comprises the initial 45e60 s after intravenous injection, and during this phase, the contrast material is predominantly intravascular. A protocol that aims to assess tissue perfusion should be comprised of a rapid sequence of images during the first circulation of an injected contrast medium, and the optimum bolus profile is an injection of 40e50 ml contrast medium (300 mg iodine/ml) at 7 ml/s. Faster dynamic acquisitions, smaller amounts of iodinated contrast medium, and sharper bolus profiles (50 ml, 6e7 ml/s) than those of previous measurements were used in the present study resulting in a suitable dedicated firstpass CT perfusion study. Furthermore, the reproducibility and measurement error of perfusion parameters was studied closely, because it is important for quantitative measurements, especially when considering therapeutic responses. Some investigators have
Table 2 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for different tumour types Perfusion parameters
Tumour type
p-Value
Adenocarcinoma (n ¼ 55)
Squamous cell carcinoma (n ¼ 30)
Small cell carcinoma (n ¼ 6)
Large cell carcinoma (n ¼ 4)
Perfusion (ml/min/ml) Median 25e75th percentile of IQR
48.2 28.4e79.5
41.4 22.3e65.3
63.3 38.9e95.4
21.9 9.2e47.2
PEI (HU) Median 25e75th percentile of IQR
46.8 25.3e69.2
36.4 23.2e58
58.1 44.3e76.8
29.4 13.5e57.6
TTP (s) Median 25e75th percentile of IQR
36 27e42.0
36.5 26.5e45
29 24.5e33
35 29e44.5
BV (ml/100 g) Median 25e75th percentile of IQR
25.2 14.7e44.5
22.7 12.9e30.4
34.6 25.9e43.2
14.6 8.5e20.3
IQR, interquartile range.
0.144
0.316
0.471
0.072
Whole tumour perfusion of peripheral lung carcinoma
Figure 1 A 67-year-old woman with an adenocarcinoma. Functional maps of perfusion (a), PEI (b), TTP (c) and BV (d) for each contiguous transverse level of the entire tumour lesion. The distribution of perfusion within the tumour is heterogeneous. The colour spectrum indicates the value of the perfusion parameters, ranging from high (red) to low (blue). The mean value of all individual sections involved for perfusion, PEI, TTP, and BV are 54.7 ml/100 g/min, 63.5 HU, 41 s, and 43.2 ml/100 g, respectively.
633
reported that intraobserver agreement was better than interobserver agreement.13e15 The authors suggested that the follow-up measurement on individual patients should be made by the original reviewer. Indeed, no two observers are likely to outline exactly the same ROI, which is very important in measuring tissue perfusion. Errors due to observer differences, even when low, may lead to great variability. Therefore, in the present study the measurements were performed by the same observer rather than by two observers, and achieved a high degree of agreement between the replicated measurements. To investigate the variation of CT perfusion parameters in lung carcinoma, the parameters for different histological types, sizes, and metastases were compared. It was found that perfusion, PEI, and BV of lung carcinomas according to pathological cell types was highest for small cell carcinomas, followed by adenocarcinomas, squamous cell carcinomas, and large cell carcinomas, but these differences were not statistically significant. This result was partly consistent with previous reports. Yamashita et al.3 demonstrated that the maximum attenuation of adenocarcinoma was higher than that of large cell carcinomas and squamous cell carcinoma, and these differences were not statistically significant. This discrepancy of relative low perfusion for large cell carcinomas in the present series might be due to the small sample size. Both the histological type of lung carcinoma and other factors could influence the enhancement characteristics of lung carcinomas. As shown in the present study, perfusion, PEI, and BV were found to be lowest in stage T2 tumours, which was in keeping with the results of Kiessling et al.16 and Miles et al.17 Their studies also indicated a trend for blood flow to be lower in the larger lung carcinomas. It is known that solid tumours proliferate automatically to a small size of a few millimetres in diameter. Further expansion requires angiogenesis, which is essential to deliver nutrients for tumour growth, invasion, and metastatic spread.18 Although a higher level of angiogenesis would be expected to increase the perfusion of tumour,2,3,19 it is unlikely that any one factor is responsible for tumour perfusion. Numerous factors might have been implicated in the process of tumour perfusion, especially in the advanced tumours. As the tumour grows, perfusion might be decreased because of a number of other biophysical parameters, such as infection, hypoxia, and necrosis.20 As a result, decrease of blood volume could lead to low perfusion status in a large tumour. Moreover, the present study demonstrated a similar high level of perfusion, PEI, and BV in both stage
634
Y. Li et al.
Table 3 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for different tumour sizes Perfusion parameter
T stage T1
p-Value T2
T3
T4
Perfusion (ml/min/ml) Median 25e75% of IQR
60.5 43e88.9
33.1 18.4e61
54.9 17.3e95.2
41.7 24.9e71.7
PEI (HU) Median 25e75% of IQR
61.2 32.7e72.5
35.2 21.1e52.4
60.2 25.8e75
44.6 24.3e69.1
TTP (s) Median 25e75% of IQR
31 26.5e37
39 26.5e47.5
36 28.5e46.5
34.5 28.5e40.5
BV (ml/100 g) Median 25e75% of IQR
32.1 22e51.9
20.1 12.6e30.9
35.6 12.6e47.9
25.2 14e53.7
T1 versus T2
T1 versus T3
T1 versus T4
T2 versus T3
T2 versus T4
T3 versus T4
0.000a
0.381
0.067
0.128
0.248
0.548
0.001a
0.625
0.110
0.078
0.419
0.548
0.089
0.111
0.163
0.784
0.507
0.784
0.000a
0.438
0.206
0.089
0.110
0.936
IQR, interquartile range. a Significant difference was found between the two groups by means of KruskaleWallis test.
T1 and T3 tumours. These results might have been due to the small sample size of stage T3 tumours. Besides, the overlap of the size between stage T1 and T3 tumours might have been responsible for the similarity. To fully understand the association between perfusion and size of tumour, data from a larger patient population with various T stages, especially advanced tumours, are needed in future studies. Regarding tumour metastases, the present results demonstrated that perfusion, PEI, and BV for tumours with distant metastasis were significantly higher than those for tumours without distant
metastasis. Nevertheless, perfusion parameters of the node metastasis group showed complete overlap with metastasis negative group. Although neovascularization has been found to correlate with the incidence of distant metastasis in nonsmall cell carcinoma21 and with relapse after surgery in adenocarcinoma,22 none have shown the relationship between lymph node metastasis and angiogenesis in lung carcinoma. Findings in some reports demonstrated that microvessel counts were not correlated with lymph node metastases.23,24 These results together with the present results, reflect the difference in the mechanism
Table 4 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for tumour metastases Perfusion parameters
Nodal metastasis Negative (n ¼ 40)
Distant metastasis Positive (n ¼ 57)
Perfusion (ml/min/ml) Median 25e75% of IQR
39 22.1e61
55 30.1e77.2
PEI (HU) Median 25e75% of IQR
40.7 20.3e59.8
49.3 26e68.3
TTP (s) Median 25e75% of IQR
36.5 27.5e46.5
33 26.5e41
BV (ml/100 g) Median 25e75% of IQR
23.2 14.1e35.2
27.2 13.9e41.3
p-Value
Negative (n ¼ 79)
Positive (n ¼ 18)
p-Value 0.025a
0.157 43.2 22.4e70.1
68.3 33.7e117.6
44.1 24.3e64.4
64.7 27.4e96.2
36 27e43
29.5 25.5e37
23.1 13.9e36.6
36.8 14.0e59.9
0.395
0.076
0.291
0.070
0.369
0.055
IQR, interquartile range. a Significant difference was found between the two groups by means of KruskaleWallis test.
Whole tumour perfusion of peripheral lung carcinoma
of producing lymphatics and that of angiogenesis, and might indicate that the incidence of nodule metastasis in lung carcinoma depends not only on tumour angiogenesis, but also on other factors, such as angiolymphatic invasion, adhesion molecules, and immunological mechanisms. Neither perfusion nor other dynamic measurement is an adequate measure of tumour metastasis. There were a few limitations to the present study. First, perfusion CT is associated with considerable irradiation to the patient, because the dataset was primarily obtained for clinical assessment of lung carcinomas. Second, the sample sizes of small cell carcinoma and large cell carcinoma were small. Data from a larger patient population with various types of primary lung carcinoma are needed in future studies. Third, the observational period was relative short, and the clinical outcomes based on the perfusion parameters obtained using dynamic CT could not be analysed. A further follow-up study with larger numbers of patients may be required to investigate the usefulness of perfusion CT in cancer prognosis assessments. In conclusion, the protocol used in the present study showed that 64-detector row CT was feasible to assess first-pass whole-tumour perfusion. One of the important applications of this technique is the indication of a highly perfused area, which could be used to guide biopsy to the site of the most active tumour, reducing the chance of sampling error. Moreover, the results of this study demonstrated that perfusion parameters of peripheral lung carcinoma were associated with tumour size and distant metastasis, which might be due to underlying tumour angiogenesis. These findings contribute valuable information for angiogenic therapeutic approaches.
References 1. Swensen SJ, Brown LR, Colby TV, et al. Pulmonary nodules: CT evaluation of enhancement with iodinated contrast material. Radiology 1995;194:393e8. 2. Swensen SJ, Brown LR, Colby TV, et al. Lung nodule enhancement at CT: prospective findings. Radiology 1996; 201:447e55. 3. Yamashita K, Matsunobe S, Takahashi R, et al. Small peripheral lung carcinoma evaluated with incremental dynamic CT: radiologyepathological correlation. Radiology 1995; 196:401e8. 4. Zhang M, Kono M. Solid pulmonary nodules: evaluation of blood flow patterns with dynamic CT. Radiology 1997;205: 471e8.
635
5. Swensen SJ, Viggiano RW, Midthun DE, et al. Nodule enhancement at CT: multicenter study. Radiology 2000; 214:73e80. 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;233:191e9. 7. Ng QS, Goh V, Klotz E, et al. Quantitative assessment of lung cancer perfusion using MDCT: does measurement reproducibility improve with greater tumour volume coverage? AJR Am J Roentgenol 2006;187:1079e84. 8. Moutain CF. Revisions in the international system for staging lung cancer. Chest 1997;111:1710e7. 9. Bland MJ, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307e10. 10. Fleiss JL, Cohen J. The equivalence of weighted kappa and the intraclass correlation coefficient as measures of reliability. Educ Psychol Meas 1973;33:613e9. 11. Miles KA. Perfusion CT for the assessment of tumour vascualrity: which protocol? Br J Radiol 2003;76:S36e42. 12. Miles KA. Ideal contrast medium bolus for perfusion measurement in dynamic CT. Radiology 1998;209:583. 13. Goh V, Halliga S, Hugill JA, et al. Quantitative assessment of colorectal cancer perfusion using MDCT: inter- and intraobserver agreement. AJR Am J Roentgenol 2005; 185:225e31. 14. Goh V, Halligan S, Hugill JA, et al. Quantitative assessment of tissue perfusion using MDCT: comparison of colorectal cancer and skeletal muscle measurement reproducibility. AJR Am J Roentgenol 2006;187:164e9. 15. Jackson A, Jayson GC, Li KL, et al. Reproducibility of quantitative dynamic contrast enhanced MRI in newly presenting glioma. Br J Radiol 2003;76:153e62. 16. Kiessling F, Boese J, Corvinus C, et al. Perfusion CT in patients with advanced bronchial carcinomas: a novel chance for characterization and treatment monitoring? Eur Radiol 2004;14:1226e33. 17. Miles KA, Griffiths MR, Keith CJ. Blood flow-metabolic relationships are dependent on tumour size in non-small cell lung cancer: a study using quantitative contrast-enhanced computer tomography and positron emission tomography. Eur J Nucl Med Mol Imaging 2006;33:22e8. 18. Folkman J. The vascularization of tumour. Sci Am 1976; 234:59. 19. Tateishi U, Kusumoto M, Nishihara H, et al. Contrastenhanced dynamic computed tomography for the evaluation of tumor angiogenesis in patients with lung carcinoma. Cancer 2002;95:835e42. 20. Costouros NG, Diehn FE, Libutti SK. Molecular imaging of tumour angiogenesis. J Cell Biochem Suppl 2002;39:72e8. 21. Jain RK, Safabakhsh N, Sckell A, et al. Endothelial cell death, angiogenesis, and microvascular function after castration in an androgen-dependent tumour: role of vascular endothelial grow factor. Proc Natl Acad Sci U S A 1998;95:10820e5. 22. Macchiarini P, Fontanini G, Hardin MJ, et al. Relation of neovascularization to metastasis of non-small cell lung cancer. Lancet 1992;340:145e6. 23. Yamazaki K, Abe S, Takekawa H, et al. Tumour angiogenesis in human lung adenocarcinoma. Cancer 1994;74:2245e50. 24. Yano T, Tanikawa S, Fujie T, et al. Vascular endothelial growth factor expression and noevascularisation in nonsmall cell lung cancer. Eur J Cancer 2000;36:601e9.
Whole tumour perfusion of peripheral lung carcinoma: evaluation with first-pass CT perfusion imaging at 64-detector row CT Y. Lia, Z.-g. Yanga,b,*, T.-w. Chena, Y.-p. Denga, J.-q. Yua, Z.-l. Lia a
Department of Radiology, and bNational Key Laboratory of Biotherapy Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China
Received 21 September 2007; received in revised form 21 December 2007; accepted 23 December 2007
AIM: To prospectively assess the feasibility of a whole-tumour perfusion technique using 64-detector row computed tomography (CT) and to analyse the variation of CT perfusion parameters in different histological types, sizes, and metastases in patients with peripheral lung carcinoma. METHODS AND MATERIALS: Ninety-seven pathologically proved peripheral lung carcinomas (less than 5 cm in largest diameter) underwent dynamic contrast-enhanced CT using a 64-detector row CT machine. Small amounts of iodinated contrast medium with a sharp bolus profile (50 ml, 6e7 ml/s), and 12 repeated fast acquisitions encompassing the entire tumour lesion were adopted to quantify perfusion of the whole-tumour during first-pass of contrast medium. Four kinetic parameters, including perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV), were measured and statistically compared among different histological types, sizes, and metastases. RESULTS: Mean values for perfusion, PEI, TTP, and BV of the 97 lung carcinomas were 57.5 45.4 ml/min/ml (range 5.9e243 ml/min/ml), 53.4 40.6 HU (range 10.3e234.4 HU), 34 11 s (range 11e60 s), and 30.1 21.7 ml/100 g (range 3.9e113.4 ml/100 g), respectively. No statistical differences were found between the histological types regarding the perfusion parameters (p > 0.05). Perfusion, PEI, and BV of stage T2 tumours were significantly lower than those of stage T1 tumours (all p < 0.05), whereas no statistically significant differences was found between other stages of tumours (all p > 0.05). Perfusion of the tumours with distant metastasis was significantly higher than that of the tumours without distant metastasis (p < 0.05), but there was no statistically significant difference between nodal metastasis positive and negative groups (p > 0.05). CONCLUSION: The present study of first-pass perfusion imaging using 64-detector row CT could provide a feasible method for assessment of whole-tumour perfusion. CT perfusion parameters of peripheral lung carcinoma may be associated with tumour size and distant metastasis. ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Lung cancer is the leading cause of cancer death in both men and women worldwide. Efforts have been made to identify contrast-enhanced computed tomography (CT) findings that can evaluate tumour vascularity based on the analyses of functional parameters, such as perfusion, permeability, and
* Guarantor and correspondent: Z.-g. Yang, Department of Radiology, West China Hospital, Sichuan University, 37# Guo Xue Xiang, Chengdu, Sichuan 610041, China. Tel./fax: þ86 28 85423817. E-mail address: [email protected] (Z.-g. Yang).
blood volume since the 1990s.1e6 Incremental acquisitions at long time intervals were used in these measurements, and the sample volume was often restricted to a single-section level. It is necessary to develop techniques that can be used to detect minor changes of tumour perfusion more accurately, rapidly, and comprehensively. The advent of new-generation helical CT machines offering high spatial and temporal resolution, versatile imaging sequences, and extensive availability, may allow the adoption of quantitative measures of whole-tumour perfusion within the first-pass of contrast medium. To the best of the authors’ knowledge, there has been no reported series of first-pass whole-tumour perfusion measurements
0009-9260/$ - see front matter ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2007.12.012
630
(i.e., dynamic CT imaging encompassing the entire tumour lesion during the first circulation of an injected contrast medium) of peripheral lung carcinomas using 64-detector row CT.1e7 The purpose of the present study was to prospectively evaluate, in patients with peripheral lung carcinoma, the feasibility of obtaining reliable data for whole-tumour perfusion by using 64-detector row CT and to analyse the variances of perfusion parameters in different histological types, sizes, and metastases.
Materials and methods Patients Between September 2006 and April 2007, 110 consecutive patients (79 men, 31 women; age range 23e81 years; mean age 57.9 years) suspected of peripheral lung carcinoma (less than 5 cm in largest diameter) underwent contrastenhanced dynamic CT perfusion imaging. Our institutional review board approved this prospective study, and informed written consent was obtained from all participants before the study. Ninety-seven primary lung carcinomas were finally diagnosed from the pathological specimens obtained at surgical resection (n ¼ 46), videoassisted thoracoscopic surgery (n ¼ 5), bronchoscopy (n ¼ 16) or CT-guided transthoracic needle aspiration (n ¼ 30). The patients ranged in age from 26e79 years (mean age 58.2 years), including 72 men and 25 women. All patients underwent diagnostic procedures, including brain CT or magnetic resonance imaging (MRI), whole-body CT, and bone scintigraphy. Each patient had one peripheral pulmonary carcinoma; 55 had adenocarcinoma, 30 had squamous cell carcinoma, four had large cell carcinoma, six had small cell carcinoma, and two had adenosquamous carcinoma. The 97 pulmonary carcinomas measured 1.8e5 cm in largest diameter (mean, 3.3 0.85 cm). For tumours that were surgically resected, the final staging was determined on the basis of surgical and pathological findings, according to the tumour-node-metastasis (TNM) classification system.8 For tumours that were not resected surgically, tumour stage was defined from the clinical data and reports of the diagnostic procedures. Thirty-two tumours were T1, 34 were T2, 13 were T3, and 18 were T4. Lymph node involvement was positive in 57 patients and negative in 40 patients. Eighteen patients had distant metastasis, such as brain, adrenal or liver metastasis.
Y. Li et al.
Dynamic CT imaging Imaging was performed using a 64-detector row CT machine (Philips Brilliance 64; Philips Medical System, Best, the Netherlands). A 19 G cannula (B. Braun, Melsungen, AG, Germany) was placed into cubital vein while the patient lay supine on the CT table. First, breath-hold helical CT (120 kV, 200 mAs) was undertaken, without intravenous contrast medium, in order to locate the tumour and plan acquisition of the CT perfusion imaging. Dynamic CT perfusion images were subsequently acquired by using 12 repeated spiral scans that encompassed the entire tumour lesion at 5 s intervals. In all patients, 50 ml iodinated contrast medium (Ultravist 300, iopamidol, Schering, Germany) was administrated as a bolus using a pump injector (Meorao-Stellant, Medrad, Inc., Pittsburgh, Pa., USA) at rate of 6e7 ml/s. The dynamic acquisition commenced 5 s after the start of bolus injection, and the total duration time varied from patient to patient, but was approximately 55 s. The following parameters were used: 120 kV; 100 mAs; rotation time 0.40 s; table speed 110 mm/s; collimation 64 0.625 mm; field of view 350 mm; matrix 512 512. In this protocol, the current used was half that used for conventional diagnostic imaging (i.e., 200 mAs) to minimize radiation dose but offer low noise. Patients were reminded to breathe gently during the dynamic scan acquisition to minimize movement. Dynamic images were reconstructed with 5 mm section thickness using a standard reconstruction algorithm without edge enhancement at a display window width of 350 HU and a window level of 40 HU. The total imaging time varied from patient to patient, but was approximately 55 s. An experienced radiologist (Z.G.Y., who had 22 years of experience in thoracic radiology) supervised the acquisition of the perfusion studies.
Data analysis On completion of the CT examination, data were transferred to the Extended Brilliance Workstation (Philips Medical System) and analysed by another radiologist (Y.L., with 3 years of experience interpreting perfusion CT images) by using commercial perfusion software (Brilliance perfusion 2.1.1). The artery input was determined by placing a circular region of interest (ROI) over the aorta or the left subclavian artery if the aorta was not included in the section. A circular or oval ROI for calculation of perfusion parameters was established around the peripheral margin of the tumour which was larger than 70% of diameter of the
Whole tumour perfusion of peripheral lung carcinoma
tumour, while avoiding atelectatic lung tissue, intratumoural necrosis, and cavitation.1,3 The analytical method used in this study was based on the maximum slope model and dedicated to perfusion measurement of lung carcinoma, yielding four major kinetic parameters: (1) perfusion (measured in ml/min/ml); (2) peak enhancement intensity (PEI, measured in HU); (3) time to peak (TTP, measured in s); and (4) blood volume (BV, measured in ml/ 100 g). Moreover, time attenuation curves (TACs) for the input artery and tumour were generated, respectively, by the software along with colour maps of the four kinetic parameters. We used a section-by-section averaging technique to evaluate the whole-tumour perfusion. First, we repeated drawing the ROI for each contiguous transverse level of the entire tumour lesion. Then, a global value representing the perfusion of the entire tumour was calculated by taking the mean value of all individual sections involved. To minimize operator-dependent bias, each perfusion study was analysed with the observer blinded to the patients’ clinical or histopathological data. To test intraobserver agreement, each case was reanalysed 1 month later. Results of the two sets of measurements were compared, and when good agreement was achieved between the replicated measurements, values of the first set of measurement were chosen to represent the perfusion parameters for lung carcinomas.
Statistical analysis All statistical analyses were performed by using commercially available software (SPSS, release 11.5; SPSS, Chicago, IL, USA). Repeatability between the two sets of measurements was assessed by methods described by Bland and Altman.9 The mean differences between measurements, SD of the differences, 95% confidence interval, and 95% limits of agreement for each of the four perfusion parameters were determined. We also used intraclass correlation coefficients (95% confidence intervals) to assess the level of agreement as described by Fleiss and Cohen.10 The perfusion parameters, including perfusion, PEI, TTP, and BV, were compared in different histological types, sizes, and metastases. Normality of distribution was firstly evaluated with the ShapiroeWilk test. Because none of the continuous variables was normally distributed, all statistical analyses were performed with nonparametric methods (KruskaleWallis test). Data were summarized with median and range from 27e75th percentile of interquartile range (IQR). A p-value of 0.05 or less was considered to indicate a statistically significant difference for all statistical tests.
631
Results Repeatability of CT perfusion parameters of peripheral lung carcinoma between the replicated measurements The mean values of first-pass perfusion, PEI, TTP, and BV were 57.5 45.4 ml/min/ml (range 5.9e243 ml/min/ml), 53.4 40.6 HU (range 10.3e 234.4 HU), 34 11 s (range 11e60 s), and 30.1 21.7 ml/100 g (range 3.9e113.4 ml/100 g), respectively, for the first set of measurements, and 56.8 43.8 ml/min/ml (range 6.3e239.4 ml/min/ ml), 53.1 38.2 HU (range 10.9e226.4 HU), 34.1 10.9 s (range 11e60 s), and 29.5 19.7 ml/ 100 g (range 3.7e87.9 ml/100 g), respectively, for the second set of measurements. Good agreements were obtained between the replicated measurements in terms of measuring four first-pass perfusion parameters (Table 1).
CT perfusion parameters of peripheral lung carcinoma The mean CT perfusion parameters of peripheral lung carcinoma according to pathological cell types are summarized in Table 2. Functional images of perfusion, PEI, TTP, and BV for each contiguous transverse level of an adenocarcinoma are shown in Fig. 1. Perfusion, PEI, and BV were highest for small cell carcinoma, followed by adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, but these differences were not statistically significant (all p > 0.05). Detailed analyses of perfusion parameters with regard to each T stage are summarized in Table 3. Perfusion, PEI, and BV of stage T2 tumours were significantly lower than those of stage T1 tumours (all p < 0.05), whereas no statistically significant differences were found between other tumours with different T stages (all p > 0.05). As for tumour metastases, perfusion of the distant metastasis group manifested significantly higher than those of the tumours without distant metastasis (Table 4, p < 0.05), whereas no significant differences were found between nodal metastasis positive and negative groups (Table 4, all p > 0.05).
Discussion A growing body of evidence indicates that contrastenhanced dynamic CT can be applied for quantitative, non-invasive evaluation of blood-flow patterns and
632
Y. Li et al.
Table 1 Repeatability for first-pass perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements Perfusion parameters
Differences between measurements (mean SD)
95% CI
Perfusion (ml/min/ml) PEI (HU) TTP (s) BV (ml/100 g)
0.66 3.56 0.3 3.82 0.08 0.34 0.55 3.41
0.06 0.47 0.15 0.13
to to to to
1.38 1.07 0.01 1.24
95% Limits of agreement
ICC (95% CI)
6.32 7.19 0.75 6.13
0.9973 0.9971 0.9995 0.9913
to to to to
7.63 7.78 0.59 7.23
(0.9962 (0.9957 (0.9992 (0.9870
to to to to
0.9983) 0.9981) 0.9996) 0.9942)
95% CI, 95% confidence interval; ICC, intraclass correlation coefficient.
angiogenesis of lung carcinomas1e7; however, the technique of previous measurements was rather limited. The sample volume was restricted to a single transverse level, which might be problematic for quantitative assessment of perfusion in lung lesions, especially in large tumours. With the development of multisection CT, the capability to cover a larger tissue volume, even whole-organ acquisition, is possible. Ng et al.7 assessed wholetumour perfusion using a helical acquisition technique and found that perfusion measurement derived from a greater tumour volume was more sensitive to variation. In the present study, spiral CT systems with acquisition capabilities of up to 64 sections per gantry rotation were introduced. The measurement of whole-tumour perfusion could be obtained using fast, repeated, spiral acquisitions and the section-by-section averaging technique. Furthermore, owing to lower noise and higher spatial resolution, perfusion images obtained using 64-detector row CT machines might be of better quality than other techniques. As the early phase of contrast enhancement seems to be very important to the characterization
of tumour blood flow pattern, a faster imaging technique that focused on the first-pass of contrast medium was attempted in the present study. According to Miles,11,12 the first-pass of contrast material through the vascular system typically comprises the initial 45e60 s after intravenous injection, and during this phase, the contrast material is predominantly intravascular. A protocol that aims to assess tissue perfusion should be comprised of a rapid sequence of images during the first circulation of an injected contrast medium, and the optimum bolus profile is an injection of 40e50 ml contrast medium (300 mg iodine/ml) at 7 ml/s. Faster dynamic acquisitions, smaller amounts of iodinated contrast medium, and sharper bolus profiles (50 ml, 6e7 ml/s) than those of previous measurements were used in the present study resulting in a suitable dedicated firstpass CT perfusion study. Furthermore, the reproducibility and measurement error of perfusion parameters was studied closely, because it is important for quantitative measurements, especially when considering therapeutic responses. Some investigators have
Table 2 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for different tumour types Perfusion parameters
Tumour type
p-Value
Adenocarcinoma (n ¼ 55)
Squamous cell carcinoma (n ¼ 30)
Small cell carcinoma (n ¼ 6)
Large cell carcinoma (n ¼ 4)
Perfusion (ml/min/ml) Median 25e75th percentile of IQR
48.2 28.4e79.5
41.4 22.3e65.3
63.3 38.9e95.4
21.9 9.2e47.2
PEI (HU) Median 25e75th percentile of IQR
46.8 25.3e69.2
36.4 23.2e58
58.1 44.3e76.8
29.4 13.5e57.6
TTP (s) Median 25e75th percentile of IQR
36 27e42.0
36.5 26.5e45
29 24.5e33
35 29e44.5
BV (ml/100 g) Median 25e75th percentile of IQR
25.2 14.7e44.5
22.7 12.9e30.4
34.6 25.9e43.2
14.6 8.5e20.3
IQR, interquartile range.
0.144
0.316
0.471
0.072
Whole tumour perfusion of peripheral lung carcinoma
Figure 1 A 67-year-old woman with an adenocarcinoma. Functional maps of perfusion (a), PEI (b), TTP (c) and BV (d) for each contiguous transverse level of the entire tumour lesion. The distribution of perfusion within the tumour is heterogeneous. The colour spectrum indicates the value of the perfusion parameters, ranging from high (red) to low (blue). The mean value of all individual sections involved for perfusion, PEI, TTP, and BV are 54.7 ml/100 g/min, 63.5 HU, 41 s, and 43.2 ml/100 g, respectively.
633
reported that intraobserver agreement was better than interobserver agreement.13e15 The authors suggested that the follow-up measurement on individual patients should be made by the original reviewer. Indeed, no two observers are likely to outline exactly the same ROI, which is very important in measuring tissue perfusion. Errors due to observer differences, even when low, may lead to great variability. Therefore, in the present study the measurements were performed by the same observer rather than by two observers, and achieved a high degree of agreement between the replicated measurements. To investigate the variation of CT perfusion parameters in lung carcinoma, the parameters for different histological types, sizes, and metastases were compared. It was found that perfusion, PEI, and BV of lung carcinomas according to pathological cell types was highest for small cell carcinomas, followed by adenocarcinomas, squamous cell carcinomas, and large cell carcinomas, but these differences were not statistically significant. This result was partly consistent with previous reports. Yamashita et al.3 demonstrated that the maximum attenuation of adenocarcinoma was higher than that of large cell carcinomas and squamous cell carcinoma, and these differences were not statistically significant. This discrepancy of relative low perfusion for large cell carcinomas in the present series might be due to the small sample size. Both the histological type of lung carcinoma and other factors could influence the enhancement characteristics of lung carcinomas. As shown in the present study, perfusion, PEI, and BV were found to be lowest in stage T2 tumours, which was in keeping with the results of Kiessling et al.16 and Miles et al.17 Their studies also indicated a trend for blood flow to be lower in the larger lung carcinomas. It is known that solid tumours proliferate automatically to a small size of a few millimetres in diameter. Further expansion requires angiogenesis, which is essential to deliver nutrients for tumour growth, invasion, and metastatic spread.18 Although a higher level of angiogenesis would be expected to increase the perfusion of tumour,2,3,19 it is unlikely that any one factor is responsible for tumour perfusion. Numerous factors might have been implicated in the process of tumour perfusion, especially in the advanced tumours. As the tumour grows, perfusion might be decreased because of a number of other biophysical parameters, such as infection, hypoxia, and necrosis.20 As a result, decrease of blood volume could lead to low perfusion status in a large tumour. Moreover, the present study demonstrated a similar high level of perfusion, PEI, and BV in both stage
634
Y. Li et al.
Table 3 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for different tumour sizes Perfusion parameter
T stage T1
p-Value T2
T3
T4
Perfusion (ml/min/ml) Median 25e75% of IQR
60.5 43e88.9
33.1 18.4e61
54.9 17.3e95.2
41.7 24.9e71.7
PEI (HU) Median 25e75% of IQR
61.2 32.7e72.5
35.2 21.1e52.4
60.2 25.8e75
44.6 24.3e69.1
TTP (s) Median 25e75% of IQR
31 26.5e37
39 26.5e47.5
36 28.5e46.5
34.5 28.5e40.5
BV (ml/100 g) Median 25e75% of IQR
32.1 22e51.9
20.1 12.6e30.9
35.6 12.6e47.9
25.2 14e53.7
T1 versus T2
T1 versus T3
T1 versus T4
T2 versus T3
T2 versus T4
T3 versus T4
0.000a
0.381
0.067
0.128
0.248
0.548
0.001a
0.625
0.110
0.078
0.419
0.548
0.089
0.111
0.163
0.784
0.507
0.784
0.000a
0.438
0.206
0.089
0.110
0.936
IQR, interquartile range. a Significant difference was found between the two groups by means of KruskaleWallis test.
T1 and T3 tumours. These results might have been due to the small sample size of stage T3 tumours. Besides, the overlap of the size between stage T1 and T3 tumours might have been responsible for the similarity. To fully understand the association between perfusion and size of tumour, data from a larger patient population with various T stages, especially advanced tumours, are needed in future studies. Regarding tumour metastases, the present results demonstrated that perfusion, PEI, and BV for tumours with distant metastasis were significantly higher than those for tumours without distant
metastasis. Nevertheless, perfusion parameters of the node metastasis group showed complete overlap with metastasis negative group. Although neovascularization has been found to correlate with the incidence of distant metastasis in nonsmall cell carcinoma21 and with relapse after surgery in adenocarcinoma,22 none have shown the relationship between lymph node metastasis and angiogenesis in lung carcinoma. Findings in some reports demonstrated that microvessel counts were not correlated with lymph node metastases.23,24 These results together with the present results, reflect the difference in the mechanism
Table 4 Perfusion, peak enhancement intensity (PEI), time to peak (TTP), and blood volume (BV) measurements for tumour metastases Perfusion parameters
Nodal metastasis Negative (n ¼ 40)
Distant metastasis Positive (n ¼ 57)
Perfusion (ml/min/ml) Median 25e75% of IQR
39 22.1e61
55 30.1e77.2
PEI (HU) Median 25e75% of IQR
40.7 20.3e59.8
49.3 26e68.3
TTP (s) Median 25e75% of IQR
36.5 27.5e46.5
33 26.5e41
BV (ml/100 g) Median 25e75% of IQR
23.2 14.1e35.2
27.2 13.9e41.3
p-Value
Negative (n ¼ 79)
Positive (n ¼ 18)
p-Value 0.025a
0.157 43.2 22.4e70.1
68.3 33.7e117.6
44.1 24.3e64.4
64.7 27.4e96.2
36 27e43
29.5 25.5e37
23.1 13.9e36.6
36.8 14.0e59.9
0.395
0.076
0.291
0.070
0.369
0.055
IQR, interquartile range. a Significant difference was found between the two groups by means of KruskaleWallis test.
Whole tumour perfusion of peripheral lung carcinoma
of producing lymphatics and that of angiogenesis, and might indicate that the incidence of nodule metastasis in lung carcinoma depends not only on tumour angiogenesis, but also on other factors, such as angiolymphatic invasion, adhesion molecules, and immunological mechanisms. Neither perfusion nor other dynamic measurement is an adequate measure of tumour metastasis. There were a few limitations to the present study. First, perfusion CT is associated with considerable irradiation to the patient, because the dataset was primarily obtained for clinical assessment of lung carcinomas. Second, the sample sizes of small cell carcinoma and large cell carcinoma were small. Data from a larger patient population with various types of primary lung carcinoma are needed in future studies. Third, the observational period was relative short, and the clinical outcomes based on the perfusion parameters obtained using dynamic CT could not be analysed. A further follow-up study with larger numbers of patients may be required to investigate the usefulness of perfusion CT in cancer prognosis assessments. In conclusion, the protocol used in the present study showed that 64-detector row CT was feasible to assess first-pass whole-tumour perfusion. One of the important applications of this technique is the indication of a highly perfused area, which could be used to guide biopsy to the site of the most active tumour, reducing the chance of sampling error. Moreover, the results of this study demonstrated that perfusion parameters of peripheral lung carcinoma were associated with tumour size and distant metastasis, which might be due to underlying tumour angiogenesis. These findings contribute valuable information for angiogenic therapeutic approaches.
References 1. Swensen SJ, Brown LR, Colby TV, et al. Pulmonary nodules: CT evaluation of enhancement with iodinated contrast material. Radiology 1995;194:393e8. 2. Swensen SJ, Brown LR, Colby TV, et al. Lung nodule enhancement at CT: prospective findings. Radiology 1996; 201:447e55. 3. Yamashita K, Matsunobe S, Takahashi R, et al. Small peripheral lung carcinoma evaluated with incremental dynamic CT: radiologyepathological correlation. Radiology 1995; 196:401e8. 4. Zhang M, Kono M. Solid pulmonary nodules: evaluation of blood flow patterns with dynamic CT. Radiology 1997;205: 471e8.
635
5. Swensen SJ, Viggiano RW, Midthun DE, et al. Nodule enhancement at CT: multicenter study. Radiology 2000; 214:73e80. 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;233:191e9. 7. Ng QS, Goh V, Klotz E, et al. Quantitative assessment of lung cancer perfusion using MDCT: does measurement reproducibility improve with greater tumour volume coverage? AJR Am J Roentgenol 2006;187:1079e84. 8. Moutain CF. Revisions in the international system for staging lung cancer. Chest 1997;111:1710e7. 9. Bland MJ, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307e10. 10. Fleiss JL, Cohen J. The equivalence of weighted kappa and the intraclass correlation coefficient as measures of reliability. Educ Psychol Meas 1973;33:613e9. 11. Miles KA. Perfusion CT for the assessment of tumour vascualrity: which protocol? Br J Radiol 2003;76:S36e42. 12. Miles KA. Ideal contrast medium bolus for perfusion measurement in dynamic CT. Radiology 1998;209:583. 13. Goh V, Halliga S, Hugill JA, et al. Quantitative assessment of colorectal cancer perfusion using MDCT: inter- and intraobserver agreement. AJR Am J Roentgenol 2005; 185:225e31. 14. Goh V, Halligan S, Hugill JA, et al. Quantitative assessment of tissue perfusion using MDCT: comparison of colorectal cancer and skeletal muscle measurement reproducibility. AJR Am J Roentgenol 2006;187:164e9. 15. Jackson A, Jayson GC, Li KL, et al. Reproducibility of quantitative dynamic contrast enhanced MRI in newly presenting glioma. Br J Radiol 2003;76:153e62. 16. Kiessling F, Boese J, Corvinus C, et al. Perfusion CT in patients with advanced bronchial carcinomas: a novel chance for characterization and treatment monitoring? Eur Radiol 2004;14:1226e33. 17. Miles KA, Griffiths MR, Keith CJ. Blood flow-metabolic relationships are dependent on tumour size in non-small cell lung cancer: a study using quantitative contrast-enhanced computer tomography and positron emission tomography. Eur J Nucl Med Mol Imaging 2006;33:22e8. 18. Folkman J. The vascularization of tumour. Sci Am 1976; 234:59. 19. Tateishi U, Kusumoto M, Nishihara H, et al. Contrastenhanced dynamic computed tomography for the evaluation of tumor angiogenesis in patients with lung carcinoma. Cancer 2002;95:835e42. 20. Costouros NG, Diehn FE, Libutti SK. Molecular imaging of tumour angiogenesis. J Cell Biochem Suppl 2002;39:72e8. 21. Jain RK, Safabakhsh N, Sckell A, et al. Endothelial cell death, angiogenesis, and microvascular function after castration in an androgen-dependent tumour: role of vascular endothelial grow factor. Proc Natl Acad Sci U S A 1998;95:10820e5. 22. Macchiarini P, Fontanini G, Hardin MJ, et al. Relation of neovascularization to metastasis of non-small cell lung cancer. Lancet 1992;340:145e6. 23. Yamazaki K, Abe S, Takekawa H, et al. Tumour angiogenesis in human lung adenocarcinoma. Cancer 1994;74:2245e50. 24. Yano T, Tanikawa S, Fujie T, et al. Vascular endothelial growth factor expression and noevascularisation in nonsmall cell lung cancer. Eur J Cancer 2000;36:601e9.