Clinical Imaging 37 (2013) 256–264

Optimization of the scanning technique and diagnosis of pulmonarynodules with first-pass 64-detector-row perfusion VCT☆,☆☆

Sheng Jie Shu⁎, Bai Lu Liu, Hui Jie Jiang

Department of CT diagnosis, Second Affiliated Hospital of Harbin Medical University, Harbin, Heilongjiang 150086, China

Received 12 April 2012; accepted 1 May 2012

Abstract

Objective: The aims of this study were to optimize the scanning technique of first-pass 64-detector-row perfusion volume computedtomography imaging, to evaluate the effectiveness and stability of this scan protocol, and lastly to evaluate the differential diagnosis ability ofperfusion imaging in solitary pulmonary nodules (SPNs). Methods: A total of 144 patients with SPNs underwent perfusion scan with 64-slice spiral CT scanner. The CT perfusion imaging was analyzed for time–density curve, perfusion parametric maps, and the respectiveperfusion parameters. We then analyzed the main factors concerning the imaging quality and evaluated the effectiveness of scan protocol bydetermining the receiver operating characteristic (ROC) curve, diagnostic efficacy, and odds ratio as well as the stability of scan protocol byconsistency analysis. Immunohistochemical findings of microvessel density measurement and vascular endothelial growth factor expressionwere evaluated. Results: The total sensitivity, specificity, accuracy, positive predictive value, negative predictive value, likelihood ratio, andthe area under ROC curve during 5–45-s scan period were 78.95%, 82.4%, 80.6%, 83.3%, 77.8%, 4.620, 0.280, and 0.840, respectively, andKappa value was 0.894. The diagnostic efficacy of CT pulmonary perfusion was significantly higher than during 0–40-s scan period. Theparameter values in different nodules were different. Conclusion: The optimized 5–45-s scan period of CT pulmonary perfusion imaging iseffective in pathologic diagnosis and has good stability, worthy of being popularized. Lung perfusion CT could be a promising and feasiblemethod for differentiation of SPNs.© 2013 Published by Elsevier Inc.

Keywords: SPN; Perfusion; Volume Computed Tomography

1. Introduction

Morphologic findings would be helpful to differentiatebenign and malignant cavities when they have typicalfeatures [1], but differentiation may be challenging.Perfusion studies have become increasingly important.Recent studies indicate that differentiation between benign

☆ This article was supported by the Institute of Youth Fund Project ofthe Second Affiliated Hospital of Harbin Medical University (Projectnumber is QN2011-03).

☆☆ Conflict of interest: none.⁎ Corresponding author. Department of CT diagnosis, Second Affiliated

Hospital of Harbin Medical University, Harbin, Heilongjiang 150086,China. Tel.: +86 13904810289.

E-mail addresses: shengjiejiayou@126.com (S.J. Shu),liubailuhmu@sohu.com (B.L. Liu), Jhj68323@yahoo.com (H.J. Jiang).

0899-7071/$ – see front matter © 2013 Published by Elsevier Inc.http://dx.doi.org/10.1016/j.clinimag.2012.05.004

and malignant nodules is feasible using dynamic computedtomography (CT) or magnetic resonance with manuallydefined regions of interest (ROIs) [2–13].

Since data acquisition protocols differ between thestudies and most of the analysis methods employed requirerather cumbersome kinetic modeling, perfusion imaginghas not been established as a routine tool used in clinicalpractice. As such, there is continued demand for animaging strategy which can be used to classify pulmonarylesions with sufficient accuracy, such that therapeuticdecisions can be made without the need for invasiveinterventional procedures.

The purposes of this study were to search for optimizationof the scanning technique of first-pass 64-detector-rowperfusion volume computed tomography (VCT) imaging, toassess differential diagnosis among benign and malignantpulmonary nodules, and to investigate the relationship

257S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

between CT perfusion imaging and tumor angiogenesis andvascular endothelial growth factor (VEGF).

2. Materials and methods

2.1. Patients

The institutional review board approved this study, andwe obtained written informed consent from all patients priorto the study. A total of 144 patients with solitary pulmonarynodules (SPNs) underwent first-pass perfusion scan with 64-detector-row VCT scanner. The CT perfusion imaging wasanalyzed for time–density curve (TDC), perfusion paramet-ric maps, and the respective perfusion parameters. We thenanalyzed the main factors concerning the imaging qualityand evaluated the effectiveness of scan protocol bydetermining the receiver operating characteristic (ROC)curve, diagnostic efficacy, and odds ratio as well as thestability of scan protocol by consistency analysis. Immuno-histochemical findings of microvessel density (MVD)measurement and VEGF expression were evaluated.

One hundred and forty-four patients with SPNs fromNovember 2009 to January 2012 were selected. All patients(38 men, 34 women, age range 28–79 years, mean age 52.8years) were included in this study. The diameters of SPNswere 0.8–3.0 cm, and the average was 2.3 cm. All tumorsunderwent surgical resection operation or biopsy, and theduration between CT examination and surgery was within 2weeks. The 144 SPNs had a total of 76 malignant nodules(36 adenocarcinoma, 28 squamous carcinoma, 8 adenosqua-mous carcinoma, 4 metastatic carcinoma), 36 inflammatorynodules (32 inflammatory granuloma, 8 suppurative pneu-monia), and 32 benign nodules (16 tuberculoma, 16hamartoma). Patients were selected according to thefollowing criteria: (a) no preoperative treatment such aschemotherapy and radiotherapy; (b) no contraindications tocontrast medium administration; (c) capability of cooperat-ing with examination; and (d) absence of severe heart, liver,or kidney disease.

Two radiologists experienced in chest CT respectivelyobserved the TDCs and perfusion parametric maps.

All 144 patients experienced the whole 0–45-s scanningtime; we then compared the blood volume (BV) andpermeability surface (PS) between the two groups of 0–40s and 5–45 s.

2.2. Equipment and method

Images of the patients’ breath-holding phase wereobtained using the 64-detector-row VCT scanner (Light-speed, GE, USA) and when they held a supine position. Atfirst, a thin-slice scanning of the lesion was performed.Then, first-pass CT perfusion scan was performed after thetumor was found. The nonionic contrast material appliedwas 50 ml of Ultravist at a rate of 5 ml/s. Perfusion scan

was in a cine pattern, with 0.5 s/coil, 5 mm sectionthickness×8, 0 mm interval, no inclination of scan frame,tube voltage of 120 kV, tube current of 60 mA, expositiontime of 40 s, data acquisition time of 45 s, and delay timeof 0 s.

2.3. Analysis of CT perfusion images

The data were input into a work station (ADW4.3), andan image construction was performed. The standard ROIwas set in the maximal section of the solid area in the softtissues (not including calcification or necrosis). Thedynamic analysis software was used to draw a TDC ofthe SPN and the aorta (or common carotid artery if theaorta was not included in the section) on the same level.The curve was analyzed, and the following dynamicenhancement parameters were studied. The parameters ofCT perfusion including BV, blood flow (BF), mean transittime (MTT), and PS were analyzed. The peak height indensity, enhancement value (EV), SPN-to-aorta enhancedratio (S/A), and time to peak (TTP) were measured.

2.4. Immunohistochemical staining

Lung tissue specimens were prepared corresponding tothe ROI of the CT imaging section and fixed with a 4%polyformaldehyde solution, embedded in paraffin, andsliced at a thickness of 5 μm. One slice was hematoxylinand eosin stained, and the other two were immunohisto-chemically stained with streptavidin peroxidase. The anti-VEGF antibodies were the first antibodies. The firstantibody kit and immunohistochemical kit were purchasedfrom the Boster Company (Boster, China). The micro-scopic brown–yellow cytoplasm indicated that the VEGFexpressed positive cells, and the intensity was subdividedinto three degrees: strong positive if the proportion ofmicroscopic (×400) positive cells ≥50%, positive if theproportion b50% and ≥25%, and negative if theproportion b25%.

A pathologist experienced in lung pathology recorded thehistologic diagnosis of each pulmonary nodule. MVDmeasurement and VEGF expression were assessed bymeans of immunohistochemistry.

2.5. Statistical analysis

The BV, BF, MTT, PS, EV, S/A, and TTP of pulmonarynodules for all four anatomic section locations available foreach group of patients were recorded. All data wereexpressed as means±standard deviation (S.D.). The compar-ison of the parameters among the three kinds of SPN wasanalyzed using the t test. χ2 test was used in expression ofVEGF. Pearson coefficients were used to represent therelationship between the perfusion value (BV, BF, MTT, PS)and MVD measurement. Comparing 0–40-s group and 5–45-s group of ROC curves, we then got the sensitivity,

Fig. 1. Carcinoma of a 59-year-old man. (A) BVmap of 0–40 s. (B) BVmapof 5–45 s.

258 S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

specificity, accuracy, and area under the curve. All statisticaldata were analyzed using the SPSS16.0 software system. APb.05 was considered statistically significant.

3. Results

3.1. Parameters and maps of SPNs

The comparison of BV and PS between two scanningperiods of all the SPNs is in Table 1 and Fig. 1. Inconclusion, the diagnostic efficacy of CT pulmonaryperfusion during 5–45-s scan period was significantly higherthan during 0–40-s scan period, and the BV and PS of themhad a statistically significant difference (Pb.01).

The total sensitivity, specificity, accuracy, positivepredictive value, negative predictive value, likelihood ratio,and the area under ROC curve during 5–45-s scan periodwere 78.95%, 82.4%, 80.6%, 83.3%, 77.8%, 4.620, 0.280,and 0.840, respectively, and Kappa value was 0.894.

The perfusion parametric maps during 5–45-s scanningperiod were as follows (Figs. 2–4). The BV, PS, and BFvalues of the benign SPN were significantly lower thanthose of the malignant and the inflammatory SPN (Pb.05)(Table 1). The MTT of the malignant SPN was higherthan that of the benign and the inflammatory SPNs, butthey did not have a statistically significant difference(PN.05). The BV, PS, BF, and MTT value of theinflammatory SPN were significantly lower than those ofthe malignant SPN. The BV and PS of them had astatistically significant difference (Pb.05), but the BF andMTT of them did not have a statistically significantdifference (PN.05).

Since the BV and PS were valuable, they were used as thestandard for the identification of benign and malignant SPNs(Table 2).

The EV, S/A, and TTP of the malignant SPN weresignificantly higher than those of the benign and theinflammatory SPN (Pb.05) (Table 1). The above threekinds of data in the benign nodules and inflammatorynodules difference were not significant (PN.05).

Thus, it may be known: The EV and S/A had theobvious remarkable difference in the benign and themalignant SPN.

Table 1SPN parameters of three different types of pathological changes and intergroup comparison and analysis (mean±S.D.)

Group BV PS BF MTT EV (HU) S/A TTP (s)

(1) Malignant (n=76) 7.89±1.78 9.16±0.82 38.92±1.78 13.16±0.43 42.46±1.31 0.14±0.02 24.5±1.29(2) Inflammatory (n=36) 5.63±0.56 5.22±0.73 24.72±4.73 12.83±1.61 31.57±10.39 012±0.03 10.60±0.89(3) Benign (n=32) 1.05±0.58 1.60±0.23 20.93±5.15 12.91±0.75 25.34±8.85 0.10±0.02 9.43±0.53

Comparison between the benign group and malignant group: BV (Pb.001); PS (Pb.001); BF (Pb.001); MTT (PN.05); EV (Pb.001); S/A (Pb.001); TTP(Pb.001). Comparison between the inflammatory group and malignant group: BV (Pb.001); PS (Pb.001); BF (PN.05); MTT (PN.05); EV (Pb.05); S/A(Pb.05); TTP (Pb.001). Comparison between the benign group and inflammatory group: BV (Pb.001); PS (Pb.001); BF (Pb.05); MTT (PN.05); EV (PN.05);S/A (PN.05); TTP (Pb.05).

Fig. 2. Squamous cell carcinoma of a 55-year-old woman. (A) BV mapshows red and yellow colors signifying high perfusion, especially in theperipheral area. (B) Tumor cytoplasm brown staining (+++) (VEGFimmunohistochemistry, original magnification×400).

Fig. 3. Inflammatory nodule in a 68-year-old man. (A) BV map shows greencolor signifying middle perfusion.

Fig. 4. Pulmonary hamartoma in a 62-year-old man. (A) BV map showsyellow color signifying middle perfusion, especially in the peripheral area.

259S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

3.2. Curve types

The TDC was divided into three types, as follows: type 1:one that increases the peak height initially, maintains aplateau for some time, and finally decreases slowly; type 2:one that increases rapidly to the peak height, then declinesquickly or maintains a plateau for a longer time than doestype I, and finally declines quickly with a later peak timethan that of type I; type 3: one that increases moderately andmaintains relatively constant CT values between each timepoint (Fig. 5).

The TTP of the malignant SPN was statisticallysignificantly longer than that of the benign and theinflammatory SPN (Pb.01) (Table 1).

3.3. Immunohistochemical results

VEGF-positive particles were mainly in the cytoplasm,with a heterogenous distribution of stronger staining in thecytoplasm and weaker staining in the marginal regions oftumor cells. Nineteen cases of 38 malignant SPNs werestrongly positive, 15 cases were positive, and 4 cases werenegative. The inflammatory SPNs indicated the existence ofa copious amount of dilated mature capillary in the nodulewithout VEGF expression except in only one case of weakpositivity. The benign SPNs indicated little stained vesselswithout VEGF-positive expression (Figs. 2–4). VEGF-positive expression of malignant SPN was significantlyhigher than that of the benign and the inflammatory SPN

Table 2Set BV and PS for the diagnosis threshold of malignant nodules and the diagnosis results

The diagnosis result BV≥6 ml·100 g−1 PS≥7 ml·100 g−1·min−1 BV≥6 ml·100 g−1 and PS≥7 ml·100 g−1·min−1

(1) Sensitivity 89.5% 84.2% 94.7%(2) Specificity 82.4% 88.2% 94.1%(3) Accuracy 18.1% 86.1% 94.1%(4) Positive predictive value 85% 88.9% 94.4%(5) Negative predictive value 87.5% 83.3% 94.7%

260 S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

(Pb.01),while there was no statistically significant differencebetween the inflammatory and benign SPN (PN.05). TheSPN with VEGF-positive expression showed significantlyhigher BV, PS, BF, MTT, EV, S/A, and TTP than those ofthe SPN with VEGF-negative expression (Table 3). Therelationship between MVD and perfusion parameters ofthree groups of SPNs with VEGF-positive and -negativeexpression is shown in Table 4.

4. Discussion

As was shown in this study, spiral CT systems withacquisition capabilities of up to 64 sections per gantryrotation were introduced. Measure of the whole-tumorperfusion could be obtained by using fast repeated spiralacquisitions. And good agreement between the replicatedmeasurements was also achieved for each of the four first-pass perfusion parameters. This may indicate that volume-based technique with 64-detector-row CT could improve thereliability of tumor perfusion assessment.

For the initial period following intravenous injection ofthe iodinated contrast medium, contrast enhancement islargely due to the presence of the contrast medium within theintravascular space, which depends on the available volumeof blood vessels within the tumor. Therefore, a fasterimaging technique that focuses on the first pass of contrastmedium is necessary to assess tumor vascular functions.

A dedicated first-pass CT perfusion study was, therefore,acquired to assess tumor vascular functions of lung SPN.

First-pass CT perfusion based on the TDC and itscorresponding perfusion parameters on the same scan oftissues was studied to evaluate BF patterns. Many studiesdone on the brain and liver have been reported [14], butstudies and their applications on SPN were minimal [15].

CT perfusion 3 exports microvascular permeabilityexpressed as PS, which is theoretically based on a boluscontrast injection and image acquisition beyond the first-passcirculation. This requirement is not always fulfilled in ourshort data acquisition protocol. Data acquisition has to belimited to 40 s, mainly determined by the possible breath-holdduration but also by efforts to reduce radiation exposure.

4.1. Feasibility and optimization the scanning technique offirst-pass 64-detector-row perfusion VCT imaging

Current researchers commonly base their research on theBF pattern obtained from a single scan of the SPN, but this

does not accurately represent the entire pattern [15]. Thisstudy was performed using the 64-detector-row CT with arapid scanning speed, high time resolution (1-s long foreach scan), lower radiate dose, a single convolutionscanning (X-ray tube spinning one ring), and the largestrange of z-axis (maximum of 3.2 cm), with continuousexposure time of 40 s providing the condition for anevaluation of the entire pattern. A dynamic CT perfusionmay explore SPN features more accurately and reliablyusing the TDC, EV, S/A, and TTP than other factors [15].In this study, 72 of the 75 cases were included; the threecases with SPNs located near the cardiac border or thediaphragm were excluded. The criteria ensured that theresults of this study are feasible and reliable, except forresults from SPN lesions near the cardiac large vessel or thediaphragm; technological improvement is still necessary inorder to test SPNs in those areas.

In this study, the contrast was injected at a fast rate (5 ml/s) with a large dosage (50 ml) that increased the perfusion ofSPN vessels and consumed more of the diffusion time thandid the others. This result was supported by Yamashita et al[16], who used a slower injection rate (2 ml/s) with a largerdosage (100–150 ml) to show the peak height of pulmonarycancer in 2–5 min. The perfusion feature of the SPN mainlydepended on the contrast's diffusion to the outer vascularspace. Another research group [15] thought that thefollowing factors were responsible for perfusion features:(a) scanning method and multislice computed tomography(MSCT) application; (b)overweight, less cardiac output, andpulmonary BF in patients; (c) small nodule and fewer vesselsdistant from the pulmonary helium. However, the aboveresearch was limited to a small population, and newtechniques and a larger population are needed in futurestudies. Our study demonstrated that the peak height ofmalignant SPNs appeared significantly later than did thepeak height of inflammatory SPNs with two different TDCs,which indicated differences in their respective bloodsupplies. The evaluation, based on the peak height time,ignored the individual differences (especially the effect ofheart function), and thus, individual mistakes would bemade. Therefore, our study directly compared the TDC of theSPN with that of the aortic artery to choose the same sliceand to avoid analytical errors from individual differences.

In this study, the diagnostic efficacy of CT pulmonaryperfusion during 5–45-s scan period was significantly higherthan that of 0–40-s scan period, and the BV and PS of themhad a statistically significant difference (Pb.01). So, we can

Fig. 5. Three types of TDC: type 1: benign nodule; type 2: inflammatorynodule; type 3: malignant nodule.

261S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

draw a conclusion that choosing 5–45-s scan period wasbetter than 0–40-s scan period for diagnosis of SPNs.

CT lung perfusion imaging techniques are the keytechnologies of lung lesions of BF patterns. Because thescanning type, parameter setting, and the postprocessingperfusion software in the current study are not the same, thedata cannot be effectively compared and merged, so most ofthe research was conducted in a small study sample, limitingthe actual application and promotion of the research results.

4.2. Clinical application of perfusion and enhancementparameters in SPN

In this study, there were four perfusion parameters: BV,PS, BF, and MTT. The BV, PS, and BF values of themalignant SPN were significantly higher than those of thebenign and the inflammatory SPN (Pb.05), of which BVand PS had the largest difference. So they were valuable.For MTT, there was no statistical significance among thethree groups of nodules. The results of this study indicatedthat there was a better consistency among BF, BV, PS,EV, S/A, and TTP in the malignant SPN, inflammatorySPN, and benign SPN cases. This result suggested that thefour perfusion parameters may better reflect bloodperfusion status and the malignant and benign degree ofpulmonary nodules.

The reason may be the fact that the vascular supply ofmalignant nodule and inflammatory nodule is moreabundant, and therefore, both blood perfusions haveincreased. The vascular supply of benign nodule is lesser;

Table 3Comparison among the parameters between VEGF-positive and -negative express

VEGF expression BV PS BF

Positive (n=70) 7.24±0.21 8.33±1.15 30.33±2.08Negative (n=74) 3.00±1.58 4.66±1.53 24.00±1.00

Comparison between the VEGF expression positive group and negative group: BVTTP (Pb.05).

therefore, its blood perfusion is obviously less than that ofmalignant nodule and inflammatory nodule. This coincidedwith the study results of Swensen et al. [5] and Zhanget al. [12], who compared pdak height (PH), or perfusionvalue of malignant, inflammatory, and benign nodule, andsuggested that malignant and inflammatory nodule showedhigher PH, PHpm/PHa or higher perfusion value thanthose of benign nodule.

The BV value (unit: ml·100 mg−1) is the total BV in thevascular system of tissue or organ. It reflects blood primingvolume of tissue or organ, and represents the quantity of thefunctional blood capillary. The BV value is also concernedwith the size of blood and the quantity of patent bloodcapillary. Owing to the stimulus from angiogenesis factor inlung cancer, the blood vessel in it grows in number.Consequently, the BV value of lung cancer obviouslyincreases rather than that of benign mass. Therefore, the BVvalue possesses significant worth to the identification ofbenign and malignant diseases.

The PS value (unit: ml·min−1·100 mg−1) is a one-waytransmission speed for contrast material to enter cell spacesby capillary endothelium, and it objectively reflects thediffusion coefficient of blood vessel endothelial interspace oftumor angiogenesis. The pathologic observations of thisstudy concluded that there were many new tumor micro-vessels in peripheral lung cancer. These tumor microvesselsdeveloped immaturely, the vessel wall was not intact, and thePS of blood capillary increased. Therefore, the PS value ofperipheral lung cancer is higher. There were straighterbranches of blood vessels, mature blood capillaries, andnear-normal PSs of blood capillaries in benign andinflammatory nodules. Therefore, PS value of inflammatorynodule and benign nodule was lower and had statisticalsignificance. The malignant SPN is found to have moredevelopmental immaturity of tumor capillary, no intactvessel wall, and increased PS of blood capillary whichresulted from VEGF and other vascular growth factors.Therefore, the PS value of the malignant SPN is higher.Therefore, the BV value possesses significant worth to theidentification of benign and malignant diseases.

Since the BV and PS were valuable, they were used as thestandard for the identification of benign and malignant SPNs.

Set BV≥6 ml·100 g−1 for the diagnosis of malignantnodules threshold. The sensitivity was high (89.5%), butthe specificity was only 82.4%. Set PS≥7 ml·100g− 1·min− 1 for the diagnosis of malignant nodulesthreshold. The specificity was high (88.2%), but thesensitivity was low (84.2%). But Set BV≥6 ml·100 g−1

ion of three groups of 144 SPN (mean±S.D.)

MTT EV (HU) S/A TTP (s)

12.97±0.58 39.37±3.06 0.14±0.03 18.67±3.2111.67±0.58 27.67±0.58 0.10±0.01 11.0±1.00

(Pb.001); PS (Pb.05); BF (Pb.05); MTT (Pb.05); EV (Pb.05); S/A (Pb.05);

Table 4Relationship between MVD and perfusion parameters of three groups of SPNs with VEGF-positive and -negative expression

Group SPN n BV BF MTT PS

MVD of VEGF-positive expression Malignant 28 r 0.536 0.502 −0.301 0.675P .005 .001 .335 .000

Inflammatory 0 r - - - -P - - - -

Benign 5 r 0.836 0.892 −0.857 0.902P .169 .049 .074 .022

MVD of VEGF-negative expression Malignant 10 r 0.120 0.020 −0.162 0.126P .790 .989 .640 .756

Inflammatory 18 r 0.138 0.106 −0.259 0.292P .662 .731 .401 .326

Benign 11 r 0.412 0.354 −0.527 0.528P .287 .335 .119 .129

262 S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

and PS≥7 ml·100 g− 1·min− 1 for the diagnosis ofmalignant nodules threshold. Its sensitivity was high(94.7%), and specificity was high (94.1%). Both sensitivityand specificity were higher than with single standard. Thisstudy indicated that the BV and PS were valuableparameters for differentiating the three kinds of SPNs.

The BF value(unit: ml·min−1·100mg−1) is the velocityof BF in the tumor or tissue. It is influenced by manyfactors such as BV, draining vein, oxygen consumption oftissue, etc. The BF value reflects assemblage rate per unitof tissue to BF in the first scans after an intravenous bolus ofcontrast medium; it is directly correlated with numbers ofblood vessels. Yamashita et al. [6] suggested that bloodperfusion of pulmonary nodule was more correlated withsmall vessels (0.02–0.1 mm), which are microvessels. Inthis study, the BF value of the malignant SPN was higherthan the inflammatory SPN and benign SPN. It indicatedthat the BF has little value as a parameter fordifferentiating the lung SPN. But we still need morecases to verify it.

The MTT value(unit: s) is the mean transmit time ofblood passing vascular system or red blood cell passingpulmonary microcirculation. It is the optimal parameter forunderstanding microcirculation. Because of different paths,transit time is different; therefore, it is demonstrated byMTT, which mainly reflected the time the contrast materialgoes through the blood capillary. Since MTT value hasmore overlap between peripheral lung cancer and benignpulmonary nodule, the practical application value of MTTis lesser.

The EV, S/A, and TTP of the malignant SPN weresignificantly higher than those of the benign and theinflammatory SPN (Pb.05). Malignant SPNs were complete-ly enhanced with a significantly higher perfusion than werebenign SPNs due to the stimulus of VEGF and a highervessel count. An inflammatory SPN could be homogeneous-ly enhanced because of its rich blood supply and lack ofnecrosis. A benign SPN could be only slightly or notenhanced because of its small blood supply. Thus, perfusionof benign SPNs was obviously lower than that of malignantand inflammatory SPNs [3].

If we combine the EV≥35 HU and S/A≥0.12, the SPNwas considered malignant. As a result, the sensitivity ratewas 89.5%, specificity rate was 88.2%, positive predictivevalue was 89.5%, negative predictive value was 88.2%, andthe diagnosis accordance rate was 88.9%. These are similarto the former results [2,20].

4.3. Applicable value of SPN pathologically changed CTperfusion TDC

The three types of TDC in this SPN all show that theblood supply of a malignant nodule is related to the aorticartery. In this study, 38 cases of malignant nodules withtype 1 curves demonstrated a synchronized lung cancerenhancement and an aortic artery, and the peak value islower than the enhanced peak value of the aortic artery.Han et al [17] have reported that most of the blood supplyof lung cancer is adequate, compliant with the bloodsupply of the bronchia artery, but the enhanced peak valueof lung cancer is not as high as that of the aortic artery.Sixteen cases of inflammatory nodules with type 2 curvesshow that the peak value of an inflammatory SPN islower than that of a malignant SPN. The enhanced lastingtime is longer than that of the malignant SPN. After sometime, the inflammatory SPN's TDC dropped dramatically.The forms of TDC of inflammatory SPNs and malignantSPNs are different. The benign lung nodule TDC isreflected by a type 3 curve that does not have an obviousenhancement, which is quite different from the curve oflung cancer. Pulmonary tuberculoma is usually type 3,which is similar to what research reveals: most tubercu-loses mostly do not have enhancement. Therefore, thisstudy indicates that CT perfusion may well explore SPN'sblood supply features, which may act as a reference indexfor distinguishing benign and malignant nodules. MostTDCs of benign SPNs were type 3 without obviousenhancement, which was significantly different from thatof malignant nodules. Tuberculoma was mostly type 3,which was close to the report of no enhancement [18–20].Through this study, it was surmised that CT perfusionshows the features of blood pattern well, and these

263S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

features could serve as a reference index to distinguishbenign and malignant SPNs.

4.4. MVD and VEGF expression of SPN and theclinical significance

Growth of tumor will suffer limitation if there isoccurrence of no angiogenesis. Tumor cells obtain nutritiononly when there is diffusion. However, growth of tumor willaccelerate once new vessels are put into tumor tissues, andthe nutrition supply of tumor transforms perfusion fromdiffusion process. Therefore, the tumor is a growth thattypically depends on supply from the blood vessels.Although there are many kinds of vascular growth factors,VEGF played a crucial role in tumor angiogenesis because ofits speciality factor in promoting tumor karyokinesis. MVDis the number of microvessel per unit area which is countedby using some specific antibodies (for instance, CD34monoclonal antibody) to mark microvessel endothelial cellsof tumor tissues.

Angiogenesis is a complicated process, as it is controlledby vascular growth factors and vascular growth inhibitorfactors. This is because the endothelial cells which haveresided in blood capillary and postcapillary vein will enteragain into cell generation cycle. As a result, there will bedegrading basal membrane under endothelia cells, and thiswill form blood capillary spore. In turn, blood capillary willmove towards peripheral tissues by a pullulate way.

A tumor is typically a vessel-dependent disease. Theseneogenetic vessels could result in changes in BV, perfusionamount, and capillary permeability; the active inflammationcould result in changes in vascular lumen size and capillarypermeability, both of which form the basis of CTenhancement. As a result, the quantitative evaluation ofthe BF pattern could help distinguish malignant andactively inflamed SPNs. Vascularization is a complicatedprocess controlled by both vascular growth and vasculargrowth inhibition factors. Although many kinds of tumorvascular growth factors have been determined, VEGF isconsidered to play the most important role in tumorvascularization because of its special emphasis on endo-thelial mitosis.

In this study, the positive VEGF expression of malignantSPNs was significantly higher than that of benign SPNs. TheSPN with VEGF-positive expression showed significantlyhigher BV, PS, BF, MTT, EV, S/A, and TTP than those ofthe SPN with VEGF-negative expression. As a result, BV,PS, BF, MTT, EV, S/A, and TTP could be used as the indexfor VEGF-related vascularization of SPN and could bepresented as having a potential value in the study of SPNvascularization. MSCT perfusion was well related to tumorangiogenesis and could reflect VEGF expression, and thuscan provide a noninvasive method to quantitatively evaluatethe BF pattern and differential diagnosis of an SPN.However, different scanning conditions might influence thechanges of CT perfusion parameters in different degrees, so

the consistency of scanning conditions should be carefullymaintained during the experimental process.

This study indicated that in the case of VEGF-positiveexpression, MVD was found to be positively correlated withBF, BV, and PS of malignant SPNs. MVD was notcorrelated with BF, BV, and PS of the three groups ofSPNs with VEGF-negative expression. The malignant SPNswith VEGF-positive expression showed significantly higherBF, BV, and PS value than those of the malignant SPNs withVEGF-negative expression and those of benign SPNs withVEGF-positive expression. It is clear that the BF, BV, andPS value can reflect MVD measurement and VEGFexpression of benign and malignant SPNs. This explainswhy BF, BV, and PS may act as angiogenesis target ofpulmonary nodules and have potential value of studiedpulmonary nodule angiogenesis.

5. Conclusion

The 64-slice spiral CT perfusion imaging closelycorrelated with tumor angiogenesis and VEGF expression.It provided a noninvasive method of quantitative assessmentfor BF patterns of peripheral pulmonary nodules.

The malignant SPN usually showed significantly higherBV, PS, EV, and S/A values than those of inflammatorynodule and benign nodule. When a pulmonary nodule showsthat BV≥6 ml·100 g−1, PS≥7 ml·100 g−1·min−1, EV≥35HU, and S/A≥0.12, it is possible that before otherpulmonary nodules are found, lung cancer may beconsidered first. On the contrary, inflammatory nodulemay be considered first.

This study demonstrated that using CT perfusion for SPNevaluation not only could show the total enhancementmanifestation of SPNs but also could allow an estimation ofthe histological features of the condition and distribution of avascularization according to BV, PS, BF, MTT, EV, S/A,TTP, and TDC, so as to determine the malignant or benignproperty of the SPNs. CT perfusion could be used as anoninvasive, rapid, and effective examination method.

Acknowledgments

The experiments were performed with the approval of thehuman subjects ethics committee of the second affiliatedhospital of Haerbin medical university.

The lung tissue specimens were obtained in this study byoperation or biopsy. This study was supported by grants fromthe Ethics Committee for the excision/biopsy of all pulmonarynodules and known not to have changed in appearance overmany years to be biopsied or excised. All patients in this studyparticipated voluntarily and agreed to this experiment checks.The feasibility of this experiment was permitted by the EthicsCommittee of Harbin medical university.

The permission was granted by the publisher and authorto reproduce any previously published figures, including

264 S.J. Shu et al. / Clinical Imaging 37 (2013) 256–264

permission to reproduce in both print and electronic formats.The patients gave informed consent for the study.

References

[1] Lee YR, Kim MG, Kang EY, Suh WH. CT evaluation of cavitary lunglesions: focused on lung cancer, tuberculosis and abscess. J KoreanRadiol Soc 1992;28:897–902.

[2] Yi CA, Lee KS, Kim EA, et al. Solitary pulmonary nodules: dynamicenhanced multi-detector row CT study and comparison with vascularendothelial growth factor and microvessel density. Radiology 2004;233:191–9.

[3] Lee YH, Kwona W, Kim MS, Kim YJ, Lee MS, Yong SJ, Jung SH,Chang SJ, Sung KJ. Lung perfusion CT: the differentiation of cavitarymass. Eur J Radiol 2009;4:1–7.

[4] Miles KA. Tumour angiogenesis and its relation to contrastenhancement on computed tomography: a review. Eur J Radiol1999;30:198–205.

[5] Swensen SJ, Viggiano RW, Midthun DE, et al. Lung noduleenhancement at CT: multicenter study. Radiology 2000;214:73–80.

[6] Yamashita K, Matsunobe S, Takahashi R, et al. Small peripheral lungcarcinoma evaluated with incremental dynamic CT: radiologic–pathologic correlation. Radiology 1995;196:401–8.

[7] Kono M, Adachi S, Kusumoto M, Sakai E. Clinical utility of Gd-DTPA-enhanced magnetic resonance imaging in lung cancer. J ThoracImaging 1993;8:18–26.

[8] Hittmair K, Eckersberger F, Klepetko W, Helbich T, Herold CJ.Evaluation of solitary pulmonary nodules with dynamic contrast-enhanced MR imaging: a promising technique. Magn Reson Imaging1995;13:923–33.

[9] Gückel C, Schnabel K, Delmling M, Steinbrich W. Solitary pulmonarynodules: MR evaluation of enhancement patterns with contrast-enhanced dynamic snapshot gradient-echo imaging. Radiology 1996;200:681–6.

[10] Schaefer JF, Schneider V, Vollmar J, et al. Solitary pulmonarynodules: association between signal characteristics in dynamic

contrast enhanced MRI and tumor angiogenesis. Lung Cancer2006;53:39–49.

[11] Xu W, Zhang L, Xie Y. Expression of vascular endothelial growthfactor in primary esophageal squamous cell carcinoma and itssignificance in angiogenesis and prognosis of the tumor. Chin Med J2001;81:860–2.

[12] Zhang M, Kono M. Solitary pulmonary nodules: evaluation of bloodflow patterns with dynamic CT. Radiology 1997;205:471–8.

[13] Yamashita K, Matsunobe S, Tsuda T, Nemoto T, Matsumoto K, MikiH, et al. Solitary pulmonary nodules: preliminary study of evaluationwith incremental dynamic CT. Radiology 1995;194:399–405.

[14] Li SJ, Xiao XS, Li HM, Liu SY, Li CZ, Zhang CS, et al. Evaluation ofblood flow pattern of solitary pulmonary nodules with dynamicenhanced multi-slice spiral CT scanning. J Clin Radiol 2003;22:748–53.

[15] Swensen SJ, Browen LR, Colby TV, Weaver AL. Pulmonary nodules:CT evaluation of enhancement with iodinated contrast material.Radiology 1995;194:393–8.

[16] Bai RJ, Cheng XG, Qu H, Shen BZ, Han MJ, Wu ZH. Solitarypulmonary nodules: comparison of multi-slice computed tomographyperfusion study with vascular endothelial growth factor and micro-vessel density. Chin Med J 2009;122:541–7.

[17] Diederich S, Theegarten D, Stamatis G, Lüthen R. Solitarypulmonary nodule with growth and contrast enhancement at CT:inflammatory pseudotumour as an unusual benign cause. Br J Radiol2006;79:76–8.

[18] Huang Y, Wu N, Lin DM, Li L, Wang JW. The diagnostic value ofdynamic CT in solitary pulmonary nodules: a prospective study. Chin JOncol (Chin) 2005;27:360–3.

[19] Chen H, Wang XH, Ma DQ, Ma BR. Neural network-based computer-aided diagnosis in distinguishing malignant from benign solitarypulmonary nodules by computed tomography. Chin Med J 2007;120:1211–5.

[20] Ma SH, Xu K, Xiao ZW, Wu Ma, Sun ZY, et al. Peripheral lungcancer: relationship between multi-slice spiral CT perfusion imagingand tumor angiogenesis and cyclin D1 expression. Clin Imaging 2007;31:165–77.