- Email: [email protected]
Dual energy spectral CT imaging for the evaluation of small hepatocellular carcinoma microvascular invasion
European Journal of Radiology 95 (2017) 222–227
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Research paper
Dual energy spectral CT imaging for the evaluation of small hepatocellular carcinoma microvascular invasion
MARK
Chuang-bo Yanga, Shuang Zhangb, Yong-jun Jiaa, Yong Yua, Hai-feng Duana, Xi-rong Zhanga, ⁎ Guang-ming Maa, Chenglong Rena, Nan Yua, a b
Departments of Diagnostic Radiology, Affiliated Hospital of Shaanxi University of Chinese Medicine, China Department of Basic Chemicals, School of Medicine, Shaanxi Institute of International Trade & Commerce, Xi'an, 712046, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Spectral CT Contrast-enhanced tomography Differentiation Hepatocellular carcinoma Microvascular invasion
Objective: To study the clinical value of dual-energy spectral CT in the quantitative assessment of microvascular invasion of small hepatocellular carcinoma. Methods: This study was approved by our ethics committee. 50 patients with small hepatocellular carcinoma who underwent contrast enhanced spectral CT in arterial phase (AP) and portal venous phase (VP) were enrolled. Tumour CT value and iodine concentration (IC) were measured from spectral CT images. The slope of spectral curve, normalized iodine concentration (NIC, to abdominal aorta) and ratio of IC difference between AP and VP (RICAP–VP: [RICAP–VP = (ICAP−ICVP)/ICAP]) were calculated. Tumours were identified as either with or without microvascular invasion based on pathological results. Measurements were statistically compared using independent samples t test. The receiver operating characteristic (ROC) analysis was used to evaluate the diagnostic performance of tumours microvascular invasion assessment. The 70 keV images were used to simulate the results of conventional CT scans for comparison. Results: 56 small hepatocellular carcinomas were detected with 37 lesions (Group A) with microvascular invasion and 19 (Group B) without. There were significant differences in IC, NIC and slope in AP and RICAP–VP between Group A (2.48 ± 0.70 mg/ml, 0.23 ± 0.05, 3.39 ± 1.01 and 0.28 ± 0.16) and Group B (1.65 ± 0.47 mg/ml, 0.15 ± 0.05, 2.22 ± 0.64 and 0.03 ± 0.24) (all p < 0.05). Using 0.188 as the threshold for NIC, one could obtain an area-under-curve (AUC) of 0.87 in ROC to differentiate between tumours with and without microvascular invasion. AUC was 0.71 with CT value at 70 keV and improved to 0.81 at 40 keV. Conclusion: Dual-energy Spectral CT provides additional quantitative parameters than conventional CT to improve the differentiation between small hepatocellular carcinoma with and without microvascular invasion. Clinical Application/Relevance: Quantitative iodine concentration measurement in spectral CT may be used to provide a new method to improve the evaluation for small hepatocellular carcinoma microvascular invasion.
1. Introduction Hepatocellular carcinoma is the most common primary malignant tumour in the liver; accounting for 19.33% of all malignant liver tumours [1]. With the improvement in imaging techniques, more and more small hepatocellular carcinomas are discovered early and treated in time. Currently, the radical resection is still the most effective way to treat small hepatocellular carcinoma, but the prognosis and long term effect varies dependent on patients. Roayaie [2] reported that up to 70% patients with small hepatocellular carcinoma would be in relapse at 5 years after hepatic resection. The main reason for liver cancer recurrence and poor prognosis is due to portal vein and tumour
⁎
Corresponding author at: Weiyang western road -2#, Xianyang, 712000, China. E-mail address: [email protected] (N. Yu).
http://dx.doi.org/10.1016/j.ejrad.2017.08.022 Received 4 May 2017; Received in revised form 19 July 2017; Accepted 22 August 2017 0720-048X/ © 2017 Elsevier B.V. All rights reserved.
microvascular invasion [3]. Researchers have shown that tumours with positive microvascular invasion can more easily invade adjacent vessels and lymphatics, cause early dissemination and metastasis, and have greater influence on prognosis. So, accurate preoperative evaluation of tumour microvascular invasion is important for the treatment plan and the prognosis. Many scholars have used the hepatic artery CT angiography, dynamically enhanced MRI and superparamagnetic contrast enhanced MRI in the preoperative evaluation of hepatocellular carcinoma microvascular invasion [4,5]. Dual-energy CT performed during two consecutive scans or with a dual x-ray source, dual-detector can provide additional information that is used for material separation at imaging [6–8]. Spectral imaging
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
2.2. Imaging methods
is also reported as an excellent qualitative as well as a quantitative tool for assessing and predicting hepatocellular carcinoma in cirrhotic patients [9]. The spectral CT is based on the principle that this technique involves scanning at distinctly different energies (most commonly used energy levels are 80 and 140 kVp). This works best for post contrast CT scans due to the use of Iodine, which has a K-edge of 33.2 keV. The higher attenuation of Iodine suggests hypervascular (Iodine rich) nodules in contrast to rest of the liver parenchyma. Therefore, we tried to applicate dual-energy spectral CT for evaluating the primary hepatic carcinoma microvascular invasion. The purpose of this study was to study the clinical value of dual-energy spectral CT in the quantitative assessment of microvascular invasion of primary small hepatocellular carcinoma.
All patients underwent the non-contrast and two-phase contrast enhanced CT scans on a Discovery CT750HD scanner (GE Healthcare, Waukesha WI USA) using the dual-energy spectral CT acquisition mode. Patients fasted for 4 h and took 800–1000ML warm water orally 5–10 min before the scanning. The scan protocol included 5 mm slice thickness, 600 mA tube current, 0.5s/rot gantry rotation speed, and 0.984:1 helical pitch. The nonionic contrast agent Ioversol (300mgI/ ml) (China, Jiangsu Hengrui medicine) was injected through the median cubital vein with a German ORICH high pressure syringe at a patient weight-dependent dose of 0.8–1.0 ml/kg and infusion rate of 4.0–5.0 ml/s. The contrast injection was followed by 40 ml physiological saline at the same injection rate. The arterial phase (AP) scan started 20–25 s after the contrast injection, and the portal vein phase (PV) scan started 30 s after AP scanning. The volumetric CT dose index (CTDIvol) was 21.8 mGy (which is comparable to the 21.5-mGy dose for the conventional contrast enhanced liver scanning for a normal size patient in our institution). Images were reconstructed at 1.25 mm slice thickness using filtered back-projection and standard reconstruction kernel and dual-energy spectral CT-specific software to generate both virtual monochromatic image sets with energy levels from 40 to 140 keV and material decomposition images with water and iodine as the basis material pair.
2. Materials and methods 2.1. General information This study was approved by our ethics committee, and patient consent was waived. We retrospectively analyzed the imaging information of 50 patients who came to our hospital between July 2013 and March 2015 for the diagnosis and treatment of primary small hepatocellular carcinoma. All patients had confirmed results of hepatocellular carcinoma by pathology. There were 36 male patients and 14 female patients with age ranged from 38 to 77 years and median age of 49 years. 34 cases had reported history of cirrhosis. All patients underwent the contrast-enhanced dual-energy spectral CT for definitive diagnosis of primary small hepatocellular carcinoma before surgical operation were enrolled in the study, if they meet the following criteria: (1) the maximum diameter of single tumour was under 3 cm; or (2) in the case of multiple tumours, the numbers were no more than two, and the maximum total diameter was below 3 cm, and there was no obvious signs of metastasis; (3) 1–3 weeks after CT examination the tumours were resected in our hospital with complete tumour resection and negative margins, and no metastasis findings in the 1 month follow up imaging; (4) imaging revealed lesions in the liver, but no obvious portal vein tumour thrombus; (5) there was no treatment for cancers before the operation. Patients with severe image artifacts were excluded. Patient information is summarized in Table 1.
2.3. Observation and analysis method 2.3.1. Image analysis Images were transferred to a GE AW4.6 Workstation equipped with Gemstone Spectral Imaging (GSI) Viewer software for post-processing and analysis. Region-of-interest (ROI) with diameter of about half of the tumour size was placed on tumour to measure the iodine concentration (IC) on the iodine-based material decomposition images and the CT value on the 101 sets of virtual monochromatic images (corresponding to photon energies from 40keV–140 keV at 1 keV interval) in both AP and VP. The copy and paste function was used to ensure measurement consistency between the two phases. To reduce measurement variation, ROI was placed 3 times on the tumour and the average of the 3 measurements was used as the final result. The iodine concentration measurement was repeated on the abdominal aorta at the same imaging level for generating normalized iodine concentration (NIC) for tumours: NIC(tumour) = IC(tumour)/IC(aorta). The ratio of the iodine concentration difference between AP and VP (RICAP–VP: [RICAP–VP = (ICAP−ICVP)/ICAP]) were calculated. The CT value measurements from the virtual monochromatic images from 40 keV to 90 keV were used to calculate the slope (λHU) for spectral HU curve (CT value as a function of virtual monochromatic energy level): λHU = [(CT(40keV)–CT(90keV)]/50. Since the 70 keV virtual monochromatic images in spectral CT have similar muscle CT value to the conventional 120kVp images in abdominal imaging, the 70 keV virtual monochromatic images in our study were used to simulate the results of conventional CT scans for comparison.
Table 1 Patient information between the two pathological groups of small hepatocellular carcinoma in Spectral CT. Parameter
Group Tumours with microvascular invasion (n = 37)
Gender (Number of patients) Male (n = 36) 21/36 Female (n = 14) 8/14 Age 53 ± 26.3 Location (Number of lesions) Left lobe of liver 16/24 Right lobe of liver 21/32 History of cirrhosis 20/34 (n = 34) Degree of differentiation well-differentiated and moderatedifferentiated HCC (n = 22) poorly differentiated HCC (n = 34)
p-value Tumours without microvascular invasion (n = 19)
15/36 6/14 55 ± 17.2
0.939 0.746
8/24 11/32 14/34
0.935 0.863
8/22
14/22
29/34
5/34
2.3.2. Pathological analysis The tissues for pathological analysis were obtained from 50 patients with primary small hepatocellular carcinoma after surgical operation. The whole process followed the protocol edited by the Department of Health of China (“Diagnosis and treatment of primary liver cancer”, edition 2011), including the location and the number for obtaining specimens. The specimen treatment and pathological analysis were performed by two experienced pathologists. The specimen treatment included formaldehyde fixation and paraffin fixation followed by section cutting with the thickness of 4 mm and conventional H–E staining. According to the Hyung standard [10], the presence of microvascular invasion was defined as satisfying the following situations after HE staining: (1) the formation of tumour thrombus were found in central
0.000*
*p < 0.05.
223
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Fig. 1. A 50-year-old male with poorly differentiated hepatocellular carcinoma. (1A) A enhanced lesion was found in the right liver lobe in the arterial phase (red arrow); (1B) the rich blood supply lesion was still found on the iodine-based material decomposition image (red arrow); (1C) the density of the lesion decreased in portal venous phase (red arrow); and (1D) the isodensity lesion was also found on the iodine-based material decomposition image (red arrow); (1E) the formation of tumour thrombus was found in central veins and small veins (blue arrows). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
statistically analyzed using the two-sample t-test. A p-value < 0.05 was considered statistically significant. Receiver operating characteristic (ROC) analysis was carried out to establish threshold values and to calculate the area-under-curve (AUC) for ROC curves, sensitivity and specificity for differentiating small hepatocellular carcinoma with and without microvascular invasion.
vein, portal vein branch or small vein under liver capsule; (2) the tumour cells were found at vascular endothelium and vascular smooth muscle; or (3) fibrin clots or red blood cells were found around the tumour cells in the vascular lumen. The histological grades were also assessed. And the pathological findings were categorized into welldifferentiated HCC, moderate-differentiated HCC and poorly differentiated HCC.
3. Results 2.4. Statistical analysis The enhanced spectral CT scans detected 56 small hepatocellular carcinomas in 50 patients. There were 37 lesions with microvascular invasion (37/56, 66.1%, Group A) (Fig. 1) and 19 lesions without microvascular invasion (19/56, 33.9%, Group B) (Fig. 2). There were no statistically significant differences in age, gender between the two groups (p > 0.05). CT imaging related measurements are listed in Table 2. As indicated in Table 2, the iodine concentration (IC), normalized iodine concentration (NIC), slope (λHU) of the spectral HU curve in the arterial phase and ratio of the iodine concentration
Statistical analysis was performed using SPSS17.0 (SPSS Inc., Chicago, IL) and all measurements were expressed as mean ± standard deviation (SD). The differences in iodine concentration (IC), normalized iodine concentration (NIC), slope of spectral HU curve (λHU), ratio of the iodine concentration difference between AP and VP (RICAP–VP), and CT value at 70 keV and 40 keV virtual monochromatic images between Group A (small hepatocellular carcinoma with microvascular invasion) and Group B (without microvascular invasion) were 224
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Fig. 2. A 64-years-old male with well differentiated hepatocellular carcinoma. (2A) A slightly high density lesion was found in the right liver lobe in the arterial phase (red arrow); (2B) the rich blood supply lesion was still found on the iodine-based material decomposition image (red arrow); (2C) the density of the lesion decreased in portal venous phase (red arrow), showing as an isodensity; and (2D) the lesion was also found on the iodine-based material decomposition image (red arrow) with the similar iodine concentration as liver; (2E) the intact endothelium can be seen under HE staining (red arrows). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
previous studies, patients with signs of microvascular invasion may have higher probability of tumour recurrence and metastasis [2,11]. Regular CT is helpful in detecting the vascular invasions to larger blood vessels in hepatocellular carcinoma, including portal vein and its branches, hepatic vein and hepatic artery. However, microscopic vascular invasions such as intratumoural microvascular invasion cannot be observed directly. Hepatocellular carcinoma is mainly supplied by hepatic artery. The hepatic lobule central veins were easily compressed by tumours due to the lack of connective tissue sheath resulting in blood flow being blocked. On the other hand, with the proliferation of cancer cells, the infiltration of surrounding matrix and vascular endothelium were induced, as well as the formation of micro tumour thrombus and the cancer metastasis through portal vein system. All the above processes are consistent with pathologic findings. Therefore, the prediction of microvascular invasion is particularly important for those patients who underwent the hepatocellular carcinoma resection surgery [12]. Previous studies reported that tumour microvascular invasion was correlated with multiple factors including tumour size, tumour number, pathologic type, pathologic stage and level of biomarkers and gene expression [13–16]. Imaging examination may also have certain value in the evaluation of tumour microvascular invasion, including contrast
difference between AP and VP (RICAP–VP) in Group A were significantly higher than those in Group B (p < 0.05). The CT values measured on the images at 70keV, which simulating the energy level of a conventional 120kVp imaging, were statistically the same between tumours with and without microvascular invasion (p > 0.05). However, CT values between the two types of tumours in the arterial phase became significantly different on images with energy levels from 40 keV to 60 keV (p < 0.05). ROC analysis results are listed in Table 3. The CT value measurement at 70 keV had low AUC value at 0.71, while CT value at 40 keV improved it to 0.81. The dual-energy spectral CT-specific parameters IC, NIC, λHU and RICAP–VP all generated higher areaunder-curve values for the ROC study than using the conventional CT value measurement (Fig. 3). Using 0.188 as the threshold for NIC, one could obtain an AUC of 0.87, sensitivity of 81.3% and specificity of 79.6% for differentiating tumours with and without microvascular invasion. 4. Discussion Microscopic vascular invasion is an important risk factor for poor prognosis of patients with hepatocellular carcinoma. According to 225
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Table 2 Quantitative assessment between the two pathological groups of small hepatocellular carcinoma in Spectral CT (mean ± SD). Parameters
Group
t-value
p-value
Tumours with microvascular invasion (n = 37)
Tumours without microvascular invasion (n = 19)
Iodine Concentration (mg/ml) Arterial Phase Venous Phase
2.48 ± 0.70 1.73 ± 0.47
1.65 ± 0.47 1.54 ± 0.35
3.605 1.248
0.001 0.222
Normalized Iodine Concentration Arterial Phase Venous Phase
0.23 ± 0.05 0.50 ± 0.09
0.15 ± 0.05 0.45 ± 0.09
4.481 1.581
< 0.001 0.125
Slope (k) of Spectral HU Curve Arterial Phase Venous Phase IC Reduction Rate (AP-VP)
3.39 ± 1.01 2.33 ± 0.63 0.28 ± 0.16
2.22 ± 0.64 2.15 ± 0.48 0.03 ± 0.24
3.721 0.867 3.415
0.001 0.393 0.002
CT value at 40 keV (HU) Arterial Phase Venous Phase
246.16 ± 64.72 177.14 ± 44.44
177.25 ± 42.66 168.39 ± 34.12
3.388 0.598
0.002 0.555
CT value at 70 keV (HU) Arterial Phase Venous Phase
100.97 ± 16.20 81.05 ± 18.67
89.16 ± 16.83 79.11 ± 15.19
1.956 0.309
0.061 0.760
Table 3 Thresholds, sensitivities and specificities for differentiating small hepatocellular carcinoma with and without microvascular invasion. Parameter
Threshold
AUC
Sensitivity
Specificity
Youden index
Iodine Concentration at AP (mg/ml) Normalized Iodine Concentration at AP IC Reduction Rate (AP-VP) Slope (k) of Spectral HU Curve CT number at 40 keV (HU) CT number at 70 keV (HU)
1.906 0.188 0.231 2.691 194.73 93.58
0.853 0.871 0.826 0.871 0.813 0.714
87.5 81.3 75.0 93.8 81.3 75.0
79.6 79.6 79.6 79.6 72.4 64.3
0.671 0.609 0.546 0.734 0.537 0.393
imaging has made this imaging modality more relevant in diagnosing diseases such as tumours. Because the absorption coefficients of substances vary from one to another in different photon energy levels, dual-energy spectral CT can be used to more effectively distinguish them based on the absorption curve. Dual-energy spectral CT also generates material decomposition images to enable concentration measurements for base materials such as iodine, water, fat and calcium. Thus, dual-energy spectral CT provides more parameters than the CT value in conventional CT, which may be helpful in clinical decision [20–24]. Iodine and water are the main basic substances that can be separated by using absorption curve. As the main ingredient of contrast medium, iodine concentration in the contrast-enhanced CT scans may reflect the blood flow and the iodine concentration change may be used to indicate the change of microcirculation. Our results indicated that the Spectral CT-specific parameters, i.e. iodine concentration (IC), normalized iodine concentration (NIC), slope of spectral curve, ratio of iodine content difference between AP and VP (RICAP–VP) as well as the CT value in 40–60 keV in the arterial phase in the group with microvascular invasion (group A) were significantly higher than those in group B without microvascular invasion (P < 0.05). However, these parameters in the venous phase showed no significant difference (P > 0.05) between group A and group B. The higher value of RICAP–VP suggested a faster clearance speed, and less contrast agent remaining in lesions in group A. Recent studies reported the evolution process from cirrhosis nodules to small hepatocellular carcinoma, and suggested the series of changes as a process of multistep carcinogenesis [25]. In this process, the formation of regenerative nodules as well as neoangiogenesis are found, accompanied by the reduction of portal vein and hepatic artery blood supply to nodules. The tumour angiogenesis is shown as immature development, disordered arrangement, increased permeability of incomplete endothelium. At the same time, the invasion of small vessels
Fig. 3. ROC curves to evaluate the diagnostic efficacy of dual energy spectral CT parameters in differentiating small hepatocellular carcinoma with and without microvascular invasion.
enhanced CT, MRI and PET-CT [17–19]. However, there are limited study on the prediction of microvascular invasion by using dual-energy spectral CT. Therefore, in this study, we assessed the capacity of spectral CT in predicting microvascular invasion in patients with hepatocellular carcinoma. The introduction of dual-energy CT changed the diagnostic mode in comparison with the conventional CT which only has single parameter: CT value. Even though in the early days, the clinical use of dual-energy CT has been hampered because of motion mis-registration, high image noise, and excessive radiation exposure caused by long acquisition times [8]. The technologic improvement in dual-energy spectral CT 226
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
[2] S. Roayaie, I.N. Blume, S.N. Thung, et al., A system of classifying microvascular invasion to predict outcome after resection in patients with hepatocellular carcinoma, Gastroenterology 137 (3) (2009) 850–855. [3] S.A. Shah, S.P. Cleary, A.C. Wei, et al., Recurrence after liver resection for hepatocellular carcinoma: risk factors, treatment, and outcomes, Surgery 141 (3) (2007) 330–339. [4] H. Kim, M.S. Park, J.Y. Choi, et al., Can microvascular invasion of hepatocellular carcinoma be predicted by pre-operative MRI? Eur. Radiol. 19 (7) (2009) 1744–1751. [5] A. Nishie, K. Yoshimitsu, Y. Asayama, et al., Radiologic detectability of mimute portal venous invasion in hepatocellular carcinoma, AJR 190 (1) (2008) 81–87. [6] J.M. Lewin, P.K. Isaacs, V. Vance, et al., Dual-energy contrast-enhanced digital subtraction mammography: feasibility, Radiology 229 (1) (2003) 261–268. [7] D.R. Holmes 3rd, J.G. Fletcher, A. Apel, et al., Evaluation of non-linear blending in dual-energy computed tomography, Eur. J. Radiol. 68 (8) (2008) 409–413. [8] A. Graser, T.R. Johnson, E.M. Hecht, et al., Dual energy CT in patients suspected of having renal masses: can virtual nonenhanced images replace true nonenhanced images? Radiology 252 (2) (2009) 433–440. [9] S.T. Laroia, A.S. Bhadoria, Y. Venigalla, et al., Role of dual energy spectral computed tomography in characterization of hepatocellular carcinoma: initial experience from a tertiary liver care institute, Eur. J. Radiol. 3 (2016) 162–171. [10] W.J. Hyung, J.H. Lee, S.H. Choi, et al., Prognostic impact of lymphatic and/or blood vessel invasion in patients with node-negative advanced gastric cancer, Ann. Surg. Oncol. 9 (6) (2002) 562–567. [11] P.P. McHugh, J. Gilbert, S. Vera, et al., Alpha-fetoprotein and tumour size are associated with microvascular invasion in expiated livers of patients undergoing transplantation with hepatocellular carcinoma, HPB (Oxford) 12 (1) (2010) 56–61. [12] S. Sumie, R. Kuromatsu, K. Okuda, et al., Microvascular invasion in patients with hepatocellular carcinoma and its predictable clinicopathological factors, Ann. Surg. Oncol. 15 (5) (2008) 1375–1382. [13] S. Equchi, M. Takatsuki, M. Hidaka, et al., Predictor for histological microvascular invasion of hepatocellular carcinoma: a lesson form 229 consecutive cases of curative liver resection, World J. Surg. 34 (5) (2010) 1034–1038. [14] M. Kaibori, M. Ishizaki, K. Matsui, et al., Predictors of microvascular invasion before hepatectomy for hepatocellular carcinoma, J. Surg. Oncol. 102 (5) (2010) 462–468. [15] S.X. Yuan, F. Yang, Y. Yang, et al., Long noncoding RNA associated with microvascular invasion in hepatocellular carcinoma promotes angiogenesis and serves as a predictor for hepatocellular carcinoma patients poor recurrence-free survival after hepatectomy, Hepatology 56 (6) (2012) 2231–2241. [16] J.H. Shim, Y.H. Chung, J.A. Kim, et al., Clinical implications of arrest-defective protein 1 expression in hepatocellular carcinoma: a novel predictor of microvascular invasion, Dig. Dis. 30 (6) (2012) 603–608. [17] C.D. Witjes, F.E. Willemssen, J. Verheij, et al., Histological differentiation grade and microvascular invasion of hepatocellular carcinoma predicted by dynamic contrastenhanced MRI, J. Magn. Reson. Imaging 36 (3) (2012) 641–647. [18] C.T. Chou, R.C. Chen, C.W. Lee, et al., Prediction of microvascular invasion of hepatocellular carcinoma by pre-operative CT imaging, Br. J. Radiol. 85 (1014) (2012) 778–783. [19] T.T. Cheung, S.C. Chan, C.L. Ho, et al., Can positron emission tomography with the dual tracers [11C] acetate and [18F] fludeoxyglucose predict microvascular invasion inhepatocellular carcinoma? Liver Transpl. 17 (10) (2011) 1218–1225. [20] N. Takahashi, T.J. Vrtiska, A. Kawashima, et al., Detectability of urinary stones on virtual nonenhanced images generated at pyelographic-phase dual-energy CT, Radiology 256 (1) (2010) 184–190. [21] G. Pache, B. Krauss, P. Strohm, et al., Dual-energy CT virtual noncalcium technique: detecting posttraumatic bone marrow lesions-feasibility study, Radiology 256 (2) (2010) 617–624. [22] J.D. Linder, J.E. Geenen, M.F. Catalano, Cyst fluid analysis obtained by EUS-guided FNA in the evaluation of discrete cystic neoplasms of the pancreas: a prospective single-center experience, Gastrointest. Endosc. 64 (5) (2006) 697–702. [23] X.Z. Lin, F. Miao, J.Y. Li, et al., High-definition CT gemstone spectral imaging of the brain: initial results of selecting optimal monochromatic image for beam-hardening artifacts and image noise reduction, J. Comput. Assist. Tomogr. 35 (2) (2011) 294–297. [24] D. Zhang, X. Li, B. Liu, Objective characterization of GE discovery CT750 scanner: gemstone spectral imaging mode, Med. Phys. 38 (3) (2011) 1178–1188. [25] M. Kobayashi, K. Ikeda, T. Hosaka, et al., Dysplastic nodules frequently develop into hepatocellular carcinoma in patients with chronic viral hepatitis and cirrhosis, Cancer 106 (3) (2006) 636–647. [26] M. Kudo, Multistep human hepatocarcinogenesis: correlation of imaging with pathology, J. Gastroenterol. 44 (2009) 112–118. [27] M. Roncalli, E. Roz, G. Coggi, et al., The vascular profile of regenerative and dysplastic nodules of the cirrhotic liver: implications for diagnosis and classification, Hepatology 30 (5) (1999) 1174–1178. [28] O. Matsui, M. Kadoya, T. Kameyama, et al., Benign and malignant nodules in cirrhotic livers: distinction based on blood supply, Radiology 178 (2) (1991) 493–497. [29] T.R. Johnson, B. Krauss, M. Sedlmair, et al., Material differentiation by dual energy CT: initial experience, Eur. Radiol. 17 (6) (2007) 1510–1517. [30] R. Iannaccone, A. Laghi, Passariello R. Multislice, CT angiography of mesenteric vessels, Abdom. Imaging 29 (2) (2004) 146–152. [31] J.R. Du, Y. Jiang, Y.M. Zhang, et al., Vascular endothelial growth factor and microvascular density in esophageal and gastric carcinomas, World J. Gastroenterol. 9 (7) (2003) 1604–1606. [32] K. Matsumoto, M. Jinzaki, Y. Tanami, et al., Virtual monochromatic spectral imaging with fast kilovoltage switching: improved image quality as compared with that obtained with conventional 120-kVp CT, Radiology 259 (1) (2011) 257–262. [33] M. Kudo, Multistep human hepatocarcinogenesis: correlation of imaging with pathology, J. Gastroenterol. 44 (2009) 112–118.
and portal vein branches induced the formation of a large number of arterial-portal shunting (A–P shunt) and the arterial supply gradually increased [26]. Roncalli [27] study showed that tumour microvessel count (MVC) can help in the diagnosis of dysplastic nodules in cirrhotic liver and small hepatocellular carcinoma at early stage by detecting the significant changes of microcirculation from dysplastic nodules to hepatocellular carcinoma. In our study, most of the patients enrolled had hepatocellular carcinoma which is the progression of chronic liver disease (34/50). During the process, the portal venous blood supply decreased from 68% to 6%, while the arterial blood supply increased from 4% to 94% [28]. These blood supply changes may be detected very sensitively by measuring the iodine concentration change in spectral CT [29]. The change of iodine content can reflect the change of tumour microcirculation indirectly, and the change of tumour microcirculation correlates with the vascular invasion. In the arterial phase, the iodine content reflects the capillary density and capillary permeability, and in the venous phase, the iodine content reflects the clearance and retention of contrast agent. Previous study has found that [30] nutrients uptake in tumour is mainly through the tumour angiogenesis and host vessels invaded by tumour. The higher malignancy of the tumour, the more abundant of tumour angiogenesis [31]. The tumour angiogenesis with incomplete endothelial cell structure leads to an increased vascular permeability, and therefore, increased iodine concentration. In our study the lesions with microvascular invasion had high value of iodine concentration, which is consistent with previous results. And it confirmed the feasibility and accuracy of dual-energy spectral CT in the assessment of microvascular invasion of hepatocellular carcinoma. Our results also indicated that the CT value at 70 keV monochromatic images had low efficacy in the diagnosis of microvascular invasion of hepatocellular carcinoma. Since the conventional 120kVp polychromatic x-ray has an average energy of about 70 keV after penetrating the abdomen of an average size adult patient [32], CT values for 70 keV and 120kVp images are similar. Our results obtained using the 70 keV CT value are indicative and consistent with the fact that conventional 120kVp CT imaging does not provide satisfactory diagnosis of microvascular invasion of hepatocellular carcinoma. On the other hand, with the increased conspicuity, lower keV images (i.e. at 40keV) improved the area-under-curve in ROC study from 0.71 with CT value at 70 keV to 0.81 with CT value at 40keV. This research still has some limitations: firstly, the microvascular invasion was confirmed by H–E staining rather than immuno-histochemical study; secondly, only hepatocellular carcinoma was included in this study, whereas focal nodular hyperplasias, hepatic adenomas or metastatic tumors were excluded. We therefore have limited experience in the diagnosis of microvascular invasion in patients with these diseases. Thirdly, our study were based on the specific patient population with HCC, so we cannot generalize the conclusion to other liver lesions with similar enhancement pattern as HCC. According to previous study [33], it is difficult in diagnosing HCC by only assessing blood supply by using dual energy spectral CT. Future investigations with more cases of different hypervascular tumors are needed to make broader conclusions. In conclusion, dual-energy Spectral CT provides additional quantitative parameters than conventional CT to improve the differentiation between small hepatocellular carcinoma with and without microvascular invasion. Conflict of interest The authors have no conflict of interest. References [1] J.G. Chen, S.W. Zhang, Liver cancer epidemic in China: past, present and future, Semin. Cancer Biol. 21 (1) (2011) 59–69.
227
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Research paper
Dual energy spectral CT imaging for the evaluation of small hepatocellular carcinoma microvascular invasion
MARK
Chuang-bo Yanga, Shuang Zhangb, Yong-jun Jiaa, Yong Yua, Hai-feng Duana, Xi-rong Zhanga, ⁎ Guang-ming Maa, Chenglong Rena, Nan Yua, a b
Departments of Diagnostic Radiology, Affiliated Hospital of Shaanxi University of Chinese Medicine, China Department of Basic Chemicals, School of Medicine, Shaanxi Institute of International Trade & Commerce, Xi'an, 712046, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Spectral CT Contrast-enhanced tomography Differentiation Hepatocellular carcinoma Microvascular invasion
Objective: To study the clinical value of dual-energy spectral CT in the quantitative assessment of microvascular invasion of small hepatocellular carcinoma. Methods: This study was approved by our ethics committee. 50 patients with small hepatocellular carcinoma who underwent contrast enhanced spectral CT in arterial phase (AP) and portal venous phase (VP) were enrolled. Tumour CT value and iodine concentration (IC) were measured from spectral CT images. The slope of spectral curve, normalized iodine concentration (NIC, to abdominal aorta) and ratio of IC difference between AP and VP (RICAP–VP: [RICAP–VP = (ICAP−ICVP)/ICAP]) were calculated. Tumours were identified as either with or without microvascular invasion based on pathological results. Measurements were statistically compared using independent samples t test. The receiver operating characteristic (ROC) analysis was used to evaluate the diagnostic performance of tumours microvascular invasion assessment. The 70 keV images were used to simulate the results of conventional CT scans for comparison. Results: 56 small hepatocellular carcinomas were detected with 37 lesions (Group A) with microvascular invasion and 19 (Group B) without. There were significant differences in IC, NIC and slope in AP and RICAP–VP between Group A (2.48 ± 0.70 mg/ml, 0.23 ± 0.05, 3.39 ± 1.01 and 0.28 ± 0.16) and Group B (1.65 ± 0.47 mg/ml, 0.15 ± 0.05, 2.22 ± 0.64 and 0.03 ± 0.24) (all p < 0.05). Using 0.188 as the threshold for NIC, one could obtain an area-under-curve (AUC) of 0.87 in ROC to differentiate between tumours with and without microvascular invasion. AUC was 0.71 with CT value at 70 keV and improved to 0.81 at 40 keV. Conclusion: Dual-energy Spectral CT provides additional quantitative parameters than conventional CT to improve the differentiation between small hepatocellular carcinoma with and without microvascular invasion. Clinical Application/Relevance: Quantitative iodine concentration measurement in spectral CT may be used to provide a new method to improve the evaluation for small hepatocellular carcinoma microvascular invasion.
1. Introduction Hepatocellular carcinoma is the most common primary malignant tumour in the liver; accounting for 19.33% of all malignant liver tumours [1]. With the improvement in imaging techniques, more and more small hepatocellular carcinomas are discovered early and treated in time. Currently, the radical resection is still the most effective way to treat small hepatocellular carcinoma, but the prognosis and long term effect varies dependent on patients. Roayaie [2] reported that up to 70% patients with small hepatocellular carcinoma would be in relapse at 5 years after hepatic resection. The main reason for liver cancer recurrence and poor prognosis is due to portal vein and tumour
⁎
Corresponding author at: Weiyang western road -2#, Xianyang, 712000, China. E-mail address: [email protected] (N. Yu).
http://dx.doi.org/10.1016/j.ejrad.2017.08.022 Received 4 May 2017; Received in revised form 19 July 2017; Accepted 22 August 2017 0720-048X/ © 2017 Elsevier B.V. All rights reserved.
microvascular invasion [3]. Researchers have shown that tumours with positive microvascular invasion can more easily invade adjacent vessels and lymphatics, cause early dissemination and metastasis, and have greater influence on prognosis. So, accurate preoperative evaluation of tumour microvascular invasion is important for the treatment plan and the prognosis. Many scholars have used the hepatic artery CT angiography, dynamically enhanced MRI and superparamagnetic contrast enhanced MRI in the preoperative evaluation of hepatocellular carcinoma microvascular invasion [4,5]. Dual-energy CT performed during two consecutive scans or with a dual x-ray source, dual-detector can provide additional information that is used for material separation at imaging [6–8]. Spectral imaging
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
2.2. Imaging methods
is also reported as an excellent qualitative as well as a quantitative tool for assessing and predicting hepatocellular carcinoma in cirrhotic patients [9]. The spectral CT is based on the principle that this technique involves scanning at distinctly different energies (most commonly used energy levels are 80 and 140 kVp). This works best for post contrast CT scans due to the use of Iodine, which has a K-edge of 33.2 keV. The higher attenuation of Iodine suggests hypervascular (Iodine rich) nodules in contrast to rest of the liver parenchyma. Therefore, we tried to applicate dual-energy spectral CT for evaluating the primary hepatic carcinoma microvascular invasion. The purpose of this study was to study the clinical value of dual-energy spectral CT in the quantitative assessment of microvascular invasion of primary small hepatocellular carcinoma.
All patients underwent the non-contrast and two-phase contrast enhanced CT scans on a Discovery CT750HD scanner (GE Healthcare, Waukesha WI USA) using the dual-energy spectral CT acquisition mode. Patients fasted for 4 h and took 800–1000ML warm water orally 5–10 min before the scanning. The scan protocol included 5 mm slice thickness, 600 mA tube current, 0.5s/rot gantry rotation speed, and 0.984:1 helical pitch. The nonionic contrast agent Ioversol (300mgI/ ml) (China, Jiangsu Hengrui medicine) was injected through the median cubital vein with a German ORICH high pressure syringe at a patient weight-dependent dose of 0.8–1.0 ml/kg and infusion rate of 4.0–5.0 ml/s. The contrast injection was followed by 40 ml physiological saline at the same injection rate. The arterial phase (AP) scan started 20–25 s after the contrast injection, and the portal vein phase (PV) scan started 30 s after AP scanning. The volumetric CT dose index (CTDIvol) was 21.8 mGy (which is comparable to the 21.5-mGy dose for the conventional contrast enhanced liver scanning for a normal size patient in our institution). Images were reconstructed at 1.25 mm slice thickness using filtered back-projection and standard reconstruction kernel and dual-energy spectral CT-specific software to generate both virtual monochromatic image sets with energy levels from 40 to 140 keV and material decomposition images with water and iodine as the basis material pair.
2. Materials and methods 2.1. General information This study was approved by our ethics committee, and patient consent was waived. We retrospectively analyzed the imaging information of 50 patients who came to our hospital between July 2013 and March 2015 for the diagnosis and treatment of primary small hepatocellular carcinoma. All patients had confirmed results of hepatocellular carcinoma by pathology. There were 36 male patients and 14 female patients with age ranged from 38 to 77 years and median age of 49 years. 34 cases had reported history of cirrhosis. All patients underwent the contrast-enhanced dual-energy spectral CT for definitive diagnosis of primary small hepatocellular carcinoma before surgical operation were enrolled in the study, if they meet the following criteria: (1) the maximum diameter of single tumour was under 3 cm; or (2) in the case of multiple tumours, the numbers were no more than two, and the maximum total diameter was below 3 cm, and there was no obvious signs of metastasis; (3) 1–3 weeks after CT examination the tumours were resected in our hospital with complete tumour resection and negative margins, and no metastasis findings in the 1 month follow up imaging; (4) imaging revealed lesions in the liver, but no obvious portal vein tumour thrombus; (5) there was no treatment for cancers before the operation. Patients with severe image artifacts were excluded. Patient information is summarized in Table 1.
2.3. Observation and analysis method 2.3.1. Image analysis Images were transferred to a GE AW4.6 Workstation equipped with Gemstone Spectral Imaging (GSI) Viewer software for post-processing and analysis. Region-of-interest (ROI) with diameter of about half of the tumour size was placed on tumour to measure the iodine concentration (IC) on the iodine-based material decomposition images and the CT value on the 101 sets of virtual monochromatic images (corresponding to photon energies from 40keV–140 keV at 1 keV interval) in both AP and VP. The copy and paste function was used to ensure measurement consistency between the two phases. To reduce measurement variation, ROI was placed 3 times on the tumour and the average of the 3 measurements was used as the final result. The iodine concentration measurement was repeated on the abdominal aorta at the same imaging level for generating normalized iodine concentration (NIC) for tumours: NIC(tumour) = IC(tumour)/IC(aorta). The ratio of the iodine concentration difference between AP and VP (RICAP–VP: [RICAP–VP = (ICAP−ICVP)/ICAP]) were calculated. The CT value measurements from the virtual monochromatic images from 40 keV to 90 keV were used to calculate the slope (λHU) for spectral HU curve (CT value as a function of virtual monochromatic energy level): λHU = [(CT(40keV)–CT(90keV)]/50. Since the 70 keV virtual monochromatic images in spectral CT have similar muscle CT value to the conventional 120kVp images in abdominal imaging, the 70 keV virtual monochromatic images in our study were used to simulate the results of conventional CT scans for comparison.
Table 1 Patient information between the two pathological groups of small hepatocellular carcinoma in Spectral CT. Parameter
Group Tumours with microvascular invasion (n = 37)
Gender (Number of patients) Male (n = 36) 21/36 Female (n = 14) 8/14 Age 53 ± 26.3 Location (Number of lesions) Left lobe of liver 16/24 Right lobe of liver 21/32 History of cirrhosis 20/34 (n = 34) Degree of differentiation well-differentiated and moderatedifferentiated HCC (n = 22) poorly differentiated HCC (n = 34)
p-value Tumours without microvascular invasion (n = 19)
15/36 6/14 55 ± 17.2
0.939 0.746
8/24 11/32 14/34
0.935 0.863
8/22
14/22
29/34
5/34
2.3.2. Pathological analysis The tissues for pathological analysis were obtained from 50 patients with primary small hepatocellular carcinoma after surgical operation. The whole process followed the protocol edited by the Department of Health of China (“Diagnosis and treatment of primary liver cancer”, edition 2011), including the location and the number for obtaining specimens. The specimen treatment and pathological analysis were performed by two experienced pathologists. The specimen treatment included formaldehyde fixation and paraffin fixation followed by section cutting with the thickness of 4 mm and conventional H–E staining. According to the Hyung standard [10], the presence of microvascular invasion was defined as satisfying the following situations after HE staining: (1) the formation of tumour thrombus were found in central
0.000*
*p < 0.05.
223
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Fig. 1. A 50-year-old male with poorly differentiated hepatocellular carcinoma. (1A) A enhanced lesion was found in the right liver lobe in the arterial phase (red arrow); (1B) the rich blood supply lesion was still found on the iodine-based material decomposition image (red arrow); (1C) the density of the lesion decreased in portal venous phase (red arrow); and (1D) the isodensity lesion was also found on the iodine-based material decomposition image (red arrow); (1E) the formation of tumour thrombus was found in central veins and small veins (blue arrows). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
statistically analyzed using the two-sample t-test. A p-value < 0.05 was considered statistically significant. Receiver operating characteristic (ROC) analysis was carried out to establish threshold values and to calculate the area-under-curve (AUC) for ROC curves, sensitivity and specificity for differentiating small hepatocellular carcinoma with and without microvascular invasion.
vein, portal vein branch or small vein under liver capsule; (2) the tumour cells were found at vascular endothelium and vascular smooth muscle; or (3) fibrin clots or red blood cells were found around the tumour cells in the vascular lumen. The histological grades were also assessed. And the pathological findings were categorized into welldifferentiated HCC, moderate-differentiated HCC and poorly differentiated HCC.
3. Results 2.4. Statistical analysis The enhanced spectral CT scans detected 56 small hepatocellular carcinomas in 50 patients. There were 37 lesions with microvascular invasion (37/56, 66.1%, Group A) (Fig. 1) and 19 lesions without microvascular invasion (19/56, 33.9%, Group B) (Fig. 2). There were no statistically significant differences in age, gender between the two groups (p > 0.05). CT imaging related measurements are listed in Table 2. As indicated in Table 2, the iodine concentration (IC), normalized iodine concentration (NIC), slope (λHU) of the spectral HU curve in the arterial phase and ratio of the iodine concentration
Statistical analysis was performed using SPSS17.0 (SPSS Inc., Chicago, IL) and all measurements were expressed as mean ± standard deviation (SD). The differences in iodine concentration (IC), normalized iodine concentration (NIC), slope of spectral HU curve (λHU), ratio of the iodine concentration difference between AP and VP (RICAP–VP), and CT value at 70 keV and 40 keV virtual monochromatic images between Group A (small hepatocellular carcinoma with microvascular invasion) and Group B (without microvascular invasion) were 224
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Fig. 2. A 64-years-old male with well differentiated hepatocellular carcinoma. (2A) A slightly high density lesion was found in the right liver lobe in the arterial phase (red arrow); (2B) the rich blood supply lesion was still found on the iodine-based material decomposition image (red arrow); (2C) the density of the lesion decreased in portal venous phase (red arrow), showing as an isodensity; and (2D) the lesion was also found on the iodine-based material decomposition image (red arrow) with the similar iodine concentration as liver; (2E) the intact endothelium can be seen under HE staining (red arrows). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
previous studies, patients with signs of microvascular invasion may have higher probability of tumour recurrence and metastasis [2,11]. Regular CT is helpful in detecting the vascular invasions to larger blood vessels in hepatocellular carcinoma, including portal vein and its branches, hepatic vein and hepatic artery. However, microscopic vascular invasions such as intratumoural microvascular invasion cannot be observed directly. Hepatocellular carcinoma is mainly supplied by hepatic artery. The hepatic lobule central veins were easily compressed by tumours due to the lack of connective tissue sheath resulting in blood flow being blocked. On the other hand, with the proliferation of cancer cells, the infiltration of surrounding matrix and vascular endothelium were induced, as well as the formation of micro tumour thrombus and the cancer metastasis through portal vein system. All the above processes are consistent with pathologic findings. Therefore, the prediction of microvascular invasion is particularly important for those patients who underwent the hepatocellular carcinoma resection surgery [12]. Previous studies reported that tumour microvascular invasion was correlated with multiple factors including tumour size, tumour number, pathologic type, pathologic stage and level of biomarkers and gene expression [13–16]. Imaging examination may also have certain value in the evaluation of tumour microvascular invasion, including contrast
difference between AP and VP (RICAP–VP) in Group A were significantly higher than those in Group B (p < 0.05). The CT values measured on the images at 70keV, which simulating the energy level of a conventional 120kVp imaging, were statistically the same between tumours with and without microvascular invasion (p > 0.05). However, CT values between the two types of tumours in the arterial phase became significantly different on images with energy levels from 40 keV to 60 keV (p < 0.05). ROC analysis results are listed in Table 3. The CT value measurement at 70 keV had low AUC value at 0.71, while CT value at 40 keV improved it to 0.81. The dual-energy spectral CT-specific parameters IC, NIC, λHU and RICAP–VP all generated higher areaunder-curve values for the ROC study than using the conventional CT value measurement (Fig. 3). Using 0.188 as the threshold for NIC, one could obtain an AUC of 0.87, sensitivity of 81.3% and specificity of 79.6% for differentiating tumours with and without microvascular invasion. 4. Discussion Microscopic vascular invasion is an important risk factor for poor prognosis of patients with hepatocellular carcinoma. According to 225
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
Table 2 Quantitative assessment between the two pathological groups of small hepatocellular carcinoma in Spectral CT (mean ± SD). Parameters
Group
t-value
p-value
Tumours with microvascular invasion (n = 37)
Tumours without microvascular invasion (n = 19)
Iodine Concentration (mg/ml) Arterial Phase Venous Phase
2.48 ± 0.70 1.73 ± 0.47
1.65 ± 0.47 1.54 ± 0.35
3.605 1.248
0.001 0.222
Normalized Iodine Concentration Arterial Phase Venous Phase
0.23 ± 0.05 0.50 ± 0.09
0.15 ± 0.05 0.45 ± 0.09
4.481 1.581
< 0.001 0.125
Slope (k) of Spectral HU Curve Arterial Phase Venous Phase IC Reduction Rate (AP-VP)
3.39 ± 1.01 2.33 ± 0.63 0.28 ± 0.16
2.22 ± 0.64 2.15 ± 0.48 0.03 ± 0.24
3.721 0.867 3.415
0.001 0.393 0.002
CT value at 40 keV (HU) Arterial Phase Venous Phase
246.16 ± 64.72 177.14 ± 44.44
177.25 ± 42.66 168.39 ± 34.12
3.388 0.598
0.002 0.555
CT value at 70 keV (HU) Arterial Phase Venous Phase
100.97 ± 16.20 81.05 ± 18.67
89.16 ± 16.83 79.11 ± 15.19
1.956 0.309
0.061 0.760
Table 3 Thresholds, sensitivities and specificities for differentiating small hepatocellular carcinoma with and without microvascular invasion. Parameter
Threshold
AUC
Sensitivity
Specificity
Youden index
Iodine Concentration at AP (mg/ml) Normalized Iodine Concentration at AP IC Reduction Rate (AP-VP) Slope (k) of Spectral HU Curve CT number at 40 keV (HU) CT number at 70 keV (HU)
1.906 0.188 0.231 2.691 194.73 93.58
0.853 0.871 0.826 0.871 0.813 0.714
87.5 81.3 75.0 93.8 81.3 75.0
79.6 79.6 79.6 79.6 72.4 64.3
0.671 0.609 0.546 0.734 0.537 0.393
imaging has made this imaging modality more relevant in diagnosing diseases such as tumours. Because the absorption coefficients of substances vary from one to another in different photon energy levels, dual-energy spectral CT can be used to more effectively distinguish them based on the absorption curve. Dual-energy spectral CT also generates material decomposition images to enable concentration measurements for base materials such as iodine, water, fat and calcium. Thus, dual-energy spectral CT provides more parameters than the CT value in conventional CT, which may be helpful in clinical decision [20–24]. Iodine and water are the main basic substances that can be separated by using absorption curve. As the main ingredient of contrast medium, iodine concentration in the contrast-enhanced CT scans may reflect the blood flow and the iodine concentration change may be used to indicate the change of microcirculation. Our results indicated that the Spectral CT-specific parameters, i.e. iodine concentration (IC), normalized iodine concentration (NIC), slope of spectral curve, ratio of iodine content difference between AP and VP (RICAP–VP) as well as the CT value in 40–60 keV in the arterial phase in the group with microvascular invasion (group A) were significantly higher than those in group B without microvascular invasion (P < 0.05). However, these parameters in the venous phase showed no significant difference (P > 0.05) between group A and group B. The higher value of RICAP–VP suggested a faster clearance speed, and less contrast agent remaining in lesions in group A. Recent studies reported the evolution process from cirrhosis nodules to small hepatocellular carcinoma, and suggested the series of changes as a process of multistep carcinogenesis [25]. In this process, the formation of regenerative nodules as well as neoangiogenesis are found, accompanied by the reduction of portal vein and hepatic artery blood supply to nodules. The tumour angiogenesis is shown as immature development, disordered arrangement, increased permeability of incomplete endothelium. At the same time, the invasion of small vessels
Fig. 3. ROC curves to evaluate the diagnostic efficacy of dual energy spectral CT parameters in differentiating small hepatocellular carcinoma with and without microvascular invasion.
enhanced CT, MRI and PET-CT [17–19]. However, there are limited study on the prediction of microvascular invasion by using dual-energy spectral CT. Therefore, in this study, we assessed the capacity of spectral CT in predicting microvascular invasion in patients with hepatocellular carcinoma. The introduction of dual-energy CT changed the diagnostic mode in comparison with the conventional CT which only has single parameter: CT value. Even though in the early days, the clinical use of dual-energy CT has been hampered because of motion mis-registration, high image noise, and excessive radiation exposure caused by long acquisition times [8]. The technologic improvement in dual-energy spectral CT 226
European Journal of Radiology 95 (2017) 222–227
C.-b. Yang et al.
[2] S. Roayaie, I.N. Blume, S.N. Thung, et al., A system of classifying microvascular invasion to predict outcome after resection in patients with hepatocellular carcinoma, Gastroenterology 137 (3) (2009) 850–855. [3] S.A. Shah, S.P. Cleary, A.C. Wei, et al., Recurrence after liver resection for hepatocellular carcinoma: risk factors, treatment, and outcomes, Surgery 141 (3) (2007) 330–339. [4] H. Kim, M.S. Park, J.Y. Choi, et al., Can microvascular invasion of hepatocellular carcinoma be predicted by pre-operative MRI? Eur. Radiol. 19 (7) (2009) 1744–1751. [5] A. Nishie, K. Yoshimitsu, Y. Asayama, et al., Radiologic detectability of mimute portal venous invasion in hepatocellular carcinoma, AJR 190 (1) (2008) 81–87. [6] J.M. Lewin, P.K. Isaacs, V. Vance, et al., Dual-energy contrast-enhanced digital subtraction mammography: feasibility, Radiology 229 (1) (2003) 261–268. [7] D.R. Holmes 3rd, J.G. Fletcher, A. Apel, et al., Evaluation of non-linear blending in dual-energy computed tomography, Eur. J. Radiol. 68 (8) (2008) 409–413. [8] A. Graser, T.R. Johnson, E.M. Hecht, et al., Dual energy CT in patients suspected of having renal masses: can virtual nonenhanced images replace true nonenhanced images? Radiology 252 (2) (2009) 433–440. [9] S.T. Laroia, A.S. Bhadoria, Y. Venigalla, et al., Role of dual energy spectral computed tomography in characterization of hepatocellular carcinoma: initial experience from a tertiary liver care institute, Eur. J. Radiol. 3 (2016) 162–171. [10] W.J. Hyung, J.H. Lee, S.H. Choi, et al., Prognostic impact of lymphatic and/or blood vessel invasion in patients with node-negative advanced gastric cancer, Ann. Surg. Oncol. 9 (6) (2002) 562–567. [11] P.P. McHugh, J. Gilbert, S. Vera, et al., Alpha-fetoprotein and tumour size are associated with microvascular invasion in expiated livers of patients undergoing transplantation with hepatocellular carcinoma, HPB (Oxford) 12 (1) (2010) 56–61. [12] S. Sumie, R. Kuromatsu, K. Okuda, et al., Microvascular invasion in patients with hepatocellular carcinoma and its predictable clinicopathological factors, Ann. Surg. Oncol. 15 (5) (2008) 1375–1382. [13] S. Equchi, M. Takatsuki, M. Hidaka, et al., Predictor for histological microvascular invasion of hepatocellular carcinoma: a lesson form 229 consecutive cases of curative liver resection, World J. Surg. 34 (5) (2010) 1034–1038. [14] M. Kaibori, M. Ishizaki, K. Matsui, et al., Predictors of microvascular invasion before hepatectomy for hepatocellular carcinoma, J. Surg. Oncol. 102 (5) (2010) 462–468. [15] S.X. Yuan, F. Yang, Y. Yang, et al., Long noncoding RNA associated with microvascular invasion in hepatocellular carcinoma promotes angiogenesis and serves as a predictor for hepatocellular carcinoma patients poor recurrence-free survival after hepatectomy, Hepatology 56 (6) (2012) 2231–2241. [16] J.H. Shim, Y.H. Chung, J.A. Kim, et al., Clinical implications of arrest-defective protein 1 expression in hepatocellular carcinoma: a novel predictor of microvascular invasion, Dig. Dis. 30 (6) (2012) 603–608. [17] C.D. Witjes, F.E. Willemssen, J. Verheij, et al., Histological differentiation grade and microvascular invasion of hepatocellular carcinoma predicted by dynamic contrastenhanced MRI, J. Magn. Reson. Imaging 36 (3) (2012) 641–647. [18] C.T. Chou, R.C. Chen, C.W. Lee, et al., Prediction of microvascular invasion of hepatocellular carcinoma by pre-operative CT imaging, Br. J. Radiol. 85 (1014) (2012) 778–783. [19] T.T. Cheung, S.C. Chan, C.L. Ho, et al., Can positron emission tomography with the dual tracers [11C] acetate and [18F] fludeoxyglucose predict microvascular invasion inhepatocellular carcinoma? Liver Transpl. 17 (10) (2011) 1218–1225. [20] N. Takahashi, T.J. Vrtiska, A. Kawashima, et al., Detectability of urinary stones on virtual nonenhanced images generated at pyelographic-phase dual-energy CT, Radiology 256 (1) (2010) 184–190. [21] G. Pache, B. Krauss, P. Strohm, et al., Dual-energy CT virtual noncalcium technique: detecting posttraumatic bone marrow lesions-feasibility study, Radiology 256 (2) (2010) 617–624. [22] J.D. Linder, J.E. Geenen, M.F. Catalano, Cyst fluid analysis obtained by EUS-guided FNA in the evaluation of discrete cystic neoplasms of the pancreas: a prospective single-center experience, Gastrointest. Endosc. 64 (5) (2006) 697–702. [23] X.Z. Lin, F. Miao, J.Y. Li, et al., High-definition CT gemstone spectral imaging of the brain: initial results of selecting optimal monochromatic image for beam-hardening artifacts and image noise reduction, J. Comput. Assist. Tomogr. 35 (2) (2011) 294–297. [24] D. Zhang, X. Li, B. Liu, Objective characterization of GE discovery CT750 scanner: gemstone spectral imaging mode, Med. Phys. 38 (3) (2011) 1178–1188. [25] M. Kobayashi, K. Ikeda, T. Hosaka, et al., Dysplastic nodules frequently develop into hepatocellular carcinoma in patients with chronic viral hepatitis and cirrhosis, Cancer 106 (3) (2006) 636–647. [26] M. Kudo, Multistep human hepatocarcinogenesis: correlation of imaging with pathology, J. Gastroenterol. 44 (2009) 112–118. [27] M. Roncalli, E. Roz, G. Coggi, et al., The vascular profile of regenerative and dysplastic nodules of the cirrhotic liver: implications for diagnosis and classification, Hepatology 30 (5) (1999) 1174–1178. [28] O. Matsui, M. Kadoya, T. Kameyama, et al., Benign and malignant nodules in cirrhotic livers: distinction based on blood supply, Radiology 178 (2) (1991) 493–497. [29] T.R. Johnson, B. Krauss, M. Sedlmair, et al., Material differentiation by dual energy CT: initial experience, Eur. Radiol. 17 (6) (2007) 1510–1517. [30] R. Iannaccone, A. Laghi, Passariello R. Multislice, CT angiography of mesenteric vessels, Abdom. Imaging 29 (2) (2004) 146–152. [31] J.R. Du, Y. Jiang, Y.M. Zhang, et al., Vascular endothelial growth factor and microvascular density in esophageal and gastric carcinomas, World J. Gastroenterol. 9 (7) (2003) 1604–1606. [32] K. Matsumoto, M. Jinzaki, Y. Tanami, et al., Virtual monochromatic spectral imaging with fast kilovoltage switching: improved image quality as compared with that obtained with conventional 120-kVp CT, Radiology 259 (1) (2011) 257–262. [33] M. Kudo, Multistep human hepatocarcinogenesis: correlation of imaging with pathology, J. Gastroenterol. 44 (2009) 112–118.
and portal vein branches induced the formation of a large number of arterial-portal shunting (A–P shunt) and the arterial supply gradually increased [26]. Roncalli [27] study showed that tumour microvessel count (MVC) can help in the diagnosis of dysplastic nodules in cirrhotic liver and small hepatocellular carcinoma at early stage by detecting the significant changes of microcirculation from dysplastic nodules to hepatocellular carcinoma. In our study, most of the patients enrolled had hepatocellular carcinoma which is the progression of chronic liver disease (34/50). During the process, the portal venous blood supply decreased from 68% to 6%, while the arterial blood supply increased from 4% to 94% [28]. These blood supply changes may be detected very sensitively by measuring the iodine concentration change in spectral CT [29]. The change of iodine content can reflect the change of tumour microcirculation indirectly, and the change of tumour microcirculation correlates with the vascular invasion. In the arterial phase, the iodine content reflects the capillary density and capillary permeability, and in the venous phase, the iodine content reflects the clearance and retention of contrast agent. Previous study has found that [30] nutrients uptake in tumour is mainly through the tumour angiogenesis and host vessels invaded by tumour. The higher malignancy of the tumour, the more abundant of tumour angiogenesis [31]. The tumour angiogenesis with incomplete endothelial cell structure leads to an increased vascular permeability, and therefore, increased iodine concentration. In our study the lesions with microvascular invasion had high value of iodine concentration, which is consistent with previous results. And it confirmed the feasibility and accuracy of dual-energy spectral CT in the assessment of microvascular invasion of hepatocellular carcinoma. Our results also indicated that the CT value at 70 keV monochromatic images had low efficacy in the diagnosis of microvascular invasion of hepatocellular carcinoma. Since the conventional 120kVp polychromatic x-ray has an average energy of about 70 keV after penetrating the abdomen of an average size adult patient [32], CT values for 70 keV and 120kVp images are similar. Our results obtained using the 70 keV CT value are indicative and consistent with the fact that conventional 120kVp CT imaging does not provide satisfactory diagnosis of microvascular invasion of hepatocellular carcinoma. On the other hand, with the increased conspicuity, lower keV images (i.e. at 40keV) improved the area-under-curve in ROC study from 0.71 with CT value at 70 keV to 0.81 with CT value at 40keV. This research still has some limitations: firstly, the microvascular invasion was confirmed by H–E staining rather than immuno-histochemical study; secondly, only hepatocellular carcinoma was included in this study, whereas focal nodular hyperplasias, hepatic adenomas or metastatic tumors were excluded. We therefore have limited experience in the diagnosis of microvascular invasion in patients with these diseases. Thirdly, our study were based on the specific patient population with HCC, so we cannot generalize the conclusion to other liver lesions with similar enhancement pattern as HCC. According to previous study [33], it is difficult in diagnosing HCC by only assessing blood supply by using dual energy spectral CT. Future investigations with more cases of different hypervascular tumors are needed to make broader conclusions. In conclusion, dual-energy Spectral CT provides additional quantitative parameters than conventional CT to improve the differentiation between small hepatocellular carcinoma with and without microvascular invasion. Conflict of interest The authors have no conflict of interest. References [1] J.G. Chen, S.W. Zhang, Liver cancer epidemic in China: past, present and future, Semin. Cancer Biol. 21 (1) (2011) 59–69.
227