Choroidal melanoma is the most common intraocular malignant tumor in adults. It threatens not only the visual function but also the patient's life. The existing methods used in diagnosis of choroidal melanoma include ophthalmolscopy, fundus fluorescein angiography, indocyanine choroidal angiography, magnetic resonance imaging, ultrasonography, etc. Some cases can be diagnosed accurately by these methods.1 However, in some cases, lesion size, clinical presentation, clarity of the refractive media, intraocular hemorrhage and other factors may cause false negative or false positive results in these examinations, which makes it difficult to diagnosis choroidal melanoma in these cases. Contrast-enhanced ultrasound (CEUS) is a real-time, non-invasive, radiation-free imaging method, which has been successfully applied in the liver and diseases of other organs, especially tumors in recent years.2 CEUS is a new method for diagnosing choroidal melanoma, and by clearly showing the perfusion in the tumor tissue, it may reduce the limitation of color Doppler ultrasound in diagnosis of intraocular tumors.3 The present study was designed to assess whether quantitative analysis of CEUS parameters may find differences between choroidal melanoma and normal tissue.
Fifty-one eyes of 51 patients with choroidal melanoma and 14 eyes of 14 patients with choroidal hemangioma were included in this retrospective study. They were presented in Beijing Tongren eye center from October 2007 to December 2008. There were 31 males and 18 females in the choroidal melanoma group. The age ranged from 18 to 78 years old. For differential diagnosis 14 patients clinically diagnosed as choroidal hemangioma were also included. This group included 9 males and 5 females. Their age ranged from 24 to 71 years old. All patients were diagnosed clinically by an experienced retina specialist and underwent detailed fundus examination, ultrasound and color Doppler flow imaging, fundus fluorescein angiography, indocyanine green choroidal angiography, optical coherence tomography, magnetic resonance imaging and CEUS.
The patients were devided into 3 groups. Group 1 included 21 patients with choroidal melanoma which was confirmed by pathology. Group 2 included 30 patients with clinically diagnosed choroidal melanoma. Group 3 included 14 patients with clinically diagnosed choroidal hemangioma.
MyLab90 color Doppler (Esaote, S.p.A. Italy) ultrasonography was used. Probe frequency of conventional ultrasound and color Doppler ultrasound were (9–18) ×106 Hz and that for CEUS was (3–9) ×106 Hz. The color Doppler flow imaging was set at its own conditions for small organ. The power remained about 20% while doing color Doppler flow imaging and the mechanical index maintained at 0.4 or less.
Two-dimension ultrasound was performed to observe the location of lesions, lesion size (largest diameter and the line perpendicular to it) and the acoustic characteristics of lesions, such as whether the internal echo in lesion was uniform or not, whether there was silent decay or not, whether there was choroidal depression or not and the differences between echo within the lesions and the orbital tissue. Blood flow within the lesion was observered by color Doppler Flow imaging.
Contrast tuned imaging (CnTI) techniques were applied for CEUS. Sulfur hexafluoride (SonoVue, Braco, Imaging B.V, Switzerland) were used as contract agent. According to the manual, 2.4 ml contrast agent was injected into the ulnar vein, and then washed with 5 ml physiological saline. Timing device was started at the same time of the injection. Then the filling of contrast agent in the lesion was observed. The entire imaging process was about 8 minutes and recorded and saved as Dicom file.
SonoLiver 1.1 (TomTec Imaging System, Germany) was used to quantitatively analyze the CEUS results. Analysis area included normal tissue region, lesion region and the core lesion region. The core area of the choroidal lesion was selected as reference tissue for image analysis because during the exam, patients' eyes might have a mild shift. A same size of orbital tissue was selected as normal tissue control. The maximum intensity (MI), rising time (RT), time to peak (TTP) and mean transit time (MTT) of the choroidal lesion and control tissue were obtained. The relationships between concentration and the filling time of contrast agent in different tissues were quantitatively analyzed.
Statistical analysis was performed using SPSS 13.0 (SPSS Inc., USA). Data were recorded as mean ± standard deviation (SD). Single-factor repeated measurements analysis of variance and paired t-test were performed. Logistic regression was used to screen independent factors. Predictive sensitivity and specificity were analyzed. P <0.05 was considered statistically significant.
CEUS image of choroidal melanoma
Typical choroidal melanoma was completely filled by contrast agent as time went by. The concentration of contrast agent within the lesions was stronger than normal orbital tissue in the beginning, and then decreased over time. Finally, it was lower than the normal orbital tissue (Figure 1). In the tumors with the largest diameters greater than 15 mm, filling of the contrast agent from the periphery to the center could be seen. In 2 of the 51 choroidal melanoma patients, contrast agent filling defect within the lesion was observed.
Quantitative analysis of choroidal melanoma using CEUS
Compare of parameters of choroidal melanoma and normal tissue
MI of choroidal melanoma was statistically stronger than normal tissue (P=0.010). RT and TTP of choroidal melanoma were earlier than the normal tissue and, but the differences were not statistically significant (P=0.088 and P=0.175). MTT of choroidal melanoma were statistically less than the normal tissue (P=0.040) (Table 1).
Compare of the parameters of choroidal melanoma and choroidal hemangioma in the lesion regions and the core lesion regions.
Both choroidal hemangioma and choroidal melanoma filled by the contrast agent during CEUS. But the contrast agent in choroidal hemangioma had higher intensity and longer duration than that in choroidal melanoma (Figure 2).
As shown in Table 2, MI of choroidal hemangioma was statistically higher than that of choroidal melanoma (P=0.000). MI in lesion area and lesion core area had statistically significant difference (P=0.001). RT of choroidal hemangioma were more than that of choroidal melanoma, but showed no statistical difference (P=0.196). RT between the core lesion regions and the lesion regions was significantly different (P=0.015). TTP of choroidal hemangioma were less than that of choroidal melanoma, but the difference was not statistically dignificant (P=0.461); TTP of the core lesion region and lesion region was significantly different (P=0.025). MTT of choroidal hemangioma were statistically longer than choroidal melanoma (P=0.010); MTT in the core lesion region and the lesion region was significantly different (P=0.035).
Analyzed data in the above tables could be shown vividly on the time-intensity curve (Figure 3).
Logistic regression analysis
Binary logistic regression was used for groups 1 and 3. The results showed that MI, MTT were independent and protective factors. The greater the value was, the less the possibility to be choroidal melanoma. The sensitivity and specificity of prediction in choroidal melanoma and choroidal hemangioma with logistic regression equations were 90.5% and 85.7%, respectively, as shown in Tables 3 and 4.
If the 30 clinically diagnosed choroidal melanoma cases that did not participate in the logistic regression were chosen as objects of prediction, the predict result were 25 (83.3%) cases of melanoma and 5 (16.7%) cases of benign tumor, as shown in Table 5.
CEUS has been widely used in the diagnosis of diseases of heart, liver, kidney, pancreas and peripheral vessels et al. It can reveal the real-time blood flow in the lesion, which makes it a revolutionary technique and many new applications.4–11 Choroidal melanoma is a life-threatening intraocular tumor and often occurs in adults. Using CEUS to diagnose choroidal melanoma is an exploratory work and a new diagnostic technique. We have performed CEUS on more than 500 cases of intraocular tumor since October 2007. The diagnosis of CEUS has been reliable and no adverse event has been observed. We consider it a safe and effective diagnostic method.12
Value of CEUS in the diagnosis of intraocular tumors
Color Doppler flow imaging has been used in diagnosis of intraocular tumor widely.13 But the imaging method has its own limitation. When the ultrasound is vertical to the blood vessels within the tumor, it may not detect blood flow and lead to false-negative diagnosis. Lacking of clinical experiences may cause misdiagnosis. In CEUS, the microbubbles of contrast agent contact with the media and provide a surface for ultrasonic reflect, which increases the signal of blood ultrasound echo, improves the contrast between the blood and the surrounding tissue and increases the sensitivity of the Doppler signal. Avoiding the false negatives diagnoses of color Doppler flow imaging in vascular diseases.
As Folkman proposed, tumor growth depends on the formation of neovascularization.14 Angiogenesis plays an important role in both tumor growth and metastasis. CEUS can show blood perfusion in tumor better. In our patients, when a typical choroidal melanoma had a maximum diameter of over 15 mm, the typical blood perfusion was generally detectable within the tumor, and the perfusing was from the periphery to the center. This is associated to the pathogenesis of malignant tumors. New vessels in the center of the tumor are induced by hypoxia, but they are mostly naive, immature. Therefore the blood supply at the periphery of the tumor is better than that at the center of the tumor. In some cases, non-perfusion area was observed in the tumor. We consider this also related to tumor pathogenesis. New vessels in the tumor do not have normal capillary structure. Due to the caliber variation, increased branching, irregular lumen, blood viscosity and flow resistance increase in new vessels. Massive arteriovenous fistula, increased vascular permeability and lack of lymphatic drainage may increase the pressure in the tumor. Defect of normal vascular structures and non-perfusion in the center of the tumor may cause hypoxia and necrosis. Therefore, we believe that the non-perfusion area on CEUS within the tumor is due to hypoxia and necrosis.
Our results showed that all the tumors were filled with the contrast agent. This suggests that there were a lot of new tumor blood vessels in choroidal melanoma.
Value of CEUS in tumor diagnosis and differential diagnosis
In this study, we used CEUS to examine 51 cases of choroidal melanoma and 14 cases of choroidal hemangioma. Statistical results showed that MI of choroidal melanoma was greater than that of normal tissue. RT, TTP and MTT of choroidal melanoma were less than normal tissue. Perfusion characteristics of choroidal melanoma and choroidal hemangioma on CEUS were different. As shown in Tables 2 and 5, the MI and MTT of choroidal melanoma was less than choroidal melanoma and the differences were statistically significant. This result showed different perfusion character of choroidal melanoma and hemangioma. Choroidal melanoma showed fast wash in and fast wash out of the contrast agent, while choroidal hemangioma showed fast wash in and slow wash out of the contrast agent. This is in accordance with the manifestation of liver tumors.
The CEUS characteristics of choroidal melanoma is related to the new vessels in the tumor.15–21 Tumor cells stimulate the small veins of the host to sprout and cause neovascularization by secreting angiogenic factors. Abnormal new vessel in malignant tumor includes arteriovenous anastomosis, vascular ring, arteriovenous fistula, venous lakes and blind-end blood vessels. The above characteristics cause faster clearance of contrast agent in the choroidal melanoma than in normal tissues, and result in the characteristics of fast wash in and fast wash out of contrast agent on CEUS. New vessels in benign tumor are tortuous and dilated, but lacking arteriovenous shunt. This elongates the transit time and caused a typical CEUS character of fast wash in and slow wash out in benign tumor.
Logistic regression showed that MI and MTT of the time-intensity curve were two markers that could differentiate choroidal melanoma from choroidal hemangioma and by regression equation, we could accurately identify choroidal melanoma and choroidal hemangioma, with sensitivity of 90.5% and specificity of 85.7%. Using this equation to the 30 clinically diagnosed choroidal melanoma cases, 25 cases (83.3%) were in accordance with choroidal melanoma and 5 cases (16.7%) were in accordance with choroidal hemangioma.
Among the 21 cases of pathologically diagnosed choroidal melanoma, 16 were spindle cell type, 3 cases were epithelial-like cells type and 2 patients were balloon-like cell type. The tumor located in the posterior pole in 15 cases, at the periphery in 4 cases, at the equator in 2 cases. The different pathological cell types might have caused the different manifestation in CEUS. As for the clinically diagnosed 30 cases of choroidal melanoma, no pathological diagnosis was made because they received treatments to retain eyeballs. Using the Logistic regression analysis, 25 cases (83.3%) of them were choroidal melanoma. That meant there were 5 cases of clinical diagnosed choroidal melanoma, which did not meet the characteristics of CEUS. Further study with more cases and more detailed classification are needed to clarify the issue.
In summary, the emergence of CEUS technology has brought us a new method for diagnosis of intraocular tumors. Combining time-intensity curve analysis can differentiate choroidal melanoma and choroidal hemangioma better. With further research, this method may offer quantitative data to help to differentiate benign and malignant tumors.
1. Yang WL, Hu SM, Wang JZ, Zheng BH. Color Doppler imaging diagnosis
of intra-ocular tumor. Chin Med J 1997; 110: 664-666.
2. Chen MH, Yang W, Dai Y, Wu W. High mechanical index post-contrast ultrasonography improves tissue structural display of hepatocellular carcinoma. Chin Med J 2005; 118: 2046-2051.
3. Li DJ, Wang YX, Yang WL, Wei N, Wang SH, Wang L, et al. Application of contrast-enhanced ultrasound in ocular occupying lesions. Ophthalmol in Chin (Chin) 2010; 19: 240-243.
4. Liu Y, Tian JW, Xu Y, Cheng W. Role of transvaginal contrast-enhanced ultrasound in the early diagnosis
of endometrial carcinoma. Chin Med J 2012; 125: 416-421.
5. Xu HZ, Qin ZY, Gu YX, Zhou P, Chen XC. Diagnostic value of contrast-enhanced intraoperative Doppler sonography for cerebral arteriovenous malformations compared with angiography. Chin Med J 2010; 123: 2812-2815.
6. Yu TF, Lu FQ, Li ZY, Liu L, Liu Q, Liu AJ, et al. Haemostatic agents of the gelatin matrix for a large liver wound by percutaneous injection without pressure under the guidance of contrast-enhanced ultrasound. Chin Med J 2011; 124: 1352-1356.
7. Zhang XL, Zheng RQ, Yang YB, Huang DM, Song Q, Mao YJ, et al. The use of contrast-enhanced ultrasound in uterine leiomyomas. Chin Med J 2010; 123: 3095-3099.
8. Dong XQ, Shen Y, Xu LW, Xu CM, Bi W, Wang XM. Contrast-enhanced ultrasound for detection and diagnosis
of renal clear cell carcinoma. Chin Med J 2009; 122: 1179-1183.
9. Dirk AC, Kerstin S, Chen MH, Zhu QL, Maximilian R. Role of contrast enhanced ultrasound in detection of abdominal aortic abnormalities in comparison with multislice computed tomography. Chin Med J 2009; 122: 858-864.
10. Dong Y, Wang WP, Cao JY, Fan PL, Lin XY. Quantitative evaluation of acute renal failure in rabbits with contrast-enhanced ultrasound. Chin Med J 2012; 125: 652-656.
11. Forte R, Cennamo G, Staibano S, De Rosa G. Echographic examination with new generation contrast agent of choroidal malignant melanomas. Acta Ophthalmol Scand 2005; 83: 347-354.
12. Yang Q, Wei WB, Yang WL, Shi XH, Zhou D, Liu Y. Experimental study on the impact of choroioretinal structure and function by SonoVue real-time contrast-enhanced ultrasonography. Ophthalmol in Chin (Chin) 2008, 17: 193-197.
13. Regan S, Egan KM, Hart L, Gragoudas ES. Color Doppler imaging of untreated and irradiated choroidal melanomas. Eur J Ophthalmol 2001; 11: 150-155.
14. Folkman J. Tumor angiogenesis: therapeutic implication. N Engl J Med 1971; 285: 1182-1186.
15. Anderson H, Price P, Blomley M, Leach MO, Workman P. Measuring changes in human tumour vasculature in response to therapy using functional imaging techniques. Br J Cancer. 2001; 85: 1085-1093.
16. Yang H, Grossniklaus HE. Constitutive overexpression of pigment epithelium-derived factor inhibition of ocular melanoma growth and metastasis. Invest Ophthalmol Vis Sci 2010; 51: 28-34.
17. Piña Y, Cebulla CM, Murray TG, Alegret A, Dubovy SR, Boutrid H, et al. Blood vessel maturation in human uveal melanoma: spatial distribution of neovessels and mature vasculature. Ophthalmic Res 2009; 41: 160-169.
18. Raoof N, Rennie IG, Salvi SM, Sisley K, Caine A, Mudhar HS. What is the significance of vortex vein invasion in uveal melanoma? Eye (Lond) 2009; 23: 1661-1667.
19. Avery RB, Diener-West M, Reynolds SM, Grossniklaus HE, Green WR, Albert DM. Histopathologic characteristics of choroidal melanoma in eyes enucleated after iodine 125 brachytherapy in the collaborative ocular melanoma study. Arch Ophthalmol 2008; 126: 207-212.
20. Guerin E, Hiscott P, Damato B. Choroidal neovascular membrane in a series of cases of malignant melanoma of the choroid. Acta Ophthalmol Scand 2006; 84: 323-327.
21. Midena E, Pilotto E, de Belvis V, Zaltron S, Doro D, Segato T, et al. Choroidal vascular changes after transpupillary thermotherapy for choroidal melanoma. Ophthalmology 2003; 110: 2216-2222.