|Year : 1998 | Volume
| Issue : 3 | Page : 153-157
Ultrasonic characterisation of malignant melanoma of choroid
S John, H Sujana, S Suresh, S Swarnamani, J Biswas, L Gopal
Medical Research Foundation, Sankara Nethralaya, India
Medical Research Foundation, Sankara Nethralaya
Source of Support: None, Conflict of Interest: None
An in-vitro study of wave spectral analysis in 8 enucleated eyes was conducted in order to differentiate histological subtypes of malignant melanoma. To obtain the backscattering coefficient for the tissues, we used a broadband focussed transducer with a frequency range of 7-12 MHz and a centre frequency of 10 MHz. Experimental measurement of backscattering coefficient and attenuation coefficient at various frequencies was done by substitution techniques. The backscattering coefficient, scatterer size, and root mean square velocity fluctuation were derived by the numerical method, while the attenuation coefficient at 1 MHz was derived from attenuation coefficient at different frequencies. This study revealed that backscattering coefficient and attenuation coefficient, over a frequency range of 7-12 MHz, show an increase in the spindle cell type compared to the mixed cell type of malignant melanoma. Particularly, the scatterer size was significantly higher in the spindle cell group (p = 0.013) in contrast to the mixed cell type. Spindle cells have uniform and compact histological pattern which contributes to an increase in scatterer size and root mean square velocity fluctuation. The ultrasonically obtained parameters have been shown to have a good correlation with the histology of malignant melanoma.
Keywords: Backscattering coefficient, malignant melanoma, spectral analysis, tissue characterisation, ultrasonography
|How to cite this article:|
John S, Sujana H, Suresh S, Swarnamani S, Biswas J, Gopal L. Ultrasonic characterisation of malignant melanoma of choroid. Indian J Ophthalmol 1998;46:153-7
|How to cite this URL:|
John S, Sujana H, Suresh S, Swarnamani S, Biswas J, Gopal L. Ultrasonic characterisation of malignant melanoma of choroid. Indian J Ophthalmol [serial online] 1998 [cited 2022 Jul 1];46:153-7. Available from: https://www.ijo.in/text.asp?1998/46/3/153/14957
|DM IS SCATTER SIZE, μ IS ROOT MEAN SQUARE VELOCITY FLUCTUATION, AND αO IS ATTENUATION COEFFICIENT AT 1MHZ.|
Click here to view
Malignant Melanoma of the choroid is one of the most common intraocular malignancies in adults. The four most important determinants of prognosis of the tumour are the size, the cell type, mitotic activity, and the presence or absence of extra-scleral extension. Of these diagnostic predictors, routine ultrasound can guide the clinician in judging the tumour size, and the presence or absence of extra-scleral extension. The segregation of the histological subtypes of malignant melanoma is not possible by routine ultrasound imaging. However, the analysis of radiofrequency echographic signals from the tissues, and spectral analysis of the same can provide information regarding the inherent ultrasonic backscatter of tissues. This information can serve as a guide to the acoustic typing of tissues. The basic interaction between tissue and ultrasound depends on the density of the tissue, the speed at which ultrasound travels through the tissue, and the presence of scattering elements. Scattering elements indicate the cellular arrangement, microvasculature and connective tissue elements found in the tissue. When the ultrasonic wave interacts with the tissue, some of it is scattered, which is measured as backscattering, while some of it is absorbed. The scattering and absorption elements together constitute attenuation.
|DM IS SCATTER SIZE, μ IS ROOT MEAN SQUARE VELOCITY FLUCTUATION, AND αO IS ATTENUATION COEFFICIENT AT 1MHZ.|
Click here to view
Scattering of ultrasound in a given tissue depends on several factors, that is, the size of the scatterer, the scattering strength of the particle, the density of the scatterers and the acoustic impedance between the scatterer and the surrounding medium. The particle size, shape, and its elastic properties determine the frequency dependence of the scatterer. Swarnamani and Singh utilised tissue parameters for ultrasonic characterisation of th structure of fibrosarcoma during tumour growth and regression in albino rats.
| Materials and Methods|| |
| Sample preparation|| |
Eight eyes enucleated for malignant melanoma of the choroid or ciliary body, and preserved in 10% formalin, were taken for the study. One half was used for routine histopathological study, while the other half was used for in-vitro ultrasonic analysis. The histopathological data of the enucleated specimens was masked from the observer performing these in-vitro experiments. The cut half of the eye was transferred to a tank filled with distilled water and placed on a flat reflector with the tumour facing the transducer [Figure - 1]. Precautions were taken to remove gaseous inclusions by applying slight pressure. The scattering volume (VS) from which ultrasonic parameters were derived was at a depth of Xo from sample surface [Figure - 1].
| Estimation of ultrasonic parameters|| |
From experimentally obtained backscattering coefficient, two important parameters, that is, maximum size of scatterer (dm) and the root mean square velocity fluctuation (μ) were obtained. From experimentally derived attenuation coefficient at various frequencies (7-12 MHz), the attenuation coefficient at 1 MHz (α0) was derived. These described quantitatively the changes in tissue structure and these principles were applied to the identification of the histological subtypes of malignant melanoma.
| Backscattering coefficient|| |
The instrumentation employed consisted of a broadband focussed transducer with a frequency range of 7-12 MHz and a centre frequency of 10 MHz (Type V Panametrics) in the pulse echo mode specified by the ultrasonic analyser (UA 5052 Panametrics) [Figure - 2]. The scattered signals were obtained from the sample located around the focal region in the far-field of the transducer (range 80-120 mm). The scattering volume was located by electronic gating. The spectrum analyser output corresponding to time sweep was fed to peak detector circuit to obtain the envelope of the spectrum. This was digitized by the A/D Card (PCL-201) and the data was stored in a PC/AT for further analysis. The spectrum of transducer echo obtained from a flat reflector, kept at various axial ranges corresponding to the chosen scattering volume VS was also digitized and stored.
The backscattering coefficient n(f) was determined for 7-12 MHz by applying the following equation: [8,9] where VS(f) is scattering volume computed at frequency f in cm3, R is axial range of scattering volume VS(f) in cm, rr is ultrasonic reflection coefficient between the plane reflector and bath, T is ultrasonic transmission coefficient between the sample and bath, and S(f) is normalized backscattering transfer function of sample at frequency independent of transducer system estimated at spectral analysis of scattered echoes from sample and flat reflector echo at the same range in dB. where Srw(f) is spectral amplitude of the reflector echo measured in water in dB, and SS(f) is spectral amplitude from sample in dB
| Attenuation coefficient|| |
Attenuation coefficient α(f) of the sample was determined by the substitution techniques. [8,9] The ultrasound pulse was directed and received from a flat reflector with and without the sample in between. The attenuation was obtained using the following equation. where Srw(f) and Srs(f) are spectral amplitude of the reflector echo before and after introduction of sample in water in dB, dS is thickness of tissue sample in cm, and f is frequency in MHz.
Attenuation and frequency relation is given by α(f)= α0fn where αo is attenuation coefficient at 1 MHz frequency and n is the frequency dependence.
| Scatterer size and root mean square velocity|| |
The relation between the two tissue parameters, namely, scatterer size (dm) and root mean square velocity fluctuation (μ) and the backscattering coefficient n(f) was given by Sehgal and Greenleaf applicable to the continuous inhomogeneous model. where k is 2πf/C and CS is ultrasonic velocity in the sample.
| Results|| |
Eight eyes containing malignant melanoma were enucleated in the two-year study period. On histopathologic examination four were found to be of the spindle cell type (A, B, C, F), and four were of the mixed cell type (D, E, G, H). The backscattering coefficient and the attenuation coefficient were plotted against the frequencies (7-12 MHz) for each of the specimens [Figure - 3] and [Figure - 4]. A distinct increase was seen in the attenuation coefficient and backscattering coefficient in spindle cell type of malignant melanoma in comparison to the mixed cell type.
The data on scatterer size, root mean square velocity fluctuation and attenuation coefficient at 1 MHz for each sample with histological diagnosis is shown in [Table - 1]. Factors for the spindle cells and mixed cells were separately analysed as shown in [Table:2].
The spindle cell type had an average scatterer size of 28.97±4.76 μm whereas the corresponding values for mixed tumours were 18.73±3.37 μm with a difference that was found to be statistically significant (p=0.013).
Root mean square velocity fluctuation (μ) for spindle cells was 5.39±1.66 whereas it was 4.23±0.26 for mixed cells. Although this was higher in spindle cell, the difference was not statistically significant (p=0.21).
Attenuation coefficient at 1MHz in the spindle cell type was found to be 0.24±0.03, which was higher when compared to the mixed cell type which was 0.19±0.02. This difference was statistically significant (p=0.039). The segregation of the two types of tumours can be clearly seen in [Figure - 5].
| Discussion|| |
Tissue characterisation is defined in its purest form as the identification of one or more physical parameters of a small but representative volume of tissue that is sufficiently well correlated with the type of the tissue so that these measurements may be effectively extrapolated for the purpose of tissue characterisation.
The scatterer size represents the dimensions of the largest scatterer that scatters back at 180° to the direction of the propagation of the incident wave. Coleman and coworkers in a study involving the ultrasonographic features of malignant melanoma of choroid showed that the scatterer size represents tumour microregions such as intervascular nests of cells, rather than the size of individual cells. We found the average scatterer size to range from 15.1-35.3 μm, which is slightly less than the average scatterer size of 30-50 μm as predicted by the model for malignant melanoma. Mixed tumours are composed of epitheloid cells and spindle cells. The epitheloid cells themselves can vary widely in size. Since they are more malignant, they have more vascularity, tumour necrosis and cystic spaces, leading to a non-uniform histologic pattern. In areas of necrosis, scatter intensity is less and hence there is a decrease in scatterer size. Spindle cells have a uniform histological pattern, without much variation in cell size with the presence of reticulin fibres which contribute to an increase in scatterer size.
Collagen has high ultrasonic velocity and is the major source of elastic variation in tissue, and elasticity variation is the major source of acoustic scattering. The organization and distribution of collagen-containing structures determine the appropriate description of tissue as an ultrasonic target. According to Fields and Dunn, Young's modulus of elasticity of collagen fibres is greater in parenchymal tissues. Since the elastic modulus is related to ultrasonic velocity, the presence of collagen in tissues would introduce a velocity fluctuation. It is therefore, reasonable to expect tissues with higher collagen concentration, to have large root mean square velocity fluctuations (μ) The higher μ of the spindle cell tumour though not more statistically significant than that of mixed cells, is thus explained by the presence of reticulin fibres in this tumour. Edema and softening of tissues with an increase in water content reduces the μ As the mixed tumours are composed of epitheloid cells in a non-cohesive matrix, with edema, tumour death, and necrosis in regions of hypovascularity, the reduction in μ is explainable.
The spindle cell type is composed of cells with scanty cytoplasm arranged in a compact manner. This leads to greater attenuation ultrasound energy than in the mixed cell tumour.
It is evident from this in-vitro study that using spectral analysis of the ultrasonic backscatter, it is possible to differentiate between spindle and mixed cell types of malignant melanoma. Specimen 'C' belonging to spindle cell type melanoma had an extremely high value of scatterer size, root mean square velocity fluctuation and attenuation coefficient at 1 MHz. To check whether this value has inflated the group mean values, statistical testing was repeated, with this specimen excluded. The p value for dm (scatterer size) was 0.019; for μ (root mean square velocity fluctuation) was 0.236; and that for αo (coefficient at 1MHz) was 0.085. As is evident, the relationship between the two groups and their statistical significance remained unaltered in the repeat analysis for scatterer size (dm) but was found to be reduced for coefficient at 1 MHz (αo). Hence at least one (and maybe two) parameters show distinctive differences between the two groups.
The shrinking effect of formalin fixative on the tissue is well known. But this affects all specimens equally and is not expected to alter the differences in ultrasound parameters between the two groups. The overall characteristics of the tissues are identified by backscattering coefficient, attenuation coefficient, the root mean square velocity fluctuation and scatterer size, which are utilised to evaluate microstructural details as seen normally in the histology of melanoma. Although the database for this study is relatively small, the results indicate that by acoustic evaluation, an idea of the malignant melanoma microstructure can be obtained.
An in-vivo study applying the information obtained from this study is underway. This should enable analysis of the gray scale of routine, B-scan ultrasound image to possibly differentiate between the two major cellular subtypes of malignant melanoma.
| References|| |
American Academy of Ophthalmology. Ophthalmic Pathology and Intraocular Tumours.
Basic and Clinical Science Course. Section 4. San Francisco: AAO; 1992. p 160.
Seddon JM, Egan KM, Gragoudas ES. Choroidal melanoma: prognosis. In: Ryan SJ, editor. Retina.
St. Louis: CV Mosby; 1989. Vol 1. p 663-71.
Colema DJ, Abramson DH, Jack RL, Franzen LA. Ultrasonic diagnosis of tumours of the choroid. Arch Ophthalmol
Byrne SF, Green RL. Intraocular tumor. In: Craven L, editor. Ultrasound of the Eye and Orbit.
Boston: Mosby Year Book; 1992. p 134-73.
Insana MF, Hall TJ. Characterizing the microstructure of random media using ultrasound. Phys Med Biol
Swarnamani S, Singh M. Ultrasonic characterization of structural parameters of fibrosarcoma during tumour growth and regression in albino rats and its comparison with laser reflectometry. Innov Tech Biology Med
Fellingham LL, Sommer G. Ultrasonic characterization of tissue structure in the in-vivo human liver and spleen. IEEE Trans Sonics Ultrasonics
Nicholas D, Hill CR, Nassiri DK. Evaluation of backscattering coefficient for excised human tissues: results, interpretation and associated measurements. Ultrasound in Medicine and Biology
Swarnamani S, Singh M. Biomechanism analysis of macro and microstructure of normal and pathological tissues by ultrasonic backscattering technique. Med Biology Engg and Comput
Sehgal CM, Greenleaf JF. Scattering of ultrasound by tissues. Ultrasonic Imaging
Chivers RC. Tissues characterization. Ultrasound in Medicine and Biology
Coleman DJ, Silverman RH, Rondean MJ, Lizzi FL, Mclean TW, Jakobiec FA. Correlations of acoustic tissue typing of malignant melanoma and histopathologic features as a predictor of death. Am J Ophthalmol
Kincaid MC, Green WR. Pathology of chorodial melanoma. In: Ryan SJ, editor. Retina.
St. Louis; CV Mosby. 1989. Vol 1. p 675-79.
Johnson RL, Goss SA, Maynard V, Brady JK, Frizzell LA, O'Brien WD. Elements of tissue characterization. In: Linzer M, editor. Ultrasonic Tissue Characterization II.
Washington: Spec. Publishers; 1979. p 19-27.
Fields S, Dunn F. Correlation of echographic visualizability of tissues with biological composition and physiological state. J Acoust Soc Am
[Figure - 1], [Figure - 2], [Figure - 3], [Figure - 4], [Figure - 5], [Figure - 6], [Figure - 7], [Figure - 8], [Figure - 9]
[Table - 1]