Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Current Directions in Biomedical Engineering

Joint Journal of the German Society for Biomedical Engineering in VDE and the Austrian and Swiss Societies for Biomedical Engineering

Editor-in-Chief: Dössel, Olaf

Editorial Board: Augat, Peter / Buzug, Thorsten M. / Haueisen, Jens / Jockenhoevel, Stefan / Knaup-Gregori, Petra / Kraft, Marc / Lenarz, Thomas / Leonhardt, Steffen / Malberg, Hagen / Penzel, Thomas / Plank, Gernot / Radermacher, Klaus M. / Schkommodau, Erik / Stieglitz, Thomas / Urban, Gerald A.

1 Issue per year

Open Access
Online
ISSN
2364-5504
See all formats and pricing
More options …

Radiopacity assessment of neurovascular implants

Thomas Hoffmann
  • Corresponding author
  • Institute for Medical Engineering, Otto-von-Guericke University Magdeburg
  • Institute of Neuroradiology, Otto-von-Guericke University Magdeburg
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Sebastian Gugel / Oliver Beuing / Georg Rose
Published Online: 2016-09-30 | DOI: https://doi.org/10.1515/cdbme-2016-0118

Abstract

An experimental method for radiopacity evaluation of neurovascular implants is presented. State of the art methods are described for cardiovasular implants. To extend the methods to neurovascular devices, imaging parameters were determined from clinical radiography data sets and protocols. To reduce background noise, a thresholding method is introduced. This method is evaluated on a setup with four different laser cut and braided stents. The results are compared to current evaluation methods without filtering. It is shown that unfiltered methods underestimate the radiopacity by 0.89 mm ± 0.21 mm (aluminum equivalent). The introduced method is highly repeatable and uses modern clinical imaging technology. Thus an image evaluation under realisitc conditions is possible. Slightly visible vascular devices and implants can be compared with high accuracy.

Keywords: ASTM F640-12; DIN 13273; image processing; neurovascular; radiopacity; standard; standard software; stent; visibility

1 Introduction

Implants for cerebral arteries, like laser cut stents or braided nitinol and flow diverter stents often show poor radiopacity in radiography. The tradeoff between delivery systems with a small diameter and material thickness as a main parameter influencing the radiopacity is the major reason for poor visibility of neurovascular implants. Furthermore, skull bones which highly absorb X-ray, may impair visibility in the target areas of the implants.

But implant visibility is a crucial factor for therapeutic success. Therefore, implant manufacturers face the problem to develop implants with adequate visibility. For this, a reproducible testing setup with realistic properties is needed to evaluate the progress. Currently no common written standard for the evaluation of X-ray visibility of vascular stents is available. Few details were published by Schmidt et al. [1] referring to DIN 13273-7:2003-08 [2] and ASTM F640-12 [3] as well as by Dobránszky et al. in [4]. Different methods for the evaluation of radiopacity of neurovascular implants were based on qualitative phantom studies or contrast measurements with individual setups [5]. These setups do not comply with the imaging standards of a neurovascular intervention and should be adapted. In the following, a test setup and an image processing method for a quantitative and repeatable evaluation will be introduced and tested on different implants.

1.1 State of the art

State of the art methods recommend the use of a flat detector and X-ray system with a tube voltage of 70 kV and a detector-to-source distance of 1150 mm [1]. A scattering body made of aluminum or polymethyl methacrylate (PMMA) is recommended to be used. As described by Schmidt et al. the scattering phantom mimics the attenuation of the human body and decreases the contrast between stent and background. The resulting image is evaluated by contrast calculations, aluminum equivalent or an image histogram. Michelson or Weber contrast is calculated by means of an area including the stent and a reference area right beside the stent. The stent surface area has a strong influence on this method. With this, contrast calculations of stents with small struts (low stent surface area) and high radiopacity will be underestimated by the arithmetic mean, because the stent structure is not separated from the surrounding.

The aluminum equivalent is determined by measuring the mean gray value of the aluminum steps. A linear regression of those values is done and the calculated mean gray value of the implant area is fitted in the regression line. This results in an equivalent aluminum thickness value, which is independent of the composition of the implant and thus enables the comparison of different stents.

Another method is based on a reference radiographic image without stent and a second image with a stent placed in a defined area. The image without stent is subtracted from that with stent. Consequently it is possible to merely calculate the gray value of the stent surface area by using a simple thresholding. The disadvantage of this method is that the distribution of noise cannot be predicted and varies substantially between different images. Since different material or geometric properties of implants only have little impact on the radiopacity, even minor changes and uncertainties in the measurement setup can influence the result.

The aim of our method is to avoid the inaccuracies and uncertainties of current methods.

2 Material and methods

2.1 Setup

Imaging parameters were determined by the observation of stent interventions. The parameters were adapted to a setup based on a clinical angiographic system (Siemens Artis zeego, Siemens Healthcare GmbH, Erlangen, Germany). The setup comprises a scattering body made of a 200 × 200 × 150 × mm3 (l × w × h) polymethyl methacrylate (PMMA) block (see Figure 1). An aluminum step wedge (see Figure 2 for details) is used as a reference object to translate stent attenuation into a universal and comparable value of equivalent aluminum thickness. The step wedge is placed in the center of scattering body and detector.

The mechanical setup consists of a tissue mimicking scattering body, an aluminum reference step wedge and a flat panel C-arm. System parameters were compared to realistic clinical scenarios.
Figure 1

The mechanical setup consists of a tissue mimicking scattering body, an aluminum reference step wedge and a flat panel C-arm. System parameters were compared to realistic clinical scenarios.

Experimental setup with defined image center (black circle), step wedge, PMMA block (APMMA), area for implant placement (AMA) and different reference areas for contrast calculation. The reference area (ASref) for stent contrast calculation has the same geometry as the stent region (AS) in length (lS) and width (dS).
Figure 2

Experimental setup with defined image center (black circle), step wedge, PMMA block (APMMA), area for implant placement (AMA) and different reference areas for contrast calculation. The reference area (ASref) for stent contrast calculation has the same geometry as the stent region (AS) in length (lS) and width (dS).

Several factors attributed to the X-ray system can strongly influence the comparability of different implants and distort the result. For instance, dead pixels of the detector system can overestimate the radiopacity, in particular of small structures. Likewise, an inhomogeneous X-ray exposure, e.g. due to the heel effect, can influence the measurement (see Figure 3). Those site depending parameters were determined and considered in our calculations.

Experimental setup with inhomogenous dose distribution on the detector caused by the Heel effect of the X-ray tube. 1 - Measuring area with nearly homogenous distribution. 2 - Investigated vertical line with listed gray scale values in diagram below. 3 - Aluminum step wedge.
Figure 3

Experimental setup with inhomogenous dose distribution on the detector caused by the Heel effect of the X-ray tube. 1 - Measuring area with nearly homogenous distribution. 2 - Investigated vertical line with listed gray scale values in diagram below. 3 - Aluminum step wedge.

According to the clinical scenario the detector is placed above the scattering body parallel to the surface of the PMMA block. Source to detector distance is adjusted to 1000 mm. The PMMA block is placed on an extra mounting which is connected to the patient table. The mounting prevents that the patient table is in the field-of-view.

2.2 Image system parameters

State of the art angiographic systems operate with complex control systems, which adapt dose, tube voltage and exposure time during the radiographic acquisition to improve the image quality and reduce the patient dose. As the X-ray parameters depend on the imaged objects, it is impossible to perform identical experiments. The eventual consequent X-ray parameter highly depend on the object imaged. Therefore it is necessary to ensure a reproducible placement of the scattering body and the aluminum reference.

Image post-processing filters (gain correction, noise reduction, edge enhancement) were disabled and the copper filter was switched off to detect the unfiltered stent characteristics without manufacturer depending image properties. The tube voltage was set to 70 kV as used in clinical practice. Radiography is done in single shot mode with a quadratic measuring field and a field of view of 200 × 200 mm2.

2.3 Image evaluation

The Michelson contrast measure is used to assess the visibility of the implant. For this purpose, a reference region close to the stent (with the same length and width as the stent region) is defined. The mean gray value GrefM of this region is calculated by the arithmetic mean method of all pixels. As distinguished from state of the art methods, the background is separated by a threshold filtering from the stent to merely calculate the mean gray value of the implant surface structure. The threshold between background and stent structure is defined as the value for the reduction of the pixel number in the reference area down to 1% for 70 kV. The remaining pixels ensure, that dead pixels have no influence to the result. The mean stent gray value GstentM is used for the calculation of the Michelson contrast C to:

C=GrefMGstentMGrefM+GstentM(1)

As reference for the measurement of the aluminum equivalent, the contrast of the single steps of the aluminum wedge is determined by calculating the mean gray values. The area right beside the 0.5 mm aluminum step is used as reference area (see Figure 2). It is identical in height and width as a single step. The respective contrast of each step is divided by the maximum Michelson contrast of 10 mm aluminum and a linear regression is calculated out of this. For calibration the exact height of the used aluminum step is measured and taken into account. The contrast value of the segmented stent is divided by the maximum contrast of the aluminum step wedge and fit into the regression line.

The majority of neurovascular stents contain marker structures made of a divergent material than the stent struts to enhance the attenuation at well-defined points. The markers were considered separately with the described methods to evaluate the stent radiopacity. For this, the area of interest is reduced to the marker structure.

The amount of pixels of the filtered stent, represented in the area of interest, is counted to obtain an estimation of the projected stent surface area. The coverage of the stent is calculated by dividing the pixel number of corpus and marker by the number of pixel in the unfiltered stent region.

2.4 Quality assurance

For the approval processes of medical devices all relevant parameters of a testing should be recorded. For this the following is necessary:

  • calibration protocol of the aluminium step wedge

  • tested detector (without dead pixels)

  • predefined area for stent positioning without disruptive tube/detector effects

  • logfile of the X-ray system

  • original DICOM-data

  • filtered DICOM files

  • diagram of contrast and aluminum equivalent with fitted stent/marker

  • number of structure representing pixels

A software demonstrator was created to perform the calibration, filtering, measurement, calculation, evaluation and data export as described above.

3 Results

Body and marker structures of four different nitinol stents were investigated with the introduced method with and without the thresholding method. One was a laser cut stent (3.5 × 25 mm2, Figure 4, 1A–C), the other three were braided nitinol stents (3.5 × 15 mm2, see Figure 4, 2A–C; 6.0 × 35 mm2, see Figure 4, 3A–C; 4.0 × 33 mm2, see Figure 4, 4A–C). Stent three and four had the largest stent areas, which find expression in the number of pixels. In relation to the size of the corresponding stent area, the relative coverage of stent four is the highest with 73.1%. The aluminum equivalent was calculated for every corpus and one marker structure of each stent (see Table 1).

Investigated Stents. Radiography (A) with the previous described parameters. Filtered stent corpus (B) and marker structures (C) from radiography images. Implant size is not comparable, due to a different scale.
Figure 4

Investigated Stents. Radiography (A) with the previous described parameters. Filtered stent corpus (B) and marker structures (C) from radiography images. Implant size is not comparable, due to a different scale.

Table 1

Evaluation of the investigated stents. Aluminum equivalent was measured for stent one to stent four for corpus (B) and marker (M) structures. The new threshold based calculation was compared to the method without filter. Pixel number and relative coverage illustrate the surface areas and geometries of each stent corpus and marker.

It was shown that unfiltered images underestimate the equivalent aluminum thickness up to 0.89 mm ± 0.21 mm.

4 Conclusion

In this paper we presented a method for radiopacity assessment of neurovascular implants, which is less prone to inaccuracies than currently used methods. The results are more precise with background filtering. By counting the pixels, representing the stent corpus and marker, information about the object size can be derived. Furthermore, the introduced method employs imaging parameters which are often used in neurointerventional practice. Consequently the method generates realistic data sets with a highly reproducible evaluation and documentation process.

Author’s Statement

Research funding: This work was partly funded by the Federal Ministry of Education and Research in Germany within the Research Campus STIMULATE under grant number 03FO16102A. Conflict of interest: Authors state no conflict of interest. Material and Methods: Informed consent: Informed consent is not applicable. Ethical approval: The conducted research is not related to either human or animal use.

References

  • [1]

    Schmidt W, Behrens P, Kamke F, Schmitz KP. Measurement of radiopacity of vascular implants. Biomed Tech (Berl). 2014;59:S1216–9. Google Scholar

  • [2]

    DIN 13273-7:2003-08 – Catheters for medical use – Part 7: Determination of the x-ray attenuation of catheters; Requirements and testing. Google Scholar

  • [3]

    ASTM F640 – 12 – Standard test methods for determining radiopacity for medical use. Google Scholar

  • [4]

    Dobránszky J, Ring G, Bognár E, Kovács R, Bitay E. New method for evaluating the visibility of coronary stents. Acta Polytechnica Hungarica. 2014;11;81–94. Google Scholar

  • [5]

    Boese A, Rose G, Friebe M, Hoffmann T, Serowy S, Skalej M, et al. Increasing the visibility of thin NITINOL vascular implants. Current Directions in Biomedical Engineering. 2015;1:503–6. Google Scholar

About the article

Published Online: 2016-09-30

Published in Print: 2016-09-01


Citation Information: Current Directions in Biomedical Engineering, Volume 2, Issue 1, Pages 533–536, ISSN (Online) 2364-5504, DOI: https://doi.org/10.1515/cdbme-2016-0118.

Export Citation

©2016 Thomas Hoffmann et al., licensee De Gruyter.. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. BY-NC-ND 4.0

Citing Articles

Here you can find all Crossref-listed publications in which this article is cited. If you would like to receive automatic email messages as soon as this article is cited in other publications, simply activate the “Citation Alert” on the top of this page.

[1]
F. Clarençon, F. Di Maria, J. Gabrieli, E. Shotar, V. Degos, A. Nouet, A. Biondi, and N.-A. Sourour
American Journal of Neuroradiology, 2017, Volume 38, Number 10, Page 1966

Comments (0)

Please log in or register to comment.
Log in