Jump to comment:

• Repeatability when measuring porosity and pore density of flow diverters: can we measure in vivo?
  Ignacio Larrabide, Demetrius Lopes
  Published on: 8 August 2018

• Published on: 8 August 2018

  Repeatability when measuring porosity and pore density of flow diverters: can we measure in vivo?

  • Ignacio Larrabide, Researcher Pladema-CONICET, UNICEN
  • Other Contributors:
    • Demetrius Lopes, NRI

  The paper by Farzin et al.[1] shows interesting results about measuring porosity of fully expanded flow diverter stents using (photographic) images of the stent being assessed. In their study, authors used 3 different methods and repeated measurements by different observers to assess the porosity of stents. According to their results, the variability when measuring porosity is so large that previous works assessing it should be questioned. On the other hand, they indicate that pore density seems to be more reliable and repeatable. The study highlights the difficulty of measuring such parameter in a controlled in vitro environment. After carefully reading the article, it became clear that the most reproducible way of measuring porosity, from the 3 options studied, was M3 (based on measuring the width and length of the struts and number of struts per reference square). Furthermore, some simple assumptions should improve the results and substantially reduce errors and variability:
  1. Wire width: the value for wire width, indicated by the manufacturer, is likely to be more accurate. If this value is no to be trusted, at least in average, then the reproducibility of the manufacturing process could not be trusted. Measuring wire width directly on the images is likely to introduce error as it might be affected by reflection/refraction of light on the wire material and wire coating, as well as lens imperfections or optical aberrations in some cases.
  2. Calculating poro...

  Show More

  The paper by Farzin et al.[1] shows interesting results about measuring porosity of fully expanded flow diverter stents using (photographic) images of the stent being assessed. In their study, authors used 3 different methods and repeated measurements by different observers to assess the porosity of stents. According to their results, the variability when measuring porosity is so large that previous works assessing it should be questioned. On the other hand, they indicate that pore density seems to be more reliable and repeatable. The study highlights the difficulty of measuring such parameter in a controlled in vitro environment. After carefully reading the article, it became clear that the most reproducible way of measuring porosity, from the 3 options studied, was M3 (based on measuring the width and length of the struts and number of struts per reference square). Furthermore, some simple assumptions should improve the results and substantially reduce errors and variability:
  1. Wire width: the value for wire width, indicated by the manufacturer, is likely to be more accurate. If this value is no to be trusted, at least in average, then the reproducibility of the manufacturing process could not be trusted. Measuring wire width directly on the images is likely to introduce error as it might be affected by reflection/refraction of light on the wire material and wire coating, as well as lens imperfections or optical aberrations in some cases.
  2. Calculating porosity per strut: using simple geometrical computations is possible measuring the amount of covered area divided by total area, namely porosity, for the rectangular region enclosed by the extremes of one strut (Figure 1 - https://www.dropbox.com/s/oxmbrbs2kv1fcov/Figure%201.pdf?dl=0). This area is equivalent to half the area of one stent cell (i.e. enclosed by 4 neighbouring struts) and the covered area is the rectangular region diagonal multiplied by the wire width, and then subtracting the overlapping region, called Area_open in Figure 1.
  3. Calculating pore density per strut: equivalently, pore density can be computed for each strut. In this scenario, partial struts should not be considered. Each strut accounts for half a pore, then a simple division and transformation on the local porosity will yield a local estimation of pore density, as described in Figure 1.
  4. Regional measurements are largely affected by FD shape and viewing angle: the further away from the point perpendicular position to the camera view point the measure is performed, the least reliable the measurement will be. Unavoidably, a reference rectangle will consider a region that is affected by this camera view point change. Under such conditions and without correcting the view angle position, only a small region of the FD nearby the region perpendicular to the camera should be considered. A correction to the proposed parameters, considering the angle between the mesh and the view point, should be considered.

  The subject of FD porosity and its relation to aneurysm occlusion is recurrently mentioned the literature but so far this subject has not been properly addressed so far. Despite its relevance and implications in treatment selection and local hemodynamics [2], its computation in vivo has not been largely stuied. The work of Boulliot et al.[3] nicely addresses this matter on stents with multiple visible wires, providing an elegant approach to measuring porosity in treated patients as well as the means to simulate the behavior of flow diverters from their geometrical description. Fernandez et al. also proposed and validated a simulation technique with such capabilities[4]. Current medical imaging modalities (3DRA, Dyna-CT/Exper-CT, CTA) provide a resolution that allows visualizing individual wires in FD stents. The combination of improved imaging resolution, improved visibility of flow diverter wires in X-ray images, advanced imaging and simulation algorithms is likely to allow assessment of porosity and pore density in vivo, in the near future.

  Ignacio Larrabide, DSc.
  Pladema Institute - CONICET – UNICEN
  Tandil, Argentina

  Demetriud Lopes, PhD. Dr.
  RUSH University Hospital
  Chicago, IL, USA

  References:
  1 Farzin, B., Brosseau, L., Jamali, S., Salazkin, I., Jack, A., Darsaut, T. E., & Raymond, J. (2015). Flow diverters: inter and intra-rater reliability of porosity and pore density measurements. Journal of neurointerventional surgery, 7(10), 734-739.
  2 Mut, F., & Cebral, J. R. (2012). Effects of flow-diverting device oversizing on hemodynamics alteration in cerebral aneurysms. American Journal of Neuroradiology, 33(10), 2010-2016.
  3 Bouillot, P., Brina, O., Ouared, R., Yilmaz, H., Farhat, M., Erceg, G., ... & Pereira, V. M. (2016). Geometrical deployment for braided stent. Medical image analysis, 30, 85-94.
  4 Fernandez, H., Macho, J. M., Blasco, J., San Roman, L., Mailaender, W., Serra, L., & Larrabide, I. (2015). Computation of the change in length of a braided device when deployed in realistic vessel models. International journal of computer assisted radiology and surgery, 10(10), 1659-1665.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top