A. Fruscalzo*† and R. Schmitz‡
Article first published online: 27 MAY 2013
Copyright © 2013 ISUOG. Published by John Wiley & Sons Ltd.
Ultrasound in Obstetrics & Gynecology
Volume 41, Issue 6, pages 712–714, June 2013
How to Cite
Fruscalzo, A. and Schmitz, R. (2013), Reply. Ultrasound Obstet Gynecol, 41: 712–714. doi: 10.1002/uog.12475
- Issue published online: 27 MAY 2013
- Article first published online: 27 MAY 2013
We are very pleased that our letter started such an interesting discussion on the possible approaches of elastography for evaluation of cervical stiffness in pregnancy. Recently, this Journal also published an Opinion paper on quantitative cervical elastography in pregnancy, in which current approaches were reviewed. In their Correspondence, Fuchs and colleagues report their own experience on cervical elastography in pregnancy using an approach similar to that used by Molina et al., but also cite the elastographic approach first presented by Świątkowska-Freund and Preis. In our opinion, some critical points of the approaches presented in their letter and in the aforementioned Opinion paper should be considered more carefully; we would like to bring some of these important details to the readers' attention.
In general, elastography enables one to assess the deformation of tissue in response to an applied stress by means of comparing radiofrequency data before and after the stress has been applied. This deformation can be generated intrinsically within the target tissue (as occurs in heart tissue during a contraction cycle) or extrinsically, for example by internal sources of motion (such as arterial or respiratory motion), or by an externally applied stress, such as a controlled or freehand compression. A further method for inducing deformation externally utilizes a focused ultrasound beam, as occurs with shear wave technology[5, 6]. This latter technology is, in theory, the only one that could permit an exact calculation of tissue stiffness, as the acoustic wave vibration force remains constant and the absorbed wave intensities depend on the tissue constitution (above all, on the water concentration). Nonetheless, to our knowledge, no such study on the uterine cervix, during pregnancy or otherwise, has been presented to date.
In the approach used by Świątkowska-Freund and Preis, elastographic images of the cervix are ‘achieved through the movement generated by the patient's breathing and arterial pulsation’, without any active manual cervical compression being exerted through the vaginal probe. Unfortunately, the way in which elastographic images are obtained is not made clear (i.e. through the patient's breathing or arterial pulsation, or perhaps both?). Additionally, claiming, as Świątkowska-Freund and Preis do, that the amount of force generated is not operator-dependent does not mean that the force is constant. For example, the amount of tissue compression could also depend on resistance exerted by the vaginal probe, which is affected by the distance of the probe from the target tissue – particularly in this case, in which the pressure exerted by the vaginal probe, being convex, is unequally distributed on the cervical tissue, which is linear. Finally, to our knowledge, no study has been conducted supporting the reliability of the method proposed by Świątkowska-Freund, in particular the operator-independence and the reproducibility of the cervical movements recorded by elastography; a reliability study appears to be absolutely mandatory before any clinical studies are reported.
The Correspondence by Fuchs et al. reports the preliminary experience of their group concerning cervical elastography in pregnancy. Details of the method used are not described extensively, although they refer to their recent study conducted on pregnant women in the third trimester of pregnancy based on real-time elastography. In this paper, Fuchs and colleagues simply stated that: ‘the Hitachi 7500 ultrasound device uses the method designed by Shiin et al. allowing for real-time assessment of the tissue elasticity’, giving no reference; after searching online we found some studies carried out by Tsuyoshi Shiina, to which we suppose Fuchs et al. were referring, that used real-time freehand elasticity imaging. If the authors used this approach, as it seems they did, no clarification is given as to how they propose to standardize the force applied. If they did so by advancing the vaginal probe by 1 cm as did Molina et al. in their study, we refer the readers to our original letter, wherein we stated that this approach does not in fact standardize the applied force.
Furthermore, in their paper to which this Correspondence refers, Fuchs et al. propose calculation of the strain ratio for assessing the stiffness of cervical tissue. In fact, as has been proposed for the fat tissue of the cutis during breast elastography, determining the strain ratio of two different regions of interest (ROI), positioned correspondingly in two different types of tissue, one of these being used as reference tissue, could solve the problem of standardization of the applied force. Nonetheless, there would be no necessity to calculate the strain ratio for cervical stiffness assessment if the elastography-based method used by Fuchs et al. was not operator-dependent. Also, positioning the ROI at different levels of the target tissue (anterior and posterior cervical lips, for example) introduces an important bias on the strain calculation. Indeed, applied force will not be comparable between the two ROIs selected, due to attenuation of the stress produced on the posterior cervical lip. In any case, the term ‘strain ratio’ should refer to a ratio calculated between two values. Thus, it seems that the authors use this term improperly when referring to the strain measured for the ROI, as positioned on the anterior cervical lip.
In addition, setting the dimensions of the ROI at a fixed value (e.g. a circular ROI, with a diameter of 1 cm, as proposed by Fuchs et al., or 6 mm, as proposed previously by Molina et al.) does not provide any advantage for standardization of the procedure, as it is a clinical fact that the dimensions of the cervix differ greatly among women. Indeed, after setting the dimensions of the ROI, one needs to decide where to place it on the coronal profile of the cervix, its position being one of the most important factors influencing the amount of tissue displacement and the quantitative stiffness measurement.
This is why we proposed standardizing the applied force during the process of raw data acquisition, by exerting movements that were directed perpendicular to the longitudinal axis of the cervix, avoiding lateral displacement of the cervix. The compressing movements should be sufficient to obtain maximum compression of the anterior portion of the cervical tissue (i.e. until the posterior part of the cervical lip is displaced axially, without obtaining further compression of the anterior lip). Additionally, in order to standardize the process of strain calculation we proposed using a ROI covering the entire thickness of the anterior cervical lip.
Finally, in the aforementioned Opinion paper on quantitative elastography in pregnancy published in this Journal, Feltovich and Hall report Molina et al.'s criticism of the method we proposed, saying that it ‘does not sound to us more reproducible’. Yet, as previously stated in our letter, the central point in the case of Molina et al.'s study is not the reproducibility of the process of measurement that was demonstrated, but rather the fact that Molina and colleagues do not seem clear about how to interpret their results, because of a lack of standardization. Now, Fuchs and colleagues state in the present Correspondence that ‘the standardization method described by Fruscalzo et al. … seems difficult to apply and reproduce consistently …’. Again, ours is a very simple method with which to assess cervical stiffness. It reflects the approach of Parra-Saavedra et al., although the experimental part of our study was carried out without knowing of the existence of their study. They state that the cervical stiffness, represented as the cervical consistency index (CCI), should be analogically calculated as a percentage, dividing the residual cervical thickness after manual compression by the initial cervical thickness (CCI = ((AP′/AP) × 100)). This is nothing more than the formula to calculate the Lagrangian strain used during elastography. Our method differs from that of Parra-Saavedra et al. in the mode of strain calculation after manual cervical compression; our use of tissue Doppler imaging (TDI)-based elastography enables automated calculation of the tissue deformation. Thus, even if the main problem associated with these methods remains the standardization of the force applied, reliability tests of the abovementioned studies[12, 13] demonstrate that this method for compressing the cervix and calculating strain seems to work.
Nonetheless, even though elastography-based strain measurement appears reliable and much easier to perform than is calculation of the CCI, there are many factors to be taken into consideration and tested in future studies. Indeed, as recently demonstrated in a further pilot study conducted in at-term pregnancies by our group, the accuracy of the strain measured, and thus the reliability of quantitative elastography, is influenced by the amount of displacement generated and depends on the type of strain considered; natural strain (based on the temporal integration of the instantaneous deformation (dϵ) of the tissue: dϵN(t) = [L(t + dt) – L(t)]/L(t) performs better than does the otherwise classically used Lagrangian strain in the presence of large tissue deformation, such as in the case of very soft cervices, as are generally found in at-term pregnancies. Additionally, as shown in a further study, currently under review and thus not yet available, the value of the derivative pitch and the phase of the cycle selected during the process of automatic strain calculation have an influence on this process and should be considered when comparing results of quantitative elastography. In the same study we also demonstrated the feasibility and reliability of TDI-based elastography for cervical quantitative stiffness assessment in a larger population setting, including all three trimesters of pregnancy.
To summarize, we are convinced that further studies evaluating the proposed different approaches for quantitative elastography in pregnancy are warranted. Particularly when publishing in a highly regarded Journal such as this, the principle of evidence-based medicine should always be taken into consideration.
†Obstetrics and Gynecology,
St. Franziskus-Hospital Münster, Münster, Germany;
‡Obstetrics and Gynecology,
University Hospital of Münster, Münster, Germany
- 4What challenges must be overcome before ultrasound elasticity imaging is ready for the clinic? Imaging Med 2011; 3: 433–444., .
- 12Reliability of cervix elastography in late first and second trimester of pregnancy. Ultraschall Med 2012; 33: 1–7., , , .