1 Medical Vision Laboratory, Department of Engineering Science, University of Oxford, Parks Road, Oxford OX1 3PJ and 2 Nuffield Department of Obstetrics and Gynaecology, The John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
3 To whom correspondence should be addressed. Email: gooding{at}robots.ox.ac.uk
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: accuracy/follicle/fluid/speed of sound/ultrasound
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
An ultrasound machine calculates the depth (or, more strictly, the distance along a scan line) from the time taken for the pulse to return. The depth is calculated using the assumption that the speed of sound is 1540 m/s in soft tissue (Gent, 1997). Therefore, a difference in the true speed of sound will have an impact on the accuracy with which measurements can be made. Thus, herein we measured the speed of sound in follicular fluid to determine the order of magnitude of any potential error. Although a 1% difference in the speed of sound from that assumed will only affect a diameter measurement by 1%, the volume estimated from a three-dimensional scan could potentially be 3% in error, assuming that follicles can be measured as spheres. It is therefore important to know the speed of sound in the tissue that is being measured, in order to ensure, or improve, the accuracy of volume measurement by ultrasound.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
To measure the speed of sound, a single element ultrasound transducer (D69759 3.5 MHz and D51709 5.0 MHz, Cygnus Instruments, UK) was used in conjunction with a pulse generator (Panametrics 500PR, Panametrics-NDT, Waltham, MA). The pulse generator was connected to an oscilloscope (Infinium 54820A, Agilent Technologies, Palo Alto, CA) to record both the transmitted and received pulses. The fluid measured was held in a machined perspex cylinder with a flat bottom, into which the transducer was screwed, allowing precise vertical positioning of the transducer to within a tenth of a millimetre. It also ensured that the element face was parallel to the cylinder bottom on which the ultrasound pulse was reflected. The cylinder and transducer were held in a temperature-controlled water bath (LVF6, Grant Instruments, Cambridge, UK) such that all of the follicular fluid within the cylinder was submerged. The temperature of the water bath varied by up to 0.1°C and reached a steady state 0.2°C below the target temperature, as indicated on the water bath display. The oscilloscope output was processed on a personal computer. The signal envelope was found from the magnitude of the analytical signal, calculated using a discrete Hilbert transform (Bracewell, 2000
). The time of flight was calculated as the mean of the time differences between matched landmarks on the envelopes of the transmitted and received pulses. The landmarks used were the points of one-third maximum power on the leading and trailing edges of the pulse envelopes, as these were found to be more stable than the maximum power itself. These landmarks were identified automatically.
The total distance travelled by the ultrasound pulse was calibrated by measuring the time of flight in pure water at 37.3°C. The speed of sound in pure water was taken from the National Physics Laboratory's interactive version (NPL 2004) of the Marczak equation (Marczak, 1997
). The distance calibration was confirmed by testing at 24.8 and 49.8°C.
Experiment 1: variation of speed with time from aspiration
For the purpose of in vitro experimentation, it was necessary to ensure any changes in the speed of sound occurring in the time after aspiration were quantified. The speed was therefore measured at 37.5°C for one sample at 24, 96 and 120 h from the time of aspiration, using the 5.0 MHz transducer.
Experiment 2: variation of speed with temperature
Although body temperature is assumed to be 37.5°C, there is some evidence to suggest that ovarian follicles normally function at a lower temperature (Hunter et al., 1997; Luck et al., 2001
). Therefore, the effect of temperature on the speed of sound in follicular fluid was studied in the range 2545°C. Measurements of the speed of sound were made at 25°C and at 2.5°C intervals from 30 to 45°C. Although the large range of temperatures is not representative of normal ovarian temperature, use of measurements over this range aids fitting of a suitable parametric model for variation of the speed with temperature. This model can be used to estimate smaller variations. Repeated measurements (varying numbers of measurements, minimum five at 45°C, maximum 26 at 37.5°C, see Table I) were made for a single patient sample at each temperature using the 5.0 MHz transducer, to ensure that the fluid had reached a steady temperature and that the average measurement would be representative despite any minor fluctuations in temperature.
|
Experiment 4: patient dependence
The composition of follicular fluid, which varies with the stage of oocyte maturity (Gerard et al., 2002) and between patients, will inevitably affect the speed of sound transmission. Although it is not possible to measure the fluid composition in vivo to correct for this factor, it is important to quantify the range of such error. To determine if the variation of the speed of sound between patients is significant, the speed of sound in samples from five patients was measured at 37.5°C.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Where the sample has not been changed, the greatest source of error comes from the temporal resolution of the oscilloscope, which was 5 x 108 s. Any variation in the time taken for the pulse can only be detected to this precision. This corresponds to a quantization on the speed of sound of 1 m/s. Changes in the speed of sound within the range cannot be detected or considered significant.
Where multiple readings have been taken, the median result has been given in the text.
Variation with time
The speed of sound in follicular fluid was 1550.5, 1550.5 and 1549.8 m/s for 24, 96 and 120 h from aspiration, respectively. Thus, no significant variation with respect to time was found.
Variation with temperature
Figure 1 shows the variation of the speed of sound in follicular fluid with temperature. A quadratic best fit curve (s=0.018t2+2.9315t+1466) has been plotted (r=0.9993) for the median values. Median values are indicated by crosses, while each data point is represented by a diamond. Table I gives the median, mean, SD and the number of samples at each temperature.
|
Variation with patient
The measurements for the five patients were 1552.1, 1550.5, 1552.1, 1546.7 and 1550.2 m/s. The difference recorded is within the experimental error range; therefore, no variation with patient could be detected.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Increasing the temperature of the follicular fluid from 25 to 45°C produced a change in the speed of sound from 1527 to 1561 m/s. The reported variation in the temperature of the ovary below normal body temperature of up to 2°C (Hunter et al. 1997) equates to an in vivo reduction in the speed of sound in the follicular fluid to 1547 m/s from 1551 m/s.
Any variation in the speed due to the time from aspiration, transducer frequency or individual patient differences was within the experimental error and could not be detected. However, the small range of transducer frequencies and the small number of patients used for these experiments mean that a strong conclusion cannot be drawn from this part of the work.
In summary, it is important to verify the speed of sound when using ultrasound for quantitative measurement, such as volume measurement. For instance, in the authors' current work on computer-assisted 3D measurement of ovarian follicles (Gooding et al., 2003), this is a potential source of error. In the case of follicular fluid, we have shown the speed to be 1550 m/s compared with 1540 m/s which is assumed in practice. The question then is, is the resulting error due to this assumption significant? With regard to follicle diameter measurement, observer error is an order of magnitude greater (Forman et al., 1991
) than the error introduced by an incorrect assumption of the speed of sound. Thus based on our findings, correction would not influence day-to-day practice. However, the improved accuracy (Kyei-Mensah et al., 1996a
) and repeatability (Kyei-Mensah et al., 1996b
) of follicle volume measurements over diameter measurement means that the error introduced by an incorrect assumption of the speed of sound is of the same order as any observer error. Therefore, our study indicates that this error should be corrected for when trying to improve the accuracy of measurements, or deriving any predictive contribution from follicular fluid volume (Saith et al., 1998
). This has implications for potential clinical use of 3D ultrasound in follicular studies in future.
![]() |
Acknowledgements |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Forman RG, Robinson J, Yudkin P, Egan D, Reynolds K and Barlow DH (1991) What is the true follicular diameter: an assessment of the reproducibility of transvaginal ultrasound monitoring in stimulated cycles. Fertil Steril 56, 989992.[ISI][Medline]
Gent R (1997) Applied Physics and Technology of Diagnostic Ultrasound. Milner Publishing Ltd, Prospect, South Australia, pp. 110.
Gerard N, Loiseau S, Duchamp G and Seguin F (2002) Analysis of the variation of follicular fluid composition during follicular growth and maturation in the mare using proton nuclear magnetic resonance. Reproduction 124, 241248.
Gooding MJ, Kennedy SH and Noble JA (2003) Volume reconstruction from sparse 3D ultrasonography. Lecture Notes Comput Sci 2879, 416423.[ISI]
Hunter RHF, Grøndahl C, Greve T and Schmidt M (1997) Graafian follicles are cooler than neighbouring ovarian tissues and deep rectal temperatures. Hum Reprod 12, 95100.[Abstract]
Kupesic S (2001) The present and future role of three-dimensional ultrasound in assisted conception. Ultrasound Obstet Gynecol 18, 191194.[CrossRef][ISI][Medline]
Kyei-Mensah A, Zaidi J, Pittrof R, Shaker A, Cambell S and Tan S-L (1996a) Transvaginal three-dimensional ultrasound: accuracy of follicular volume measurements. Fertil Steril 65, 371376.[ISI][Medline]
Kyei-Mensah A, Maconochie N, Zaidi J, Pittrof R, Cambell S and Tan S-L (1996b) Transvaginal three-dimensional ultrasound: reproducibility of ovarian and endometrial volume measurements. Fertil Steril 66, 718722.[ISI][Medline]
Luck MR, Griffiths S, Gregson K, Watson E, Nutley M and Cooper A (2001) Follicular fluid responds endothermically to aqueous dilution. Hum Reprod 16, 25082518.
Marczak W (1997) Water as a standard in the measurements of speed of sound in liquids. J Acoust Soc Am 102, 27762779.[CrossRef][ISI]
National Physics Laboratory, NPL Acoustics, 2004 http://www.npl.co.uk/acoustics/techguides/soundpurewater/content.html.
Penzias AS, Emmi AM, Dubey AK, Layman LC, DeCherney AH and Reindollar RH (1994) Ultrasound prediction of follicle volume: is the mean diameter reflective? Fertil Steril 62, 12741276.[ISI][Medline]
Raine-Fenning NJ, Campbell BK, Clewes JS and Johnson IR (2003) The interobserver reliability of ovarian volume measurement is improved with three-dimensional ultrasound, but dependent upon technique. Ultrasound Med Biol 29, 16851690.[CrossRef][ISI][Medline]
Saith RR, Srinivasan A, Michie D and Sargent IL (1998) Relationships between the developmental potential of human in-vitro fertilization embryos and features describing the embryo, oocyte and follicle. Hum Reprod Update 4, 121134.
ter Haar-Romeny BM, Titulaer B, Kalitzin S, Scheffer G, Broekmans F, Staal J and te Velde E (1999) Computer assisted human follicle analysis for fertility prospects with 3D ultrasound. Lecture Notes Comput Sci 1613, 5669.
Submitted on May 11, 2004; resubmitted on August 25, 2004; accepted on October 4, 2004.
|