Hum. Reprod. Advance Access originally published online on December 23, 2004
Human Reproduction 2005 20(2):497-500; doi:10.1093/humrep/deh591
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Measurement of the speed of sound in follicular fluid
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 |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Measurement of ovarian follicles by ultrasound is common practice in fertility treatment. However, the effect of the speed of sound is not taken into account. We present results from a study aimed at measuring this. METHODS: The speed of sound was measured in samples of follicular fluid aspirated from patients undergoing fertility treatment. The transmitted and received pulses from a single element ultrasound transducer were recorded using a digital oscilloscope for a pulse passed through a sample of the fluid. The distance over which the pulse travelled was known from calibration with pure water. Variation with temperature was investigated in the range 25–45°C. Dependence on ultrasound frequency, patient and time from aspiration were also investigated. RESULTS: The speed of sound in follicular fluid was found to be 1550±3 m/s at 37.3°C using 5.0 MHz ultrasound. The speed varied from 1528±3 m/s at 24.8°C to 1561±3 m/s at 44.8°C. Variation with patient, time and frequency were not detected. CONCLUSION: The speed of sound in follicular fluid at body temperature is 1550 m/s. This small difference from the speed assumed by the ultrasound machine corresponds to the systematic bias in volume measurement evident in previously published results.
Key words: accuracy/follicle/fluid/speed of sound/ultrasound
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Measuring the size of ovarian follicles plays a vital role in the management of treatments such as IVF and ovulation induction. Currently, standard practice is to determine a mean follicular diameter based upon measurements in two dimensions. However, the follicular volume has been proposed as a more useful measurement (Saith et al., 1988
; Penzias et al., 1994; Kupesic, 2001). These volume measurements have been obtained both manually (Kyei-Mensah et al., 1996a,b; Raine-Fenning et al., 2003) and automatically (ter Haar-Romeny et al., 1999; Gooding et al., 2003).
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 |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Follicular fluid was collected at the time of ultrasound-guided, transvaginal, oocyte recovery during five IVF treatment cycles. Care was taken to minimize contamination with flushing media or blood, and to this end only the fluid from the first few follicles aspirated was collected, after the removal of any oocytes. The fluid was stored in a sealed container at room temperature until it was processed. The fluid from several follicles was pooled for each patient to produce a volume in excess of the minimum of 10 ml required for the measurement apparatus.
To measure the speed of sound, a single element ultrasound transducer (D69759
[GenBank]
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 25–45°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 3: variation of speed with frequency
Follicular scanning is performed at a wide range of frequencies, typically 5–12 MHz. To investigate if the frequency affects the speed of sound, measurements were taken at 37.5°C using the 5.0 and 3.5 MHz transducers.
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 |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Calibration results and error analysis
The distance travelled by the pulse was 0.0623 m for the 5.0 MHz transducer and 0.0570 m for the 3.5 MHz transducer. The distance could be adjusted to a precision of±0.0001 m each time the sample was changed. At 1550 m/s this range in distance corresponds to a range of about±3 m/s in the corresponding speed measurement. Therefore, when the sample has been changed, for the patient and frequency dependence experiments, any changes of <3 m/s cannot be detected, or considered significant.
Where the sample has not been changed, the greatest source of error comes from the temporal resolution of the oscilloscope, which was 5 x 10–8 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 frequency
The measurements for the speed of sound were 1551.7 m/s (1551.3–1552.1 m/s) and 1547.2 m/s (1545.6–1547.2 m/s) for the 5.0 and 3.5 MHz probes, respectively. This difference is within the experimental error range; therefore, no significant variation with frequency could be detected.
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 |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Three-dimensional ultrasound scanning is becoming increasingly utilized in the clinical setting, and its use in follicular measurement is gradually developing. The reasons for this transition to 3D include the improved measurement accuracy and repeatability that come from analysis of an entire object rather than just a single plane (Kupesic, 2001
). If quantitative accuracy is desirable, then it is imperative that any sources of error are eliminated. One such error comes from assuming an incorrect speed of sound for use in image reconstruction. The speed of sound in follicular fluid at 37.3°C was found to be 1550±3 m/s using a 5.0 MHz ultrasound transducer, not the value of 1540 m/s that is typically used by computer software in ultrasound machines when calculating volumes. The difference translates into measurements that underestimate ovarian follicular volumes by up to 2%, which corresponds to the systematic bias of 0.11 ml for follicles of mean volume 6.8 ml, evident in a previously published report (Kyei-Mensah et al. 1996a).
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 |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The authors wish to thank Dr P.Probert Smith (Sensor Systems Group, Engineering Science, University of Oxford) for assistance in setting up experiments and loan of equipment, Dr C.Bain and Ms K.Wilkinson (Physical and Theorical Chemistry, University of Oxford) for the loan of equipment, the patients and staff of the John Radcliffe Hospital Fertility Unit (Oxford) for careful acquisition of the follicular fluid, and Dr J.Bamber (Joint Department of Physics, The Royal Marsden NHS Trust) for advice on relevant texts. M.J.G. is funded by the EPRSC as part of the MIAS-IRC. (GR/N14248).
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Bracewell RN (2000) The Fourier Transform and its Applications. McGraw-Hill Book Co, Singapore, pp. 359–367.
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, 989–992.[Medline]
Gent R (1997) Applied Physics and Technology of Diagnostic Ultrasound. Milner Publishing Ltd, Prospect, South Australia, pp. 1–10.
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, 241–248.[Abstract]
Gooding MJ, Kennedy SH and Noble JA (2003) Volume reconstruction from sparse 3D ultrasonography. Lecture Notes Comput Sci 2879, 416–423.
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, 95–100.[Abstract]
Kupesic S (2001) The present and future role of three-dimensional ultrasound in assisted conception. Ultrasound Obstet Gynecol 18, 191–194.[CrossRef][Web of Science][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, 371–376.[Web of Science][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, 718–722.[Web of Science][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, 2508–2518.
Marczak W (1997) Water as a standard in the measurements of speed of sound in liquids. J Acoust Soc Am 102, 2776–2779.
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, 1274–1276.[Web of Science][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, 1685–1690.[CrossRef][Web of Science][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, 121–134.
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, 56–69.
Submitted on May 11, 2004; resubmitted on August 25, 2004; accepted on October 4, 2004.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||