Ultrasound induced improvement in optical coherence tomography (OCT) resolution
Abstract
Optical coherence tomography (OCT) is a rapidly emerging technology for high-resolution biomedical imaging. The axial resolution of this technology is determined by the bandwidth of the source. Commercial sources generally provide resolutions of 10–20 μm whereas laboratory-based solid state lasers have resolutions of ≈4 μm. The resolution in tissue depends almost exclusively on detecting single scattered events. However, the phenomenon known as multiple scattering results in a deterioration of resolution as a function of depth. In this study, OCT was combined with ultrasound in an attempt to reduce the effect of multiple scattering. The theory is that, with parallel ultrasound and OCT beams, multiply scattered light with a momentum component significantly perpendicular to the OCT beam will be reduced because the light is Doppler shifted outside the bandpass filter of the OCT detection electronics. A 7.5-MHz ultrasound transducer was used to introduce the photon/phonon interaction. A reflecting metal plate was placed within biological tissue, and the point spread function (PSF) was assessed off the reflector. The PSF was determined in the presence of no ultrasound, pulsed ultrasound, and continuous-wave (CW) ultrasound. CW ultrasound resulted in a 17% improvement (P < 0.001) in resolution and pulsed ultrasound resulted in 8% (P < 0.01). Image noise reduction could also be noted. Combining OCT with a parallel ultrasound beam results in an improvement in resolution through a reduced effect of multiple scattering due to photon/phonon interaction. With higher frequencies, better control of the acoustical beam, and tests in media with higher rates of multiple scattering, improved results are anticipated.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Optical coherence tomography (OCT), a recently developed technology based on low coherence interferometry, is a rapidly emerging technology for high-resolution biomedical imaging (1, 2). OCT is analogous to ultrasound, measuring the back-reflection of infrared light rather than sound. The disorders where OCT is demonstrating its greatest potential include the prevention of myocardial infarction, the early detection of cancer, the early diagnosis of osteoarthritis, and the guidance of microsurgery (3–6). OCT has several advantages for biomedical imaging. First, the resolution of OCT at 1,300 nm is between 4–20 μm, depending on the source used. Second, the acquisition rate for biomedical imaging is in the range of 4–8 frames per second (7). Third, OCT is based on fiber optic technology, so inexpensive catheters and endoscopes can be microengineered to very small cross-sectional diameters. Finally, OCT can be combined with a range of spectroscopic techniques including polarization and Doppler spectroscopy (8, 9).
With OCT, broad bandwidth infrared light from the source is split by a beam splitter, part to the source and part to the reference arm. Light reflects from within the tissue and off of the mirror in the reference arm and is recombined by the beam splitter. If the optical group delays in both arms match to within the coherence length, interference will occur. The interference signal is used to assess back-reflection intensity. The detected light is separated from noise via a low pass filter, demodulation, and bandpass filter (to remove the carrier frequency induced by reference mirror motion).
OCT imaging in nontransparent tissue is typically performed around 1,300 nm because an optical window exists at this wavelength, minimizing both scattering and absorption, which increases penetration (10). One important factor determining the axial resolution with OCT is the bandwidth of the source. With all other parameters constant, the greater the source bandwidth, the higher the axial resolution in free space. Commercially available diode sources generally have axial resolutions in the range of 10–20 μm, but are expensive, trade off power for resolution, and are in limited supply. Investigators have used solid-state lasers to increase the resolution to less than 10 μm, such as the Kerr lens, mode locked, chromium forsterite laser (7). However, these lasers require considerable expertise to use, and generally costs are over $100,000.
For the resolution of OCT to be maximal, only single scattered photons should be detected in the interferogram. Multiply scattered light results in an accurate indirect measurement of time of flight. Whereas the electromagnetic theory behind multiple scattering is a complex phenomenon, which has been reviewed elsewhere, in general, if we envision multiply scattered photons as “bouncing around,” then it is apparent that the time of flight does not provide accurate ranging information (11). This result leads to a deterioration of the resolution, which worsens as a function of depth and intensity.
If multiply scattered light can be reduced in the image, then image resolution can be improved. The approach of this work is that, because most multiply scattered light typically has some momentum component perpendicular to the OCT beam (whereas single scattered light has a minimal component), this Doppler shift can be used to control the multiply scattered light. By using a sound wave parallel to the OCT beam to induce a Doppler shift in light momentum components perpendicular to the OCT beam (and potentially diffraction), at least some multiply scattered light in theory should be frequency shifted outside the bandpass filter. The sound waves induce the frequency shift by transferring energy to the multiply scattered light through photon–phonon interaction. The theory will be tested by imaging a metal reflector, in the absence and presence of an ultrasound beam, embedded in biological tissue.
Methods
OCT imaging was performed as previously described (1, 2). A 1300-nm diode source (AFC, Toronto, Canada) was used with an axial resolution in free space of over 24 μm with an intensity of 12 mW. A grating based optical group delay line was used with an acquisition rate of 4 frames per second (7). The lateral resolution was 30 μm, and the signal-to-noise ratio was 101 dB. The OCT imaging optics were scanned across the sample by using a translation stage (Physik Instrumente, Karlsruhe/Palmbach, Germany) with 0.1 μm resolution.
The detection system used to separate the interference envelope from the carrier and noise has been previously described. After detection (conversion from an optical to electronic signal), transimpedence amplification, bandpass filtration (passive Butterworth filter), and demodulation, the signal is low pass filtered to remove the carrier frequency. Through photon–phonon interaction, it is postulated that a Doppler shift can be induced in multiply scattered light that results in the photons not being detected because they are outside the low pass filter.
A 7.5-MHz ultrasound transducer (Advanced Technical Laboratory, Bothell, WA) was placed approximately in parallel to the OCT system. The system was used because it is a commercially available device. Because of the size of the ultrasound transducer, an angle of approximately 10° existed between the ultrasound and OCT beam. The ultrasound beam was brought into direct contact with the tissue with ultrasound transducing medium (AccuGel; Lynn Medical, Bloomfield Hills, MI). The OCT imaging was performed with the ultrasound beam in three settings: ultrasound off, pulsed ultrasound, and continuous-wave (CW) ultrasound. The CW was performed at a power of 10.6 mW, beam diameter of 0.15 cm, and focal length of 2.1 cm. The pulsed ultrasound was performed with an average power of 17.8 mW, beam diameter of 0.24 cm, and a focal length of 2.1 cm. The ultrasound beam is being used over 1.5 cm proximal to the focus, resulting in an essentially collimated beam (12). The pulse repetition rate was 1.3 ms, with an average pulse intensity of 225 W/cm2.
Imaging was performed of Melanogrammus aeglefinus (haddock), immersed in saline, which was used as a scattering medium. A metal reflector was placed ≈1 mm below the surface to measure the point-spread function (PSF).
Resolution was assessed by measuring the PSF off the reflector at 10 distinct points at the three ultrasound settings through the scattering media. Means were calculated, and groups were compared by using paired t tests.
Results
Table 1 illustrates the average measured point spread functions. When no ultrasound was present, the mean PSF was 24 ± 2 μm. When pulsed ultrasound was used, the mean PSF was 22 ± 2 μm; and with CW ultrasound, the mean PSF was 20 ± 1 μm. This result represents a 17% improvement in resolution. The resolution with CW ultrasound was significantly greater than both no ultrasound (P < 0.001) and pulsed ultrasound (P < 0.01). Pulsed ultrasound was significantly greater than no ultrasound (P < 0.001).
Table 1
| Ultrasound mode | Resolution |
|---|---|
| No ultrasound | 24 ± 2 μm |
| Pulsed | 22 ± 2 μm (P < 0.01) |
| CW | 20 ± 1 μm (P < 0.001) |
Fig. 1 shows an image of the tissue in the presence (Upper Left) and absence (Upper Right) of CW ultrasound . In the enlarged sections in the lower sections of the figure, it can be seen that random signal between scatterers is reduced.
Figure 1
Fig. 2a shows an example of PSF at the three different ultrasonic conditions whereas Fig. 2b is the Fourier transform of the PSF. In Fig. 2a, imaging done in the presence of CW ultrasound shows a smaller PSF than pulsed ultrasound, whereas pulsed was superior to no ultrasound at all. In Fig. 2b, the Fourier transform of the PSF is shown for two reasons. First, the broader spectrum of the CW ultrasound is confirmed. Second, a slight increase in the spectral wings is seen in the ultrasound spectrums, consistent with Doppler-shifted multiply scattered light still within the bandpass filter (yellow arrow).
Figure 2
Discussion
In this manuscript, it is demonstrated that the use of ultrasound with OCT results in a small but significant improvement in resolution as measured by the PSF. Changes seen in the corresponding OCT images were noted but were small (seen in enlargement but not obvious in total image). One reason why this may be is that a 17% improvement is not sufficient to see detectable changes in most images. Better effects may be noted with higher frequencies, higher intensities, or a smaller frequency range in the bandpass filter. A second reason may be the wings noted in Fig. 2b. These may represent multiply scattered infrared light that has undergone a Doppler shift, consistent with the proposed photon–phonon interaction, but the shift is insufficient to move it outside the bandpass filter. These partially Doppler-shifted photons may result in increased noise in the image and can be compensated for by either decreasing the width of the bandpass filter or compensating for the wings with image processing. A third reason may be a limitation of the study. By using a commercial ultrasound transducer at this low frequency, the ultrasound is slightly off axis from the OCT beam, complicating the results. Finally, tissue that generates even greater multiple scattering, such as calcified arterial tissue, could have an even more dramatic effect.
It is believed that the effect is due to a frequency shift in multiply scattered light from photon–phonon interactions, although some degree of diffraction outside the detection optics through the same mechanism may contribute. The frequency shift would then move the multiply scattered light outside the frequency range of the bandpass filter. Multiple groups have investigated classical models of the interaction of sound and light in tissue, which has included Bragg and Raman-Nath Scattering mechanisms (13–19). Notable work with multiple scattering includes that of Wang, who has modeled the interaction of multiply scattered light with acoustical waves and the contribution from changes in refractive index vs. displacements of scatterers (18–19). The authors of this manuscript in general have not attempted to relate the results seen to macroscopic phenomena, such as refractive index gradients and movement of scatterers from the modeling previously described, because the relative contribution of these factors depends on the medium (e.g., liquid vs. solid), characteristics of the ultrasound beam, and characteristics of the OCT beam. Future studies varying these parameters should provide better insight into underlying mechanisms. However, irrespective of the underlying macroscopic description, the vector transfer of momentum between the photon and phonon adequately describes the results obtained and is also the mechanism for the described macroscopic phenomena.
That said, the question has been raised as to why harmonic motion of the scatterers in the solid does not result in a Doppler shift in the “ideal” single scattered OCT photons to outside the bandpass filter and therefore reduce performance. The authors can speculate on several reasons, although do not claim that this was directly examined in the data. First, in a solid where scatterers are typically bound, energy from the sound wave may not produce motion sufficient to significantly alter single backreflected photon. Second, because the interaction of multiply scattered light with scatterers is by definition greater than that of single scattered light, any effect would be greater for multiply scattered light.
Future studies will focus on increasing ultrasonic frequencies, improving the alignment of the OCT/ultrasound beam, and comparisons of ultrasound beams in the far and near fields. Higher frequencies and intensities will be examined in future studies. It is assumed that higher frequencies will result in a greater number of multiply scattered photons being shifted outside the bandpass filter. This result occurs in theory because the degree of frequency shift depends not only on the frequency of the ultrasound beam, but also on the vector angular momentum of the multiply scattered light. The closer the scattered light is aligned with the original OCT beam, the smaller the frequency shift induced by the sound at a given frequency. The greater the frequency of the ultrasound beam, the greater the likelihood that the multiply scattered light at low angles will be outside the bandpass filter. In addition, at higher frequencies, diffraction of the multiply scattered photons outside the collection optics will likely become a more important phenomenon (13). It should be noted, however, that, at higher frequencies, the wing phenomena seen in Fig. 2b may become exaggerated, requiring the width of the bandpass filter to be reduced.
As stated, future work will also look at improving the alignment between the infrared light and ultrasound beam. This result will be achieved with custom designed ultrasound transducers. One intriguing design would be to use a transparent (to near infrared light) ultrasound transducer and perform OCT imaging through the transducer. Future work will also examine imaging of phantoms and tissues of different characteristics to have a better understanding of the macroscopic mechanisms, such as a refractive index grating, that may contribute to the improved resolution.
This work demonstrates that ultrasound can improve and control the resolution of OCT within a scattering medium. Future studies will modify the approach to further increase the improvement in resolution.
Abbreviations
- OCT
- optical coherence tomography
- CW
- continuous-wave
- PSF
- point spread function
Note
This paper was submitted directly (Track II) to the PNAS office.
Acknowledgments
We thank Dr. Michael Fiddy, Nirlep Patel, and Dr. Ika Rogowska for their insights into this work. This research is sponsored by the National Institutes of Health, Contracts NIH-RO1-AR44812-02, NIH-1-R29-HL55686-01A1, NIH R01 AR46996-01, NIH R01-HL63953-01, NIH-1-R01-HL55686-07, and R01 EB000419-01.
References
1
D Huang, E A Swanson, C P Lin, J S Schuman, W G Stinson, W Chang, M R Hee, T Flotte, K Gregory, C A Puliafito, J G Fujimoto Science 254, 1178–1181 (1991).
2
M E Brezinski, G J Tearney, B E Bouma, J A Izatt, M R Hee, E A Swanson, J F Southern, J G Fujimoto Circulation 93, 1206–1213 (1996).
3
M E Brezinski, G J Tearney, N J Weissman, S A Boppart, B E Bouma, M R Hee, A E Weyman, E A Swanson, J F Southern, J G Fujimoto Heart 77, 397–403 (1997).
4
M E Brezinski, G J Tearney, S A Boppart, E A Swanson, J F Southern, J G Fujimoto J Surg Res 71, 32–38 (1997).
5
J M Herrmann, C Pitris, B E Bouma, S A Boppart, C A Jesser, D L Stamper, J G Fujimoto, M E Brezinski J Rheumatol 26, 627–635 (1999).
6
C Pitris, A Goodman, S A Boppart, J J Libus, J G Fujimoto, M E Brezinski Obstet Gynecol 93, 135–139 (1999).
7
G J Tearney, M E Brezinski, B E Bouma, S A Boppart, C Pitris, J F Southern, J G Fujimoto Science 276, 2037–2039 (1997).
8
W Drexler, D Stamper, C Jesser, X Li, C Pitris, K Saunders, S Martin, M B Lodge, J G Fujimoto, M E Brezinski J Rheumatol 28, 1311–1318 (2001).
9
J A Izatt, M D Kulkami, S Yazdanfar, J K Barton, J Welch Optics Lett 22, 1439–1441 (1997).
10
P Parsa, S Jacques, N Nishioka Appl Optics 28, 2325–2330 (1989).
11
M Born, E Wolf Principles of Optics (Cambridge Univ. Press, 7th Ed., Cambridge, U.K.), pp. 708–710 (1999).
12
P J Benkeser, A L Churchwell, C Lee, D M Abouelnasr J Am Soc Echocardiogr 6, 158–165 (1993).
13
A Korpel Acousto-Optics (Dekker, 2nd Ed., New York, 1997).
14
L Wang, S L Jacques, X Zhao Optics Lett 20, 629–631 (1995).
15
S L Jacques, S Kirkpatrick Optics Lett 23, 879–881 (1997).
16
A Holm, H W Persson, K Lindstrom Ultrasound Med Biol 17, 505–512 (1991).
17
H M Colbert, K L Zankel J Acoust Soc Am 35, 359–363 (1963).
18
L V Wang Phys Rev Lett 87, 043903 (2001).
19
L V Wang Optics Lett 26, 1191–1193 (2001).
Information & Authors
Information
Published in
Classifications
Copyright
Copyright © 2002, The National Academy of Sciences.
Submission history
Received: March 15, 2002
Published online: July 15, 2002
Published in issue: July 23, 2002
Keyword
Acknowledgments
We thank Dr. Michael Fiddy, Nirlep Patel, and Dr. Ika Rogowska for their insights into this work. This research is sponsored by the National Institutes of Health, Contracts NIH-RO1-AR44812-02, NIH-1-R29-HL55686-01A1, NIH R01 AR46996-01, NIH R01-HL63953-01, NIH-1-R01-HL55686-07, and R01 EB000419-01.
Authors
Metrics & Citations
Metrics
Altmetrics
Citations
Cite this article
Ultrasound induced improvement in optical coherence tomography (OCT) resolution, Proc. Natl. Acad. Sci. U.S.A.
99 (15) 9761-9764,
https://doi.org/10.1073/pnas.142155899
(2002).
Copied!
Copying failed.
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.