Paired-angle-rotation scanning optical coherence
tomography forward-imaging probe
Jigang Wu, Michael Conry, Chunhui Gu, Fei Wang, Zahid Yaqoob, and Changhuei Yang
Department of Electrical Engineering, California Institute of Technology, Pasadena, California, 91125
Received December 22, 2005; accepted January 31, 2006; posted February 14, 2006 (Doc. ID 66880)
We report a novel forward-imaging optical coherence tomography (OCT), needle-probe paired-angle-rotation
scanning OCT (PARS-OCT) probe. The probe uses two rotating angled gradient-index lenses to scan the out-
put OCT probe beam over a wide angular arc (?19° half-angle) of the region forward of the probe. Among
other advantages, this probe design is readily amenable to miniaturization and is capable of a variety of scan
modes, including volumetric scans. To demonstrate the advantages of the probe design, we have constructed
a prototype probe with an outer diameter of 1.65 mm and employed it to acquire four OCT images, with a
45° angle between adjacent images, of the gill structure of a Xenopus laevis tadpole. The system sensitivity
was measured to be 93 dB by using the prototype probe with an illumination power of 450 ?W on the
sample. Moreover, the axial and the lateral resolutions of the probe are 9.3 and 10.3–12.5 ?m, respectively.
© 2006 Optical Society of America
OCIS codes: 170.4500, 170.3880, 120.5800, 170.2150, 110.2350, 120.3890.
Over the past decade the development of various en-
doscopic optical coherence tomography (OCT) probes
has greatly extended the application range of this
Most probe implementations can be divided into two
groups based on their scan modes—side imaging1–3
and forward imaging.4,5In side-imaging OCT probes
the actual scan actuators can generally be located far
from the probe tips, which enables very narrow bore
side-imaging OCT probes to be built; the smallest re-
ported side-imaging probe2has an outer diameter of
0.4 mm. We note that all side-imaging OCT probes
are essentially 2D in nature, although the combina-
tion of side imaging and the back-and-forth transla-
tion of the probe may be used to generate 3D images.
In comparison, forward-imaging probes are generally
more complicated in design and require the location
of the actuators to be at or very near the probe tips.4,5
Unsurprisingly, forward-imaging OCT probes that
have been reported are relatively large in probe di-
probe has a diameter of 2.4 mm and is capable of ren-
dering 2D scans.4Therefore the research pursuit of a
narrower (diameter of 1.5 mm or smaller) forward-
imaging OCT probe design with a wide field of view is
a worthy one. Such a probe can be applied in a wide
range of clinical procedures, e.g., anesthesiology pro-
cedures and breast core biopsies,6where having a for-
ward depth-resolved image ahead of a surgical needle
can greatly aid clinicians in guiding the needle to the
In this Letter we present a new design for a
forward-imaging OCT needle probe—the paired-
angle-rotation scanning OCT (PARS-OCT) probe.
This probe design utilizes a pair of angle-cut rotating
gradient-index (GRIN) lenses to deflect and scan the
OCT probe beam across the forward region ahead of
the probe tip. In this design the scan actuation sys-
tem may be located away from the probe tip, much
like in the case of the side-imaging OCT probe de-
scribed in Ref. 1; thus, the miniaturization of the
probe is straightforward. In addition, this probe de-
sign can achieve a large ratio of the forward-scan arc
length to the probe diameter. This parameter is espe-
cially relevant for clinical probe considerations, as a
clinician will desire as wide a scan range as possible
and as small a probe size as possible. We note that
the use of a small GRIN lens or a GRIN fiber as a
probe has been reported by several other groups.7,8
Our probe system is unique in that it permits large-
angle directional scanning of the output probe beam
collection numerical aperture. As a point of reference,
a well-designed probe based on a pair of GRIN lenses
with 112 ?m diameter with a maximum deflection
angle of 20° can have an initial beam width that is
34% of the probe diameter.
The operating principle of the PARS-OCT probe is
as follows. The PARS-OCT probe channels the input
OCT probe light from a single-mode fiber through the
first GRIN lens [see Figs. 1(a) and 1(b)]. The light
beam exits from the other face of the GRIN lens,
which is cut at an angle ?. The beam then enters the
second GRIN lens through an identically angle-cut
face of the GRIN lens. Finally, the beam exits the sec-
ond GRIN lens and focuses at a point ahead of the
probe. The exact focal point is determined by the
pitches of the two GRIN lenses. Amongst other good
choices, an appropriate choice will be to have a
slightly longer than a 1/4 pitch GRIN lens for the
first and a shorter than a 1/4 pitch GRIN lens for the
second; this choice results in a weakly focused beam
between the two GRIN lenses and a tightly focused
exit beam with the desired working distance. For
completeness, we shall define the orientations of the
two GRIN lenses by angles ?1and ?2, the angles be-
tween the projections of vectors r ˆ1and r ˆ2, respec-
tively, in the image plane and the x axis [see Fig.
1(a)]. We shall define the direction of the output light
beam by its polar angle ? that it makes with the z
axis and its azimuthal angle ?; an angle of ?=0 im-
plies that the exit beam propagates along the z axis.
A fan sweep of the output beam in the x–z plane
[shown vertically in Fig. 1(d)] can be performed by
May 1, 2006 / Vol. 31, No. 9 / OPTICS LETTERS
0146-9592/06/091265-3/$15.00 © 2006 Optical Society of America
simply rotating the two GRIN lenses in opposite di-
rections at the same angular speed from the starting
position where the two GRIN lenses are oriented,
such that ?1=0, ?2=180°. This scan pattern can be
understood by taking a closer look at the initial out-
put beam orientation when ?1=0, ?2=180° [see Fig.
1(a)]. In this case the exit beam from the first GRIN
lens is deflected toward the positive x-axis direction.
Note that the second GRIN lens with ?2=180° further
deflects the beam in the same direction. In Fig. 1(d),
this is shown as upward deflection. When we rotate
the two GRIN lenses by an equal and opposite
amount ???1=−??2?0?, the upward deflection of the
exit beam from the first GRIN lens will lessen, and
the beam will lean to the right if we view the probe
head on. The shifting of the beam to the right will be
compensated by the second GRIN lens, which con-
veys an equal but opposite shift to the beam; the
upward-deflection contribution of the beam by the
second GRIN lens will lessen as well. The net effect is
a smaller upward deflection of the beam and little or
no horizontal shifting. Continued rotations of the
GRIN lenses will eventually result in the GRIN
lenses orientation of ?1=?2=90°. In this configuration
[see Fig. 1(b)] the two GRIN lenses compensate for
each other’s deflection of the beam and result in an
output beam that is undeflected. Further rotation of
the GRIN lenses will then deflect the beam down-
ward. A complete 180° rotation of the GRIN lens will
therefore result in a vertical sweep of the output
beam from its up position to its down position—a fan
sweep or an effective OCT B scan.
There are three major advantages associated with
this probe design. First, by attaching the GRIN
lenses to separate concentric needle shafts, we can
actuate the GRIN lenses rotations by simply turning
the needle shafts. This can be done with actuators
that are located far from the probe tip. Second, be-
cause of the novel scanning mechanism employed,
this probe design can more fully use the effective op-
tical channel area of the probe than any other re-
ported probe design. For example, in our prototype
PARS-OCT probe design, the ratio of arc length at
the working distance to the probe diameter is 0.56.
We note that our prototype is not optimally designed
to achieve the best ratio. In an optimal design with
the same scan angle and beam width characteristics,
this ratio can reach 3.57 if the GRIN lenses used are
swapped for a pair with diameter 112 ?m and a 31-
gauge needle is used for the outer needle shaft. Fi-
nally, this probe design allows us to obtain a complete
volumetric scan of the forward region, because the
exit beam direction can be changed freely by rotating
the two needles independently.
Based on this probe design, we fabricated a proto-
type PARS-OCT needle probe and provided a proof-
of-principle demonstration of the PARS-OCT capabil-
ity. The prototype consists of an inner needle (18XTW
gauge) and an outer needle (16TW gauge) that are
cut from standard hypodermic tubings (Poppers &
Sons, Inc.) and attached to a pair of angled GRIN
lenses. The overall probe diameter is 1.65 mm. The
exposed needle length is 1 cm in this prototype and
can be easily adjusted to be up to 3 cm long. Both
GRIN lenses (1 mm in diameter) are angle cut and
polished at an angle ? of 22° at a suitable length. The
measured working distance is 1.4 mm (from the exit
face of the probe, when the exit beam is straight), and
the focused exit beam has a measured focal spot size
of ?10.3 ?m. When the exit beam is tilted, the work-
ing distance will be slightly shorter ??1.27 mm?, and
the focal spot size will be slightly larger ??12.5 ?m?.
Two dc motors attached to the proximal end of the
probe to provide the rotation actuation. Waterproof-
ing the probe for surgical work can be done through
numerous approaches. We evaluated the use of den-
tal epoxy as a way to seal the probe, and it proved to
be sufficient; immersing a treated probe in water did
not change its performance.
The relationship among ?, ?1, and ?2defies a simple
analytical expression. MATLAB simulation of the
probe trajectory during B scans [see Fig. 2(a)] shows
a consistent up-and-down sweep trajectory with ac-
ceptable deviation (the ratio of the maximum angle
deviation to the maximum sweep angle is 1.2%). An
analytical expression of the ? as a function of ?1, ?2,
assuming ?−? is small, can be derived as
? = − N0?A sin?Z?A?d tan????? − ?? + N0cos?Z?A??1,
??? − ??sin ? cos??1− ?2? +?
and ?=sin−1?N0sin ??. N0and?A are the on-axis re-
fractive index and the index gradient constant of the
sin2? + 2?
? − ?
imaging needle probe. (a) The case when there is an angle
between the two angled surfaces of the GRIN lenses; the
exit laser beam is tilted; (b) The case when the two angled
surfaces of the GRIN lenses are parallel; the exit laser
beam is undeviated; (c) PARS-OCT probe setup; (d) profile
of PARS-OCT B-scan mode. SMF, single-mode fiber; GL1,
GRIN lens 1; GL2, GRIN lens 2; IN, inner needle; ON,
outer needle; MS, metal sleeve; M, motor; G, gear.
(Color online) Schematic of the forward-cone-
OPTICS LETTERS / Vol. 31, No. 9 / May 1, 2006
GRIN lens, respectively. Z is the length of the second
GRIN lens, and d is the diameter of the GRIN lens.
We experimentally verify the agreement of the per-
formance of the prototype probe with this simple for-
mula by experimentally measuring ??? versus ??1−?2?.
The cases for ?1−?2?0 and ?1−?2?0, both shown in
Fig. 2(b), indicate very good agreement between
theory and experiment.
We next employed the PARS-OCT prototype probe
to image the gill structure of a euthanized stage 54
Xenopus laevis tadpole. The OCT engine employed in
this experiment is based on a swept laser source with
center wavelength1300 nm,
power 2.5 mW, sweep rate 250 Hz, and a theoretical
signal-to-noise ratio (SNR) of 125 dB. A dual bal-
anced scheme is used to suppress excess noise.9The
measured SNR of the PARS-OCT probe system is
93 dB, and the illumination power on the sample is
450 ?W. The drop in the SNR is due to the coupling
loss within the probe and the beat noise caused by
the interference between the reference beam and in-
ternal reflections within the probe.
In the demonstration, we rotated the two needles
with equal and opposite angular speeds ??21 rpm?
and acquired a single B-scan image from the speci-
men. We then rotated both needles by 45° increments
and acquired the second, third, and fourth B-scan im-
age [see Fig. 3(b)]. Figure 3(a) shows the photograph
of the needle and the tadpole when the images were
being acquired. The scanned locations are shown in
Fig. 3(b). The acquired images are displayed in Fig.
3(c)–3(f). Each image has 350 A scan lines and is ac-
quired in 1.4 s. We can clearly discern the gill pockets
in the images. The scan depth in the image is
2.3 mm, and the largest scan half-angle is 19°.
The PARS-OCT probe design is capable of perform-
ing volumetric scans with very few modifications. By
simply incrementally shifting the starting orienta-
tions of the two GRIN lenses while performing
B-scans, we can acquire volumetric scans. A simpler
implementation will be to introduce a slight offset to
the relative rotation scan velocities. In this case, the
acquired B-scans will automatically sweep through
the entire volume scan space. Our acquisition of
B-scans in different direction with the prototype
probe demonstrates the simplicity by which volumet-
ric scans may be performed.
In summary, we have demonstrated a forward-
imaging OCT needle probe with a 1.65 mm outer di-
ameter based on the use of two rotating GRIN lenses
for probe beam scanning. The probe is readily minia-
turizable and capable of performing volumetric scans
with very few modifications. The probe can be poten-
tially used in needle surgical procedures to provide
high-resolution 3D tomographic images of the targets
forward of the probe.
This research was supported by funding from the
National Institutes of Health, 5R21EB004602-02.
J. Wu’s e-mail address is email@example.com.
1. G. J. Tearney, S. A. Boppart, B. E. Bouma, M. E.
Brezinski, N. J. Weissman, J. F. Southern, and J. G.
Fujimoto, Opt. Lett. 21, 543 (1996).
2. X. D. Li, C. Chudoba, T. Ko, C. Pitris, and J. G.
Fujimoto, Opt. Lett. 25, 1520 (2000).
3. P. R. Herz, Y. Chen, A. D. Aguirre, K. Schneider, P.
Hsiung, J. G. Fujimoto, K. Madden, J. Schmitt, J.
Goodnow, and C. Petersen, Opt. Lett. 29, 2261 (2004).
4. X. M. Liu, M. J. Cobb, Y. C. Chen, M. B. Kimmey, and
X. D. Li, Opt. Lett. 29, 1763 (2004).
5. T. Q. Xie, H. K. Xie, G. K. Fedder, and Y. T. Pan, Appl.
Opt. 42, 6422 (2003).
6. T. A. King and G. M. Fuhrman, Semin Surg. Oncol. 20,
7. W. A. Reed, M. F. Yan, and M. J. Schnitzer, Opt. Lett.
27, 1794 (2002).
8. J. C. Jung and M. J. Schnitzer, Opt. Lett. 28, 902
9. M. A. Choma, M. V. Sarunic, C. H. Yang, and J. A.
Izatt, Opt. Express 11, 2183 (2003).
the focal plane of the exit beam. (b) Calculated and mea-
sured exit beam polar angle, ???, versus the difference be-
tween the orientation angles of the two GRIN lenses,
(a) Calculated B-scan mode profile as projected in
pus laevis tadpole. (a) Photograph of the probe and the tad-
pole when the images are acquired. (b) Indication of scan
locations (c)–(f) in the tadpole. (c)–(f) OCT images acquired
by the PARS-OCT probe; g, gill pockets.
OCT images of the gill pockets of a stage 54 Xeno-
May 1, 2006 / Vol. 31, No. 9 / OPTICS LETTERS