Li et al. demonstrated the use of OR-PAM/OCT for evaluating microcirculation in the mouse ear in 2009 [3]. The depth range was estimated to be 1.5 mm for OR-PAM, which was comparable to the 1.8 mm depth for OCT. While this was the first effort to show the feasibility of a multimodal combination with OCT, the transmission (forward) mode configuration required the sample to be very thin. In addition, a mechanically translated objective meant very slow acquisition time (2.5 frames/s), which further limits the application of this configuration. Despite these restrictions, the system was later used to look at neovascularization of a tissue scaffold in anesthetized animals, with the relatively thin mouse ear again used as a model [1].

An effort by Jiao et al. in 2009 looked at the microvasculature of the mouse ear using backward mode detection with an unfocused 15 MHz ultrasound transducer [2]. The authors presented synchronized simultaneous acquisition, though the measurement rate was limited by the excitation pulse repetition rate of the PAM system to 1,024 Hz, whereas a 24 kHz A-scan acquisition would have been possible given their OCT CCD integration time. Liu et al. applied this system for spectroscopic measurements (at 570 nm, 578 nm, 588 nm, and 590 nm) to measure oxygen saturation (sO₂) with a custom-built unfocused 35 MHz needle transducer [4]. They used these sO₂ levels together with Doppler OCT blood flow velocity and total hemoglobin concentration (from laboratory-based blood tests) to measure the metabolic rate in the mouse ear.

Because one of the most common clinical applications of OCT imaging is in ophthalmology, a natural extension of multimodal PAM/OCT imaging is for in vivo retinal imaging. Access to absorption-based contrast could be used to evaluate the retinal pigment epithelium in diseases like age-related macular degeneration, and sO₂ measurements might be used to assess blood vessel status in diabetic retinopathy. Jiao et al. presented a combined PAM/OCT ophthalmoscope system that used an unfocused transducer placed directly on the sclera to acquire preliminary in vivo images of the rat retina at a 93 Hz frame rate (256 A-scans/frame; volumetric image in 2.7 s) [17]. The pulse energy of the excitation laser source was 10 μJ at the output of the laser and estimated to be less than 40 nJ/pulse at the sample. In a study published in 2012 by Song et al., this PAM/OCT system was further extended to include additional modalities like scanning laser ophthalmoscopy and fluorescein angiography and again demonstrated in the rat retina [5]. Despite these preliminary demonstrations in animals, it is unclear whether the pulse energy necessary to generate a PA signal is within the maximum permissible exposure limitation for the human retina.

Using a single light source for PAM excitation and OCT imaging would have several advantages including increased portability and simpler co-registration and synchronization. The first demonstration of a single pulsed light source for PAM/OCT by Zhang et al. (2012) is promising but has some important limitations [7]. Because the same light source and delivery system were used, the images were laterally co-registered. However, the source operates in the visible light range (λ_o = 580 nm, Δλ = 20 nm), where exposure limits are lower than at near-infrared (NIR) wavelengths and there is shallower penetration into tissue. In addition, the spectrum was very noisy and the quality of the OCT images suffered as a result. The source was also not operated at the maximum pulse repetition rate (30 kHz) but instead at 5 kHz. Considering these factors and the limited source bandwidth and axial resolution, this solution is suboptimal for most biological applications of OCT.

Lee et al. have presented PAM/OCT imaging of black and white human hair using a single nanosecond-pulsed supercontinuum NIR laser source [9]. The output range was 600–1,700 nm, though the spectrometer limited detection to between 650 and 900 nm, with the dominant wavelength band beyond 1,064 nm, making this wavelength range more appropriate for OCT imaging. However, many endogenous molecules that would provide spectroscopic contrast for PA imaging, such as hemoglobin, do not exhibit maximum absorb in this wavelength range. Furthermore, the forward mode configuration for PAM detection would limit any potential in vivo applications to thin samples, and the sample is translated mechanically, limiting the imaging speed.

A step towards intraoperative PA and OCT imaging was made by Yang et al., who developed a miniaturized probe for endoscopic measurements of PAM/OCT signals along with conventional pulse-echo ultrasound [6]. The excitation pulse repetition rate was 15 Hz, which limited the speed of acquisition, but they presented sequential imaging of ex vivo of ovarian tissues at 740 nm with excitation energy of 1 mJ/pulse. The endoscope had a 5 mm diameter. Xi et al. also recently demonstrated PAM/OCT detection using a smaller endoscopic delivery probe with a diameter of 2.3 mm. Their system consisted of a low-frequency unfocused 10 Hz transducer with time-domain OCT detection at 1 kHz [10]. If scanning speed can be improved, these delivery systems may further extend the clinical utility of multimodal PA/OCT imaging.

A major drawback to several of the multimodal approaches presented above – and the reason many of them have only been demonstrated in the mouse ear – is their forward mode detection configuration. Backward mode optical detection methods may eventually offer a more practical alternative with respect to clinical translation.

Noncontact photoacoustic sensing by optically probing tissue has recently been demonstrated [18–24]. These methods optically detect the displacement of tissue that occurs upon an acoustic wave reaching the surface. Especially promising for combination with OCT are approaches that use low-coherence interferometry (LCI) for acoustic sensing. Wang et al. presented a method for PAM using homodyne LCI detection of surface displacement [24]. This approach, however, is very sensitive to ambient vibrations, which can change the optical path length more than the path length change that is expected from a PA disturbance. To account for this, the authors synchronized acquisition to the point at which the interferometer is most sensitive. One drawback to their approach is that their configuration only allows for combination with time-domain OCT detection, limiting their acquisition speed. The approach by Blatter et al. demonstrated simultaneous PA/OCT imaging using phase-sensitive swept-source OCT, which could potentially be much faster [18]. Currently, however, their implementation is limited to mechanically scanning the sample.

The next section (Sect. 52.2) describes the transduction of PA signals using an optical detection mechanism based on an FPI sensor and its combination with OCT as presented by Zhang et al. in 2011 [25]. To date, this is the only method to have been applied for in vivo human imaging.

Beard and colleagues first described the fundamental principles of backward mode ultrasonic detection using a FPI sensor with an interrogation laser in 2004 [26]. There are multiple advantages associated with this optical approach over conventional sensing techniques for multimodal imaging. The sensor:

This section describes sensor transduction of ultrasonic signals in the context of one implementation of a PAT system combined with a spectral-domain OCT system at 1,050 nm [8].