High-Resolution Photoacoustic Tomography for Early-Stage Cancer Detection and Its Clinical Translation

Abstract

Diagnosing cancer during early stages can substantially increase the cure rate, decrease the recurrence rate, and reduce health care costs. Over the past few decades, the continual development of new medical imaging modalities has been an important factor for diagnosing cancer, selecting therapies, and monitoring response to treatment. Photoacoustic tomography (PAT) is a hybrid imaging modality combining optical contrast from absorption of light with the outstanding spatiotemporal resolution of US imaging, providing biomedical morphologic and functional information of early-stage cancer. In this review, the basics and modalities of PAT, as well as a summary of its state-of-art applications in early-stage cancer (breast cancer, melanoma, and prostate cancer) detection and treatment guidance will be introduced. The potential clinical translation in cancer detection of PAT and prospects for the possibilities to lead to further clinical breakthroughs will also be discussed.

Keywords: Molecular Imaging-Cancer, Photoacoustic Imaging

© RSNA, 2020

Summary

With the continuous advancement of technology, photoacoustic tomography features rapid image acquisition, high detection sensitivity, fine spatial resolution, and miniature probe size, translating it to clinical practice.

Essentials

• ■ Photoacoustic tomography (PAT) provides not only quantitative structural information, but also functional and metabolic parameters, corresponding to the comprehensive progression of malignant tumors at various stages.

• ■ Emerging PAT modalities for lesion characterization include photoacoustic microscopy, photoacoustic endoscopy, and photoacoustic CT.

• ■ PAT enables early cancer diagnosis, precise positioning, and treatment outcome evaluation for research study and bedside operations.

Introduction

According to GLOBOCAN 2018, the number of newly discovered cancer cases in the world will increase from 18.1 to 29.5 million from 2018 to 2040 (1). It is estimated that the cost of cancer treatment will increase to $173 billion in the United States by 2020 (2). Another report also showed that the overall cancer death rate has declined by approximately 1.5% per year since the early 1990s due to advances in early cancer detection and subsequent treatment, including targeted therapies (3). The continuous improvement of surgery, radiation therapy, chemotherapy, and medical imaging technologies has played an increasingly important role in achieving the evaluation of tumor morphology, physiologic characteristics, and therapeutic response.

With growing clinical needs, many new medical imaging technologies have flourished in the clinic over the last decade (4). Different imaging modalities, such as PET, SPECT, CT, MRI, US imaging, and optical imaging, have enabled imaging of human anatomy and function in the clinic. With the aid of these medical imaging methods, conditions involving different features of metabolism, blood flow, and oxygen saturation can be detected (5). However, each modality has its own limitations such as cost, portability, sensitivity, spatial resolution, or imaging depth. For example, the cost of MRI in a single examination is relatively high compared with US and optical imaging, increasing costs of longitudinal monitoring. In addition, strong scattering of light in soft tissue usually limits high-resolution optical imaging to very superficial clinical applications.

Photoacoustic tomography (PAT) combines the advantages of both conventional optical imaging and US imaging, providing a hybrid modality that has broken the penetration limits of traditional high-resolution optical imaging. Applications of PAT have grown in the past decades, providing unprecedented high-resolution imaging and functional information at depths far beyond the optical diffusion limit (approximately 1 mm) of ballistic optical imaging. PAT utilizes ultrashort laser pulses to illuminate tissue. The absorbed light energy is then converted into thermal energy, causing thermoelastic expansion of nearby tissue to emit ultrasonic waves. The much lower scattering effect of sound than light (two to three orders of magnitude) in tissues contributes to greater penetration with a scalable spatial resolution than that of optical microscopy (6). PAT has a relative sensitivity to minor variations in optical absorption, which is limited by incident laser fluence, absorption characteristics of the target, and detection efficiency of the ultrasonic transducer. With optimization of these conditions, thermal noise from the detection system and in the medium becomes the fundamental limit. Furthermore, exogenous contrast agents can characterize early-stage tumor markers and changes in tumor microenvironments. There are three main advantages of PAT, which can be summarized as: (a) high-resolution images at depths from microscopic to macroscopic scales as compared with pure optical imaging; (b) noninvasive structural and functional imaging by using endogenous or exogenous contrast agents; and (c) high compatibility with US imaging, fluorescence imaging, and optical coherence tomography. These advantages promise to expand applications of PAT in clinical imaging.

To date, PAT is playing an increasingly important role in the fields of life science and medicine. Reviewing the basic principles, current applications, and promising breakthroughs will inform instrument engineers, probe chemists, application users, and radiologists about potential applications of the technology. In this review, we categorize different PAT instruments and discuss key application trends in early-stage cancer detection from bench to bedside. We also conclude with complementary functions of PAT to other existing imaging modalities, as well as the possibility of widespread clinical applications.

Principles and Representative Instruments of PAT

Signal Generation and Detection

Although there are diverse applications of PAT, the generation and detection mechanisms of photoacoustic signals are consistent. At first, a pulsed laser is usually used to excite a transient wideband ultrasound wave. A simplified formula illustrates this effect (6):

where P₀ is defined as the generated initial acoustic pressure, Γ is the Grüneisen parameter (composed of a series of inherent parameters of the tissue), n_th is the percentage of absorbed light energy that is converted into heat, μ_a is the tissue’s absorption coefficient, and F is the incident light energy. The formula illustrates that P₀ is related to thermal, mechanical, and optical properties of substances. However, it is generally assumed that the optical absorption of tissue components is the dominant factor. Therefore, PAT has a relatively high detection sensitivity to optical absorption, which means that the small percentage changes of optical absorption coefficient leads to variations in photoacoustic amplitude.

The initial instantaneous photoacoustic signal is considered as the sound source and further transmitted omnidirectionally. The acoustic signal that the biologic tissue reflects is detected by an ultrasonic probe or transducer array. Finally, a three-dimensional image that reflects the absorption characteristic distribution inside the tissue is reconstructed. All tissue components have original absorption spectrum for photoacoustic differentiation. For example, in the visible spectral range (400–700 nm), hemoglobin and melanin have dominant optical absorption coefficients that are orders of magnitude larger than those of other absorption tissue components (collagen, lipids, and water). However, water and lipids have characteristic absorption peaks in the near-infrared region (900–1800 nm), which can also be imaged with suitable wavelengths. Accordingly, PAT is able to measure vascular parameters (diameter, density, speed of blood flow, etc) and melanin distribution without exogenous labels. Moreover, the tissue oxygenation can be calculated by separating the photoacoustic contributions of oxygenated hemoglobin and deoxygenated hemoglobin with multiple wavelengths illumination, which is then obtained by spectral-unmixing techniques (7). To date, there are three main imaging modalities of photoacoustic tomography in biomedical imaging, which are categorized as photoacoustic microscopy, photoacoustic endoscopy, and photoacoustic CT.

Photoacoustic Microscopy and Photoacoustic Endoscopy

Compared with conventional optical microscopy, photoacoustic microscopy accomplishes label-free imaging in biologic tissues without endogenous contrast agents. At the microscopic level, depending on the resolution, photoacoustic microscopy is separated into optical-resolution photoacoustic microscopy, acoustic-resolution photoacoustic microscopy, and photoacoustic endoscopy.

Optical-resolution photoacoustic microscopy achieves lateral optical diffraction-limited resolution by focusing the excitation light. Because a laser pulse must be emitted at each scanning position, the imaging speed of the optical-resolution photoacoustic microscopy is always limited by the laser repetition rate and raster scanning speed. Conventional photoacoustic microscopy obtains images by raster (line-by-line) scanning, which accounts for prolonged imaging times with large field of view imaging.

With the help of a water-immersible microelectromechanical system scanning mirror, the volumetric imaging speed has been substantially increased compared with raster scanning. Lin et al (8) developed a high-speed optical-resolution photoacoustic microscopy system as shown in Figure 1, A, which provided a B-scan rate of 400 Hz over a 3-mm scanning range at 3-mm lateral resolution. Recently, they improved their system to a B-scan rate of 900 Hz over a 12-mm scanning range at 10-mm lateral resolution (9). The three-dimensional imaging speed of 1 × 1-mm² region is at least 10 times faster than their microelectromechanical system–based photoacoustic microscopy system. An ultraviolet photoacoustic microscopy system enabled label-free, specific imaging of fixed, unprocessed breast tissue and achieved multilayered histology-like imaging of the tissue surface (10). The lateral imaging resolution was about 330 nm, approaching the diffraction limit. Miniaturization of probes is also an important issue to advance the technology. Recently, handheld optical-resolution photoacoustic microscopy was presented in 80 mm × 115 mm × 150 mm, which has potential for skin tumor applications (11). Additionally, Chen et al presented an ultracompact optical-resolution photoacoustic microscopy probe (Fig 1, B), which only weighs 20 g with a size of 22 mm × 30 mm × 13 mm (12). At present, photoacoustic microscopy is moving toward rapid imaging, large scale, high resolution, and probe miniaturization.

Figure 1:

The schematics of representative optical-resolution photoacoustic microscopy, acoustic-resolution photoacoustic microscopy, and photoacoustic endoscopy systems. A, High-speed optical-resolution photoacoustic microscopy with 3-mm lateral resolution and a millisecond-level cross-sectional imaging speed over a millimeter-level field of view (8). B, Optical-resolution photoacoustic microscopy system with a miniaturized probe (12). C, The schematic of 1064-nm acoustic-resolution photoacoustic microscopy; the US transducer is housed in a condenser for photoacoustic signal acquisition (13). D, A dual-view system increases the resolution isotropy of traditional acoustic-resolution photoacoustic microscopy system (14). E, An illustration of dual-mode, multiwavelength endoscopy. The system records and displays a set of dual wavelengths photoacoustic and US B-scan images in real time (15). F, A schematic representation of all-optical forward-viewing photoacoustic endoscopy probe with Fabry-Pérot (FP) sensor (16). AC = aluminum coating, AL = acoustic lens, AMP = amplifier, BC = beam combiner, CL = collimation lens, DAQ = data acquisition, FC = fiber collimator, FG = function generator, FL = focusing lens, MEMS = microelectromechanical systems, MS = motorized stage, PA = photoacoustic, PC = personal computer, UST = US transducer, UT = ultrasonic transducer. (Reprinted, with permission, from references 8,12–16.)

In acoustic-resolution photoacoustic microscopy, the lateral spatial resolution is determined by a relatively large area of acoustic focusing compared with optical-resolution photoacoustic microscopy, which enables more laser energy to reach deeper tissue. The penetration depth is greatly enhanced at several centimeters at the expense of spatial resolution. Thus, the axial resolution is always finer than lateral resolution. Near-infrared window II (1000–1700 nm) has been recently used, and a dark-field confocal 1064-nm acoustic-resolution PA microscopy system was developed, which achieved an imaging depth of 11 mm with a lateral resolution of 130 mm in Figure 1, C (13). Combined with the near-infrared laser source, acoustic-resolution photoacoustic microscopy enables high-resolution imaging at depth with much improvement in lateral resolution. Vienneau et al presented a dual-view acoustic-resolution photoacoustic microscopy system (Fig 1, D) that enhanced the resolution isotropy and improved image quality (14). Quantitatively, the left-view and right-view images offer lateral resolutions of 288 and 310 mm, and axial resolutions of 98 and 78 mm, respectively. The average resolution isotropy is 29% in a single view. In contrast, the dual-view image has a lateral resolution of 114 mm and an axial resolution of 108 mm, increasing the resolution isotropy from an average of 29% to 95% without introducing artifacts.

Photoacoustic endoscopy is considered as a miniature variant of photoacoustic microscopy for imaging internal organs, which has the potential for diagnosing early gastrointestinal diseases. Compared with clinical endoscopy, photoacoustic endoscopy provides additional functional information and is essentially compatible with US imaging, permitting multimodal imaging with complementary contrast. On the basis of these advantages, simultaneous photoacoustic and ultrasonic dual-mode endoscopy was developed for imaging internal organs in vivo (Fig 1, E) (15). The system presented a set of photoacoustic and US B-scan images in real time during 4-Hz rotation scanning. With the help of a mechanical device, the probe moves at a speed of approximately 200 µm per second, which allows three-dimensional imaging inside a blood vessel. Due to mechanical limitations of electrical US transducers, most previously reported photoacoustic endoscopy devices are based on side-viewing probes. An all-optical forward-viewing photoacoustic probe has been developed based on the Fabry-Pérot sensor (Fig 1, F) (16). One limitation of this system is long acquisition time for an image (approximately 25 minutes). In summary, photoacoustic endoscopy provides a label-free and functional imaging method for vascular and organ imaging with deeper imaging (more than 1 mm) than conventional optical endoscopy. However, the size of the probe needs to be further miniaturized (less than 1 mm) for more clinical applications.

Photoacoustic CT Imaging

Photoacoustic CT is the fastest growing and clinically popularized PAT modality. Photoacoustic CT is expected to be an optional imaging system for early screening of breast cancer in clinical medicine within the coming years. Different from mechanical raster scanning detection in photoacoustic microscopy, photoacoustic CT can simultaneously acquire ultrasonic waves by using multiple ultrasonic probes with different angles, realizing deeper volume imaging quickly. There are several common effective photoacoustic CT modalities as shown in Figure 2. The commercial bowl-shaped system (Nexus 128; Endra, Ann Arbor, Mich) is presented in Figure 2, A. A total of 128 single transducers at a center frequency of 5 MHz are spirally distributed on the inner surface of the hemispherical bowl to maximize acquisition of photoacoustic signals. The parallel acquisition system can simultaneously collect signals at different positions and then transfer data to a computer for further analysis (17).

Figure 2:

Representative schematic of various photoacoustic CT systems. A, A commercially available fully three-dimensional photoacoustic imaging system (17). B, A handheld photoacoustic and US dual-mode probe (18). C, A whole-body single-impulse panoramic photoacoustic CT with 512 elements full-ring acoustic detection (19). D, A typical photoacoustic CT system uses a two-dimensional Fabry-Pérot interferometer as photoacoustic signal detector (20). DAQ = data acquisition, FP = Fabry-Pérot, PA = photoacoustic. (Reprinted, with permission, from references 17–20.)

Another type of photoacoustic CT (Fig 2, B) (18), a linear-array US transducer−based handheld photoacoustic CT probe, has potential to adapt to different clinical scenarios. Two fiber bundles emit pulsed lasers from both sides, and the line-array ultrasonic probe in the middle can collect the generated photoacoustic signals in real time. Whole-body imaging of intact small animals is a critical step in preclinical studies of human disease and drug development. Single-impulse panoramic photoacoustic CT for small animals was developed by Li et al (19). With 512-element full-ring acoustic detection, the system achieved 125-mm in-plane resolution within 50 msec per frame (Fig 2, C). For photoacoustic CT systems based on a Fabry–Pérot interferometer as the acoustic sensor (Fig 2, D) (20), the photoacoustic waves are recorded by raster scanning of a probing laser beam over the surface of the interferometer. The spatial resolution can be well maintained with dense spatial sampling over the field of view. The imaging speed is limited by the point-by-point scanning of the probing beam.

Currently, many groups are working on various commercial photoacoustic CT systems that enable preclinical and clinical imaging. As the first Conformité Européenne–approved linear-array–based photoacoustic CT system, Imagio (Seno Medical Instruments, San Antonio, Tex) has been used in several prospective clinical trials in the United States. Vevo LAZR (Fujifilm VisualSonics, Toronto, Ontario, Canada) is equipped with probes across various frequencies (from 9 to 55 MHz), which are suitable for application requirements in small-animal imaging. Different from the linear-array–based probe, a cup-shaped handheld multispectral optoacoustic tomography probe is another commercially available preclinical system (iThera Medical, Munich, Germany). The Table shows the comparison of various typical modalities of photoacoustic CT.

Comparison of Various Typical Modalities of Photoacoustic Tomography

Endogenous Photoacoustic Contrast and Early-Stage Cancer Detection

Angiogenesis plays an essential role in cancer development, invasion, and metastasis, which results in a locally abnormal vascular morphology and increased density. Structural and/or functional changes of these vessels and associated deficits in oxygenation are crucial indicators of early-stage cancer and potential targets for therapy. PAT can specifically image hemoglobin based on its particular absorption spectrum, which is a powerful tool for label-free angiography. Quantitative analysis of the spatial distribution, morphologic changes, and density of vessels can be achieved from the high-contrast imaging results. In addition, functional features such as blood flow velocity and blood oxygen saturation can be obtained with PAT. The superb sensitivity and specificity of PAT provide a basis for the early-stage diagnosis and treatment evaluation of tumors.

Breast Cancer

According to estimated new cancer cases and deaths stratified by sex in United States (2018), breast cancer is the most commonly diagnosed cancer (32.9%) and the second leading cause of cancer death (14.3%) in women, following lung cancer (28.6%) (32). Currently, common clinical diagnostic imaging methods include mammography, US, MRI, PET, and SPECT. Among them, mammography is the first priority for current diagnosis. However, the high false-negative rate remains a challenge for this widely used technique (33). The imaging sensitivity of PAT is hardly affected in dense breast tissue, but the mammography sensitivity decreased from 70% (heterogeneously dense breasts) to 45% (extremely dense breasts) (34). These unsatisfactory rates indicate that more biopsies than necessary are performed in patients, which leads to increased health care costs and psychological stress for patients. Compared with other human tissues, breasts have lower vascular density and usually provide less light attenuation. The increase of vascular density and blood oxygen saturation change is a key marker for breast malignancies. Therefore, PAT holds great potential to overcome these limitations and be the universal method for screening breast cancer.

Early results for PAT in breast cancer screening are exciting. Methods for imaging breast cancer with US have been reported since 1999 (35). Since then, substantial progress has been made in clinical breast cancer screening with PAT. Oraevsky et al built a photoacoustic CT system dedicated to mammography, called LOUISA-3D, that achieved clear imaging of breast vessels and tumors simultaneously. The scan duration took approximately 10 minutes with “single pulse” illumination in 20 steps of each wavelength (Fig 3, A) (36). Total data acquisition time was limited by the selection of rotation step, number of wavelengths, laser pulse repetition rate, and pulse numbers in each position. In addition to imaging of vascular density, detecting relative changes of tissue components caused by tumors is another advantage of PAT for breast cancer screening. Diot et al performed a pilot study and made a precise separation analysis of four different tissue components of breast tissue with a handheld device for multispectral optoacoustic tomography (29). By analyzing changes, such as the ratio of the different forms of hemoglobin and relative blood oxygenation, the spatial heterogeneity of a breast tumor can be further quantified. Heijblom et al utilized the Twente photoacoustic mammoscope (University of Twente, Enschede, the Netherlands) to image breast cancer in 31 patients (with a total of 33 malignant tumors) with representative images as shown in Figure 3, B. The results showed that 32 of 33 malignant tumors could be visualized with high contrast and small size deviation regardless of the density of breast tissue (37). The data acquisition time was determined by selection of their spiral scanning pattern (maximum radius varies from 24 to 96 mm). More laser pulses and increased data acquisition time were needed for the large spiral scanning pattern. In this work, a large field of view (9 × 8 cm²) was imaged in 10 minutes. In a recent study, volumetric imaging of human breast cancer was achieved in 15 seconds by using a single-breath-hold photoacoustic CT system (28). The depth-encoded photoacoustic angiograms of cancerous breasts were clearly revealed with detailed angiographic structures as shown in Figure 3, C. The improved imaging depth reached 4 cm with a spatial resolution of approximately 255 mm and a frame rate of approximately 10 Hz. This system can potentially monitor the therapeutic response to neoadjuvant chemotherapy by acquiring information similar to that of contrast material–enhanced MRI, yet with finer spatial resolution, and higher imaging speed. Additionally, Wong et al demonstrated the first label-free ultraviolet photoacoustic microscopy, using ultraviolet light for histology-like imaging without staining of breast tissue (10). This result shows that fast, multilayered histology-like imaging can be performed intraoperatively, enabling immediate-directed re-excision and reducing second surgeries.

Figure 3:

Various breast cancer detection images by using photoacoustic tomography. A, Example of image where both the blood vessels and the tumor region in the breast can be detected (36). B, Example of breast imaging with photoacoustic mammoscope; the location of the lesion in the photoacoustic volume was perfectly colocalized with the mammography result. The three-dimensional photoacoustic volume crossing region of interest is also shown (37). C, Left to right: Mammograms of the affected breasts, depth-encoded photoacoustic angiograms of whole cancerous breasts, sagittal plane image across tumor region; higher blood vessel densities region was associated with tumors (dotted circle) (28). RCC = right craniocaudal, RML = right mediolateral. (Reprinted, with permission, from references 28,36,37.)

Melanoma

Skin and subcutaneous tissue are perhaps the most easily detectable targets for PAT. Melanoma is the most aggressive type of skin cancer and usually associated with approximately 75% of skin cancer–related deaths (38). Melanoma initially grows along with the epidermis at an early stage and progressively invades into the dermis (39). The vertical depth of melanoma is the most relevant factor for prognosis (40). Early detection of malignant melanoma and subsequent precise surgical resection reduce mortality from this deadly cancer. Current clinical standard treatment of melanoma requires full surgical removal of the tumor. Dermatoscopy is a commonly used clinical measurement device for examining melanoma, but the diagnostic accuracy is mostly limited by clinician’s experience and the patient’s skin color (41). Resecting the tumor region while retaining the adjacent healthy tissue will reduce the cost of surgery and improve patient comfort. Noninvasive quantitative depth assessment imaging modalities are urgently needed for appropriate surgical treatment.

The promising skin cancer detection and quantification abilities of PAT have already been exhibited in several animal models. For instance, Zhou et al utilized handheld photoacoustic microscopy to detect the boundary of melanoma and determine the depth of tumor invasion in nude mice (Fig 4, A) (42). Thicknesses of 4.1 mm and 3.7 mm were successfully detected in phantom and in vivo melanoma models, respectively (42). Additional clinical applications are continually being explored. Kim et al reported results on a resected cutaneous melanoma lesion that was approximately 1.5 square inches in size from the heel of a patient. The three-dimensional reconstruction photoacoustic image of the lesion matched well with the pathologic examination (Fig 4, B) (43). Furthermore, Zhou et al (44) extended their work to patients with melanoma by a linear-array–based handheld photoacoustic probe. The cutaneous PAT-based depth (0.62 mm) is more accurate than the provisional incisional biopsy depth (0.48 mm) according to histologic findings (0.78 mm). Another promising application in human melanoma is to detect the metastatic status of sentinel lymph nodes. In a first-in-human study, multispectral optoacoustic tomography was used to study sentinel lymph nodes ex vivo and in vivo in patients with melanoma (45). The metastatic detection rate was significantly improved compared with conventional protocols (22.9% vs 14.2%). The detection depth of sentinel lymph nodes was further improved to 5 cm with indocyanine green contrast agents. However, a high false-positive rate caused by other endogenous absorbers still remains a problem to be further investigated.

Figure 4:

Representative photoacoustic tomographic images for melanoma diagnosis in a, A, preclinical and, B, clinical sample. A, Photoacoustic (PA) microscopic images of melanoma in a nude mouse model clearly show both the top and bottom boundaries of the tumor. The red dashed line outline of the tumor shows 3.66-mm depth (42). B, A resected cutaneous melanoma lesion from a patient and its three-dimensional reconstruction photoacoustic image. The melanoma regions are represented by dark red to bright yellow color (arrowheads), and the marking-pen regions are represented by dark green to bright green color (arrows) (43). (Reprinted, with permission, from references 42,43.)

These pilot studies support development of PAT as a preoperative screening tool to guide skin cancer resection. Although there are still technical limitations and a lack of clinical equipment standardization, a PAT system integrating portable, tunable multiwavelength, fast scanning can further expand to a wide variety of preclinical and clinical skin cancer applications.

Prostate Cancer

Prostate cancer has the highest number of new cases and is the second leading cause of cancer death in men (46). Although there are many diagnostic methods (tissue biopsy, digital rectal examination, prostate-specific antigen detection, transrectal US imaging, MRI, etc) for prostate cancer, there are still high false-positive and false-negative rates (47). Therefore, development of methods for diagnosing prostate cancer earlier and more accurately is still an area of investigation.

With structural, functional, and molecular imaging characteristics, PAT is expected to be an effective imaging modality for the early diagnosis and treatment evaluation of prostate cancer. In an ex vivo prostate cancer study (48), Dogra et al utilized a multispectral PAT system to differentiate malignant prostate tissue, benign prostatic hyperplasia, and normal human prostate tissue from excised prostate sections from 30 patients. Using four wavelengths, the tissue components of deoxyhemoglobin, oxyhemoglobin, lipid, and water were extracted by chromophore analysis on photoacoustic images as shown in Figure 5, B. A total of 13 of 16 malignant and 25 of 26 nonmalignant prostate samples were detected correctly. Additionally, there was a statistical difference between malignant tumors and nonmalignant tumors in mean intensity of deoxyhemoglobin. Another in vivo pilot study verified the feasibility of PAT for angiogenesis of prostate cancer (31). Three index tumors in patients were first recognized with MRI and also identified with transrectal US of a hypoechoic lesion. The intensity of photoacoustic signals corresponded to vascular parameters (vessel length and vascular density) of histopathologic results in normal prostatic tissues and index tumors.

Figure 5:

Representative clinical images of ex vivo prostate cancer tissue from a multispectral photoacoustic tomography system. A, Photograph and histopathologic findings of prostate cancer (malignant region in yellow circle). B, Photoacoustic image acquired at 760 and 850 nm (first row); photoacoustic images of deoxyhemoglobin (dHb) and oxyhemoglobin (Hbo₂). (Reprinted, with permission, from reference 48.)

The applications of PAT are not limited to the above-mentioned scenarios. PAT is also being applied in thyroid cancer (49,50), ovarian cancer (51) and cervical cancer (52) detection, moving toward clinical translation. Therefore, the development of standardized, commercial, and user-friendly clinical PAT systems will further broaden applications.

Sensitivity and Function Enhancement with Exogenous Photoacoustic Contrast Agents

Oxygenation in Tumor Microenvironments

Monitoring biodistribution of oxygen in living organisms is of great significance for disease diagnosis and treatment (53). Tumor hypoxia is a key characteristic of most tumor microenvironments because of dysregulated tumor metabolism. Methylene blue is a water-soluble dye with Food and Drug Administration (FDA) approval that has been used as an effective photosensitizer for clinical photodynamic therapy. Methylene blue is also a potential oxygen-sensitive contrast agent that plays an important role in photoacoustic sensing of tissue oxygen partial pressure. Photoacoustic lifetime imaging technique was used for three-dimensional oxygen measurements in conjunction with methylene blue (Fig 6, A) (54). The method enables the noninvasive measurement of oxygen levels within the physiologic partial pressure of oxygen range (0–100 mm Hg). Given associations of hypoxia with aggressiveness of cancer and resistance to therapy, translating this technique to clinical applications promises to improve noninvasive analysis of prognosis and treatment efficacy.

Figure 6:

A, Photoacoustic lifetime sensing of oxygen partial pressure. B, In vivo photoacoustic tomographic images of the sentinel lymph node and needle (54). C, In vivo photoacoustic monitoring of brain injury and rehabilitation by modified Prussian blue particles–labeled mesenchymal stem cells (56). D, Photoacoustic and MRI of all resected nodes after superparamagnetic iron oxide (SPIO) nanoparticle injection. E, F, Photoacoustic-guided surgery with indocyanine green–SPIO nanoparticle clusters (59). BMSC = bone mesenchymal stem cells, NIR = near infrared, PA = photoacoustic, PBP = Prussian blue particle. (Reprinted, with permission, from references 54,56,59.)

Visualization of Biologic Processes

Visualizing cellular and biologic processes is one of the most sought-after abilities in biomedical research. Photoacoustic molecular imaging enables noninvasive sensing of various disease processes in conjunction with exogenous photoacoustic contrast agents (55). Like methylene blue, another effective FDA-approved Prussian blue particle was designed to label stem cells (Fig 6, B and C). Photoacoustic molecular imaging was applied to noninvasively visualize traumatic brain injury and treatment by injecting labeled cells. The process of stem cell homing and therapeutic effect was monitored by using noninvasive photoacoustic imaging (56). Similar methods could be employed to visualize trafficking of cell-based therapies for cancer, such as chimeric antigen receptor T cells. In addition, photoacoustic microscopy is also used as a powerful tool for drug delivery and chemotherapy response monitoring (57).

Disease Detection and Intraoperative Staging

Capitalizing on a wide range of exogenous contrast agents, photoacoustic imaging can detect subtle lesions at multiscale, providing multiparameter information for diagnosis, staging, and treatment of diseases. Accurate intraoperative staging of metastatic malignancies is essential for disease detection, especially for lymphadenopathy.

Superparamagnetic iron oxide (SPIO) nanoparticles have been used as a sensitive photoacoustic agent for lymph node staging, which showed a clear distinction between normal nodes and metastatic nodes (58). In a preclinical mouse model, macrophage activation and immune response were accomplished by injection of incomplete Freund adjuvant. SPIO nanoparticles were then injected in the footpads of hind legs 7 days later. All nodes were excised for photoacoustic imaging. Photoacoustic images of lymph nodes illustrated the clear signal of SPIO nanoparticles in the periphery of the lymph nodes, while no signal increase was shown in the control group. MR images demonstrated a strong correlation with photoacoustic results (Fig 6, D) (58). These experimental results illustrated that SPIO nanoparticles can be used for lymph node staging and analysis in metastatic malignancies. The research creates potential opportunities for intraoperative lymph nodal staging to enhance surgical decision making.

Tumor Resection and Surgical Navigation

Complete resection is important to reduce brain tumor recurrence and raise survival rates. Preoperative detection and intraoperative navigation play important roles in accomplishing total surgical resection. Sensitive translatable contrast agents and accurate multimodal imaging technologies have attracted immense attention from scientists and surgeons. Indocyanine green and SPIO nanoparticles, two FDA-approved compounds, were assembled into multifunctional clusters (indocyanine green–coated SPIO-nanoparticle clusters) for MRI-guided preoperative detection and photoacoustic-guided intraoperative resection (59). The contrast-enhanced tumor and adequate margins can be delineated clearly based on MRI and photoacoustic imaging after the administration of the multifunctional clusters (Fig 6, D–F). Figure 6, D, illustrates that multifunctional clusters could be used for photoacoustic imaging–guided resection to detect residual tumor cells. Compared with animals with microscopic-guided resection, animals with photoacoustic imaging–guided resection displayed increased survival.

Conclusion

Accurate localization and characterization of tumors are crucial to provide timely intervention and effective treatment. PAT is suitable for multidimensional imaging of anatomy, molecules, metabolism, and genetics in biologic systems, enabling disease diagnosis from macroscopic to microscopic scales. PAT obtains single or multiple parameters, information about structure and function, and qualitative and quantitative images. New imaging methods such as PAT promise to improve early detection of cancers with resultant improvements in survival. Bench-to-bedside breakthrough in PAT promise a broad range of clinical applications for this technology in cancer and other diseases.