The importance of phantoms in optical imaging systems derives from their significant role in performance evaluation, calibration, quality control, comparison and standardization, and troubleshooting. Essentially, phantoms are standardized tools used to verify and evaluate the performance of optical imaging systems, ensuring that they operate with high accuracy based on optical-image analysis. First, phantoms are essential for the performance evaluation of optical imaging systems [1–5]. Using phantoms, one can objectively measure an image’s resolution, accuracy, sensitivity, and contrast [1, 3, 5–11]. This is vital for confirming that a system operates as designed and provides reliable data for research or clinical diagnostics [6, 8, 12–19]. For instance, in medical imaging systems phantoms are used to evaluate imaging performance, ensuring patient image quality [8, 20–24]. Second, phantoms are used to calibrate a system’s measurements to match standard values. This ensures that the system provides accurate data [25–31], essential for obtaining reliable results in research or clinical settings. This calibration process is critical in scientific research or medicine, where precise measurements are required. Third, phantoms can be used to compare and standardize the performance of different measurement systems. By using phantoms, the performance of various diagnostic devices can be evaluated based on the same standards, making it possible to compare and analyze results from optical images. This process enhances research consistency and helps to obtain reliable data in environments where multiple systems are used.

In this review, we discuss the recent progress and remaining challenges of phantoms used in advanced optical measurement systems, aiming to advance phantom research. Specifically, we focus on optical coherence tomography (OCT), photoacoustic imaging (PAI), digital holographic microscopy (DHM), optical diffraction tomography (ODT), and oximetry tools. In Chapter II, we briefly introduce the fundamental operating principles of each technology, and then highlight the key factors for phantom development and evaluation criteria. In each first subchapter of Chapters III–Ⅳ, we summarize the current status of phantoms for performance evaluation of OCT and PAI. In each second subchapter of Chapters III–Ⅳ, a review of the key factors driving phantom technology guides us to recent developments in creating realistic biological tissue (or cell)-mimicking phantoms. In Chapter V, we introduce the phantom to validate the accuracy of DHM and ODT, such as spatial resolution and refractive properties. In Chapter VI, we cover the utmost recent research advances in phantom development for oximetry tools, including examples of how these phantoms have been applied in practical biomedical and clinical studies.

OCT is an imaging technology that uses low-coherence light to obtain high-resolution three-dimensional (3D) images within an optical-scattering medium, such as biological tissue [1–5, 32–40]. Because OCT’s image acquisition is fast, real-time imaging is possible, and measurements can be performed noninvasively. Various applications of OCT have been developed by analyzing scattering changes in the sample, with representative examples including OCT angiography (OCTA) [6, 41, 42], which shows blood vessels, and optical coherence elastography (OCE) [36], which can measure tissue elasticity. Due to these advantages, OCT is widely used as a diagnostic and monitoring tool, particularly in ophthalmology, and shows promise as a diagnostic tool in dermatology and dentistry. Beyond clinical applications, it is used in industrial applications such as preserving art, detecting defects in materials, and identifying foreign substances in displays, and can be optimized to suit each requirement.

PAI is a hybrid imaging technique that combines optical and ultrasound imaging concepts [43]. It relies on the photoacoustic effect, where absorbed optical energy (usually from a pulsed laser) is converted into ultrasonic waves through thermoelastic expansion. These ultrasonic waves are then detected and used to generate images of the tissue or material under study [22]. The rapid growth of hardware and software for photoacoustic technologies has facilitated the establishment of dedicated tools for standardization and performance assessment of the system. Phantoms play a significant role in the development and evaluation of PAI, by providing flexible ways to simulate the complex optical and acoustic properties of biological tissues.

DHM and ODT are optical 3D imaging systems that reconstruct spatial phase information through 2D interference patterns [44, 45]. DHM leverages holography principles to produce high-resolution, three-dimensional images of microscopic structures [46–50]. A significant advantage of DHM is its ability to perform quantitative phase imaging, measuring changes in optical path length to determine sample thickness and refractive-index variations. DHM is noninvasive, label-free, and suitable for real-time applications, making it ideal for live-cell imaging and dynamic process monitoring. It is also employed in materials science and industrial inspection, to characterize surface profiles and detect defects. ODT is a tomographic technique that reconstructs a sample’s three-dimensional refractive-index distribution by analyzing light diffraction [25, 45, 51, 52]. By capturing multiple 2D images from various angles around the sample and applying inverse-scattering algorithms, ODT reconstructs the 3D structure. This technique provides high-resolution, quantitative imaging of transparent or semitransparent samples, making it possible to study biological cells and tissues without labeling [27–31].

Oximetry is a noninvasive technique for measuring blood oxygen saturation [53–57]. It quantifies oxygen levels in real time by exploiting the differential absorption of light by oxygenated and deoxygenated hemoglobin at specific wavelengths. This method is essential in clinical settings for monitoring respiratory and circulatory status, and has also been widely used in physiological research and sports science.

OCT phantoms have been developed to measure changes in intensity of light that has been back-reflected and back-scattered from the internal microstructure of objects. They are made by incorporating nanoparticles or microparticles (below the resolution unit) into base materials such as hydrogels (agar and gelatin), polydimethylsiloxane (PDMS), silicone, or polymers (epoxy), which do not have inherent optical-scattering properties. Additionally, to validate a dedicated OCT system, the level of rigidity can be modified, or a microtube can be inserted to provide fluid flow. The fabrication of these phantoms involves using spin coating or various molds designed for characterization of the target sample.

3.1. Phantom for OCT Image-quality Test and Calibration

The phantoms developed for OCT system verification provide essential benchmark measurements that can quantify the quality of measured OCT systems and the results of functional OCT measurements. The phantom for fundamental verification parameters of OCT systems defines the depth-dependent spatial resolution (axial and lateral resolution) and image contrast, as shown in Fig. 1. Figure 1(a) shows a phantom fabricated by mixing nanoshell particles in UV-cured epoxy, and a comparison of the evaluation of four OCT systems using a single-layer phantom [1]. The spatial resolution at the image volume was investigated, and the optimization of the system was determined using the observed values. Additionally, single-layer or multilayer thin-film phantoms have been fabricated, to measure axial resolution and verify the system performance within a specific depth range [3–5]. The homogeneous single-layer phantom, manufactured with uniform scattering concentration, enables analysis across the entire depth range. Moreover, it exhibits high reliability and is thus suitable as a standard for comparing OCT systems, or image-reconstruction algorithms. Figure 1(b) shows a multilayer phantom consisting of bright and dark layers that resemble the bars of the USAF 1951 resolution chart [3]. The fabrication of multilayered phantoms requires achieving consistent scattering properties across each layer, while also ensuring a specified thickness. To produce microscale multilayer structures, the spin-coating technique has been widely used to deposit mixtures with different concentrations repeatedly. Six phantoms of varying thickness were made by using PDMS and barium sulfate powder. Phantoms consisting of nanoparticles or multiple layers have been utilized to verify the spatial resolution of an OCT system. However, these phantoms were insufficient to verify 3D volume data. Therefore, some phantoms with 3D structures were developed by manufacturing molds to match the target of each system [32–34]. Structured 3D phantoms can quantitatively evaluate spatial resolution, speckle contrast, sensitivity across the imaging volume, and imaging distortion. In addition, a laser-inscribed phantom with a laser-engraved pattern to evaluate OCT imaging performance was introduced [34]. Using a laser-inscribed phantom, the SNR was calculated and compared in depth to evaluate the sensitivity region of OCT systems at different wavelengths. To measure the spatial resolution and 3D image analysis of endoscopic OCT, a cylindrical phantom was fabricated [35]. A common square phantom is not suitable for evaluating endoscopic OCT, due to the high-speed rotation of the probe during circular scanning, which can cause nonuniform rotational distortion. Therefore, a cylindrical phantom was created using a cylindrical mold with PDMS containing gold microspheres, for endoscopic OCT evaluation.

Figure 1. Phantoms of optical coherence tomography (OCT) for image calibration and functional quality test. (a) The OCT image and graphs of a single-layer phantom. The nano phantom consists of ultra violet (UV)-curing epoxy and nano-shells to compare the performance of the four OCT systems. Reprinted with permission from A. Fouad et al. Biomed. Opt. Express [1]. Copyright © 2014, Optica Publishing Group. (b) Schematic and OCT images of multi-layer phantoms. Reprinted with permission from A. Agrawal et al. Biomed. Opt. Express [3]. Copyright © 2013, Optica Publishing Group. (c) Wave propagation pattern of phantom from optical coherence elastography. Reprinted from S. Song et al. J. Biomed. Opt. 2013; 18; 21509. Copyright © 2013, SPIE [60]. (d) Spectroscopic OCT image and map of spectroscopic metrics of the phantom. See the main text for details. Reprinted with permission from V. Jaedicke et al. Biomed. Opt. Express [37]. Copyright © 2013, Optica Publishing Group.

The development of various functional extensions of OCT, such as OCTA and OCE, has enabled the measurement of information related to blood vessels, blood flow, and tissue specificity, beyond offering morphological information about the sample. Accordingly, a phantom for functional OCT was also developed to assess the specific performance of these systems. A tube phantom was used to verify the OCTA algorithm [41]. To evaluate OCTA and Doppler OCT systems, which are representative functions of the OCT system, a tube is used to represent blood vessels, and a syringe pump is used to generate a controlled flow within the tube. Additionally, a microtube phantom was used to evaluate the performance of microbubbles or contrast agents in enhancing contrast [12, 58, 59]. Phantoms with tubes are simple to manufacture and are suitable for evaluating the performance of a system, depending on the fluid and flow rate. Figure 1(c) shows wave-propagation-pattern images of phantoms with agar concentrations of 0.5%, 0.75%, 1%, and 2%, obtained by OCE [60]. The shear wave is measured differently depending on the concentration of agar, which indicates that the stiffness of the phantom is different. These phantoms can provide measurement estimates for comparing the relative values of shear modulus between different materials. Shear waves are measured using phase-sensitive OCT or OCE to evaluate tissue elasticity. Phantoms with various concentrations of agar/gelatin were used to mimic tissues, and impulse generators (stimulation, piezoelectric actuators, etc.) were used to generate shear waves [36, 60, 61]. Moreover, if pieces of agar with different concentrations are embedded in the phantom, disease or specific parts can be mimicked.

Figure 1(d) shows the structure of the phantom and the areas that were quantitatively analyzed by spectroscopic OCT (SOCT). SOCT can analyze the optical-scattering properties of tissue by quantifying depth-resolved spectra [37]. A phantom with different scattering properties was developed, to compare the measurements and quantitative methods of the SOCT system. Dry-form microspheres 1 µm and 3 µm in diameter were embedded in silicone foils with a thickness of 100 µm and scattering foils are stacked to create phantoms with a variety of structures. By evaluating different combinations of analysis methods on the same phantom, the clustering accuracies of the algorithms were able to be quantitatively compared.

3.2. Tissue-mimicking OCT Phantom

OCT is widely acknowledged as a significant tool in the field of tissue engineering, due to its ability to acquire and analyze the three-dimensional developmental processes of targeted tissues or organs. Technical verification and feasibility testing of the developed system are necessary steps for measuring and analyzing targeted tissues. Therefore, it is recommended to use a realistic phantom that accurately mimics an organizational structure and provides consistent results. Phantoms that mimic tissues for OCT include skin, fingerprints, cerebral cortex, esophagus, eye (cornea, retina, retinal cone), bladder, colon, and artery phantoms [6, 12, 39, 40, 42, 62–74].

As ophthalmology is the largest field of clinical application for OCT, there are many phantoms that mimic the cornea and retina. A corneal phantom was developed to verify the measurement of the corneal layer thickness of various commercial systems, and for the convenience of developers and field users (R/D engineers and field-service technicians) [63]. It was manufactured considering the anatomical structure and thickness of the cornea, and the optical properties of the cornea were imitated using polysiloxane and polymethyl methacrylate (PMMA) acrylic materials. Retinal phantoms have been developed to mimic the multiple layers of the retina and include its morphological and functional characteristics.

Figure 2(a) shows a phantom exhibiting a bullseye pattern, and another phantom mimicking multiple layers of the retina [64]. These phantoms have been fabricated using 3D-printing technology. The bullseye-pattern phantom can verify a large-field-angle system and measure the system’s distortions and field of view (FOV). The retina-mimicking phantom consisted of scattering layers 60 µm and 120 µm thick, and was used to verify the thickness distortion of the OCT. To mimic a more realistic retinal shape, phantoms representing the foveal fit or optic nerve have been developed, using laser etching or various molds [65–67]. In the layering process, glass-bead molds were used to mimic the tapering form of the fovea fit of the retina, and a phantom was developed using PDMS and TiO₂ [66]. Furthermore, retinal phantoms that exhibit retinal detachment and drusen disease were also fabricated, as shown in Fig. 2(b). The retinal phantom was introduced to replicate cone cells in the outer segment (OS) for a high-resolution adaptive-optics optical coherence tomography (AO-OCT) system [68]. Titanium dioxide (TiO₂)–doped photomaterial was affixed to a microfluidic channel to mimic the OS region of human retinal cone cells, and AO-OCT images of phantoms with different concentrations were obtained. A phantom was also developed to express not only the morphology but also the function of the retina by mimicking the retinal vessels and lipofuscin in the retinal pigment epithelium (RPE) layer, as shown in Fig. 2(c) [6]. The main material of this phantom is a mixture of PDMS and TiO₂, and such mixtures were laminated through spin coating to mimic retinal layers. The channels were fabricated through a photolithography process, and fluorescent microbeads were used to mimic autofluorescence. Because fluid flowed through the phantom channel, OCTA evaluation was possible. The curvature of the retina was also reproduced using a curvature mold. Therefore, images with a curvature ratio comparable to the actual retina could be obtained without the need for lens correction.

Figure 2. Phantoms of mimicking eye. (a) Schematic and optical coherence tomography (OCT) images of two phantoms. Reprinted from A. Corcoran et al. J. Mod. Opt. 2015; 62; 1828–1838. Copyright © 2015, Taylor & Francis [64]. (b) Picture, OCT image, enface OCT-angiography, and fluorescence anigiography of mimicking eye phantom. Reprinted from H.-J. Lee et al. Proc. SPIE XVII; Copyright ©2024, SPIE [6]. (c) Illustration images and OCT images of mimicking eye phantom. Reprinted from G. C. F. Lee et al. J. Biomed. Opt. 2015; 20; 085004. Copyright ©2015, SPIE [66].

The development of endoscopic OCT has made imaging of internal organs possible, and has shown promise as an assistive technology to white-light endoscopy. The phantom for endoscopic OCT imaging has a three-dimensional shape and represents healthy and diseased tissues. Phantoms were developed to mimic the internal organs of the bladder and colon [69–71]. Figure 3(a) shows a bladder phantom formed in 3D [70]. The phantom was shaped using a 3D-printing mold, and optical scattering of tissue was expressed using PDMS and TiO₂. For white-light imaging, ink was applied to the surface of PDMS to mimic blood vessels, and spin-coating technology was used to mimic the multilayers of the bladder. The healthy and diseased phantoms were clearly distinguished through OCT images. The colon phantom was fabricated using a mixture of PDMS with different concentrations of TiO₂, and dysplastic lesions were represented in 3D shape by using a mold, as shown in Fig. 3(b) [71]. Layered with different scattering intensities to represent the muscle layer, submucosa, and mucosa of the colon, the fabricated rectangular layer was fixed to mimic the colon-circumference geometry, and a block mimicking a dysplastic lesion such as an adenoma was attached. An OCT catheter was placed on the phantom surface by verifying its placement using a video of white light, and images of the healthy and diseased phantoms were obtained.

Figure 3. Phantoms mimicking the bladder, colon, and artery tissue. (a) Picture and optical coherence tomography (OCT) images of bladder phantom. Reprinted from K. L. Lurie et al. J. Biomed. Opt. 2014; 19; 036009. Copyright © 2014, SPIE [70]. (b) Picture and endoscopic-OCT images of colon phantom. Reprinted with permission from N. Zulina et al. Biomed. Opt. Express [71]. Copyright © 2021, Optica Publishing Group. (c) Picture, OCT, and Doppler image of phantoms. Reprinted from N. R. Munce et al. J. Biomed. Opt. 2010; 15; 011103. Copyright © 2010, SPIE [72]. (d) Schematic and intravascular-OCT images of artery phantom. Reprinted with permission from C.-É. Bisaillon et al. Biomed. Opt. Express [73]. Copyright © 2011, Optica Publishing Group and from C.-É. Bisaillon and G. Lamouche, J. Biomed. Opt. 2013; 18; 096010. Copyright © 2013, SPIE [74].

Intravascular OCT (IVOCT) is a great tool for visualizing healthy or diseased arterial tissue without perforating the arterial wall. Intravascular OCT is also useful for diagnosing arterial plaques that are prone to rupture. Furthermore, the incorporation of Doppler OCT enables the identification of blood flow. Figure 3(c) shows an arterial-occlusion phantom and a narrowed-blood-vessel phantom, using a mixture of PDMS with TiO₂ in a Teflon^TM or polycarbonate tube [72]. The flow of fluid within the phantom was visualized by flowing a 1% mixture of Intralipid and saline through a syringe pump. Doppler OCT allowed visualization of various fluid flows, and showed the potential to identify the region of arterial lesions. Healthy coronary arteries are composed of three structures: intima, media, and adventitia, with the intima being the thinnest. The diseased-artery phantom mimicked lipid plaques by using a flat groove of the shaft, as shown in Fig. 3(d) [73, 74]. A mixture of PDMS with alumina powder was used to express the artery’s optical scattering and a rotating shaft was used to stack a multilayer artery phantom. IVOCT clearly visualized phantoms of normal and diseased arteries, thereby demonstrating its potential for quantitatively characterizing the optical properties of plaque.

Chang et al. [38] demonstrated a birefringent tissue phantom for polarization-sensitive optical coherence tomography (PS-OCT). PS-OCT requires a phantom that can perform 3D imaging with birefringence properties. To verify the PS-OCT, a phantom that mimics the cross-sectional structure and birefringence properties of healthy and diseased bladder tissue was developed. Variations in birefringence were confirmed depending on the curing ratio of PDMS, and the concentration of the scattering agent (TiO₂) was adjusted to mimic tissue. OCT intensity and retardation images of normal and diseased phantoms were obtained, and it was possible to distinguish between normal and diseased regions.

PAI phantoms have comprehensive criteria that evaluate image quality, quantitative accuracy, penetration depth, reproducibility, and reliability. First, the factors for evaluating image quality include spatial resolution, contrast, and signal-to-noise ratio. Especially for PAI systems, it is essential to evaluate the systems that can distinguish between different tissue types and detect tiny details, such as arteries or tumors. Second, the factors that evaluate quantitative accuracy refer to the ability to accurately quantify physiological parameters, such as oxygen saturation, blood volume, and tissue composition. Thus, phantoms with these properties are useful for validating these quantitative factors. Third, assessment of penetration depth is important because a PAI system’s ability to measure at sufficient depths within biological tissues is critical. Finally, in terms of reproducibility and reliability, phantoms enable analysis of the stability of the laser source, detector sensitivity, and overall system robustness.

4.1. Evaluation of Photoacoustic Imaging Systems

The phantom for evaluating PAI performance is shown in Fig. 4. Ratto et al. [7] developed a multilayered phantom containing indocyanine green (ICG), gold nanorod (GNR), and red blood cell (RBC) absorbers to mimic skin, fat, fibroglandular tissue, and blood vessels, and assessed the system’s multiwavelength and multimodal imaging capabilities [24]. The tissue-mimicking materials were molded into the breast’s expected shapes. This phantom enable comprehensive evaluation of various aspects of PAI systems through methods such as confocal fluorescence and bright-field transmission, as shown in Fig. 4(a). It also allows the advantages of multimodal application. Additionally, it includes sections for optical-extinction spectra and ultrasound, enabling evaluation of PAI performance through various methods. The phantom also allows for the measurement of changes in PAI-system performance over time.

Figure 4. Phantom for photoacoustic imaging system (PAI) evaluation. (a) Multilayered phantom containing indocyanine green (ICG), gold nanorods (GNRs), and red blood cells (RBCs) for photoacoustic imaging test. Reprinted with permission from F. Ratto et al. Biomed. Opt. Express [7]. Copyright © 2019, Optica Publishing Group. (b) Images of the 3D printed lobular shaped mold and comparison of a healthy volunteer and phantom with skin and without skin images using photoacoustic tomography. Reprinted with permission from M. Dantuma et al. Biomed. Opt. Express [8]. Copyright © 2019, Optica Publishing Group.

The latter three phantoms are manufactured from unique polyvinyl chloride (PVCP) formulations and properly doped with additives, to provide tissue-like acoustic and optical characteristics. The PVCP materials are encased in a silicon layer that mimics the skin [20]. Figure 4(b) shows 3D-printed lobular molds and blood-vessel models, along with photoacoustic tomography (PAT) measurements obtained from a photoacoustic microscopy (PAM) device [8]. The color coding indicates depth, with white representing superficial areas and red representing deeper areas. The unique PVCP formulation, combined with additives, provides tissue-like acoustic and optical properties. PAI measurements show that the phantom mimics features such as the visibility of skin and blood vessels, while signals from fibroglandular and fat tissues do not appear. This suggests that the developed PAI system can effectively be evaluated for its performance in actual in vivo experiments.

4.2. Required Conditions for Phantoms for PAI

PAI systems need to accurately mimic the optical and acoustic properties of biological tissues, to provide meaningful and reliable evaluation of imaging performance. PAI phantoms can be classified by characteristics as optical or acoustic. First, PAI phantoms designed for measuring optical properties should involve characteristics such as the absorption coefficient μ_a, scattering coefficient μ_s, and anisotropy factor g [75, 76]. They are generally made from materials like intralipid, gelatin, or polyvinyl alcohol, mixed with absorbing dyes or particles to mimic tissue chromophores such as hemoglobin. In contrast, PAI phantoms for measuring acoustic properties aim to replicate the speed of sound c, acoustic impedance, and acoustic attenuation [20, 75]. Common materials include agar, gelatin, and silicone, often embedded with scatterers like glass beads to simulate the acoustic heterogeneity of biological tissues. Recently, PAI phantoms that combine both optical and acoustic properties to provide a more comprehensive model for validating the entire photoacoustic imaging process, have been introduced [11, 21, 24].

Figure 5 shows a PAI phantom made from polyacrylamide (PAA) hydrogel, with biologically relevant optical and acoustic properties analyzed over a wide range of system design parameters [9]. Hariri et al. [9] evaluated the applicability of PAA phantoms for performing image-quality assessments on three PAI systems: A custom system, AcousticX (Cyberdyne Inc., Tsukuba, Japan), and Vevo LAZR (FUJIFILM Visual Sonics Inc., Ontario, Canada).

Figure 5. Phantoms mimic the optical and acoustic properties of biological tissues. (a) Phantoms consist of polyacrylamide gels and silicon tube, ultrasound and photoacoustic images of the resolution phantom and ultrasound and photoacoustic images of penetration phantom filled with India ink. Reprinted from A. Hariri et al. Photoacoustics 2021; 22; 100245. Copyright © 2021, A. Hariri et al. [9]. (b) Construction of the 3D printed phantom consists of 12 tubes filled with dye solution and ultrasound and photoacoustic image. Reprinted from S. J. Arconada-Alvarez et al. Photoacoustics 2017; 5; 17–24. Copyright © 2017, S. J. Arconada-Alvarez et al. [76].

PAI phantoms for breast-fat and parenchyma tissue-mimicking (TMM) materials are based on silicon oil and ethylene glycol emulsions in polyacrylamide hydrogel. Solid target inclusions were created using 3D-printed molds, giving them a more filled-in appearance than conventional PAI phantoms [22, 43]. Palma-Chavez et al. [43] attempted to create target TMMs (T-TMM) for imaging targets that simulate clusters of subresolution vasculature.The community has access to several phantoms, each with unique benefits and limitations. A 3D-printed tool for holding plastic tubes with liquid contrast agent may be used rapidly, conveniently, and consistently. When placed in a beaker or evaporation dish, several tissue-simulating materials, ranging from saline to Liposyn or India ink, can be used to modulate the optical and acoustic properties. The raw material for 3D printing is polylactic acid (PLA), a biodegradable polymer. Arconada-Alvarez et al. [76] used a MakerBot Replicator 2 Desktop 3D printer (MakerBot Industries LLC., NY, USA), as shown in Fig. 5(b). The PA images were acquired using a Vevo LAZAR system (Fujifilm Visual Sonics Inc., Ontario, Canada) [76].

In early research, DHM phantoms were fabricated using lithography technology. Lithography, capable of forming delicate, high-resolution structures, was well-suited to produce DHM phantoms [44, 77, 78]. These phantoms are used to recover information about three-dimensional-shape patterns and reconstruct high-refractive-index objects embedded within media, as shown in Fig. 6. Cuche et al. [77] emphasized that DHM is a powerful three-dimensional profiling and tracking tool. They demonstrated that DHM could accurately recover various forms of microstructures, all based on the USAF 1951 resolution target. Mainly, these reflected-type DHM phantoms are suitable to verify and optimize the spatial phase difference by reconstruction of the optical path difference [44, 46]. Kang and Hong [78] studied the measurement of the three-dimensional shape of a micro-Fresnel lens using inline phase-shifting DHM. Furthermore, Yu et al. [46] provided a comprehensive review of DHM for three-dimensional profiling and tracking, showing its effectiveness in reconstructing high-refractive-index microspheres.

Figure 6. Lithography-based digital holographic microscopy (DHM) phantoms for 3D profiling and microspheres reconstruction. (a) USAF resolution target-based microstructure phantoms and 3D unwrapped phase map. Reprinted with permission from E. Cuche et al. Opt. Lett. [77]. Copyright © 1999, Optica Publishing Group. (b) Transmittance type photolithography DHM phantom for evaluating the quantitative refractive index with reconstruction distance. Reprinted from I. H. Kwon et al. Appl. Phys. Lett. 2024; 124; 093701. Copyright © 2024, AIP Publishing Group [50].

Recently, advancements in photolithography processes have enabled the utilization of nanoscale-step phantoms for the study of refractive-index-measurement accuracy in digital holography [47–50]. The precise control of nanoscale structures via photolithography allows researchers to create phantoms with well-defined height information and refractive properties, which are critical for calibrating and validating digital holographic systems. A recent study by Kwon et al. [50] highlights the use of nanoscale-step borosilicate plate phantoms fabricated through photolithography, as shown in Fig. 6(b). They demonstrated changes in spatial phase information due to subnanometer reconstruction distances, and they corrected reconstruction distances using a numerical autofocusing method. The standardized nanoscale steps of the transmission phantom produced in this process were used in a method to suggest the range of refractive-index errors that may occur during the reconstruction process. This study underscores the importance of precise nanoscale phantoms in enhancing the accuracy of refractive-index measurements in digital holography, paving the way for improved material characterization and more accurate holographic imaging systems.

In contrast, ODT phantoms are designed for analyzing the three-dimensional refractive-index distribution inside objects [45, 51, 79, 80] as shown in Fig. 7. For instance, ODT phantoms utilizing grating structures are focused on measuring and reconstructing diffraction patterns from various angles in Fig. 7(a) [25, 52]. Additionally, phantoms could help to evaluate the accuracy of internal-structure reconstructions, using a rotational tomographic system with the ODT reconstruction method [26, 27]. A cell phantom was used to validate an autonomous tomographic reconstruction workflow. The proposed method demonstrated significant improvements in axial resolution and accuracy of intensity distribution, compared to conventional illumination scanning tomography. Sun et al. [27] showed that the new method accurately reconstructs the three-dimensional refractive-index distribution of the cell phantom in Fig. 7(b), aligning closely with the ground truth. Multicore-structure ODT phantoms include internal refractive-index variations in three dimensions, which is helpful in validating the quality of reconstructed images. Figure 7(c) shows an ODT phantom that was designed to represent the phase-only scattering potential of nearly spherical cells, allowing for accurate simulations of the optical-scattering process and validation of the tomographic reconstruction methods. It includes randomly distributed subcellular organelles modeled as 3D ellipsoids, generated using stochastic sampling based on measured optical properties of cells [28].

Figure 7. Optical diffraction tomography (ODT) phantoms for 3D refractive index profiling and multicore optical fibers. (a) Grating structures for diffraction pattern measurement. Reprinted from S. O. Isikman et al. PLoS One 2012; 7; e45044. Copyright © 2012, PLOS Publishing Group [25]. (b) ODT phantoms for improved internal structure accuracy using rotated optical fiber. Reprinted from J. Sun et al. Nat. Commun. 2024; 15; 147. Copyright © 2024, J. Sun et al. [27]. (c) ODT phantom for internal reflective index analysis using multi-core micro-spherical structure. See the main text details. Reprinted with permission from B. Bazow et al. Opt. Express [28]. Copyright © 2023, Optica Publishing Group.

ODT phantoms that mimic cell structures are designed not only to resemble the shapes of cellular structures, but also to include materials with different refractive indices at various depths, as shown in Fig. 8 [29]. One method for fabricating cell-mimicking phantoms using multimaterial two-photon polymerization is as follows: First, prepare multiple photosensitive resins. Next, use a phase modulator to control the phase of the laser beam, and move the substrate in three dimensions to write complex patterns layer by layer with the laser. After fabrication, remove the unpolymerized resin and strengthen the structure through thermal or UV curing. These phantoms are used to evaluate the accuracy of phase information recovered by ODT systems.

Figure 8. Manufacturing process for cell mimicking optical diffraction tomography (ODT) phantom: (a) Multi-material printing methodology utilizing phase maps for alignment. Reprinted from E. Wdowiak et al. Addit. Manuf. 2023; 73; 103666. Copyright © 2023, E. Wdowiak et al. [29]. (b) 3D model of the cell phantom internal features, their sizes, and refractive index values. Reprinted with permission from P. Zdańkowski et al. Biomed. Opt. Express [30]. Copyright © 2021, Optica Publishing Group and from M. Ziemczonok et al. Sci. Rep. 2019; 9; 18872. Copyright © 2019, M. Ziemczonok et al. [31]. (c) Numerical reconstruction and the experimental results of the 3D refractive index measurements. Cross-sections along the white dotted lines show a comparison between both reconstructions and the model based on the reference geometry and the refractive index (RI) data. Reprinted from M. Ziemczonok et al. Sci. Rep. 2019; 9; 18872. Copyright © 2019, M. Ziemczonok et al. [31].

Ziemczonok et al. [31] developed a 3D-printed biological cell phantom to test the performance of 3D quantitative phase-imaging systems. This phantom is designed to simulate the complex morphology and refractive-index distribution of biological cells, serving as a useful tool for system calibration and performance evaluation. Additionally, the 3D-printed biological-cell phantom was used to quantify the performance of holographic tomography systems. This study emphasized the importance of phantom design in accurately reproducing the phase information of complex cell structures [31]. Krauze et al. [81] developed a 3D scattering microphantom sample to assess quantitative accuracy in tomographic phase-microscopy techniques. This sample serves as a benchmark for system performance, accurately measuring and reconstructing the phase information of microscopic structures [81, 82]. These cell-mimicking ODT phantoms are essential tools for calibrating and validating the performance of ODT and related quantitative phase-imaging systems. They not only replicate the shapes and refractive-index variations of biological cells, but also ensure that ODT systems can accurately phase information.

The reliability and reproducibility of biomedical oximetry devices are paramount for correct decision-making in clinical settings [13–15, 83, 84]. Here, optical phantoms emulating tissue morphology or tissue properties have been extensively utilized to develop and validate oximetry devices. The term “oximetry” refers to a technique for measuring oxygen saturation of hemoglobin in a tissue, based on diffuse optical spectroscopy [16, 83]. In this section, recent applied studies using representative tissue-simulating optical phantoms in oximetry are reviewed, including more accurate detection [16, 17, 85, 86], development of a robust algorithm [14], and calibration [18, 19, 53, 87, 88] by considering the influence of tissue geometry, e.g. multilayered heterogeneous structure [13–16, 19, 54, 88] and blood-vessel caliber [17, 85, 89], tissue optical properties [16, 86], pigmentation (e.g. skin-tone effect and racial bias) [87, 90], and illuminated-wavelength combination [17].

6.1. Finger & Forearm

Through finger-clip-on photoplethysmography (PPG) validation testing, Rodriguez et al. [91] reported that 3D-printed phantoms produced much stronger signal-to-noise ratio (SNR) of pulsatile waveform than did PDMS-molded ones. The proposed finger phantom can be used to investigate origins of inaccuracies in other pulse-oximetry devices.

As opposed to manual venipuncture, He et al. [92] demonstrated that a laser depth-sensing venipuncture robot, with an integrated near-infrared imager, achieved high positioning accuracy and repeatability, and thus a high success rate of venipuncture when a forearm phantom was punctured. Using the proposed automated-venipuncture machine and a forearm-mimicking phantom with an artificial vessel embedded, a total of 30 experiments could be safely carried out to characterize the venipuncture robot’s capability.

6.2. Placenta

Wang et al. [16] presented noninvasive monitoring of blood oxygenation in human placentas, using integrated frequency-domain diffuse optical spectroscopy (FD-DOS) and ultrasound (US) imaging in real time. To evaluate the hybrid instrument’s performance, they fabricated and tested with two kinds of phantoms: A water-based liquid phantom, and a liquid/solid bilayer phantom, which together mimic the human placenta and the above layers (i.e. adipose and rectus or uterus). Tissue-phantom experiments showed that the hybrid instrument has a large dynamic range and sufficient SNR to perform accurate FD-DOS measurements up to 5 cm below the skin’s surface. For the liquid/solid phantom, three kinds of tests were carried out: (1) An absorption-titration experiment, to assess the instrument’s sensitivity to the controllable absorption coefficients; (2) A depth-changing experiment, to estimate sensitivity to varying superficial-layer thickness; And (3) a deep bilayer-phantom experiment, to verify the instrument’s capability. The results of the phantom testing showed that the hybrid system accurately extracted optical properties at the deeper layer, with errors of less than 10% in absorption and less than 15% in scattering, for a superficial-layer thickness of 4.3 cm.

6.3. Brain

Using cerebral oximeters, brain-phantom studies have been widely conducted, ranging from adults [13, 93] to neonates [14, 18] and premature infants [94, 95] in the intensive-care unit (ICU). Moreover, multilayer interference [19, 54], pigmentation bias [90], and algorithm improvement [14] are factors that must be addressed for accurate estimation of tissue oxygen saturation (S_tO₂) in the brain region. For an adult’s brain, using portable diffuse optical tomography (DOT), Huang et al. [93] showed that the volumetric distribution map of variations in optical properties were reconstructed during task-induced stimulus. To validate the device’s capability, a solid phantom mimicking an adult head was fabricated.

Sudakou et al. [54] developed a bilayer blood/lipid liquid phantom mimicking superficial and deep layers such as scalp and brain, controlling independently each of the S_tO₂ values in human blood at each layer. They thoroughly analyzed the effect of varying the superficial layer’s thickness on the recovered S_tO₂ value for the deep layer, using a multiwavelength time-domain near-infrared-spectroscopy (MW-TD NIRS) tissue oximeter.

It is well known that different oximeters extract different S_tO₂ values, even on the same phantom simulating the neonatal head. Thus Kleiser et al. [18] provided an equation for conversion between commercial oximeters for calibration, using a blood-lipid liquid phantom. In addition, they confirmed the dependence of S_tO₂ values on the total concentration of hemoglobin (ctHb), which varies substantially between oximeters. As an extension, Kleiser et al. [95] applied the basic concept above to a liquid phantom characterizing the head of preterm infants in the ICU, comparing therapeutic intervention thresholds and uncertainty ranges of S_tO₂ values obtained from several devices. Likewise, Izzetoglu et al. [13] conducted a quantitative comparative study of performance assessment between commercially available cerebral-tissue oximeter (CTO) devices, using multilayer dynamic adult-head phantoms in terms of precision, correlation, accuracy, sensitivity, repeatability, and reproducibility.

Ensuring that the performance of biomedical devices does not significantly depend on skin pigmentation is essential for enhancing public health in a diverse population [90]. Thus, fabricating multilayer cerebral phantoms with adjustable pigmentation levels for neonatal, pediatric, and adult brains, Afshari et al. [90] reported that cerebral-oximeter outcomes exhibited a consistent decrease in saturation level as simulated melanin content increased.

Kovacsova et al. [14] developed a new, robust algorithm to retrieve more reliable S_tO₂ values for the newborn brain in an ICU, using simultaneously a reference time-resolved NIRS and a relatively cheap, continuous-wave multidistance NIRS (CW-MD-NIRS). The algorithm was a hybrid of a broadband fitting approach with spatially resolved spectroscopy (SRS), which is based on the diffusion approximation for light transport in a tissue. Numerical simulation and a liquid phantom emulating neonatal head layers were used to validate and compare the two algorithms above to the proposed one, in terms of dynamic range, estimated error, and outcomes. Their findings highlight the impact of S_tO₂-algorithm selection on oxygenation recovery.

Otic et al. [94] developed cerebral multiwavelength multidistance diffuse correlation spectroscopy (MW-MD-DCS) for the monitoring of cerebral hemodynamics in premature infants at risk. This approach is capable of simultaneously quantifying optical properties, S_tO₂, cerebral blood flow (CBF), and cerebral metabolic rate of oxygen (CMRO) in real time. A liquid-phantom experiment with absorption/scattering titrations, forearm-cuff occlusion tests on healthy adults, and data acquisition on two preterm infants were conducted to evaluate the system’s performance.

Rackebrandt and Gehring [19] developed a head-simulating phantom that included cerebral efferent blood circulation. Using a CW-MD-NIRS, the oxygen-saturation value in only the vessel was extracted, even under a four-layer structure (scalp, skull, vessel, and brain, in depth order). Oxygen saturation was adjusted from 99% to 20% by modulating the gas supply. The estimated original value was calibrated via a reference CO-oximetry tool.

6.4. Eye

As a standardized calibrator for retinal-vessel oximetry, Chen et al. [87] proposed a fundus-simulating phantom that mimics tissue reflectance, pigmentation, vascular perfusion, and blood oxygenation, with adjustability and reproducibility. They showed that in a human retinal vessel, the estimated value of oxygen saturation in hemoglobin (SO₂) using the proposed phantom-calibration method was consistent with that using empirical data calibration. The suggested approach has the advantage of easily controlling the SO₂ value in the RBC solution, compared to the conventional inhaled-gas intervention.

Akitegetse et al. [15] evaluated the performance of a commercially available ocular-oximetry device by using a Monte Carlo simulation and an eye phantom. A fundus-tissue phantom was used to investigate the impact of scattering, blood volume fraction, and lens yellowing on the estimated oxygen saturation, while a Monte Carlo model was used to analyze the effect of the fundus’s layered structure. They could also quantify the influence of choroidal circulation on the accuracy of measurement. In addition, they found that decreasing choroidal melanin concentration led to greater deviations in calculated SO₂ values, compared to the expected values.

Damodaran et al. [17] developed a scanning laser ophthalmoscope (SLO)-based oximetry imager, and validated the device using a retinal-tissue phantom with a two-dye mixed artificial blood vessel. Moreover, they theoretically analyzed error propagation and the vessel-packaging effect in estimating oxygen saturation. They highlighted an optimal wavelength combination for estimating a more accurate SO₂ value in the phantom test. In a review paper, MacKenzie et al. [96] summarized several ocular phantoms in oximetry, which were used in studies published from 2009 to 2016. Among them, notably Mordant et al. [89] fabricated a model eye with varying vessel diameters and different background-reflectance ratios. They demonstrated the validity of a hyperspectral fundus camera using the phantom.

The development of optical phantoms plays a critical role in evaluating and validating the performance of various optical systems. We have outlined the types and diagnostic applications of OCT, PAI, DHM, ODT, and tissue-mimicking optical phantoms. OCT phantoms are used to assess and calibrate the quality of OCT systems. They are typically made from materials such as hydrogels, silicone, and polymers that incorporate nanoparticles and do not have inherent optical-scattering properties. These phantoms measure depth-dependent spatial resolution, point-spread function (PSF), and image contrast. Various structures, from single-layer to multilayer thin-film phantoms, have been developed to verify OCT-system performance [1, 3, 6]. PAI phantoms are designed to evaluate image quality, quantitative accuracy, penetration depth, reproducibility, and reliability. These phantoms assess the ability of PAI systems to distinguish between different tissue types, and to detect subtle changes in morphological structure or concentration of chromophores. Multilayer phantoms enable comprehensive evaluation of system performance, and the measurement of performance changes over time [7, 76]. DHM phantoms emphasize changes in surface morphology and are useful for accurately reconstructing and measuring high-resolution surface details [50]. ODT phantoms analyze the three-dimensional refractive-index distribution within objects, and measure and reconstruct diffraction patterns from various angles [51, 80]. Additionally, phantoms created using multimaterial two-photon polymerization techniques mimic cellular structures, and are crucial for evaluating system performance by reproducing the shape and refractive-index variations of complex cell structures [27, 30, 31]. Tissue-mimicking optical phantoms are widely used to ensure the reliability and reproducibility of oximetry devices. These phantoms replicate the optical and mechanical properties of different tissues, enabling the evaluation of device performance in realistic clinical settings. Phantoms must remain stable over time and under various environmental conditions, to ensure reproducibility in long-term studies [75]. They should be consistently manufacturable with the same characteristics, allowing for comparative studies and standardization across research organizations. Phantoms need to be customizable to depict various anatomical conditions, such as blood vessels, tumors, and layered structures, facilitating more comprehensive testing of future optical-measurement systems. Cost-effective and simple manufacturing processes are essential for wide adoption in regular production and evaluation needs. Moreover, phantoms should be compatible with multiple imaging modalities, to offer a comprehensive evaluation platform. Including elements that mimic physiological changes, such as blood flow or oxygenation, can create more realistic testing environments. Advanced phantoms with realistic anatomical features, complex tissue structures, and heterogeneous compositions are critical for pushing the boundaries of optical technology, facilitating its transition from experimental to clinical applications. These phantoms are expected to play a vital role in advancing optical technologies and promoting their clinical application.

In part by the Korea Medical Device Development Fund grants funded by the Korean government (Ministry of Science and ICT, Ministry of Trade, Industry and Energy, Ministry of Health & Welfare, Ministry of Food and Drug Safety) (Project No. KMDF_PR_20200901_0024 and KMDF_PR_20200901_0026); In part by the National Research Council of Science & Technology (NST) grant by the Korea government (MSIT) (GTL24021-500); Development of Measurement Standards and Technology for Biomaterials and Medical Convergence funded by the Korea Research Institute of Standards and Science (Grant no. KRISS-GP2024-0007).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data sharing is not applicable to this article, as no new data were created or analyzed in this study.