Photonic-Metamaterial-Based, Near-Field-Enhanced Biosensing Approach for Early Detection of Lung and Ovarian Cancer

Abstract

Early detection of lung and ovarian cancers relies heavily on identifying tumor biomarkers, but current methods require large blood samples and complex genetic testing. This study presents a novel photonic-metamaterial-based biosensing approach that leverages near-field radiative enhancement to detect cancer biomarkers (CA 125, CEA, and CYFRA 21-1) with high sensitivity. By utilizing structured photonic metamaterials, we optimize specific wavelengths to identify these biomarkers in interstitial fluid, which can be easily collected via minimally invasive microneedle arrays. Integrating near-field interactions with wavelength-selective metamaterials amplifies the thermal response at the nanoscale, allowing for the detection of deficient concentrations of biomarkers. This photonic metamaterial technique provides a faster, more accessible, and affordable alternative to conventional blood-based methods, significantly improving early detection and monitoring of cancer. Ultimately, this approach offers a transformative tool for clinical and research applications in cancer diagnostics.

1. Introduction

Lung cancer leads to cancer-related mortality for both men and women worldwide. Non-small-cell lung cancer (NSCLC) is the primary type of lung cancer, and it accounts for over 85% of diagnosed cases [1]. It is commonly diagnosed at advanced stages, and the available systemic therapies are primarily palliative [2]. More than half of NSCLCs will test positive for a biomarker, which can be treated with an approved therapy [3,4]. The most common biomarkers include epidermal growth factor receptor (EGFR, a gene involved in promoting cell growth and proliferation) and anaplastic lymphoma kinase (ALK, a gene that contributes to cell proliferation and tumor survival) [5,6,7]. The detection of genes is based on genomic testing, like sequencing. Due to the requirement of tumor tissues to identify biomarkers, serum and interstitial biomarkers have been investigated [8], and the most common circulating biomarkers are carcinoembryonic antigen (CEA) and cytokeratin nineteen fragments (CYFRA 21-1). CEA is a glycoprotein with relatively high sensitivity for many advanced adenocarcinomas. CYFRA 21-1 may be helpful because patients with CYFRA 21-1 >2 mg/L have a 95% probability of having NSCLC. If CEA is higher than 10 mg/L, this likely indicates either adenocarcinoma or large-cell lung cancer [8].

Ovarian cancer is a prevalent gynecological tumor, yet its early detection is challenging, and the prognosis for advanced stages is poor, making it the leading cause of mortality among gynecological tumors. In China, the annual incidence of ovarian cancer ranks third among female reproductive system tumors, following cervical cancer and uterine body malignancies, with a noticeable upward trend. Its case fatality rate tops that of all female reproductive tract malignancies, posing a significant threat to women’s health. Furthermore, the survival rate drops significantly after stage 3: approximately 80% for stage 2 and only 30% to 40% for patients in stage 3/4 over five years, with most succumbing to tumor recurrence and drug resistance. Apart from genetic factors, the etiology of most ovarian cancers remains unclear. Although combined examinations using transvaginal ultrasound, pelvic examination, and serum tumor markers are available, there is a lack of effective early screening methods. Moreover, the disease is often asymptomatic in its early stages, leading to the discovery of 70% of cases being at an advanced stage, with a mere 40% survival rate within five years [8].

The cancer antigen CA 125 (CA 125) is the most frequently utilized tumor marker for ovarian cancer, particularly the marker of choice for serous cancers. The incidence of elevated CA125 levels correlates with the tumor stage and histological type. Patients with advanced and serous cancers exhibit a significantly higher incidence of elevated CA125 levels compared to those with early-stage and non-serous cancers (approximately 43.5% to 65.7% for early-stage ovarian cancer and 84.1% to 92.4% for advanced ovarian cancer) [9].

The early detection of lung and ovarian cancers has become a crucial area of focus due to the limitations of current serum-based diagnostic methods, which are time-consuming, tedious, and require high biomarker concentrations [10]. Our proposed work addresses these challenges by utilizing near-field thermal enhancement and photonic metamaterials to significantly improve the detection sensitivity for low concentrations of biomarkers like CYFRA 21-1, CEA, and CA 125. Our approach offers a faster and less invasive alternative to conventional serum-based tests by integrating wavelength-selective nanostructures with minimally invasive microneedles for sampling interstitial fluid. Unlike recent technologies such as PRAM and electrochemical biosensors, which often require tagging or complex preparation [11,12], our biosensor directly enhances the thermal response at specific wavelengths for real-time detection of cancer biomarkers. This innovation simplifies the diagnostic process and holds great potential for early cancer detection, making our technology a transformative advancement in biosensing [10].

2. Research Background

The rapid development of nanoscale technologies, small-scale thermal transport, and nano-optics can tremendously impact various engineering and biological systems. The last few years have seen impressive advances in understanding non-classical effects on radiative transfer [13,14,15,16,17,18]. We learned from theory that radiative transfer at nanoscale gaps between parallel surfaces that support surface polaritons could have greater significance than blackbody radiation by a few orders of magnitude [19,20,21]. We first developed a general theory to evaluate the radiative entropy associated with near-field radiation, which is used to assess the thermodynamic limit of small-scale energy conversion [22]. Furthermore, we studied the patch contribution of near-field radiative energy and momentum transfer between closely separated objects [23]. Previous studies investigated the wavelength selectivity of surface grating thin films doped with nanoparticles [24] and designed and fabricated a wavelength-selective thermal device composed of my-metamaterial thin films [25]. During the past few years, this has laid the groundwork for the proposed research by developing a general theory and a numerical approach for investigating the near-field thermal effects of arbitrarily shaped objects and studying the thermal conductive and radiative properties of nanostructures and van der Waals/Casimir interactions with biological tissues such as cancer cells and proteins. The team’s research experience in studying small-scale thermal transport phenomena, integrating sensor systems, and exploring the applications to biosensing directly translates to the proposed project—the design and fabrication of a photonic-metamaterial-based, near-field-enhanced biosensor for the early detection of lung and ovarian cancer biomarkers in human serum/interstitial samples.

3. Methods

Different chemicals reveal their presence at various wavelengths through the emission of thermal radiation [26,27]. Knowledge about the presence and concentration of each component can be invaluable in medical diagnostics, cancer treatment, defense, and security. Recognition and diagnostics of the proposed biomarkers indicative of lung and ovarian cancer presence are the most valuable approaches for diagnostics of lung and ovarian cancer at the initial stages. This proposal focuses on the fundamental study of the near-field mediated wavelength selectivity of wavelengths that correspond to lung and ovarian cancer biomarkers and its application to the early detection of the small number of markers in serum/interstitial samples.

3.1. Developing Highly Sensitive and Selective Nanoscale Photonic Metamaterials with Tunable Narrowband Wavelength Selectivity for Lung and Ovarian Cancer Biomarkers

Near-field radiative thermal transport has been extensively investigated theoretically between objects of various shapes, with most studies focusing on two flat surfaces. These investigations encompass near-field radiative energy, momentum, and entropy transfer between multilayered films. In this proposal, the configuration of interest, shown in Figure 1, consists of a surface grating with defined width $w$, depth $a$, and period $l$, combined with a thin film of high-emissivity material. These three geometric parameters of the surface grating structure can be tuned and determined to enhance near-field thermal radiation between the coated surface grating structure and nanoparticle-embedded thin film [28,29,30]. This thin film, with a thickness $t$, is embedded with nanoparticles of polar materials such as silicon carbide (SiC) and boron nitride (BN), characterized by their radius $r$ and volume fraction $vf\%$. The grating period is chosen to align with the wavelength of interest for specific biomarkers like CEA, ensuring efficient coupling of electromagnetic waves, typically around a few hundred nanometers for infrared applications. Around 25 nm, the nanoparticle radius provides strong near-field enhancement, maximizing sensitivity for low-concentration biomarkers. Lastly, the grating depth, approximately 300 nm, balances the spectral bandwidth and emissivity, offering narrow peaks for selective detection [31,32,33]. $vf\%$ is in the 10–30% range, exhibiting excellent wavelength selectivity [26]. The gap $d$ between the two layers, approximately 50–80 nm, is optimized for near-field effects while considering the sizes of the coated antibodies (~10 nm) and antigens (~5 nm) used for detecting lung and ovarian cancer biomarkers, such as CEA, CA125, and CYFRA21-1. The near-field-enhanced, wavelength-selective emitter is composed of two far-field wavelength-selective thermal emitters, and the goal is to manipulate the magnitude and position of wavelength-dependent emission peaks by adjusting materials and geometries. This selective emission enhances the sensitivity for biomarker detection, making the structure highly effective for biosensing applications.

3.1.1. Design and Characteristics of Tunable Wavelength Selectivity of Photonic Metamaterials

Classical wavelength-selective emitters have found many applications, such as solar power conversion [34,35,36,37] and thermophotovoltaics (TPV) [38,39,40,41,42,43,44]. Most naturally occurring materials do not exhibit sufficient selectivity without further processing. As such, selective emission is achieved by creating (1) periodic thin films [45,46,47], (2) 3D photonic crystals [48,49], and (3) photonic metamaterials [50,51,52]. All these configurations rely on the interactions with EM waves to give rise to a narrow emission spectrum. For example, BNNPs can be dispersed in a polymer sheet, providing selective emission behavior [53].

Considering the complexity of microfabrication processes, the proposed device consists of two classical ones—one with a thin film of polystyrene of thickness 500 nm doped with BN/Au NPs of radius 25 nm and volume fraction 30%, and the other with a 3D-patterned surface grating of polystyrene with a thin film coating. Since the emission from the substrate should not overwhelm that from the thin film, the substrate should be of high reflectivity. A thick layer of Au 1 µm is deposited on a substrate. Conducting a lab test using Fourier transform infrared spectroscopy (FTIR) is easy. The infrared wavelength spectrum of the proposed photonic metamaterials is monitored through an integrated measurement system. For theoretical analysis, we consider the case of thermal emission from a layer of polymer of thickness l deposited on an Au-coated substrate. The semi-spherical emissivity [54,55] can be calculated by

$$e\left(\lambda \right)={\displaystyle \frac{{\lambda }^2}{{4\pi }^2}}\sum _{\mu =s,p}\left({\int }_0^{{\displaystyle \frac{2\pi }{\lambda }}}\left[1-{\left|R^{\left(\mu \right)}\right|}^2-{\left|T^{\left(\mu \right)}\right|}^2\right]d{k}_{\rho }\right)$$

(1)

where $e\left(\lambda \right)$ is the wavelength-dependent emissivity; $\lambda$ is the wavelength of the emitted radiation; $k_{\rho }$ is the in-plane wavenumber; the superscript $\left(\mu \right)$ is the polarization of electromagnetic waves: transverse electric $\left(s\right)$ and transverse magnetic $\left(p\right)$; $R^{\left(\mu \right)}$ is the Fresnel reflection coefficient; and $T^{\left(\mu \right)}$ is the Fresnel transmission coefficient [56,57,58].

The optical properties of doped polystyrene with different NPs of various radii are plotted in Figure 2.

Near-Field-Enhanced Selectivity of Photonic Metamaterials

Due to the coupling of surface electromagnetic waves and polaritons, radiative transport between classical emitters is enhanced dramatically as spacing decreases. The aim is to manipulate the magnitude of wavelength-dependent emissivity and the position of the emission peak by adjusting the geometries of configurations. Rigorous Coupled-Wave Analysis [59,60,61,62] is applied to solve the scattering properties from periodic surface grating. This can compute the reflection, transmission, and emission spectra of structures composed of periodic and patterned layers. This method has been used for designing selective photonic metamaterials to improve energy conversion of TPVs [39] and biosensing for harmful algae blooms in coastal oceans [25]. Figure 3a shows the designs and enhanced spectral emissions of three near-infrared, wavelength-selective photonic metamaterials of multilayered and surface grating structures. In the lower wavelength region (visible and near-infrared), the wavelength selectivity of polystyrene thin film doped with SiC/Au NPs of various volume fractions has been investigated for the detection of CEA, CA 125, and CYFRA 21-1 antigens, indicators of lung and ovarian cancer cells, as shown in Figure 3b.

Photonic-Metamaterial-Based Wavelength-Selective Device

This study aimed to design a wavelength-selective device with a critical feature of 200 µm × 200 µm nanoscale photonic metamaterials consisting of two classical thermal emitters that form a chamber in between. As shown in Figure 1, the top half of the structure is a periodic surface grating structure (green grating) with a thin coating of SiC/BN. The bottom half comprises Au + polystyrene thin films (thin orange film) of thickness 1 µm + 300 nm embedded with polar SiC/BN NPs (red spheres) of radius 25 nm and volume fraction 2.5–20%.

3.1.2. Fabrication of Selective Thermal Emitter and System Setup

Micro/Nanofabrication of Selective Emitters

In this experiment, we utilized a well-controlled optoelectronic system to validate the enhanced wavelength-selective thermal emission of the fabricated photonic metamaterial-based biosensor. The setup includes the following:

• Near-Infrared (NIR) Source: This provides the excitation necessary to induce thermal emission from the nanostructured surfaces.
• IR Spectrometer (Photodetector): Integrated into the system to detect the wavelength-specific emissivity changes caused by the interaction of cancer biomarkers with the coated surfaces. The spectrometer records the emission spectrum across the desired wavelength range.
• Signal Conditioning Circuits: These are included to filter, amplify, and process the signals generated by the IR spectrometer, ensuring accurate data acquisition.

Additionally, we employed Fourier transform infrared (FTIR) spectroscopy to measure the spectral characteristics of the thermal emission. The experimental samples, fabricated using e-beam lithography and spin coating, were mounted between the IR source and spectrometer, as depicted in Figure 4a. The precise control of the gap between the surfaces (~50–80 $\mathsf{\mu }\mathrm{m}$) allows for near-field enhancement, which boosts sensitivity to biomarker concentrations.

A schematic of the near-field-enhanced, wavelength-selective structure is shown in Figure 4a. This can be fabricated using cleanroom facilities involving photolithography (Karl Suss MJB-3 UV300, SUSS MicroTec, Garching, Germany), e-beam lithography (LEO 1530 VP, Carl Zeiss SMT, Oberkochen, Germany), an e-beam evaporator (Lesker Lab 18 deposition system Kurt J. Lesker Company, Jefferson Hills, Pennsylvania, USA) of Au thin film, and spin coating (Cee Model 100 Brewer Science, Rolla, Missouri, USA) of the PS layer (Figure 4b). Figure 4c shows three fabricated samples of 1 cm × 1 cm made of 1 µm Au film + 300 nm polystyrene film doped with SiC/BN NPs of radius 25 nm and volume fraction 10%. SEM images of the fabricated emitters are shown in Figure 4d,e. This yields a higher spectral emissivity, ελ ≈ 0.95 or higher, compared to ελ ≈ 0.4 in the state of the art at the desired wavelengths, which are thermally sensitive to lung and ovarian cancer biomarkers of interest.

Setup of an Optoelectronic System

Demonstrating the power and capabilities of the proposed device through a series of well-controlled validation measurements in the long term requires the setup of the optoelectronic system, including the integration of a selective sensor, near-infrared (NIR) source, IR spectrometer (photodetector), and signal conditioning circuits, as illustrated in Figure 1. The IR source and IR spectrometer are highly specialized components that are not commonly available. However, the Fraunhofer Institute for Physical Measurement Techniques IPM and VIGO Systems are leading organizations that manufacture and sell IR components in the United States [63].

3.2. Evaluating and Demonstrating the Photonic-Metamaterial-Based, Near-Field-Enhanced Biosensor by Integrating with a Sample Collection Microneedle Array

During the past few decades, microneedles have been used for dermal and transdermal delivery of a broad range of drugs, such as small-molecular-weight drugs, DNA, proteins, and inactivated viruses [64]. Figure 5 presents the working principle and microstructure of the microneedles. Hollow microneedles have been explored for blood, interstitial, and lymph sample collection to minimize the pain associated with traditional sample collection. For example, a blood test is indispensable for health examination and monitoring.

3.2.1. Detection of Selected Early Lung and Ovarian Cancer Biomarkers

Coating with Capturing Antibodies for Biomarkers

The antibodies were coated against CEA, CA 125, and CYFRA 21-1 onto the biosensor surface, followed by blocking and antigen incubation. Spectrum changes were detected as above, and wavelengths that generated the highest spectrum change were selected for each biomarker detection. Special treatment and coating are necessary on the face-to-face surfaces of the proposed structures. The surfaces are coated with capturing antibodies and integrated with sample collection hollow microneedles, as shown in Figure 6. Recombinant lung and ovarian cancer biomarkers CEA, CA 125, and CYFRA 21-1 were purchased from MyBioSource (#MBS537751 San Diego, CA, USA), R&D Systems (5609-MU Minneapolis, MN, USA), and ProSpec (#PRO-287 Rehovot, Israel), respectively. They were spiked into a phosphate-balanced solution (pH ≈ 7.4) at a broad concentration range from 0.1 ng/mL to 100 ng/mL. The capturing antibodies were obtained from Creative Diagnostics (Clone N466 Shirley, New York, NY, USA). Antibodies at 1 µg/mL were used for coating onto the polystyrene surface at 4 °C overnight. The preparation process were performed by graduate students in collaboration with Dr. Chen’s laboratory at the College of Pharmacy.

Fabrication of Sample Collection Hollow Microneedles

Hollow pyramid-shaped microneedles (Figure 6a, 10 × 10 in the area of 5 × 5 mm²) are fabricated by UV lithography in a microfabrication facility. Each microneedle’s base diameter and height are 250 µm and 600 µm, and the inner diameter of the hollow area is 100 µm. In brief, a female PDMS mold was first fabricated by photolithography with a photomask that defines microneedle pyramid tips. Then, this mold generated male hollow microneedle molds in two lithography steps—(1) allowing cross-linking of SU-8 photoresist in microneedles except hollow parts, and (2) allowing cross-linking of microneedle bases. A microfluidic PDMS mold with the same pattern of microneedles was also generated. The hollow microneedle and microfluidic mold were aligned and subjected to oxygen plasma bonding. The final product allowed interstitial samples collected from hollow microneedles to converge and flow to the sample chamber. The collected sample was about penny size, 10 mm in diameter, and 100 µm in depth. Due to capillary attraction, it was drawn into microchannels and delivered to the proposed near-field-enhanced, selective thermal emitter chamber. Coating antibodies captured biomarkers.

3.2.2. Validation Testing Using EGFP Protein and Anti-EGFP Antibody

Model fluorescent protein EGFP and the anti-EGFP antibody were used to prove that coating antibodies onto biosensor surface followed by antigen binding will generate different spectra that allow us to identify specific wavelengths with the highest sensitivity for EGFP detection. Briefly, the anti-EGFP antibody was coated onto the thin film of the biosensor overnight at 4 °C in bicarbonate/carbonate buffer (pH ≈ 9.6), in which the antibody has the most robust binding to the polystyrene sensor surface. Figure 7 shows that a total of 6 samples were tested, including 2 control samples (the buffer only and the buffer block with non-fat milk) and 4 experimental samples with different combinations of the protein EGFP and the anti-EGFP antibody, to evaluate the sensitivity of our photonic-metamaterial-based biosensor. All experiments were conducted at room temperature (approximately 25 °C), and the samples were incubated on the sensor surface for 12–24 h to ensure consistent binding of the biomarkers [33]. These wavelengths that generate maximal changes with modification of EGFP concentrations were used in the following studies to detect EGFP in complex systems. Biosensors were then blocked with 3% non-fat milk in a peripheral blood smear (PBS) and further incubated with a series of concentrations of EGFP in PBS. Sensors without antibody coating but with non-fat milk blocking served as controls. Spectrum change were detected after EGFP incubation, as shown in Figure 7. A Nikon Eclipse E600 upright microscope and JASCO FT/IR-4100 spectrometer were used in the testing, and the fluorescent image was captured using Nikon’s NIS-Elements software version 4.x (inset of Figure 7). We observed a dose-dependent spectrum change at specific wavelength ranges from 5 to 7 $\mathsf{\mu }\mathrm{m}$ and 8.5 to 9.5 $\mathsf{\mu }\mathrm{m}$. These wavelengths that generate maximal changes with modification of EGFP concentrations were used in the following studies to detect EGFP in complex systems. Due to the specific antibody-antigen interaction, the biosensor coated with anti-EGFP antibody will have similar detection specificity and sensitivity in complex systems, like a human serum. To assess this, different concentrations of EGFP were added to human serum (H4522, sigma) and then incubated with an anti-EGFP antibody-coated biosensor followed by spectrum reading.

4. Conclusions and Discussion

This study presents a new way to spot lung and ovarian cancers using a biosensing method enhanced by photonic metamaterials and near-field technology. Current cancer detection techniques use serum samples, often requiring lots of blood and complex genetic tests. These methods can take a long time and depend on having specific medical tools, like DNA sequencers. On the other hand, our process uses interstitial fluid collected without pain through tiny needle arrays, making it easier for patients to access and less invasive.

Using the close-range heat boost and specific wavelength choices of tiny structured photonic metamaterials allows us to spot cancer markers when they are present in minimal amounts. These photonic metamaterials’ chosen light matches how key lung and ovarian cancer markers like CEA, CYFRA 21-1, and CA 125 absorb light. This improved ability to detect and identify specific markers represents a significant step forward in finding cancer. It opens up the possibility of diagnosing patients earlier and treating more patients.

Fundamental breakthroughs in this research include the application of wavelength-selective photonic metamaterials, which researchers can adjust to boost detection sensitivity, and the addition of microneedle tech for painless sample gathering. These advances combined impact cancer diagnostics by offering an affordable, responsive, and expandable tool for clinical and research uses.

Our biosensor exhibited a superior resolution in detecting low-concentration cancer biomarkers, showcasing its precision in identifying biomarkers like CEA, CA125, and CYFRA21-1 at early stages. The near-field enhancement and wavelength selectivity significantly improve the resolution compared to traditional optical and electrochemical sensors. This high resolution demonstrates the novelty and impact of our approach, especially in offering a minimally invasive, real-time detection method. However, some limitations of our technique should be acknowledged, such as the complexity of fabricating the metamaterials and challenges in scalability for mass production. Future work should address these limitations to improve this technology’s practicality and clinical application.

Ultimately, this new method shows great potential for improving patient results by spotting cancer earlier, leading to faster treatments and a higher chance of patients getting better. It marks a giant leap forward in mixing nanotechnology and bioengineering for medical testing. It can boost the accuracy, speed, and reach of cancer screenings worldwide, as shown in

Table 1. Comparison of three standard detection methods with the photonic metamaterial biosensor [13,14,65,66].

Method	Sample Size	Sensitivity	Invasiveness	Time Requirement	Cost
Traditional Serum Testing	Large (blood)	Moderate	Invasive (blood draw)	Long (lab process)	Moderate to High (lab cost)
Genomic Sequencing	Large (tumor)	High	Highly Invasive	Long	High (sequencing)
Electrochemical Biosensor	Small (serum)	Moderate to High	Minimally Invasive	Moderate	Moderate
Photonic Metamaterial Biosensor (this work)	Extreme Small	High	Minimally Invasive	Short (real-time)	Moderate to Low

.