Evanescent Wave-Based Photonic Biosensors for the Design and Evaluation of Advanced Cancer Immunotherapies.

Introduction
Cancer therapy is undergoing a profound
transformation driven by
advances in the deep knowledge of immunology, genomics, and cellular
biology. Immunotherapies are reshaping the therapeutic landscape by
harnessing the patient’s immune system to specifically recognize
and eliminate malignant cells. This broad class of treatments encompasses
molecular immunomodulators, adoptive cell transfer (ACT) strategies,
and emerging gene- and epigenetic-regulation approaches, offering
unprecedented precision in targeting tumor-specific cellular pathways
(Figure
). By integration
of these modalities, immunotherapies have the potential to enhance
therapeutic efficacy, improve the durability of clinical responses,
reduce systemic toxicity, and enable personalized treatment regimens
tailored to individual patient profiles.
Molecular immunotherapies, including immune checkpoint
inhibitors
(ICIs) targeting programmed cell death protein 1 (PD-1), programmed
death-ligand 1 (PD-L1), and cytotoxic T-lymphocyte-associated antigen
4 (CTLA-4), together with tumor-directed monoclonal antibodies, have
reshaped clinical oncology by restoring immune surveillance and inducing
durable responses across multiple malignancies.
,
In
parallel, cell immunotherapies, and in particular chimeric antigen
receptor (CAR)-T cells, have demonstrated remarkable clinical success
in hematologic malignancies, producing high rates of durable remission
in relapsed or refractory patients. Other
ACT strategies, including T-cell receptor (TCR), tumor-infiltrating
lymphocyte, and natural killer cell therapies, offer additional promise,
particularly for solid tumors, as they can target a broader range
of tumor-associated antigens (TAAs), exploit endogenous T-cell specificity,
and harness innate cytotoxicity to overcome the challenges of heterogeneous
and immunosuppressive tumor microenvironments.
−

In addition,
highly innovative in vivo CAR-T approaches using lipid nanoparticles
or viral vectors are emerging as a transformative strategy to generate
functional immune responses directly within patients, potentially
providing scalable and more accessible alternatives to ex vivo manufacturing.

Complementing these advances, interventions
targeting genetic and
epigenetic regulation, such as DNA methylation, alternative splicing
(AS), and noncoding RNAs, provide powerful tools to further modulate
both tumor and immune cell function.
,
Epigenetic
interventions, including DNA methyltransferase inhibitors and chromatin-modifying
drugs, can restore the expression of silenced tumor suppressor genes
and enhance immune cell activity. Splicing
modulators and antisense approaches allow correction of aberrant isoforms
that drive tumor progression or mediate immune evasion, while RNA-based
therapeutics, including microRNA and long noncoding RNA regulators,
can fine-tune key signaling pathways in both cancer and immune cells.
−

By enabling precise, multilayered control over gene expression and
cellular behavior, these gene-regulatory strategies can enhance the
potency, specificity, and durability of immunotherapies. Together,
the integration of molecular-, cellular-, and gene-regulation-based
approaches defines a new therapeutic paradigm that could substantially
increase the range of treatable tumors and improve patient outcomes.
Despite their transformative potential, these therapeutic modalities
face major scientific and translational hurdles. Immune-related adverse
events, tumor heterogeneity, and the immunosuppressive tumor microenvironment
remain significant barriers to consistent efficacy, particularly in
solid tumors. Moreover, challenges related to off-target effects,
gene-editing safety, and the high cost and complexity of manufacturing
cell-based products continue to limit their broad clinical implementation.
Addressing these limitations requires an integrated framework that
combines biological innovation with advanced analytical technologies
capable of providing real-time, quantitative, and mechanistic insights
into cellular and molecular dynamics.
The development and optimization
of next-generation cancer therapies
depend critically on analytical tools that can monitor treatment-induced
changes in cell phenotypes, signaling networks, and gene expression
with high spatial and temporal resolution. Traditional analytical
techniques, such as flow cytometry, fluorescence, and confocal microscopy
and molecular assays like ELISA or Western blotting, remain essential
for characterizing immune cell populations, visualizing cellular interactions,
and quantifying secreted factors. However, these methods typically
provide population-averaged or end-point readouts, which makes it
difficult to track the temporal evolution of individual cell states
or capture the heterogeneity that emerges within phenotypically similar
subsets. In addition, most of these techniques require fluorescent
or biochemical labels, which can constrain the number of biomarkers
detected simultaneously and frequently measured across separate assays
or sample aliquots, preventing their direct correlation within the
same single cell and further limiting our ability to map coordinated
molecular changes. On the other hand, next-generation sequencing and
epigenetic assays, including chromatin immunoprecipitation (ChIP)
sequencing and bisulfite sequencing, have revolutionized our understanding
of cellular heterogeneity and gene regulation, yet their high cost,
technical complexity, and lack of real-time capability constrain their
use in dynamic therapeutic monitoring.
To overcome these analytical
barriers, label-free photonic biosensors
based on the evanescent field sensing principle can become powerful
tools to investigate molecular interactions and cellular processes
in real time and under physiologically relevant conditions. By detecting subtle changes in optical properties
such as refractive index, intensity, or resonance wavelength within
the near-surface region of confined optical modes, these platforms
enable noninvasive, continuous monitoring of biomolecular events without
the need for fluorescent or radioactive labels. Plasmonic and integrated
silicon photonic biosensors, in particular, combine high surface sensitivity
with compatibility with microfluidic integration, miniaturization,
and multiplexed analysis, making them especially suited for translational
applications in cancer research.
In this perspective, we discuss
the emerging role of evanescent
wave-based photonic biosensors, with a specific focus on plasmonic
and integrated silicon photonic platforms, as enabling technologies
in the design and evaluation of advanced cancer immunotherapies. We
highlight how these platforms can bridge the current analytical gap
between molecular understanding and therapeutic translation by enabling
real-time, quantitative assessment of immune and cancer cell regulation
at the molecular and cellular levels. In addition, we critically examine
the main technological and translational bottlenecks that currently
limit the broader adoption of these optical biosensors and outline
emerging strategies and research directions to address these challenges.
Ultimately, such advances are expected to accelerate the rational
design of safer, more effective, and truly personalized cancer immunotherapies.

Evanescent Wave-Based Photonic Biosensors in Biomedicines
Photonic biosensors based on the evanescent wave principle have
become indispensable analytical platforms in biomedical research,
enabling label-free, quantitative, and noninvasive monitoring of biomolecular
interactions. These devices integrate a biorecognition element, such
as an antibody or nucleic acid probe, with an optical transducer that
converts specific biological interactions into measurable optical
signals (Figure
A).
By detecting variations in light properties within the evanescent
field region, they allow for the direct identification and quantification
of analytes without the need for fluorescent or colorimetric labels,
removing the dependency on fluorophores and the multiplexing constraints
associated with their limited spectral bandwidth. In these platforms,
sensing relies on the interaction between light confined within a
nanostructure or waveguide and the surrounding medium, where an evanescent
electromagnetic field, extending several hundred nanometers beyond
the sensor surface, responds sensitively to local refractive index
changes upon analyte binding. This shift alters the propagation or
resonance conditions of the optical mode, providing a direct and quantitative
measure of biomolecular events. Among these technologies, plasmonic
and integrated silicon photonic biosensors represent the most mature
and widely adopted platforms, offering high sensitivity, versatility
across a wide range of targets, and strong potential for multiplexing
and integration into compact, user-friendly analytical devices.
Plasmonic biosensors, and in particular SPR systems,
have long
been the benchmark in label-free biomolecular analysis, with mature
commercial platforms demonstrating their robustness, reproducibility,
and quantitative capabilities. SPR relies on detecting refractive
index changes at a metal–dielectric interface, typically by
monitoring variations in the intensity or angle of the reflected light,
which allows precise quantification of biomolecular binding kinetics
and affinities (Figure
B). This robust and well-standardized workflow has made SPR a cornerstone
in pharmaceutical and biomedical research, with established applications
in drug discovery, target validation, and quality control. Recent advances in nanoplasmonic architectures,
including nanoparticle arrays and nanohole grids, have expanded SPR
sensing toward LSPR and EOT modalities. These nanostructured platforms preserve evanescent-field sensing
while enabling alternative optical readouts, facilitating integration
into compact or on-chip formats. Such developments have pushed detection
limits toward the single-molecule regime and enabled multiplexed analysis
in miniaturized systems. Nevertheless, the need for sophisticated
nanofabrication, precise biofunctionalization, and seamless integration
of optical and microfluidic components continues to pose challenges
for large-scale clinical deployment.
In parallel, integrated
silicon photonic biosensors have emerged
as a powerful alternative, combining high analytical performance with
the scalability of semiconductor manufacturing. Devices based on ring resonators, Mach–Zehnder interferometers
(MZI), and BiMW interferometers exploit the optical properties of
silicon waveguide materials to achieve the sensitive detection of
refractive index variations induced by biomolecular binding (Figure
C). In resonant systems,
these interactions lead to measurable shifts in the resonance wavelength,
while in interferometric devices, phase changes alter the interference
pattern. The integrated nature of these silicon photonic platforms
enables dense multiplexing, compact footprints, and compatibility
with standard complementary metal-oxide-semiconductor (CMOS) fabrication,
supporting large-scale manufacturing and system integration. These
attributes position silicon photonic biosensors as attractive candidates
for next-generation biomedical and diagnostic platforms. Nonetheless,
practical challenges remain, particularly in integrating sample handling,
automation, and user-friendly interfaces, to translate laboratory
prototypes into clinically validated bioanalytical tools.
In
the last decades, research on evanescent wave-based photonic
biosensors has increasingly focused on their application and validation
in clinical diagnostics, targeting a broad range of analytes relevant
to precision medicine. These include conventional and emerging cancer
biomarkers (such as proteins, antibodies, microRNAs, and DNA mutations),
infectious pathogens (viruses and bacteria), and therapeutic drug
monitoring for personalized patient follow-up. A critical challenge in these applications is the control
of nonspecific adsorption and background signals arising from complex
biological matrices, such as serum, plasma, or blood, which can compromise
specificity despite high intrinsic sensitivity. To address this issue,
substantial advances have been made in surface biofunctionalization
strategies, including the use of oriented and high-density capture
layers, optimized linker chemistries, and affinity probes with improved
selectivity, as well as in the development of antifouling coatings
based on polymer brushes, zwitterionic materials, and mixed self-assembled
monolayers. These approaches have enabled
direct or minimally diluted analysis of complex biological fluids
while preserving the advantages of real-time, label-free detection.
Across numerous small-scale clinical validation studies, label-free
biosensors have frequently demonstrated equivalent or superior analytical
performance compared to traditional diagnostic assays, including ELISA,
chemiluminescence immunoassays (CLIA), and polymerase chain reaction
tests.

Despite these advances,
adapting these biosensors to the study
of advanced therapies remains highly challenging due to the complexity
and dynamic nature of living systems. Cell-based analyses involve
multifactorial and transient processes, such as receptor signaling,
morphological adaptation, and epigenetic modulation, that occur across
multiple spatial and temporal scales. Capturing these subtle events
requires biosensors with exceptional sensitivity and specificity,
sufficient to enable fully direct, label-free detection without the
need for secondary antibodies or sandwich assays while ensuring compatibility
with physiologically relevant environments and maintaining cell viability.
The need for multiplexed detection further increases system complexity
as advanced therapies often engage multiple targets and pathways simultaneously.
Integration with microfluidic platforms offers a promising path forward,
enabling precise control of cell culture conditions, automated delivery
of stimuli, and high-throughput, multidimensional measurements. Such
convergence of label-free photonic sensing and microfluidic technology
provides a powerful framework for real-time functional analysis of
immune or tumor cells, bridging molecular detection with the evaluation
of therapeutic efficacy in next-generation cancer treatments. In the
following sections, we highlight representative biosensor platforms
and methodologies with potential relevance to the design and evaluation
of advanced cancer therapies, emphasizing their key advantages, current
limitations, and future perspectives.

Applications in Molecular Immunotherapies
The clinical
success of ICIs and monoclonal antibodies (mAbs) has
positioned these molecular immunotherapies as well-established, routinely
administered cancer treatments. By targeting regulatory molecules
such as PD-1, PD-L1, and CTLA-4, checkpoint inhibitors restore antitumor
immune activity and have become standard-of-care across multiple malignancies,
including melanoma, lung, renal, and urothelial cancers. Similarly,
mAbs directed against TAAs, such as rituximab and trastuzumab, mediate
direct cytotoxicity and can synergize with chemotherapy or targeted
agents to improve efficacy while reducing systemic toxicity. Despite
these achievements, challenges remain in extending these therapies
to broader tumor types, optimizing combination strategies, and overcoming
resistance mechanisms that limit durable responses.
In the design
and development of these therapies, two main factors
must be considered: the profiling of TAAs and immune checkpoint expression
in either tumor or immune cells and the affinity and specificity of
the therapeutic molecules toward the targeted receptor. The latter
has been extensively addressed with diverse types of evanescent wave-based
photonic biosensors by implementing simple biomolecular interaction
assays for the screening of different candidates, identifying specific
epitopes, and evaluating the interaction dynamics. These assays typically
provide quantitative binding parameters, such as association (k
on) and dissociation (k
off) rates and equilibrium dissociation constants (K
D), which are essential for ranking therapeutic
candidates based on affinity and expected in vivo performance. For
example, high-throughput SPR platforms have been used to generate
epitope community maps integrated with affinity data across panels
of therapeutic mAbs, enabling the identification of high-affinity
lead candidates. Puopolo et al. also
reported a protocol for the implementation in commercial SPR instruments
to evaluate the effect of small molecules as PD1/PD-L1 checkpoint
inhibitors. They proved comparable results
to standard techniques like ELISA or cell-based assays (IC50 values around 85 and 650 nM) with shorter experimental runs and
lower sample volume requirements. Nonetheless, interesting approaches
are also being developed to enhance the biomimetic characteristics
of the sensor assay. In this regard, Batool et al. published an innovative
biomimetic sensor based on nanoplasmonic arrays coated with a functionalized
lipid bilayer, simulating an artificial tumor cell membrane expressing
PD-L1 molecules (Figure
A). They demonstrated high sensitivity
for the monitoring of PD-L1/PD-1 interactions, achieving an LOD of
6.7 ng/mL (0.2 nM) as well as accurate evaluation of specific mAbs
as checkpoint inhibitors (IC50 = 0.43 nM). Unlike standard
surface chemistry procedures for bioreceptor immobilization, the functional
lipid membrane enables lateral diffusion of the attached receptors,
facilitating the formation of cell-interaction clusters. Another example
is the work from Álvarez Freile et al., who utilized spherical
optical resonators coated with specific tumor receptors (Epidermal
Growth Factor Receptor, EGFR) and immune checkpoints (PD-L1) to imitate
a cancer cell and monitor the interaction with specific therapeutic
mAbs (i.e., cetuximab for EGFR and atezolizumab for PD-L1) (Figure
B). Their biosensor offers a versatile platform for the screening
and evaluation of different molecular immunotherapies in a three-dimensional
(3D) assay format.
Conversely, there is currently very little research
focused on
the study and identification of tumor antigens or immune receptors
directly in cells using evanescent wave-based photonic biosensing.
Despite numerous studies having been published for the detection of
soluble markers, exosomes, or circulating tumor cells, the design
and evaluation of effective immunotherapies critically require an
accurate profiling of the expression of such molecules within the
tumor microenvironment. This study is commonly performed with fluorescence-based
methods, and it would greatly benefit from a highly sensitive technique
that allows for fast detection and quantification in a noninvasive
and multiplexed format. In this direction, Shanehbandi et al. proposed
a SPR immunosensor for the identification of CD20-positive lymphoma
cells through their interaction with specific antibodies immobilized
on the surface. In this proof-of-concept
study, SPR signals correlated with the density of CD20 expression
on cells, and the biosensor was able to distinguish CD20-positive
from CD20-negative cells with a LOD of around 104 cells/mL,
demonstrating feasibility for direct cellular receptor profiling.
It represents a proof of concept of the potential use of SPR biosensors
in the analysis of cell receptor expression. A more advanced study
has recently been published by Ahmed et al., introducing a plasmonic-based
surface-enhanced Raman scattering (SERS) technology for deciphering
the heterogeneity of diverse immune checkpoints expression in lung
cancer cells. This mesoporous gold SERS
platform enabled label-free spectral discrimination of immune checkpoint
expression at the single-cell level, quantifying differential expression
patterns across individual cells within patient samples and revealing
inter- and intratumoral heterogeneity; the authors reported statistically
significant differences in SERS signal intensities corresponding to
variable levels of checkpoint proteins such as PD-L1 and CTLA-4. These
examples underscore how quantitative cell-level measurements of antigen
or receptor expression could inform the rational design and personalized
selection of immunotherapies.
Future efforts should aim to translate
optical biosensing from
simplified biomolecular assays to more physiologically relevant models.
The integration of biomimetic interfaces, multiplexed detection, and
microfluidic handling will be key to capturing the dynamic and heterogeneous
interactions that define the tumor–immune microenvironment.
Combining these technologies with single-cell or patient-derived analyses
could enable the rapid and quantitative profiling of immune checkpoints
and tumor antigens, providing actionable insights for the design of
tailored molecular immunotherapies.

Applications in Cell Immunotherapies
The clinical success
of CAR-T cell therapies in hematologic malignancies
has driven widespread interest in ACT approaches as a groundbreaking
strategy for cancer immunotherapy. Currently, these therapies rely
on the ex vivo engineering or expansion of patient-derived immune
cells, typically T lymphocytes, to recognize and eradicate tumor cells
through antigen-specific mechanisms. Upon reinfusion, the modified
cells can proliferate, persist, and mediate durable antitumor responses,
as demonstrated by the impressive remission rates achieved in B-cell
leukemias and lymphomas. Despite these breakthroughs, translating
ACT therapies to solid tumors remains a major challenge due to the
high heterogeneity and complexity of the tumor microenvironment, which
includes immunosuppressive signaling, physical barriers that limit
immune-cell infiltration, and dynamic antigen loss or modulation that
favors immune evasion.
Here, the role of evanescent wave-based
photonic biosensors can
be directed to facilitate and improve both the analysis of immune-tumor
cell interactions and the evaluation of effective activation of immune
response mechanisms in different tumor environments. An important
limitation in studying cell–cell binding dynamics with these
technologies is their on-surface sensing mechanism, which does not
allow for probing interactions occurring farther than 300–500
nm from the sensor surface. Nonetheless, strategies can be developed
to functionalize the sensor surface with specific tumor ligands and
analyze the interaction with immune cells. In this sense, an SPR biosensor
coated with an artificial cell membrane expressing TAAs (through a
recombinant peptide Major Histocompatibility ComplexpMHC)
was applied for the quantitative, two-dimensional affinity analysis
of tumor-specific T-cell receptors (TCRs) (Figure
A). The platform
discriminated high-affinity from low-affinity TCRs with picomolar-to-nanomolar
effective affinity differences, showing a strong correlation with
flow cytometry while providing higher sensitivity and improved resolution
compared to conventional SPR assays using immobilized proteins.
For the analysis of immune activity and signaling
events, immunoassay-based
techniques such as ELISpot and FluoroSpot have become valuable tools,
providing secretion detection and functional assessment at single-cell
resolution. However, these assays are limited by end point measurements,
labor-intensive protocols, and the inability to recover live cells
for downstream analyses, which restricts real-time monitoring and
longitudinal studies of the same cells. To address these limitations,
plasmonic biosensors have been increasingly applied to study immune
cell secretion and signaling with high sensitivity and temporal resolution.
Li et al. developed an optofluidic nanoplasmonic sensor capable of
real-time quantification of cytokine secretion at the single-cell
level, achieving detection limits in the low pg/mL range and temporal
resolution on the order of seconds (Figure
B). This concept
was later expanded from a single-plex to a high-throughput format
through integration with microwell microfluidics (Figure
C).
,
These platforms revealed pronounced functional heterogeneity in
immune cell populations, including asynchronous secretion onset times,
burst-like release patterns, and cell-to-cell variability in cumulative
cytokine output, features that are inaccessible to bulk or end point
assays. Similarly, Juan-Colás et al. employed resonant photonic
crystal sensors combined with hyperspectral imaging to achieve single-cell
cytokine mapping, providing quantitative measurements alongside spatial
distribution profiles. Despite these
advances, label-free approaches capable of simultaneously detecting
a broad panel of cytokines relevant to antitumor immune responses
are still lacking, highlighting an important opportunity for biosensor
technologies in immunotherapy development.
In addition, these
technologies could support the development of
novel in vivo CAR-T strategies, such as lipid-nanoparticle-mediated
delivery, by enabling early functional assessment of T-cell activation,
antigen recognition, and preliminary safety evaluation in biomimetic
models. By providing quantitative, real-time readouts of cytokine
secretion, receptor engagement, and cell–cell interactions,
evanescent wave-based photonic biosensors can reveal critical information
about CAR-T behavior before in vivo administration. Such applications
may help optimize delivery systems, guide CAR design to enhance efficacy
and specificity, and detect potential off-target or overstimulation
effects early in the development pipeline. Furthermore, coupling these
biosensors with high-throughput or multiplexed platforms could accelerate
iterative testing of multiple CAR constructs, nanoparticle formulations,
or costimulatory signals, providing rapid feedback on therapeutic
potential and informing rational decisions prior to clinical translation.
Taken together, evanescent wave-based photonic biosensors offer
a versatile platform for improving the analysis, optimization, and
evaluation of cell immunotherapies, bridging the gap between experimental
studies and clinical translation. Looking ahead, research should focus
on expanding the capabilities of these platforms to more comprehensively
model and monitor complex tumor–immune interactions. Strategies
to overcome the spatial limitations of surface-based sensing, such
as three-dimensional biomimetic scaffolds, microfluidic cell coculture
systems, or engineered extracellular matrices, could enable the study
of immune-cell infiltration and synapse formation in physiologically
relevant contexts. In parallel, the development of multiplexed biosensing
platforms capable of simultaneously tracking multiple cytokines, immune
checkpoints, and receptor–ligand interactions at the single-cell
level will be critical for capturing the dynamic heterogeneity of
antitumor immune responses. Ultimately, advancing label-free biosensor
technologies in these directions has the potential to accelerate the
rational design, evaluation, and clinical translation of next-generation
cell immunotherapies while providing deeper mechanistic insights into
immune function and tumor biology.

Applications in Gene Regulation Strategies
Gene regulation
at multiple levels, including epigenetic modifications,
chromatin accessibility, AS, and noncoding RNA regulation, is emerging
as a critical determinant in cancer progression, immune cell function,
and therapeutic response. Dysregulation in these pathways contributes
to tumor heterogeneity, therapy resistance, and the variable efficacy
of immunotherapies. Epigenetic alterations, such as DNA methylation,
can influence the expression of key immune checkpoints and the activity
of CAR-T cells.
,
Hypermethylation of promoter
regions in tumor cells often leads to the silencing of antigen-processing
machinery and downregulation of TAAs, reducing immune recognition.
Conversely, hypomethylation within regulatory regions of genes such
as PD-L1 can increase checkpoint ligand expression, driving T-cell
exhaustion and enabling immune escape. In engineered immune cells,
targeted modulation of methylation states has been shown to stabilize
effector phenotypes, enhance memory formation, and reduce the acquisition
of exhaustion-associated transcriptional programs, thereby improving
the in vivo persistence of CAR-T cells. Aberrant splicing and noncoding
RNA expression further modulate both tumor and immune cell behavior.
,
For example, AS generates isoforms of PD-1, CTLA-4, CD28, and CD19,
with distinct signaling capacities, some of which can attenuate T-cell
activation or enable tumor immune evasion. Similarly, modulating splicing
factors such as SRSF2 or hnRNPA1 can tune cytokine production, exhaustion
profiles, and metabolic fitness of T cells. Several microRNAs, such
as miR-155 or miR-17–92, can enhance T-cell activation, proliferation,
and resistance to exhaustion, while long noncoding RNAs regulate chromatin
accessibility and transcriptional programs tied to cytotoxicity memory
formation or tumor-induced dysfunction. Manipulating these DNA- and
RNA-based regulatory networks through genetic editing, epigenetic
drugs, or delivery of synthetic RNA modulators offers complementary
strategies to boost the therapeutic efficacy, durability, and resilience
of immunotherapies. Therefore, analyzing and monitoring these dynamic
regulatory events in real time could provide valuable insights for
the design, evaluation, and optimization of next-generation cancer
therapies.
Evanescent wave-based photonic biosensors have been
demonstrated
as versatile tools for detecting DNA methylation and balancing sensitivity,
specificity, and sample preparation requirements. For example, Yoon
et al. integrated bisulfite conversion, amplification, and detection
into a lab-on-a-chip platform with microring resonators, enabling
rapid and sensitive detection of as little as 1% methylated DNA in
mixed samples and input DNA amounts in the low nanogram range (Figure
A). An alternative strategy that bypasses bisulfite conversion
is the use of antibody-based biosensors, which enable rapid detection
by directly capturing methylated cytosines, typically achieving nM
sensitivity, although they generally lack sequence specificity. Direct
recognition of double-stranded DNA using triplex-forming hairpin probes
or CpG-binding proteins has also been demonstrated, providing real-time,
label-free quantification of methylated regions (Figure
B). In particular, this biosensor enabled discrimination of methylated
versus unmethylated DNA fragments with detection limits in the low
nanomolar range and without enzymatic preprocessing while preserving
sequence selectivity. In addition, optical biosensors can discriminate
cytosine oxidation products, such as 5-hmC and 5-fC, through surface
functionalization with selective antibodies or chemical receptors,
with reported sensitivities down to the sub-nM range, enabling differentiation
between closely related epigenetic modifications. These advances illustrate the potential of label-free optical
platforms to provide rapid, sensitive, and increasingly sequence-
or modification-specific DNA methylation analysis, paving the way
for applications in cancer epigenetics and therapy monitoring.
These optical biosensors can also be used for the
study of AS events
and their modulation in cancer and immunotherapy. AS regulates proteomic
diversity and influences key immune and oncogenic pathways, making
its precise monitoring crucial for understanding the therapy response.
By facilitating rapid screening of splice-modulating drugs or CRISPR-based
editing strategies, such platforms could provide actionable insights
for optimizing therapeutic efficacy. Huertas
et al. achieved highly selective detection of splice variants from
the Fas and Bcl-x geneskey regulators of apoptosisusing
different optical biosensors capable of distinguishing isoforms with
negligible cross-reactivity (<5%), within 20 min, at pM to nM detection
limits, depending on probe design and target length (Figure
C).
,
Precleaving long mRNA molecules before hybridization can further
enhance accessibility and sensitivity, yielding up to 1 order of magnitude
improvement in the signal-to-noise ratio and detection limits, making
the performance comparable to that of standard PCR-based assays. Furthermore, the integration of photonic biosensors
with microfluidic systems enables single-cell trapping and analysis,
which is a highly relevant approach for studying AS dynamics in immune
cells during immunotherapy.
Beyond epigenetic and splicing regulation,
noncoding RNAs (ncRNAs)
play a pivotal role in modulating immune cell differentiation, effector
functions, and resistance to immunotherapy. Aberrant ncRNA expression
can influence immune checkpoints, cytokine signaling, and the cytotoxic
potential of T cells, thus affecting the outcome of treatments such
as checkpoint blockade or CAR-T cell therapy. Therapeutic approaches
targeting ncRNAs, such as using miRNA mimics to restore tumor-suppressive
functions, antagomirs to silence oncogenic miRNAs, or engineering
immune cells with defined ncRNA profiles, are being actively explored
to enhance immune responses and overcome resistance. In this context,
label-free photonic biosensors could be used to monitor ncRNA-mediated
regulatory processes in real time, bridging the gap between the mechanistic
understanding and therapeutic application. For example, integrated
photonic devices such as BiMW, achieving attomolar detection limits
for miRNAs directly in biofluids without labeling or amplification
(Figure
D), could be exploited to track miRNA signatures
that modulate T-cell exhaustion, antigen presentation, or cytokine
signaling during immunotherapy. Similarly, the direct, amplification-free
detection of circulating miRNAs at fM to aM concentrations, as demonstrated
by Calvo-Lozano et al., illustrates how these sensors could monitor
therapy-induced molecular changes within analysis times below 30 min
dynamically and noninvasively, supporting personalized treatment adjustment. The same sensing principles can be extended
to long noncoding RNAs (lncRNAs), which are emerging as critical regulators
of chromatin remodeling and immune checkpoint expression. Photonic
biosensors employing microring resonators have already demonstrated
multiplex detection of lncRNAs targets with pM sensitivity and high
sequence selectivity in cancer models, underscoring their adaptability
to complex RNA targets (Figure
E). Integrating these sensing
architectures with microfluidic or single-cell systems could allow
time-resolved monitoring of ncRNA-driven transcriptional reprogramming
in immune cells exposed to therapeutic stimuli or gene-editing interventions.
Finally, the ability of optical biosensors to directly quantify exosomal
ncRNAs in minimally processed samples could open new opportunities
to study intercellular communication between tumors and immune cells.
Since exosome-associated ncRNAs can mediate resistance to checkpoint
blockade or modulate immune activation, real-time monitoring of their
dynamics using photonic biosensors could provide functional insights
inaccessible to conventional molecular assays. Overall, these advances
highlight how label-free optical biosensing platforms can evolve from
diagnostic tools to mechanistic probes, enabling the dynamic, high-resolution
interrogation of ncRNA-mediated regulation in next-generation immunotherapies.
By providing a unified, label-free, and dynamic framework, optical
biosensors allow comprehensive profiling of gene-regulatory states
in tumor cells, engineered immune cells, or patient-derived samples,
offering a direct link between molecular interventions and their functional
impact on immune responses and guiding the development of more precise
and effective immunotherapy strategies. Future research should focus
on expanding evanescent wave-based photonic biosensors to enable multiplexed,
single-cell monitoring of dynamic gene-regulatory events in tumor
and immune cells. Integration with microfluidic, organoid, or patient-derived
systems could provide real-time insights into how therapeutic interventions
reshape cellular function and immune responses. Enhancing sensor sensitivity
and specificity will allow simultaneous detection of multiple regulatory
targets and functional readouts, such as cytokine secretion or T-cell
activation, offering a holistic view of therapy efficacy and safety.
These advances have the potential to accelerate the rational design,
optimization, and personalization of next-generation immunotherapies
through dynamic, mechanistic feedback.

Major Challenges and Opportunities
While promising
development has been made, the integration of plasmonic
and silicon photonic biosensors into the study and evaluation of advanced
immunotherapies remains at an early stage and is currently constrained
by several technological and translational bottlenecks. At present,
many of the reported applications remain proof-of-concept or early
translational studies in which performance metrics are typically obtained
under controlled or simplified experimental conditions; therefore,
more extensive validation using clinically relevant samples and larger
cohorts will be required to rigorously establish sensitivity, specificity,
and robustness for routine biomedical implementation. Besides, the
complexity of immunotherapeutic mechanisms, ranging from molecular
checkpoint modulation to the reprogramming of immune and tumor cells
through gene-regulatory interventions, poses significant technical
and analytical challenges that must be addressed through targeted
methodological and engineering advances for these platforms to reach
their full potential.
Among the most critical bottlenecks for
the practical deployment
of evanescent wave-based photonic biosensors is their performance
in complex biological samples. Biosensors must operate reliably in
matrices such as blood, tumor interstitial fluid, or complex culture
media where the abundance of nonspecific biomolecules can mask or
distort the signal from target analytes. This is particularly problematic
when studying low-abundance regulatory molecules such as specific
splice variants, ncRNAs, or methylated DNA regions, which often require
detection limits in the femto- to attomolar range. Maintaining such
sensitivity and selectivity in clinically relevant samples remains
a central challenge. At the same time, this challenge opens opportunities
for the development of innovative surface chemistries, including antifouling
polymer brushes, zwitterionic coatings, and biomimetic membranes that
preserve sensor sensitivity while minimizing background noise. Coupling these with automated lab-on-a-chip
modules for rapid sample preparation, such as on-chip plasma separation,
nucleic acid extraction, or exosome isolation, could substantially
enhance applicability.
,
These strategies can preserve
sensor specificity and sensitivity in complex biological fluids, enabling
direct, label-free detection in a real-time format.
Another
crucial limitation lies in multiplexing capability. Immunotherapies
act on intricate molecular networks involving immune checkpoints,
cytokines, transcriptional regulators, and epigenetic modifiers. Capturing
this complexity requires the simultaneous detection of multiple molecular
and cellular biomarkers in a single assay. Although photonic architectures
such as microring resonators and bimodal interferometers have demonstrated
multiplexed detection at the research level, expanding this capability
to complex samples without cross-reactivity or signal interference
is still technically demanding. Promising research directions include
the use of advanced computational strategies based on wavelength division
or spectral deconvolution to expand multiplex capacity without compromising
sensitivity.
,
Advances in integrated photonics,
such as on-chip tunable lasers or broadband frequency combs, could
further support parallel detection of large biomarker panels relevant
to immune activity.
,

Integration with cell-based
and functional assays represents an
additional major challenge and opportunity. The effectiveness of immunotherapies
depends not only on molecular signatures but also on the dynamic behavior
of immune cellssuch as activation, exhaustion, or cytotoxic
functionwhich evolve over time in response to therapeutic
interventions. Combining label-free optical biosensing with high-throughput
microfluidic cell-trapping arrays or lab-on-chip platforms could enable
real-time monitoring of immune cell interactions, signaling events,
and phenotypic transitions under controlled conditions, providing
a direct link between molecular regulation and functional immune outcomes. However, realizing this potential requires overcoming
substantial engineering and analytical challenges. Maintaining cell
viability, ensuring precise analyte delivery, and achieving stable
and reproducible optical readouts within complex microenvironments
remain key obstacles. These issues also highlight opportunities for
next-generation microfluidics, including single-cell encapsulation,
dynamic perfusion systems, and programmable microvalve networks that
precisely modulate the cellular microenvironment.
,,
Furthermore, dynamic and longitudinal monitoring
of the therapy response introduces additional layers of complexity.
Immunotherapies often induce time-dependent molecular and cellular
adaptations, necessitating biosensors capable of continuous or repeated
measurements over extended periods. Achieving this will depend on
advances in sensor robustness, automated microfluidic control, and
data acquisition systems with emerging opportunities in the development
of self-calibrating optical architectures, photonic materials with
improved thermal and mechanical stability, and microfluidic reservoirs
that support long-term culture and stimulation cycles.
Additionally,
the integration of 3D-printed microfluidics or organ-on-chip
(OoC) modules could further enable more physiologically relevant in
vitro models for evaluating immunotherapy performance. Moreover, following the recent FDA announcement
formally recognizing organ-on-chip platforms as valid disease models
for regulatory testing, the combination of OoC systems with integrated
biosensing further underscores their potential to bridge the gap between
conventional in vitro assays and clinical outcomes. These biosensors allow continuous, noninvasive monitoring
of molecular and cellular events, such as cytokine secretion, receptor–ligand
interactions, or immune cell infiltration, directly within the microfluidic
compartments of the OoC system. By providing real-time data without
the need for fluorescent labels, label-free biosensors can capture
dynamic biological processes, link molecular signatures to functional
outcomes, and facilitate mechanistic studies of therapeutic responses.
This approach could accelerate preclinical testing, support the optimization
of dosing and treatment schedules, and provide actionable insights
for the rational design of next-generation immunotherapies under conditions
that closely mimic human tissue physiology.
Equally important,
and often underestimated as a bottleneck, is
the need for advanced computational frameworks to interpret the vast
data sets generated by multiplexed, time-resolved biosensing. These
data sets, which capture kinetic binding events and cellular dynamics
in real time, require sophisticated algorithms for noise filtering,
kinetic modeling, and correlation with clinical or biological outcomes.
The integration of artificial intelligence and machine learning could
significantly enhance this process by automatically detecting subtle
patterns and correlations across large multidimensional data sets,
identifying predictive biomarkers of therapy response, and optimizing
experimental parameters such as analyte concentrations or timing of
interventions. AI-driven approaches can also support predictive modeling
of immune dynamics, enabling simulation of therapeutic outcomes and
guiding rational experimental design. Furthermore, by combining biosensor
readouts with clinical and omics data, data-driven analytics can inform
personalized treatment regimens, adapt dosing schedules in real time,
and uncover novel mechanistic insights into tumor–immune interactions.
Ultimately, the convergence of optical biosensing, microfluidics,
and data-driven analytics holds great promise for developing adaptive,
feedback-informed immunotherapies capable of responding to the evolving
landscape of tumor–immune interactions in real time.
Beyond analytical performance, the broader adoption of evanescent
wave-based photonic biosensors in cancer immunotherapy research will
critically depend on progress in standardization, large-scale validation,
and economic scalability. At present, the technological maturity of
the different platforms remains heterogeneous. Classical SPR systems
benefit from relatively established commercial instrumentation and
partial standardization in biochemical analysis, yet their translation
into routine clinical workflows remains limited. Emerging nanoplasmonic
configurations offer enhanced miniaturization and multiplexing potential
but frequently rely on custom nanofabrication processes and laboratory-specific
optical setups that hinder interlaboratory reproducibility and cost
comparability. A similar variability is observed in integrated silicon
photonic technologies that, although compatible with CMOS manufacturing
and theoretically amenable to large-scale production, are still predominantly
implemented as research prototypes with diverse surface chemistries,
coupling schemes, and microfluidic integrations. Consequently, reported
fabrication costs, instrumentation requirements, and operational expenses
vary widely and are rarely assessed through standardized technoeconomic
analyses, making direct cross-platform comparisons premature. Equally
important is the limited availability of multicenter validation studies
using clinically relevant cohorts, as most reported applications remain
as proof-of-concept demonstrations or small-scale laboratory investigations.
Establishing shared reference materials, harmonized biofunctionalization
protocols, interlaboratory benchmarking initiatives, and regulatory-aligned
performance criteria will therefore be essential steps toward transforming
these highly sensitive analytical technologies from promising experimental
tools into robust and standardized platforms capable of supporting
the development and optimization of next-generation cancer immunotherapies.
Overall, while evanescent wave-based photonic biosensors have already
proven their sensitivity, specificity, and versatility for molecular
detection, their broader application to immunotherapy evaluation and
optimization is still limited by identifiable technological and translational
bottlenecks and defines clear research directions for the field. Systematically
addressing these challenges through coordinated advances in biofunctionalization,
microfluidic and system integration, industrialization, and data-driven
analytics will not only enhance our ability to study immune mechanisms
in real time but also establish a new generation of analytical tools
capable of guiding personalized, adaptive immunotherapies.

Outlook
Photonic biosensors are evolving from analytical
tools into translational
interfaces that bridge molecular readouts with therapeutic decision-making.
By enabling real-time, label-free monitoring of immune cell interactions,
cytokine secretion, receptor–ligand dynamics, and ncRNA signatures,
these platforms can provide quantitative feedback on therapy specificity,
efficacy, and patient-specific responses. Such capabilities are particularly
relevant for advanced cancer immunotherapies, where dynamic cellular
and molecular changes often determine therapeutic success or resistance.
The integration of photonic biosensors with microfluidic and lab-on-a-chip
systems further enhances their potential, allowing high-throughput,
single-cell, and longitudinal analyses under physiologically relevant
conditions. This convergence creates opportunities to link molecular
regulation, cell function, and treatment response directly, enabling
the rapid optimization of molecular immunotherapies, CAR-T cell designs,
and gene- or epigenetic-modulatory interventions. Moreover, multiplexed
and high sensitivity biosensing can capture the complexity of immune
and tumor signaling networks, providing a holistic view of therapeutic
mechanisms that is difficult to achieve with conventional assays.
Looking forward, the combination of label-free optical biosensing
with automated microfluidics, advanced biofunctionalization, and machine-learning-driven
data analysis promises to transform the preclinical and clinical evaluation
of cancer therapies. These integrated platforms could facilitate adaptive,
feedback-informed immunotherapy strategies, guide personalized treatment
regimens, and accelerate the translation of next-generation therapies
from the bench to bedside. Ultimately, photonic biosensors are poised
to become a cornerstone of precision oncology, providing the dynamic
mechanistic insights required to realize the promise of immunotherapy
in modern cancer care.