FBG is an intrinsic fiber element inside a short segment of an optical fiber that has a certain range of reflected and transmitted wavelengths of light. This is achieved by periodic modulation of the RI inside the core by illuminating it with an ultraviolet (UV) light to build grating structure. As light enters the core of the fiber and travels through FBG structure, the wavelength is selectively reflected by Bragg gratings (Mishra et al., 2011). Packaged FBG sensors used gratings as active sensing elements in several works. FBG at a much lower wavelength was generated by Arasu et al. using white light source with further gold coating for achieving surface plasmon resonance (SPR) effect (Arasu et al., 2016). The gratings generated by this way refract the light to the cladding, but the light does not escape from the fiber due to total internal reflection (TIR) producing evanescent waves which then can be used as SPR. Gold coated tilted FBG (TFBG) packaged in a hollow cylindrical needle were used as a sensing tool for cancer biomarker testing in tissue (Ribaut et al., 2017). Another study used FBG as a second element: temperature compensation during DNA measurements with an active element being micro-structured optical fiber (MOF).

The D-shaped plastic optical fiber (POF) sensors have also been developed and packaged for biosensing (Hu et al., 2023). The sensors were produced using multimode POF which are first fixed either by encapsulating the fiber within a solid block of resin or by placing it in the V-grove with the shape that exactly matches a semi-circle and then polishing it, so it exposes the core, resulting in a D-shaped profile. The polishing can be done using a computer numerical control (CNC) micro-milling machine or by manually grinding using a sandpaper (Bo et al., 2023; Seitz, 2024). The D-shaped POF for respiration monitoring was integrated the sensor into an elastic band. Another POF with a side-polish length of 50 mm and a polish depth of 7 mm was used to detect anti-coagulant in a sample circulation unit (Wang Y. L. et al., 2021).

D-shaped sensors were further modified in some cases for producing biosensors. For instance, the exposed core is drilled with microholes that have a diameter of around 600 µm and a depth of 500 µm. The sensor contains a SPR probe and a capture platform including three microholes containing receptor solutions. Molecular interactions may be detected by observing the changes in light propagation in the POF core caused by the binding of target molecules in the microholes (Seitz, 2024). D-shaped fiber was also used as platform for subscribing gratings with subsequent etching to produce highly sensitive RI sensors (Tahhan et al., 2017).

One of the simplest ways for fabricating optical fiber sensor components is based on tapering a relatively small piece of optical fiber with length varying from sub-millimeter to tens of millimeters (Korposh et al., 2019). Tapered optical fiber provide several advantages for sensor development, including wide evanescent fields, flexibility, and compactness and is produced by heating a specific segment of the fiber while drawing the two ends of the fiber together (Kamil et al., 2019; Korposh et al., 2019). For some very fragile sensors such as tapered optical fibers, microfluidic chip or a flow cell is needed for protection of the sensor (Tian et al., 2011). The geometrical structure of tapered optical fiber is nonuniform within the sensing region, liquid handling needs an improvement because during each measurement it is crucial to ensure that the liquid is removed completely from the optical fiber. With multiple measurement, it is difficult to handle the tapered optical fiber due to its delicate structure (Abdul Hamid et al., 2017).

The taper fiber or microwire employed single-mode fibre-28 (SMF-28) with two 12 mm concatenated tapers separated by 10 mm. The first down taper zone had no full cladding-core mode recoupling and provided an evanescent field at the slop. Both propagating core mode and stimulated cladding modes displayed a differential phase shift in the up-taper region, demonstrating interference at the transition. If transition slopes were fast enough, cladding-core mode recoupling was poor, and light passed through. A concatenated tapered microwire’s output spectrum demonstrated spectral modulation with regularly spaced notch times inversely linked to tapered region separations (Shrivastav et al., 2019).

The thin-core microfiber (TCF) was developed by splicing it between two SMFs using a fusion splicer. A nonadiabatic taper was formed on the TCF by using a fusion device to induce an arc discharge. The taper was improved by subjecting it to heat from a hydrogen flame and stretching the fiber using an optical fiber pulling machine. The ultimate TCF possessed a waist diameter of around 6.4 µm and a waist length of about 2,357 μm, leading to a transmission spectrum with a free spectral range of 35.28 nm (Chen et al., 2023).

A ball resonator sensor (Shaimerdenova et al., 2020) is another type of optical fiber packaged for biosensing. It operates as a weak interferometer, and exhibited superficial fringe patterns and a spectrum that showed characteristics to a random signal. It can be manufactured using a commercially available CO₂ splicing machine. The sensitivity of the resonators’ RI might be assessed by measuring either the shift in wavelength or the change in amplitude. The fiber possesses a ball lens on its tip with diameter of around 500 nm.

Another sensor is a single-multi-mode fiber optic coupler (SMFC) which was employed to transmit incident light and capture Fresnel reflected light, simplifying the system by eliminating the requirement for optical separation components and improving light transmission efficiency. This layout significantly enhanced the sensitivity of the all-fiber Fresnel reflection microfluidic biosensor (FRMB). In the bio-probe made of multimode fiber, the incident light that enters the fiber is reflected at the far end, and the strength of the reflected light changes depending on the local RI. The binding of target biomolecules to the bio-probe leads to a rise in RI, resulting in a decrease in the intensity of reflected light. This allows for the identification of target molecules without the need for labeling, based on their concentration (Xu et al., 2021). A Fabry-Pérot (F–P) sensor was developed by employing two-photon photopolymerization (TPP) technique to construct a microcantilever structure by photolithography located at the tip of an optical fiber (Li et al., 2022).

Optical fiber sensors in different packages for various applications are shown in the middle layer of Figure 1 and vary from flow cells and microfluidic chips to medical and wearable devices. Flow cells are typically utilized where a constant sample flow is applied to the sensor surface for applications requiring continuous monitoring and dynamic measurement (Liu et al., 2015; Arasu et al., 2016). The enormous use of microfluidic chips in biosensing applications prevails because they can accommodate large volumes of fluids with high precision (Bo et al., 2023). In this same microfluidic chip, optical fiber sensors can be integrated to monitor real-time changes in the properties of the fluid being handled (Niu et al., 2022). Furthermore, optical fiber sensing is a known technology incorporated within syringes, catheters, and other medical diagnostic and treatment monitoring devices (Sypabekova et al., 2022; Zhou et al., 2022). These devices can be pre-made or custom-modified to include optical sensing capabilities (Bekmurzayeva et al., 2022). Wearable optical fiber sensors can be sewn into fabrics or flexible bands to form a smart patch or elastic band. It is designed to monitor physiological parameters such as heart rate, respiration, and body glucose level by continuous monitoring (Wang Y. L. et al., 2021; Lo Presti et al., 2022; Han et al., 2023).

The development of OFS packaging relies on specific design requirements and intended applications. For microfluidic chips, photolithography and soft lithography methods are used (Soni et al., 2021). A silicon wafer is first used to create detailed microchannel patterns, which serve as molds to cast polydimethylsiloxane (PDMS). This layer is then bonded to a glass slide, forming the final device. Flow cells are fabricated through precision machining or micro-molding, often using transparent materials like glass and polymethyl methacrylate (PMMA) (Liu et al., 2015; Arasu et al., 2016). In most medical devices, optical fibers are passed through pre-existing channels or grooves in needles and catheters (Fan et al., 2016). Structures can be modified to fit application needs; for instance, needles can be thermo-molded and encased in thermoplastic, with a window for sensor exposure (Ribaut et al., 2017). Modifications include attaching microdialysis membranes for continuous protein measurement (Ngernsutivorakul et al., 2017) or making holes in catheters for blood-mimicking fluid experiments (Myrkhiyeva et al., 2024). In wearables, optical fiber sensors are integrated into textiles or flexible polymers using sewing, knitting, and lamination techniques (Han et al., 2023).

Different fiber optic sensor packages vary in design and materials based on their functionalities and purposes. Flow cells, for instance, can be rectangular, cylindrical, cuvette-shaped, or spiral to enhance path length or fiber positioning. They may include static or inline mixers, observation windows for spectroscopic analysis, electrodes for electrochemical studies, and membranes for specific reactions (Liu et al., 2015; Kaushik et al., 2019; Shen et al., 2020; Soni et al., 2021). A spiral flow cell for cell culture could incorporate membrane filtration to prevent clogging (Yin et al., 2021).

Microfluidic chips, typically thin and flat with etched microchannels, can feature complex shapes with multiple inlets, outlets, and chambers. They facilitate droplet generation, efficient fluid mixing through micromixers, and fluid control via micromechanical valves. (Liu et al., 2018; Wang H. et al., 2021). Medical device packages, such as long flexible tubes, vary in diameter and tip design like catheters. They might include expandable balloons for angioplasty, side ports for drug delivery, pressure or temperature sensors, and multiple lumens for delivering various medications. For example, a urinary catheter might have a simple drainage port, while a central venous catheter could have multiple lumens (Poeggel et al., 2015; Zhou et al., 2022). Syringes range from pre-filled types for convenience to insulin syringes for precise dosing and catheter-tipped syringes for accurate placement (Wu et al., 2015). Needles, used in various medical applications, come in different bevel angles, lengths, and gauges, including hypodermic, spinal, butterfly, and blunt needles (Hu et al., 2018; Hsieh et al., 2022). Wearable fiber optic probes, such as smart patches and elastic bands, are designed to be flexible and conform to body contours, integrating sensors for vital signs, biochemical markers, and drug delivery via microneedles. Some also include electrodes for muscle or nerve stimulation and wireless communication for data transmission (Lo Presti et al., 2022).

Packaging materials for optical fiber sensors need to be transparent, biocompatible, and mechanically stable. Table 2 shows common materials used for building microfluidic chips for packaged OFB. PMMA is preferred for its optical clarity and mechanical properties, while PDMS is favored for microfluidic chips due to its flexibility and ease of fabrication (Abdul Hamid et al., 2017; Shakarim et al., 2022). Transparency is crucial for biochips used in cell viability studies, often using materials like glass or water-transparent membranes (Huang et al., 2020). For fluorescence-based biosensors, materials with low background fluorescence are chosen (Ngernsutivorakul et al., 2017).

When it comes to medical devices, the material is usually biocompatible and approved for use in humans. These materials include stainless steel for needles. Silicone and rubber are used in industries to make syringes and catheters for medical purposes (Zhou et al., 2022). However, the main challenge with these devices is that their structure should sometimes be modified to both fit the optical fiber sensor and the final application. Another thing to consider when it comes to needles/catheters is their gauge/diameter. Thus, a larger fiber (400 µm) where encased in 21-gauge needles while fibers with smaller diameter (100 µm) fibers in 31-gauge needles (Hsu et al., 2017). This also applies to the intended blood vessel (for in vivo applications): smaller blood vessels will require larger gauge needles/catheters and vice versa.

The setup, especially the optical part of the setup will depend on the sensor type as can be seen from Figures 2–4. Thus, a typical setup for plasmonic TFBG, for example, includes a broadband light source for excitation of the sensor, a linear polarizer and polarization controller for adjusting and orienting the polarization state of the light which passed through the FBG, a sensor and a downstream OSA as an interrogator of tunable linearly polarized light. The inlet and outlet of the analytes in the microfluidic system is before and after the sensor respectively as in these works (Guo et al., 2014; Guo et al., 2016). For LPFG in DCF, a broadband optical source to launch the light and an OSA downstream can be used as in the study by Hsu et al. (Hsu et al., 2011). For a tapered sensor, a setup could include an optical sensing interrogator such as Micron Optics si720 which would excite the fiber from one end and monitor the transmission from the other end of the fiber while the sensing region is integrated into a microfluidic chip (Tian et al., 2011). An optical backscatter reflectometry can be used both to launch the light and measure the reflected light through a ball resonator sensor inserted into a plastic tube via a catheter/cannula system (Zhang et al., 2019). Another option is using a light source-sensor (Verma and Gupta, 2013). An experimental setup for monitoring chemotherapy-induced apoptosis involved a fiber-optic biosensor with a customized sensing probe made of optical fibers and microcapillaries. The system included light sources connected to the optical fibers, a spectrophotometer to analyze the collected spectra, and a syringe pump to deliver the fluorescent dyes (Fan et al., 2016).

Securely fixing sensors is important in the fabrication of packaged biosensors, particularly those integrated into microfluidic chips, to ensure their optimal performance. One of the methods used for this purpose involves using of UV-sensitive adhesives (Guo et al., 2014). Employing a UV-curing adhesive (NOA 61) to secure a WGM fiber probe inside the microfluidic channel used by capillary action, securely attaches the probe and effectively connects the passage to prevent any potential leakage. The WGM fiber probe is firmly placed under UV light to provide a one-stop assay. This method effectively integrates sample handling, reaction, separation, and detection within the adaptable structure of the microfluidic channel (Niu et al., 2022). Similarly, optical fibers are secured in position by using optical glue (NOA 68, Norland) within V-shaped grooves located on 2 bars of the apparatus which have a depth of around 0.3 mm. The glue undergoes polymerization when exposed to UV light, resulting in a solid state within a time frame of about 15–20 min. This process leads to the stable alignment of the fibers (Esposito et al., 2021).

Representative ligand-analyte pairs which were used in building packaged OFB include antibody-antigen, receptor-ligand (estrogen receptor and estradiol), enzyme-substrate (glucose oxidase-glucose), aptamer-target, ssDNA probe-target DNA, MIP-target. Figure 6 shows the variety of analytes measured by the optical fiber bio-sensors integrated with external packaging. This includes small molecules, proteins, viruses, bacteria, cells and even tissues.

Accurate temperature measurement is crucial in the domain of packed biosensors since it directly impacts the biochemical processes and sensor responses in microfluidic systems. Diverse techniques have been utilized to precisely monitor and regulate temperature in these environments. A technique involves utilizing FBG, which can do detection in real-time and track temperature fluctuations in the solution as it flows through the microchannel. This technology is competent not only in measuring temperature but also in identifying DNA hybridization events (Hu et al., 2023).

Temperature regulation can be achieved by employing a thermistor, thermocouple with an interconnected thermometric feedback control unit which is integrated into the flow cell or microfluidic channel (Aray et al., 2016; Khatri et al., 2018). For example, in a particular design, both bars of the device have a lateral hole with a 2 mm diameter, which is appropriate for accommodating a thermocouple. The thermocouple, located at the bottom bar, is an integral component of a temperature control system. It sends feedback to an electronic driver which in turn regulates the Peltier cells. On contrast, the thermocouple on the upper bar is linked to a thermometric device (Lutron TM-917) that gauges the temperature within the fluidic channel. This configuration is especially vital in experiments that involve solutions sensitive to temperature, as it guarantees that the solutions attain the required temperature specified by the control system, even if they are originally held at lower temperatures, such as in a refrigerator (Esposito et al., 2021).

The change of response for every unit of an analyte concentration for a biosensor is called sensitivity. The sensitivity (s) is calculated as the ratio of the signal amplitude (y) to the unit concentration (x) of the response slope, which is given by the slope of the linear area in the fitting curve. A greater y value for a comparable x value signifies superior sensitivity performance of the biosensor (Prabowo et al., 2021).

Sensitivity of the biosensor highly depends on the type of optical fiber which is employed as a sensor. For instance, a higher sensitivity of Fresnel reflection microfluidic biosensor was due to four grades higher interaction area compared to SMF which in turn leads to increased interaction of the target molecules and the immobilized ligand. Moreover, compact structure of the fiber enhances the effectiveness of optical transmission and collection and lowers light loss (Xu et al., 2021).

For some biosensors, the packaging did not affect its sensitivity but enhanced analyte detection range. The research results by Nguyen et al. suggested that the microfluidic chip had a sensitivity that is nearly identical to the LPG glucose biosensor when tested outside the chip (Nguyen et al., 2017). Furthermore, the chip had a broadened detection range spanning from 2 μM to 10 µM. The microfluidic device had the ability to accurately detect glucose concentrations as low as 1 nM. The integration of the sensor into the microfluidic chip increased its detection range from 2 μM to 10 µM (Yin et al., 2016). Upon placement into the microfluidic chip, the response time was significantly reduced from 6 min to 70 s.

One work presented a new sensing chip that incorporated three microholes, which detects variations in light propagation characteristics caused by the binding of targets and receptors, which was likely due to local temperature changes resulting from the binding response. The concentration of receptors in the microholes may be readily increased in contrast to typical sensors where receptors are bound to a solid substrate. This design greatly improves the sensitivity compared to traditional SPR and MIP-based SPR sensors. It enables the detection of interactions between target molecules and receptors at concentrations as low as aM levels, which is equivalent to a few hundred molecules in a 20 µL sample that is applied to the D-shaped plastic POF and diffused into the microcuvettes filled with receptors (Cennamo et al., 2024).

For other packaged OFB, the sensitivity was enhanced by the means of materials. The sensitivity of plasmonic FBG was enhanced using nanostructured graphene oxide coating (Arasu et al., 2016). Titanium dioxide nanostructured metal oxide was coated over the etched D-shaped FBG sensors for the purpose of increasing the sensing area of the sensor which ultimately led to the increase of its sensitivity to RI in a dynamic setup (Tahhan et al., 2017). The experimental findings demonstrated that the SDSMF-LPG biosensor, employing a hybrid sensing film, had the capability to detect glucose concentration with sensitivity, reaching as low as 1 nM (Yin et al., 2016).

According to IUPAC, LOD is defined as the lowest concentration of a substance that can be confidently detected, signified by a specific value that indicates the minimum measurable concentration within a given confidence level (Prabowo et al., 2021). A crucial factor in assessing the performance of DNA biosensors (and other biosensors) was their specific LOD. A lower detection limit is anticipated compared to MIP-based devices due to the potentially higher affinity of bioreceptors in aqueous solutions (homogeneous phase chemical reactions) compared to the binding sites in MIP (adsorption reactions) (Cennamo et al., 2024). Performance-wise, some of such biosensors could achieve very low LOD such was the case in the following studies: 0.24 p.m. (Wang et al., 2023) and 1.87 nM (Hu et al., 2023) for sensing DNA hybridization, detection of proteins at 9.3 fM in dynamic conditions (Bekmurzayeva et al., 2022).

Repeatability/reproducibility of biosensors is crucial for ensuring consistent performance and reliability in various applications. A microfluidic channel, among other advantages, is a useful tool to evaluate the performance of the developed biosensor, namely, testing its repeatability. The movement of the artificial urine sample containing p-cresol through the cell allowed studying the stability and repeatability of the probe. The sensor was also tested in more extreme non-laboratory conditions and was proved to measure p-cresol rapidly and accurately due to the stability of the ligand (Usha and Gupta, 2018). The effect of pH on the performance of a plasmonic biosensor, stability, and repeatability of the biosensor in several cycles of 0 and 10 mM of cholesterol was also performed (Semwal and Gupta, 2018). For assessing the reliability and stability of the sensor in experiments involving RI sensing, identical tests were performed ten times. During these tests, spectral fluctuations were recorded for different RI values. To tackle temperature sensitivity crosstalk to the RI, a matrix was created to assess temperature sensitivity and RI sensitivity simultaneously. The sensor exhibited excellent stability and reliability in repetitive sensing trials, with no impact from liquid velocity and placement angle, thereby showing great potential for on-chip and in situ optofluidic detection (Li et al., 2022). The repeatability and reproducibility of the integrated microchannel PCR equipment were comprehensively evaluated resulting in a strong correlation between the fluorescence data and sensor measurements validated the device’s dependable functionality and its capacity for repeated usage with thorough washing (Nguyen et al., 2017).

To evaluate the repeatability of samples, Bo et al., conducted a series of tests where a 10⁻⁷ mol/L solution of R6G was introduced into a microfluidic channel integrated with a D-shaped optical fiber SERS probe. After recording the R6G Raman signals, the channel was flushed with anhydrous ethanol to clean it, and the process was repeated five times demonstrating reliable repeatability of the microfluidic SERS probe (Bo et al., 2023).

The precision and repeatability of the BP-functionalized sensor was evaluated through a series of stability tests involving tracking the wavelength response of cDNA at a concentration of 1 nM. The results showed that the peak wavelength stabilized approximately 30 min into the measurement. Subsequent readings displayed remarkably consistent wavelength with the maximum discrepancy between these shifts, significantly lower than the typical random wavelength variations of the sensor. This consistency underscores the sensor’s reliable repeatability, affirming its effectiveness for precise applications (Wang et al., 2023).

In a study by Soni et al., the repeatability of microchip fabrications was tested by creating two separate batches of chips, each containing six devices with the same design. These devices were assessed for bulk RI sensitivity by exposing them to various concentrations of glycerol solutions, with the results measured over a consistent flow rate and duration. Results indicated a high degree of uniformity within each chip, with minimal measurement error. However, a notable variation among the devices suggested potential inconsistencies due to the variability in the on-campus fabrication facilities (Soni et al., 2021).

Sensor reproducibility, reusability, and stability were evaluated by testing different sensor chips functionalized and used on various days. Across five independent experiments with at least ten biosensors each, an average absolute resonance shift response was measured indicating a consistent bacteria concentration detection. Up to three regeneration cycles (using 10 mM glycine–HCl pH two solution) was shown to not affect the sensor’s performance. Storing the sensors at 4°C in phosphate-buffered saline (PBS) showed to be the most effective method, maintaining sensor functionality close to that of freshly prepared sensors. In contrast, accelerated aging at 35°C led to an increase in resonance shifts, likely due to antibody precipitation on the plasmonic surface, indicating less stability under these conditions (Soni et al., 2021).

Thus, the following strategies for testing the repeatability of packaged biosensors include: 1) preparing several biosensors and assessing their response to target analytes; 2) regenerating the biosensor’s surface to remove bound analytes, then repeating the measurements over several cycles. This approach ensures that the biosensors maintain consistent performance and reliability across multiple uses, providing a thorough validation of their operational stability and repeatability.

Specific sensors are those which, in an ideal world, recognize only its target analyte (Peveler et al., 2016). For this, optical fiber sensors, after being fabricated, are functionalized with ligands specific to the target of interest (antibodies, aptamers, molecularly imprinted polymers (MIP) or enzyme). However, achieving this ideal is difficult because of the similarity of analytes or the absence of specific ligands against some targets (Peveler et al., 2016). Being selective is one of the main strong features of biosensors compared to other analytical methods. This property allows the biosensor to discriminate the analyte of interest among other entities in a complex mixture with no pre-separation (Bucur et al., 2021).

The selectivity/specificity of the built biosensors can be measured in several strategies some of which were used in case of packaged OFB as shown in Table 3. One of the strategies is incubating fully functionalized biosensor with other analytes and measuring the obtained signal change. Usually, these negative controls are related to the target analyte in some way or another (also expresses protein on its surface; can be widely found in the biological fluid where target is tested) and the same concertation is measured. The obtained signal change (wavelength shift or amplitude change) is expected to be lower when controls are measured. This might be due to external interference or a minimal amount of non-specific adhesion (Song et al., 2016). Highly selective interferometry biosensor to detect a common pesticide was developed thanks to MIP. The probe demonstrated maximum selectivity for parathion methyl (PM) samples due to the complementary PM sites in the MIP nanoparticle layer, which restricts binding of other molecules (Shrivastav et al., 2019). Some works used two types of strategies (Luo et al., 2018; Bekmurzayeva et al., 2022) to study specificity of the built biosensors.

A highly specific detection of scDNA using an OFB in a microfluidic chip was demonstrated (Hu et al., 2023) where dual parameter detection of DNA hybridization and temperature using an optofluidic FBG and micro-structured optical fibers (MOF) with a single hole and dual core was demonstrated.

While the glucose/galactose binding protein (GGBP) protein is highly specific for glucose, high concentrations of fluorescent compounds near the read area can interfere with optical signal transduction, unlike electrochemical sensors. In vitro data demonstrated that this interference was due to direct exposure to common serum interferents. However, in vivo sensing, where the protein is contained within a biocompatible matrix with size-exclusion properties and concentrated near the optical fiber, is likely to reduce such interference. Additionally, increasing the GGBP protein level can enhance acrylodan emission, potentially eliminating tetracycline interference at most physiologically relevant concentrations (Weidemaier et al., 2011).

Cennamo et al. proposed an SPR biosensor which exploits specific sensing due to the device itself along with specificity of the MIP. The proposed device leverages the variation of light modes in a multimodal POF, part of the SPR platform, caused by target-receptor interactions in microholes. This differs from traditional SPR-POF-ER sensors, which rely on the RI variation of the dielectric. The receptor-analyte binding in the microholes alters the optical properties of the POF, affecting the SPR phenomena. This method, utilizing bioreceptors in an aqueous solution, is expected to achieve a lower LOD compared to MIP-based devices due to higher affinity of bioreceptors in a homogeneous phase (Cennamo et al., 2024).

Most biosensors are tested in simpler environments like PBS (Ribaut et al., 2017; Zhou et al., 2018; Kaushik et al., 2019; Bekmurzayeva et al., 2021; Niu et al., 2022; Hu et al., 2023). But for a more accurate reflection of their performance in real-world conditions, it is crucial to evaluate them in complex biological fluids such as serum and urine. This approach helps in understanding how these sensors would behave in a natural biological context, where a multitude of factors can affect their functionality. Studies were conducted using diluted or undiluted serum/plasma to simulate a more realistic biological environment, thereby enhancing the relevance and applicability of the findings to actual physiological conditions (Luo et al., 2018; Bekmurzayeva et al., 2022).

The biosensor showed LOD of 0.35 μg/mL in undiluted plasma and achieved a total time-to-result of 12 min (Qu et al., 2022). Another biosensor detected CRP with an LOD of 0.009 mg/L in serum in a label-free manner (Aray et al., 2016). Some studies used artificial biological fluids for testing the performance of biosensors (Qiao et al., 2022; Zhou et al., 2022; Han et al., 2023; Kundu et al., 2024) which offer such benefits in biosensing studies as the capacity to regulate and establish uniform test conditions, hence guaranteeing consistency across the investigations. The sensor probe was tested for p-cresol concentrations ranging from 0 μM to 1,000 µM in artificially prepared urine to mimic real application scenarios. Artificial urine was prepared with specific components and adjusted to a pH between five and 7, with p-cresol samples homogenized within it. The results showed that the peak absorbance wavelength shift varied with the pH of the analyte, with maximum interaction efficiency at pH 6 (Usha and Gupta, 2018).

Dynamic and continuous measurements are crucial for packaged biosensors to accurately capture real-time changes in biological samples. This factor is especially important in situations driven by fast biological processes or where sample characteristics might vary rapidly. To enable such measurements, several pumping methods are utilized to guarantee a consistent and regulated movement of liquids through the biosensor setup. Peristaltic pumps are frequently utilized because of their capacity to deliver uniform flow rates, which is essential for preserving the accuracy of the measuring procedure (Hsu et al., 2011; Hu et al., 2023). In addition, syringe pumps have been used for their accuracy in regulating fluid flow, particularly in small quantities. Another alternative is an electronically controlled pump, which is crucial in minimizing potential environmental factors during the measurement of bio-samples (Guo et al., 2014).

Measuring biomarker levels in flow conditions presents a possibility for continuous monitoring of biomarkers in clinical environments and/or has potential for detecting circulating biomarkers which are found in low concentrations. The presence of CD44 in diluted serum was measured in a blood flow mimicking experiments using a ball resonator sensor inserted into a catheter (Bekmurzayeva et al., 2022). The packaging was later improved in a follow-up work by the authors (Myrkhiyeva et al., 2024). To replicate the dynamic flow of blood through a vein, in vitro measurements of CD44 protein were performed using a syringe pump at a flow rate of 20 mL/min via a tube with a diameter of 1 mm. The small size of optical fibers enabled their integration into small packaging, which is essential for applications such as embedding them in catheters for analyzing dynamic blood flow.

For monitoring of chemotherapy-induced apoptosis, in vitro studies involved a biomimetic 3D cell distribution simulator, while in vivo experiments utilized MDA-MB 231 xenografts in mice (Fan et al., 2016). This configuration enabled the evaluation of the biosensor’s efficacy in both controlled research environments and live animal models for the detection of cell death induced by chemotherapy. The detection process consisted of insertion of the sensor probe into the tumor and sequential administration of FluoMb and FM 1–43. The apoptotic activity is calculated based on the fluorescence ratio. The sensor successfully distinguished between control and chemotherapy-treated groups.

Plasmonic immunosensor showed specificity and selectivity to cytokeratin 17 inside a porous polyacrylamide gel matrix, and its effectiveness was confirmed by analyses of human lung biopsy (Ribaut et al., 2017). This indicates a notable advancement in the field of minimally invasive in vivo medical diagnostics and the identification of tissue biomarkers. This showed novel opportunities for the in situ and on-line identification of biomarkers in tissues, which is essential for non-invasive medical diagnosis and research.

Packaged OFB was also explored in theranostics when an innovative method for detecting cancer in vivo and treating employing fiber-optic interstitial needles was proposed (Ran et al., 2022). The fiber-optic needles are comprised of optical fibers with small diameters measuring in the range of several hundred microns. The fibers are arranged in close contact inside a syringe needle used for commercial purposes, with a diameter that is strictly controlled to be less than 1 mm. This design allows for easier interstitial navigation between tissues. The sensor was engineered for the purpose of detecting cancer in vivo and facilitating photothermal treatment. The method exhibited successful tumor therapy, resulting in substantial suppression of tumor development and necrosis in the treated mice as compared to the control groups. The temperature was measured with a FBG that was engraved in the rare-earth-doped fiber. This FBG functions as a comprehensive and rapidly responsive temperature monitor, offering immediate feedback during photothermal therapy.

In vivo continuous glucose monitoring was conducted in live rats using an optical fiber based LSPR sensor integrated with a µD probe to prevent fouling. The µD probe filtered larger molecules, allowing only glucose to pass through for detection. The sensor was calibrated against manually collected blood glucose measurements, demonstrating linear response and stability, with an overall sensitivity of 0.0354 a. u./mg/dL and an LOD of 50.89 mg/dL (Kundu et al., 2024).

Only a subset of discussed biosensors were tested using real clinical samples. The clinical application of the FRMB was demonstrated by measuring SARS-CoV-2 IgM and IgG antibodies in spiked serum samples, achieving recoveries of 80.6%–119.7% and 80.2%–120.1% with relative standard deviations below 9.6%. This method proved capable of accurately detecting these antibodies in serum samples with minimal interference, offering a simpler, faster, and more cost-effective alternative to traditional methods like ELISA and lateral flow assays. The proposed biosensor demonstrated excellent clinical performance, showing only slight redshifts for Toxoplasma gondii negative serum samples and significant redshifts for protozoa positive serum samples, indicating its high specificity. This specificity is due to the binding of T. gondii antibodies in the positive samples with antigens on the biosensor surface. The sensor’s ability to be reused after HNO₃ treatment highlights its excellent repeatability and recovery, making it highly promising for biosensing applications (Chen et al., 2023).

Multiplexing offers the advantage of obtaining data for various analytes simultaneously and minimizes errors by reducing the need for extensive sample handling (Kasera et al., 2014) In order to examine the ability of the microfluidic-integrated D-shaped optical fiber SERS probe to detect several molecules, a mixture of three solutions (malachite green oxalate, crystal violet, and Rhodamine 6G) were introduced into a multiplexed microchannel via various inlets, and Raman signals were acquired in the detecting region resulting in clear peaks for each of the three molecules (Bo et al., 2023).

Sensing multiple parameters or having multiple sensing points on one sensor is also a useful characteristic for a biosensor. A study introduced a novel catheter structure utilizing an FBG optical fiber sensors for urodynamic analysis, offering potential for advanced diagnostics and multiple sensing points within the bladder while ensuring biocompatibility and high accuracy (Poeggel et al., 2015). A soft biosensor that can concurrently measure the respiratory (RR) and heart (HR) rates was also presented (Lo Presti et al., 2022). A fiber optic skin-interfaced biosensor, or “smart patch,” can measure local ribcage strain brought on by respiration and heartbeat to estimate HR and RR. This was reported to be the first smart patch for cardiorespiratory monitoring that is skin-mountable and based on FBG. Another on-chip optofluidic sensing platform with packaged OFS showed high sensitivities to temperature, concentration, and RI (Li et al., 2022). Although many studies claim the multiplexing capability of sensors and/or packaging, only a limited number of studies were able to demonstrate this capability.

A highly sensitive sensor with real-time detection capabilities is essential for microfluidic chip applications in biochemical experiments, such as monitoring glucose concentration in a microfluidic chip with multiple inlets and reaction tanks (Tyagi et al., 2018; Li et al., 2022). Flow of the sample (artificial urine containing p-cresol) through the cell allowed a constant interaction between the sensor and the target molecules enabling continuous monitoring and real-time detection (Usha and Gupta, 2018). A study with a needle-type biosensor was designed for immunological applications, with a specific emphasis on its capabilities for real-time monitoring, in situ measurement, and high sensitivity (Tseng et al., 2007).

To showcase its clinical application potential, real-time Fresnel reflection light intensity was monitored by FEMB for 300 s (SARS-CoV-2 IgG) or 400 s (SARS-CoV-2 IgM). The bio-probe surface was then regenerated and washed, making it ready for subsequent tests. This real-time monitoring method demonstrated high sensitivity, specificity, and stability, making it a valuable tool for COVID-19 diagnosis and population immunization evaluation (Xu et al., 2021).