Solid Fiber Inside of Capillary and Modified Fusion-Spliced Fiber Optic Microneedle Devices for Improved Light Transmission Efficiency

Abstract

Convection-enhanced delivery (CED) is a drug delivery technique used to deliver therapeutics directly to the brain and is a continually evolving technique to treat glioblastoma. Early versions of CED have proven to result in inadequate drug volume dispersed (V_d), increasing the likelihood of tumor recurrence. Fiber optic microneedle devices (FMDs) with the ability to deliver fluid and thermal energy simultaneously have shown an ability to increase V_d, but FMDs have historically had low light transmission efficiency. In this study, we present a new fabrication method, solid fiber inside capillary (SFIC) FMD, and a modified fusion splicing (FS) method with the goal of increasing light delivery efficiency. The modified FS FMD resulted in an increase in light transmission efficiency between 49% and 173% compared to previous prototypes. However, the FS FMD resulted in significantly lower transmission efficiencies compared to the SFIC FMD (p ≤ 0.04) and FS FMDs perform much worse when light-absorptive materials, like black dye, are placed in the bore. The light absorption of a candidate cytotoxic agent, QUAD-CTX, appear to be similar to water, and light delivery through FS FMDs filled with QUAD-CTX achieves a transmission efficiency of 85.6 ± 5.4%. The fabrication process of the SFIC FMDs results in extremely fragile FMDs. Therefore, the use of a modified FS FMD fabrication process appears to be better suited for balancing the desire to increase light transmission efficiency while retaining a sturdy FMD construction.

Introduction

Patients diagnosed with glioblastoma, an aggressive form of malignant brain tumor, have about a 7% 5-year survival rate despite receiving the standard care consisting of tumor resection, radiation, and chemotherapy [1–3]. Most methods of chemotherapy are ineffective due to the blood-brain barrier [4]. The noncurative nature of glioblastoma has inspired experimental forms of treatment, such as convection-enhanced delivery (CED).

Bobo et al. first proposed CED for transferring therapeutic agents to the brain to treat tumors and other neurological conditions [5]. CED takes advantage of pressure-driven flow to deliver therapeutics through a small catheter placed directly in the brain, bypassing the blood-brain barrier. CED treatment has been explored to treat malignant gliomas and Parkinson's Disease (PD). Preclinical [5–7] and early clinical trials [8–12] showed promise for CED, but Phase I/II clinical trials for the treatment of PD with a glial line-derived neurotrophic factor failed to show clinical benefit over a placebo [13]. The PRECISE trial, the only Phase III CED clinical trial for treatment of malignant gliomas, failed to achieve the clinical endpoint chosen by the trial sponsor, while illustrating significant clinical benefit comparable to an FDA-approved, diffusion-based Gliadel Wafer [14]. Moreover, retrospective studies found that drug volume dispersed (V_d) was unsatisfactory in both trials. Between 2 and 9% of the putamen was likely covered in the PD trial [15], and only 20% of the 2 cm penumbra around the resection cavity was covered in the PRECISE trial [16]. These results are deeply disappointing for the CED treatment of malignant gliomas, as tumor recurrence is inevitable and generally occurs within 2 cm of the original tumor resection cavity [17], implying the inevitable impact of failing to treat the entire tumor volume.

Improvement in the drug delivery capabilities of CED catheters has been the goal of many recent studies. Most notably, focus has been placed on catheter geometry changes to create reflux preventing catheters [18–21] and the implementation of multiport catheters [22,23]. Infusion protocols such as “infuse-as-you-go” [24], controlled reflux [25], and continuous catheter movement [26] have also been examined as methods to increase V_d for CED. Additionally, fiber optic microneedle devices (FMDs) have the potential to be a delivery platform for not only CED but also a broad range of other clinically relevant problems. Hood et al. Reported the application of fusion spliced FMDs in plasmonic photothermal therapy [27] as well as application for use in laser-induced thermal therapy [28]. FMDs were also shown to be beneficial in treatment using CED, where V_d was found to increase by 60–80% in a rodent brain model compared to a treatment without codelivery control [29]. FMDs allow for codelivery of both a therapeutic agent and light energy, causing a photothermal tissue response. Previous fabrication methods of FMDs include epoxy-bonding a standard step-indexed optical fiber to the outside of a capillary tube [30,31], and fusion splicing an optical fiber to the annular wall of a light-guiding capillary tube [28,29]. Using the fusion splicing method, previous fluoroptic temperature sensing and histopathological assessments in healthy rat brain demonstrated that a laser power of 200 mW results in steady-state temperatures of 42.0 ± 0.9 °C, causing sublethal brain hyperthermia. These preliminary FMD CED experiments confirmed the feasibility of augmenting fluid dispersal using slight photothermal heat generation, demonstrating the FMD's potential as a way to increase the efficacy of CED [28].

While both FMD designs enabled the codelivery functionality desired; they each had some limitations. The epoxy-bonded FMDs could be susceptible to delamination, which could be exasperated by the friction caused on deployment of FMDs using the previously developed arborizing catheter [22,32]. If delamination occurs, the fragile optical fiber could break inside of the brain, causing serious complications in treatment. The fusion spliced FMDs typically had low light transmission efficiency, ranging from 30 to 55% [28,33]; until recently, when splice efficiency of greater than 75% have been reported using similar methods expanded upon in this study [34,35]. The low efficiency previously reported could cause unwanted heating of other catheter components and degrade the FMD performance over time; especially given the typical length of CED treatments, which could last from hours to several days [36]. In this study, we explore the optical transmission efficiency of a new FMD fabrication process where a solid optical fiber is secured inside the bore of a standard capillary tube (SFIC FMD) and compare it to a modified fusion spliced (FS) FMD.

Methods

Fabrication of SFIC FMD.

Solid fiber inside capillary (SFIC) FMDs were fabricated by feeding a solid core step-index silica fiber (100 μm core/110 μm cladding/130 μm buffer, ASF100/110/130T, Fiberguide Industries, Inc., Sterling, NJ) through the bore of a thin-walled silica capillary tube (250 μm inner diameter (ID)/375 μm outer diameter (OD), TSP250350, GS-Tek, Newark, DE). First, the capillary tubing is assembled into a needle capable of dispensing fluid. The capillary tubing is cleaved from the spool, and each of the end faces is manually polished flat using polishing paper of sequentially decreasing grit sizes (30 μm, 6 μm, 3 μm, and 1 μm). The resulting 72 cm long capillary tubes are then inserted approximately 1 cm and epoxy-bonded within a 22-gauge plastic dispensing needle using a medical grade two-part epoxy (EA M-31CL, Henkel Corporation, Dusseldorf, Germany). Approximately 1 cm of the solid fiber's buffer is stripped using a warm sulfuric acid bath. The stripped end of the fiber is cleaved flat and inserted through the bore of the capillary tube, such that the solid fiber terminates just beyond the end of the needle, as shown in Fig. 1(a). Terminating the solid fiber beyond the capillary tube was done with the goal of increasing the likelihood of directly irradiating the tissue and reducing the interaction of the light with the infusate inside of the bore of the capillary tubing, as excessive heating of the infusate may be undesirable and may reduce light penetration in the tissue. A luer lock t-connector (PN 80061, Qosina Corporation, Ronkonkoma, NY) is placed over the fiber and connected to the needle via the mechanical luer lock. The solid fiber port of the luer connector is sealed with medical grade UV cure acrylic (AA 3926, Henkel Corporation, Dusseldorf, Germany). The open end of the t-connector allows the delivery of fluid into the FMD, as shown in Fig. 1(b). A standard 2.5 mm OD ceramic ferrule (126 μm bore diameter) is then epoxy-bonded to the proximal end of the solid fiber extending approximately 1 m from the luer lock to allow coupling to a laser source.

Fig. 1.

(a) Tip of solid fiber inside of capillary (SFIC) FMD showing solid fiber protruding from the capillary and (b) cross-sectional schematic of SFIC FMD (not to scale)

To enable repeatable attachment and detachment of the SFIC FMD to the laser without requiring realignment for each prototype, the SFIC FMDs attach to a patch cable terminated with a 2.5 mm OD ceramic ferrule. A 50 μm core diameter was chosen for the patch cable because the nonstandard size of the solid fiber (110 μm cladding diameter) allows for approximately 16 μm of core misalignment when terminated within a standard 126 μm ID ferrule. While the 50 μm cable will result in under filling of the solid fiber core, it will not cause as great a loss in optical power as that caused by the coupling of two misaligned fibers. The total theoretical power loss can be determined from the general loss equation shown in Eq. (1) [37]

$$L_{\mathrm{total}}=\alpha x+L_c+L_s$$ (1)

where L_total is the total loss in dB, α is the fiber attenuation, x is the length of the fiber in km, L_c is the connector loss, in dB, and L_s is the splice loss in dB.

Fabrication of FS FMD.

The preparation of the fusion spliced (FS) FMD is similar to that previously detailed [28]. Briefly, approximately 1 cm of the polyimide coating on both sides of a solid core step-index fiber (50 μm core/70 μm cladding/85 μm buffer, FIP050070085, Polymicro Technologies, Phoenix, AZ) and 1 cm of the polyimide coating on one side of a light guiding capillary tube (150 μm ID/350 μm OD, LTSP150375, Polymicro Technologies, Phoenix, AZ) are stripped using a warm sulfuric acid bath. One end of the solid fiber is cleaved flat, and the other side is prepared for ferrule termination. Since there is no standard ferrule bore diameter close to that of the solid fiber used (70 μm), the termination end of the solid fiber was modified to allow near concentric mating with an easily obtainable 160 μm bore diameter ceramic ferrule. To achieve this, the stripped portion of one side of the fiber is coated with an anaerobic curing epoxy (Loctite 648 442-21443, Henkel Corporation, Dusseldorf, Germany) then inserted into the bore of a small capillary tube (75 μm ID/150 μm OD, 1068150018, Polymicro Technologies, Phoenix, AZ). The capillary tube is cleaved slightly shorter than the bare length of the fiber, such that the fiber protrudes from the capillary when fully inserted. The assembled capillary tube is then terminated as normal with a ceramic ferrule (160 μm ID, 2.5 mm OD). After curing, the capillary tubing and bare solid fiber that protrudes from the ferrule are cleaved, being careful to not crack the solid fiber below the connector. The termination is then polished according to standard practice [38].

The light guiding capillary is polished flat on both ends using an electric end polisher (ULTRAPOL 1200, ULTRA TEC Manufacturing, Inc., Santa Ana, CA), using sequentially decreasing grit sizes (30 μm, 6 μm, 3 μm, and 1 μm). The splicing instrumentation and parameters have been modified from previously reported FS FMD fabrication techniques [28] to improve their light transmission. To splice the solid fiber to the light guiding capillary, the unterminated polished end of the solid fiber and the polished end of the light-guiding capillary tube is placed into a fusion splicer (S178 LDF, Furukawa Electric Co., LTD, Tokyo, Japan). The splicer automatically aligns the fibers center to center, then the solid core fiber is manually moved by 110.8 μm, using an electric linear stage, such that the core of the solid fiber is aligned to the light guiding core (annulus) of the capillary, as shown in Fig. 2(a). The splice is then created using a custom designed splicing protocol. Based on recommendations for splicing hollow core photonic crystal fibers (PCF) to solid core single mode fiber [39], the electrodes are biased by 10 μm toward the larger light guiding capillary. This allows more energy to reach the capillary than the solid fiber, reducing the chance of overheating the smaller solid fiber. A two-step arc duration is applied with the first step 1 s long and the second step 2 s long. Arc length was chosen according to results reported by Kato et al. in which arc lengths of longer than 1 s were found to increase tensile strength of fusion spliced single-mode fibers [40]. The first-step arc power is 40 AU and the second-step arc power ramps from 40 to 70 AU (power is a proprietary Furukawa unit). Further, the z-push distance is 20 μm, arc offset is −10 μm, and the gap offset is 5 μm. The theoretical efficiency of a spliced fiber, considering both the core diameter and axis offsets, but ignoring factors such as cleave quality, core noncircularity, dust contamination, and insertion loss can be determined by Eq. (2) [41]:

$$L_s=-10\log \left[{\left(\frac{2AB}{A^2+B^2}\right)}^2\exp \left(-\frac{2d^2}{A^2+B^2}\right)\right]\hspace{0.17em}$$ (2)

Fig. 2.

(a) Cross sectional schematic of fusion spliced (FS) FMD and (b) cross section of FS FMD after adding a protective capillary sleeve over the splice junction and attaching to at-connector

where L_s is the splice efficiency in dB, A is the mode field diameter of the solid fiber in μm, B is the mode field diameter of the light guiding capillary in μm, and d is the core offset between the solid fiber and the capillary in μm.

After splicing, the splice junction is supported with a protective sleeve by fixing a larger capillary tube (450 μm ID, 670 μm OD, TSP450670, Polymicro Technologies, Phoenix, AZ) over the splice junction with UV cure acrylic (AA 3926, Henkel Corporation, Dusseldorf, Germany). The entire spliced fiber is then inserted into a t-connector and furcation tubing, and UV cure acrylic is used to fix the spliced FMD into the connector, as shown in Fig. 2(b).

Fiber Optic Microneedle Device Light Transmission Efficiency Measurement Procedure.

Due to the high penetration depth of near infrared light in both healthy [42] and neoplastic [43] brain tissue, a 1000 mW 1064 nm diode-pumped solid-state laser (LRS-1064-PFM, Laserglow Technologies, Toronto, ON) was used to test the transmission efficiency of the two alternative FMD designs. The laser connects to a patch cable (M14L02, Thorlabs Inc., Newton, NJ), which was modified to have a 2.5 mm OD, 126 μm ID ferrule at the emitting end of the cable. Next, the ferrule was connected to the corresponding ferrule on the FMD using a ferrule mating sleeve (ADAF, Thorlabs Inc., Newton, NJ). A 100 mW output from the patch cable was enabled by an analog voltage output (NI9264/NI USB-9162, National Instruments, Austin, TX) set in a custom LabView program. In order to differentiate between insertion loss and splice efficiency, the power of the solid core fibers for the FS FMDs were measured after fiber termination but before splicing using an integrating sphere detector (918D-SL-OD1, Newport Corporation, Irvine, CA), connected to a power meter (model 1931-C, Newport Corporation, Irvine, CA), which recorded data directly to the LabView program. The FS FMD was then spliced as described previously. Three FMDs of each design were fabricated for testing, and power was measured in the same manner for both the SFIC and the FS FMDs. First, the emitting end of the FMD was inserted into a bare fiber terminator which was inserted into the integrating sphere detector. The bare fiber terminator ensured the FMDs were consistently placed in the same position inside of the detector. The power was measured initially with air in the bore of the FMD, then de-ionized (DI) water was flushed through the FMD, and power was measured again. Finally, the DI water was removed, and black ink (Speedball Super Black India Ink, Speedball, Statesville, NC) was used to flush the FMD, and power was measured once more. For each trial, the laser emitted light for 10 min. In order to allow the laser power to stabilize, power was measured every second during the last 30 s of the 10-min trial and averaged over the 30 s period. At the conclusion of the experiment, the SFIC and FS FMDs were compared for different transmission efficiencies using a two-sample t-test. Further, a paired t-test was used to compare the total efficiency, including both insertion loss and splice efficiency, with the splice efficiency.

After the first set of experimental results was obtained, three additional FS FMDs were then fabricated in order to test the light transmission efficiency with a candidate drug in the bore of the FMD. For this experiment, 6.4 μg of the targeted cytotoxic drug candidate, QUAD-CTX, was obtained from Wake Forest University, and its formulation has been reported in detail [44,45]. The QUAD-CTX was then diluted in 4 mL of gadolinium-labeled albumin (Galbumin, Bio-PaL, Inc., Worcester, MA) at a concentration of 25 mg/mL. This resulted in a final QUAD-CTX concentration of 1.6 μg/mL, the maximum dose used in the drug's Phase I clinical trial [46]. Three additional FS FMDs were fabricated, with their pre- and postsplice efficiencies measured. No SFIC FMDs were tested with QUAD-CTX due to the infeasibility of this design found during the first experiment. Due to the possibility of fluid retention inside the bore of the FMDs, no power transmission was acquired with DI water in the bore out of concern of small amounts of DI water remaining in the bore of the FMD, resulting in further dilution of the QUAD-CTX with the DI water, potentially skewing results. Power transmission was then recorded in the same manner as before with presplice, postsplice with air in the bore, and finally postsplice with QUAD-CTX in the bore. The three FS FMDs were compared for different transmission efficiencies using a paired t-test. The difference in efficiency was compared to the first FS group using a two-sample t-test.

Finally, in order to test for potential cable and splice degradation over time, two more SFIC and FS FMDs were manufactured. The first of each FMD type was used to emit approximately 200 mW for 8 h, and the second of each FMD type was used to emit 500 mW for 8 h. The actual output power for each power level was slightly higher than the listed 200 mW and 500 mW values. This is due to the challenge of creating perfect calibration curves for the laser that ensure that the nominal power is always achieved. Therefore, in order to minimize the likelihood that true output power would dip below the nominal rating, power was intentionally set slightly above the nominal rating to ensure a worst-case scenario for the experiment. As a reference to the stability of the laser, the power output of a patch cable was also measured for 8 h. Power was measured using a similar setup as before, with data being collected every minute for the full 8-h test. An 8-h test was chosen due to practical limitations of the experiment and an inability of the lasers to remain on outside of normal lab hours.

Results

Theoretical Efficiency.

The length of the cable used for the SFIC FMD is around 2 m, and the attenuation of fiber optic cables is of the order of dB per km; therefore, the expected attenuation resulting from fiber length is negligible. The ferrule mating sleeve has a rated typical insertion loss of less than 1.0 dB, so the connector loss can be estimated at 1 dB. The splice loss from the SFIC FMD is 0 dB since no splice was used in manufacturing the FMD. Using Eq. (1), we can find a total expected loss of less than 1 dB, which corresponds to an expected efficiency of greater than 79% for the SFIC FMD.

The FS FMD efficiency can be calculated similarly to the SFIC FMD. The length of the fiber is again negligible, the connector loss will also be less than 1 dB, and the splice loss can be determined using Eq. (2). We set the mode field diameter of the solid fiber to the core diameter of the solid fiber and the mode field diameter of the light guiding capillary to the core radius of the capillary (71.5 μm). Then, two cases were analyzed: (1) where there is no core misalignment (d = 0 μm) and (2) where the core misalignment is determined as if the top edge of the solid fiber is aligned to the top edge of the light guiding capillary (16.75 μm). The first case yields a splice loss of 0.54 dB, and the second case results in a splice loss of 0.87 dB. The splice efficiency of the FS FMD should be between 82 and 88%. Adding the 1 dB loss from the connector, the total efficiency for the FS FMD should be between 65 and 70%.

Experimental Fiber Optic Microneedle Device Efficiency.

The SFIC FMD had a significantly higher transmission efficiency with air, DI water, and black ink in the bore compared to the FS FMD (p = 0.04, p = 0.04, and p = 0.003, respectively). The average efficiency for SFIC FMD with air, DI water, and black ink in the bore was 95.1 ± 2.7%, 91.7 ± 8.5%, and 83.1 ± 10.8%, , respectively, whereas the efficiencies for the FS FMD were 79.6 ± 8.6%, 72.8 ± 6.8%, and 27.2 ± 10.2%, respectively. Figure 3 shows the total efficiency of both the SFIC and the FS FMDs.

Fig. 3.

Plot of the transmission efficiency (%) dependence on the media inside of the bore of the FMD for both the solid fiber inside of capillary (SFIC) and the fusion spliced (FS) FMDs

After 8-h testing of each of the FMDs, both the SFIC and FS FMDs had greater percent deviation from the mean measured power than the patch cable, as shown in Fig. 4. The average output of the 200 mW FS FMD was 193.8 ± 2.0 mW, whereas the average output of the SFIC FMD was 202.9 ± 2.0 mW. The average output of the 500 mW FS FMD was 494.8 ± 8.1 mW and the average output of the 500 mW SFIC FMD was 508.9 ± 2.1 mW. The deviation from the average power of the FS 200 mW FMD was 5.1% compared to a deviation of 5.3% for the SFIC FMD. The deviation of the 500 mW spliced FMD was 8.2% compared to 2.3% for the SFIC FMD. The deviation of the patch cable over the 8-h span was 2.1%.

Fig. 4.

Normalized power plotted against time for the patch cable, 200 mW solid fiber inside of capillary (SFIC) FMD, 500 mW SFIC FMD, 200 mW fusion spliced (FS) FMD, and 500 mW FS FMD

A fiber microscope (FS201, Thorlabs Inc., Newton, NJ) was used to record en face images, of the fiber optic ferrule connectors of the FS FMDs, before and after delivery of light for 8 h at 200 and 500 mW (Fig. 5). Figures 5(a) and 5(d) show the ferrule connector before 8-h delivery of light at 200 and 500 mW, respectively. Figures 5(b) and 5(c) show the same connectors after the 8-h delivery of light and reveal a dark circular ring around the connector, suggesting some damage may have occurred throughout the light delivery window. However, these changes are not permanent as cleaning with isopropyl alcohol (Figs. 5(c) and 5(f)) showed an almost perfect return to the predelivery condition of the ferrule.

Fig. 5.

Images of fiber optic ferrule connectors after 8-h light delivery experiments for fusion spliced (FS) FMDs (a) before 200 mW light delivery, (b) directly after 200 mW light delivery (c) after 200 mW light delivery and after cleaning ferrule with isopropyl alcohol, (d) before 500 mW light delivery, (e) directly after 500 mW light delivery, and (f) after 500 mW light delivery and after cleaning ferrule with isopropyl alcohol

Figure 6 shows the efficiency of the three additional FS FMDs that were fabricated in order to test the light transmission efficiency with the candidate drug, QUAD-CTX in the bore of the FMD. The average presplice efficiency was 94.6 ± 0.25%, the average efficiency with air in the bore was 84.4 ± 5.0%; and finally, the average efficiency with the QUAD-CTX compound in the bore was 85.6 ± 5.4%. There was no statistical difference in the efficiencies of the FS FMD with air inside the bore or with QUAD-CTX inside the bore. When comparing the second experimental FS group with the first experimental FS group, there was no statistical difference in presplice efficiency (p = 0.09) or air inside bore efficiency (p = 0.45). When the efficiency of the first FS group with water in the bore was compared to the second group with QUAD-CTX in the bore, a larger total efficiency was seen with the QUAD-CTX, but no significant difference was found (p = 0.06). When comparing black ink in the bore to QUAD-CTX in the bore, there was a significantly lower transmission efficiency with the black ink in the bore (p < 0.001). Finally, the average efficiency of all FS FMDs with air in the bore (n = 6) was 82.0 ± 6.8%.

Fig. 6.

Plot of transmission efficiency of three additional FS FMDs before splicing (pre-splice), with air inside of the bore (air), and with the QUAD-CTX compound inside the bore (QUAD)

Discussion

The measured efficiency in air for the SFIC FMD is about 7% higher than the theoretical efficiency. The measured total efficiency for the FS FMD is between 14 and 20% higher than the theoretical total efficiency. Considering the SFIC FMD efficiency and the efficiency of the FS fiber before splicing the total connector loss is around 0.17 dB, which is much closer to the manufacturer reported connector losses of other standard connectors that do not use a ferrule sleeve. This suggests that insertion loss caused by the ferrule sleeve adapter is much lower for the fiber size and wavelength used in this study compared to that used in the testing conducted by the manufacturer.

The splice efficiency is between 0.1% greater and 7% lower than the theoretical splice efficiency. The splice efficiency is likely at the low end of the theoretical calculations due to imperfect alignment between the two spliced components and slight angular misalignments between the two fibers caused by variations in the cleave and polishing angles.

The transmission efficiency of all FS FMDs, with air in the bore, is nearly 82% which is between a 49% and 173% improvement over previously published efficiencies [28,33]. Since the transmission efficiency of both the FS and the SFIC FMDs have a larger deviation over 8 h than just a patch cable, the discrepancy is likely not the result of splice or fiber degradation, but more likely caused by the coupling of the patch cable to the respective FMDs. But, since the 500 mW FS FMD has a much larger deviation than any other group, it is possible that some thermal degradation is occurring at high powers. However, transmission power will likely be limited to 200 mW or lower in practice as power above 200 mW has been shown to result in thermal damage in the tissue, which is not desirable [28]. Although we did not find a significant reduction in power transmission after 8 h, CED infusions can last as long as a few days [36]. While we do not anticipate the use of codelivery for time periods exceeding 8 h if it ever becomes necessary, longer duration testing should be conducted. The appearance of the fiber optic ferrule connectors changes during long periods of light delivery. While this change does not appear to be permanent, given the effectiveness of cleaning the ferrule with alcohol, it may result in significant changes in light delivery. This change may be responsible for the observed deviation in light delivery over time during the 8-h experiments. The FS FMD at 500 mW resulted in a decrease in light output with a minimum power output around 400-min, and a subsequent increase in power through the remainder of the experiment. The 200 mW FS FMD, on the other hand, appears to begin decreasing around 300-min, then it reaches a minimum, and then increases at a lower rate through the rest of the experiment. Looking at the appearance of the fiber optic ferrule in Fig. 5, the 500 mW ferrule appears to exhibit much less change between the three images whereas the after-light-delivery image of the 200 mW fiber ferrule appears quite different from the other two 200 mW ferrule images. This could indicate that the change in the condition of the ferrule is causing a loss in light throughput. One explanation for this change in ferrule condition could be the connection between the patch cable and the FMD. Given the simple friction fit of the ferrule mating sleeves, it is possible a small air gap exists between the two components, and some fraction of light may interact with the epoxy used to secure the fibers to the ferrules, resulting in off-gassing that could obscure the ferrule in a similar way shown in Fig. 5(b). Further work should be conducted to confirm this hypothesis; nonetheless, a better method of securing the patch cable to the FMD should be explored as tension on the FMD is possible, especially during insertion of the FMD or during FMD repositioning. This tension could create an undesirable gap between the FMD and the patch cable. The ferrule mating sleeve connection has been explored because of its compatibility with magnetic resonance imaging (MRI). However, other commercial or custom nonmetallic fiber termination and coupling options that create a more secure connection should be explored in detail.

The transmission efficiency of the SFIC FMD is superior to that of the FS FMD, especially when an extremely light absorptive infusate, such as black dye is used. Since the SFIC FMD uses a standard step index fiber with cladding on the outside of the fiber, no absorption of light was expected along the axis of the FMD. However, the light guiding capillary does not have cladding along the inner lumen of the tube, allowing light to be absorbed by the media inside of the bore. But, with a CED candidate drug, QUAD-CTX, placed in the bore of the FS FMD, the light transmission efficiency is nearly equal to FS FMDs with air and DI water in their bore. This result is intuitive since the quantity of the drug in solution is extraordinarily small and the carrier (Galbumin) is 25% gadolinium labeled albumin suspended in water. Therefore, the QUAD-CTX compound should have efficiency results similar to water. So, while the FS FMD is susceptible to light loss due to a lack of internal cladding, it may not be an issue depending on the infusate. Exploration of the effects of each candidate drug on light transmission and the effect of the heating on the potency of specific drugs should be further studied.

The current experiments are performed with 1064 nm radiation due to its high penetration depth in biological tissues. Using other wavelengths of light could affect the efficiency of the overall CED process by changing the interaction effects with both the infusate and the tissue. The infusate can absorb light within the device depending on its specific design. Additionally, the infusate, tissue, and their potential interaction, such as refractive index matching, can alter the absorption and scattering properties of the tissue and the overall penetration depth of the light in the tissue [47].

While there are many benefits including higher light transmission efficiency of the SFIC FMD, two factors reduce its practical efficacy. First, adding the solid fiber (130 μm) to the bore of the capillary (250 μm ID) reduces the hydraulic diameter of the FMD by ∼52%. The hydraulic resistance of laminar flow according to the Hagen-Poiseuille law is proportional to the diameter raised to the fourth power. By decreasing the hydraulic diameter, significantly more pressure would be needed to create similar flow rates. But, given the low flow rates ranging from 0.5 to 10 μL/min [48] typically associated with CED, the SFIC FMD will likely not be flowrate limited, as catheter internal diameters as low as 50 μm have been used previously [49]. The second factor which reduces the practical implementation of the SFIC FMDs is that they are prone to breakage at bend radii of about 19.5 mm as determined by wrapping the SFIC FMDs around pin gauges of increasing diameter. The likely cause of failure is the process where the solid fiber is inserted into the bore of the capillary tube. During this time, the solid fiber may contact the inner wall of the capillary, which could cause scoring of the capillary tube. This scoring, if deep enough (approximately 1 μm), could cause the capillary to be easily cleaved. Figure 7(a) shows an example of scoring within the bore of the capillary resulting from insertion of the solid fiber, and Figs. 7(b) and 7(c) show needle failure due to fracture caused by the scoring.

Fig. 7.

(a) Scribing inside of capillary possibly caused by fiber insertion, (b) heavy cracking near capillary failure point, and (c) two halves of cracked capillary

Conclusion

We present two alternative designs and methods of FMD fabrication that can improve light transmission efficiency compared to previously reported designs. The SFIC design has greater light transmission efficiency than the FS design; however, the FS design has increased fluid conductance and greater mechanical strength. The improvement in efficiency of both designs allows for the codelivery of both fluid and laser energy with a reduced chance of degradation of the device and excessive heating of the delivered infusate. Also, the light transmission efficiency of the FS FMD with a candidate drug in the bore of the FMDs resulted in a similar transmission as FMDs with water in the bore. Future work should focus on the absorptivity of additional potential drugs for CED treatment and the interaction of the drug with the addition of thermal energy.

Funding Data

• National Institutes of Health and the National Cancer Institute (Grant No. P01 CA207206-01; Funder IDs: 10.13039/100000002 and 10.13039/100000054).

Nomenclature

A =

mode field diameter of solid fiber (μm)

B =

mode field diameter of light guiding capillary (μm)

CED =

convection-enhanced delivery

d =

core offset between the solid fiber and light guiding capillary (μm)

FMD =

fiber optic microneedle device

FS =

fusion spliced

L_c =

optical transmission loss due to fiber connector (dB)

L_s =

optical transmission loss due to splice (dB)

L_total =

total optical transmission loss (dB)

PD =

Parkinson's disease

SFIC =

solid fiber inside of capillary

x =

length of optical fiber (km)

α =

optical fiber attenuation coefficient