2.1.

Microneedles

The microneedle arrays used in this study were fabricated using previously reported potassium hydroxide wet-etching techniques.³¹ Sample images of the microneedles used are shown in Fig. 1 .

Fig. 1

(a) Optical image showing a microneedle array. (b) SEM of microneedles.

The starting material was a $525\text{-}\mu \mathrm{m}$ -thick, $100\text{-}\mathrm{mm}$ -diam monocrystalline silicon wafer that is orientated in the ⟨100⟩ direction, on which marks denoting this crystal alignment have already been etched. A layer of $350\text{-}\mathrm{\AA }$ silicon oxide is grown using thermal oxidation as a stress relief layer before a subsequent low-pressure chemical vapor deposition of $1000\text{-}\mathrm{\AA }$ silicon nitride. A positive photoresist layer is then deposited and patterned in square masks using standard photolithography techniques; the dimensions of these square masks determine the array pitch and needle geometry.³¹ The mask pattern is etched into the nitride layer using a plasma etch process before the resist is stripped and the oxide layer then removed in the open areas using hydrofluoric acid (HF).

The patterned silicon wafer is then etched using a 29% w/v aqueous KOH solution at a temperature of $79°\mathrm{C}$ . Needle formation is based on the anisotropic etch behavior of monocrystalline silicon in KOH, a property of the crystal structure that causes each group of crystal planes to etch at a different rate. The sides of the square nitride mask are precisely aligned to the particularly slow-etching ⟨h11⟩ plane; the faster-etching ⟨h12⟩ planes are exposed to the KOH at the convex corners of the square. As two of these fast-etching planes are etched from each corner, an octagonal needle shape is generated when the eight planes meet. The final needle is comprised of eight ⟨312⟩ planes, a base of ⟨121⟩ planes, and has a height:base diameter aspect ratio of 3:2. The array used in this study consisted of a $10\times 10$ arrangement of needles located at a pitch of $1\mathrm{mm}$ on a $12\times 12\text{-}\mathrm{mm}$ die. The height of the needles was $280\mu \mathrm{m}$ .

2.2.

Optical Coherence Tomography System

This work employs the use of a commercial OCT systems developed by Bioptigen Incorporated, Research Triangle Park, North Carolina. The system is a spectral Fourier domain OCT (SFD-OCT) and is equipped with an $840\text{-}\mathrm{nm}$ light source. This system offers an axial resolution of $3.5\mu \mathrm{m}$ and a lateral resolution of $15\mu \mathrm{m}$ , and operates with an A-scan rate of $\mathrm{17,000}\mathrm{Hz}$ , producing images with a volume of $8\times 8\times 2.707\mathrm{mm}$ $(500\times 500\times 1024\text{pixels})$ in $14\mathrm{s}$ .

3.1.

In-Vivo Penetration Depth of Microneedles

The procedure was extended to image the micropores in vivo. These tests were performed on a $23\text{-year}$ -old Caucasian male subject with informed consent. Two separate locations were imaged: the volar forearm and the fingernail fold. The skin was not pretreated prior to needle insertion. The subject was asked to report on the level of discomfort/pain experienced as the needles were inserted.

The microneedles were placed over the insertion point, then positioned under the OCT imaging system to allow for immediate scanning. The needle array was pressed into the forearm using the forefinger with a moderate level of pressure and held for $10\mathrm{s}$ . After application, the needle array was removed from the skin and the region was imaged using the OCT (see Fig. 3 ).

Fig. 3

OCT imaging of microneedle array into volar forearm. (a) Optical image of microneedle application site (marked by square) after insertion. (b) OCT of microneedle pores immediately after insertion into the forearm for the region indicated in (a). (c) 2-D slice showing needles fully penetrating the epidermal layer along the line X-X indicated in (b). The arrows indicate the needle holes in the skin.

The distance to the dermal-epidermal junction (DEJ) was determined by averaging 10 B-scans to reduce the speckle noise present in OCT images. The surface of the skin was manually tracked. Regions containing hair follicles and micropores were deemed invalid and were not tracked [see Fig. 4 ]. The valid regions of the images were flattened along the tracked surface and the average A-scan determined [see Fig. 4]. The distance to the DEJ was manually measured by a skilled operator. The DEJ was measured to be at a depth of $161\pm 13\mu \mathrm{m}$ from a sample size of 163 at varying locations along the volar forearm. When the DEJ depth was compared to the averaged A-scan, it coincided with the distance from the entrance peak to the midpoint of the first minimum and the second peak [see Fig. 4]. It was not possible to measure the stratum corneum for the forearm using the OCT system, as the thickness has been previously reported to be $10\mu \mathrm{m}$ in this region.³²

Fig. 4

Epidermal thickness measurement of the forearm; (a) 2-D OCT slice, (b) 2-D slice with an average of 21 B-scans, and (c) flattened image showing average A-scan (red) and epidermal thickness (orange) (scale bars $500\mu \mathrm{m}$ ).

By using the previously outlined minimum projection technique, the depth of each micropore was measured. When the microneedles were inserted into skin, it was possible that the underlying tissue could compress/deform under the pressure of the microneedles. It was therefore important to determine if any epidermal tissue remained under the tip of each micropore. To achieve this, an averaged A-scan under the micropore was calculated. The averaging was performed on the plane perpendicular to a line through the tip of the micropore for $\pm 4$ voxels in both the $x$ and $y$ directions [see LM in Fig. 5 ]. The presence of a second peak in the A-scan was indicative that the DEJ remained under the micropore tip [see Fig. 5]. This was done for each projection axis (0, 45, 90, and $135\mathrm{deg}$ relative to the B-scan axis), and the average value calculated. This was repeated for each micropore and the results are indicated in Fig. 6 . As before, each micropore was identified by its grid location [see Fig. 3].

Fig. 5

Measurement of the epidermal thickness under the micropore. (a) OCT scan of the region. (b) Averaged A-Scan showing the location of the micropore tip and the dermal epidermal junction (DEJ).

Fig. 6

In-vivo measured depth of each micropore in the skin (dark gray) and the epidermal thickness measured under each micropore (light gray) for a microneedle array in the forearm [see Fig. 3 for grid location]. The measured epidermal thickness $\left(161\mu \mathrm{m}\right)$ in the region is indicated by the horizontal line.

The micropores were measured to penetrate an average depth of $179\pm 14\mu \mathrm{m}$ into the skin. This result would indicate that the microneedles had penetrated fully through the epidermis. However, epidermal tissue can still be measured under each micropore due to tissue compression/deformation. The average epidermal thickness measured under each micropore was $64\pm 19\mu \mathrm{m}$ , which indicates that the microneedles penetrate on average 61% of the epidermis. It must also be noted that the subject did not indicate any pain or discomfort during the microneedle insertion.

3.2.

Penetration Depth of Microneedles in Fingertip

The fingertip could prove a very elegant region for the administration of drugs using microneedles, given its easily accessible location. Hence this area was also assessed as an area for microneedle insertion. For this region, we aimed to examine how long the micropores would remain open. To achieve this, a scan was repeated every $5\min$ for $85\min$ after the needles were inserted. To prevent the finger from moving between subsequent scans, a mold of the fingertip was made. This was produced by pressing the finger into a large block of Blu-Tack^® and molding it around the finger. This minimized finger movement during 3-D scans and allowed the finger to be placed back into the same location after each scan.

As before, the finger was placed under the OCT system prior to needle insertion. The microneedles were then pressed into the dorsal aspect of the finger with a moderate level of pressure just above the nail-fold plexus and held for $10\mathrm{s}$ . The microneedles were removed and the location imaged using OCT (see Fig. 7 ). As before, the subject did not report any pain or discomfort during microneedle insertion. Again, the thickness of the epidermis was measured using the OCT system. The full thickness of the stratum corneum was determined to be $146\pm 19\mu \mathrm{m}$ for the fingernail plexus. This value was calculated using a refractive index of the stratum corneum previously reported as a value of $1.51\pm 0.02$ .³³ The micropores were each labeled (see Fig. 7), manually segmented, and their penetration depth determined. The results show that the average penetration depth of the needles has been measured as $158\pm 20\mu \mathrm{m}$ , similar to that obtained in the forearm region. On average, $52\pm 11\mu \mathrm{m}$ of stratum corneum remains under the micropores. This shows that the microneedles have penetrated through approximately 64% of the stratum corneum, indicating an enhanced transdermal diffusion rate of pharmaceutical agents via microneedle usage. Even though the diffusion rate would be limited by the remaining stratum corneum under the micropore, it must be noted that the stratum corneum in this particular region is relatively thick. Hence this pore depth would sufficiently overcome the relatively thinner stratum corneum $\left(10\text{to}15\mu \mathrm{m}\right)$ found in most other regions of the body.

Fig. 7

Microneedle insertion into the fingertip just above the nailfold plexus. (a) Optical image showing region where microneedles were inserted. (b) OCT scan of microneedle pores for the region marked in (a). (c) Enlarged region of a single microneedle pore in finger tissue as marked in (b). (d) Enlarged region showing a cross section of microneedle pore along the X-X line indicated in (c).

3.3.

Micropore Closure Rate

While stratum corneum disruption using microneedles has been extensively studied and it is accepted that skin healing takes place on a time scale of days,^{34, 35} we are unaware of any work that quantitatively describes the time-dependent depth kinetics and closure rate of the needle-induced micropores. These characteristics will be particularly relevant during the development of future transdermal products to assess required drug delivery volumes, dosage calculations, and transdermal diffusion rates. The OCT imaging technique that we present here is real time, and we demonstrate that it is therefore possible to examine the changes in the depth of the pores over time in more detail than was previously possible.

The data show that the pores remain open, although the micropore depth has diminished to $76\pm 19\mu \mathrm{m}$ , while the measured stratum corneum under the micropores has increased to $92\pm 11\mu \mathrm{m}$ . This indicates that the micropores still penetrate 36% of the stratum corneum thickness. It is not yet clear whether the reduction in pore depth is due to commencement of the healing process, or to a relaxation of the tissue at needle array removal due to the elasticity of the skin. Furthermore, it is expected that topical application of a medicant could prevent early commencement of the healing process and allow diffusion pathways to remain open for a considerable time. OCT scans comparing the depth of the pores immediately after microneedle insertion and $85\min$ after insertion are illustrated in Fig. 8 .

Fig. 8

Magnified view of the microneedle holes on the fingertip: (a) and (b) immediately after application and (c) and (d) $85\min$ after application.

The data show that the needle holes are still clearly visible in the OCT images $85\min$ after insertion. The evolution of the micropores over time can be examined (see Fig. 9 ). The graph shows the average micropore depth for each row of microneedles (i.e., A is the average of A1, A2, A3, and A4) and also the average stratum corneum thickness remaining under the tips of the microneedles. The data show that the pores reduce from a depth of $158\pm 20\mu \mathrm{m}$ down to a depth of $76\pm 13\mu \mathrm{m}$ in a time frame of $85\min$ . The data also show that the measured stratum corneum under the micropores increases from $52\pm 11\mu \mathrm{m}$ up to $92\pm 11\mu \mathrm{m}$ in the same time frame. This is a key result, as it shows the slow time frame over which the micropores close, which is important because it gives an indication of whether there would be sufficient time for a chemical agent to diffuse across the tissue.

Fig. 9

Dynamics characterization of microneedle closure. Graph shows the change in average penetration depth for micropores in the fingertip (a) and change in stratum corneum (SC) thickness under each micropore (b). (Dataset A represents average of A1, A2, A3, and A4. See Fig. 7 for grid reference.)