2.1.

Fiber Fabrication and Simulation

Figure 1 presents images of a fabricated diffusing fiber tip for endometrial treatment. For simple and reliable machining purposes, a 1-mm core diameter, synthetic-fused silica (Suprasil 312, Heraeus Quarzglas GmbH & Co. KG, Germany) was selected to transmit the visible laser light ($\lambda =532\mathrm{nm}$). Initially, the fiber cladding was removed mechanically, and the surface of the fiber core at the 25 mm distal end was circumferentially micro-machined by a 30 W ${\mathrm{CO}}_2$ laser ($\lambda =10.6\mu \text{m}$, Synrad, Mukilteo, WA, USA) in a predetermined zigzag pattern.²¹ A series of scattering segments (approximately 50 μm in size) were created on the fiber surface to diffuse the laser light radially in all directions. In order to achieve the constant size of the scattering segments, the fiber tip was tapered down to 0.5 mm in diameter at the distal end [i.e., minimum diameter for currently reliable manufacturability; Fig. 1(a)]. Then, the diffusing tip was covered with a 27-mm long glass (2 mm outer diameter) cap to attain wide and uniform light diffusion as well as to protect the bare fiber tip during laser treatment, as shown in Fig. 1(b).

Fig. 1

Diffusing fiber tip for endometrial treatment: (a) 25 mm long tapered bare core (fused silica, 1 mm in diameter) after ${\mathrm{CO}}_2$ laser micro-machining (each scattering segment of $\sim 50\mu \text{m}$) and (b) glass-capped diffusing fiber tip for wide and uniform photon distribution.

In an attempt to predict photon distribution from the designed diffusing tip, optical simulation (Radiant Zemax, Redmond, WA, USA) was conducted to demonstrate light intensity and its spatial distribution at various distances. Two fiber conditions were compared: a bare diffuser tip (without a glass cap) and a glass-capped diffuser tip. To model the light scattered from the fiber surface, a Lambertian diffuser model was used with one million of rays and a light source with uniform angular distribution.²² The diffusing tip was exclusively modeled with surface scattering (i.e., $\sim 50\mu \text{m}$ size scattering segment). The applied wavelength was 532 nm with the input power of 120 W, and the entire fiber length was 1.5 m including a 25 mm diffusing part at the tip. A $40\times 50\mathrm{mm}$ planar detector was placed underneath the diffusing fiber at distances of 1, 5, and 10 mm to identify light propagation and the spatial distribution of the scattered photons in two-dimensional (2-D). The profiles of light intensity (i.e., irradiance, $\mathrm{W}/{\mathrm{cm}}^2$) were also measured and quantitatively compared between the two fiber conditions.

2.2.

In Vitro Experiments

Bovine liver tissue was used as a tissue model for in vitro tests with the designed diffusing fibers, in that the chromophores, such as dead endothelial cells and blood vessels, would still be able to absorb the visible laser light ($\lambda =532\mathrm{nm}$) significantly.^23,24 The liver specimens were acquired from a local slaughter house, and they were cut into $5\times 7\mathrm{cm}$ segments in size (approximately 1 cm thick) and stored at 4°C prior to the experiments. Figure 2(a) illustrates an experimental set-up of photocoagulation tests with the fibers. A circular tissue holder (7 cm in diameter and 1 cm in thickness) was prepared, and a 1 cm thick tissue sample was placed at the bottom of the holder along the curvature. The curved surface of the tissue sample partially reflected the transverse anatomic features of human uterus [Fig. 2(a)]. The diffusing fiber was located 1 cm above the tissue surface, so the incident light was uniformly irradiated on the tissue due to the curved geometry of the sample. A conventional clinical laser ($\lambda =532\mathrm{nm}$, quasi-cw mode, High Performance System, CA, USA) was employed for laser coagulation, and the input power was maintained at 120 W in order to entail tissue coagulation through the diffusing tip. The average irradiance on the tissue surface was calculated to be approximately $3.8\mathrm{W}/{\mathrm{cm}}^2$. Three fiber conditions were tested: bare diffusing fiber, capped diffusing fiber, and capped diffusing fiber with a 1 mm thick layer of polyurethane (PUR). As PUR is a raw material for balloon catheters for endoscopic applications,²⁵ so the last condition represented the final device design to incorporate the capped diffusing fiber into a balloon catheter. However, for the sake of experimental simplicity, a thin layer of PUR was placed on top of the target tissue [Fig. 2(a)] instead of using a balloon catheter. During the tests, a PUR layer was superimposed upon the tissue sample as shown in Fig. 2(a) (left image). Prior to coagulation tests, the optical transmission of each fiber (bare diffusing, capped diffusing, and capped diffusing with a PUR layer) was roughly measured with a photodetector to be 99%, 97% and 94.5%, respectively. Figure 2(b) presents the intensity profile along the capped fiber tip that was measured every 5 mm. In addition, the transmission loss of PUR at 532 nm was measured with the detector to be less than 2.5%, which experimentally verified negligible light absorption (i.e., optically transparent with no light scattering) at 532 nm by the PUR layer. The coagulation threshold was preliminarily determined for the three fiber conditions by varying irradiation times from 2 to 8 s (1 s increments; a sample per each condition). The onset of visible discoloration on tissue surface was considered the physical evidence of coagulation. A number of irradiation times (4, 7, 15, 30, 60, 90, 120, 150, and 180 s) were then evaluated for the three conditions to identify the temporal evolution of photocoagulation on the tissue surface. Due to difficulty of preparing fresh tissue specimens with a large uniform surface area, each condition was tested only with a single liver sample for evaluating surface coagulation. All the specimens were placed under saline environment (room temperature) during the laser irradiation, and the temperature of the liver tissue was maintained at approximately 20°C. Postexperimentally, the radial depth (i.e., into the tissue) and 2-D area of coagulation on the tissue surface were quantified and compared with an image processing software tool (Image J, National Institute of Health, MD, USA). The image in Fig. 2(a) (bottom of the right image) showed a cross-sectional part of the photocoagulated tissue. In order to measure coagulation thickness, each irradiated specimen was cross-sectioned into three pieces [Fig. 2(a)], and the coagulation thickness of each piece was measured five times ($n=15$). The discolored zone (tan color) represented coagulation necrosis and the red color was the preserved native tissue. A two-tailed Student’s $t$-test was used for statistical analysis and $p<0.05$ meant statistically significant.

Fig. 2

In vitro photocoagulation test: (a) experimental set-up (cross-sectional view of diffusing light set-up, left image; tissue encased by tissue holder and fiber placed along tissue, right image; cross-sectional image ($A\text{-}{A}^{\prime }$) of coagulated tissue, bottom of right image) and (b) intensity profile measured from 25 mm capped diffusing fiber tip.

2.3.

In Vivo and Cadaver Experiments

Three mature female Saanen goats were used for in vivo retrograde laser coagulation studies. Animal procedures and care were conducted in accordance with a protocol approved by American Preclinical Service (APS) Institutional Animal Care and Use Committee (IACUC). Experiments, necropsy, and histology were performed at APS,¹² and all surgical procedures were performed with the animals under general endotracheal anesthesia. A caprine uterus is typically bicornuate so two prominent uterine horns come together to form a short uterine body.^26–28 Thus, six caprine uteri in total were tested for the current photocoagulation tests. Similar to a human uterus, the caprine uterine wall consists of two major tissue layers: endometrium and myometrium.^27,28 The endometrium is a pseudostratified layer of epithelium on the luminal surface of the uterus, containing richly vascular loose connective tissue along with fibroblasts, macrophages, and mast cells. The myometrium is two layers of smooth muscle separated by the stratum vasculare, which is a zone of large vessels (arteries, veins, and lymph vessels). For the current in vivo studies, the prototype device was evaluated to photocoagulate solely the endometrial layer, in that any thermal injury to the myometrium would adversely affect fertility. The prototype device consisted of a capped diffusing fiber, a PUR balloon catheter, a customized inflating tube (1 cm outer diameter and 8 cm long), and a customized inflating pump (variable pressure levels from 1 to 7 psi). The 4-cm long catheter device was inserted into the animal uterus, and the balloon catheter was distended with saline until it securely held the uterine wall (i.e., approximately 3 cm in balloon diameter at 5 psi). A 532 nm clinical laser was used with the input power of 120 W, and the irradiation time was approximately 30 s, based upon in vitro results (i.e., applied $\text{energy}=3600\mathrm{J}$ and $\text{irradiance}=3.2\mathrm{W}/{\mathrm{cm}}^2$ assuming a 3-mm thick endometrium for photocoagulation). Postoperatively, all the animals were euthanized 2 h after the tests by euthasol injection. Immediately after euthanasia, each uterine horn was removed, fixed in 10% neutral buffered formalin, and embedded in paraffin for hematoxylin and eosin (H&E) staining. From histology images, the thickness of coagulation necrosis was measured with Image $\mathrm{J}(n=18)$ and evaluated quantitatively.

A human uterus was donated by a 59-year-old postmenopausal patient for research at APS after a radical hysterectomy. The cadaveric uterus was used to evaluate the feasibility of the prototype device in terms of light leaking and deployment of fiber and balloon during laser irradiation. The device was inserted through the cervix for minimally invasive uterine access, and a 5-cm long balloon catheter was distended at 4 psi by saline (approximately 1.8 cm in balloon diameter). The applied power of 120 W was irradiated on the uterine wall for approximately 20 s (i.e., applied $\text{energy}=2400\mathrm{J}$ and $\text{irradiance}=4.2\mathrm{W}/{\mathrm{cm}}^2$) as the distance between the fiber and tissue was closer than the in vivo condition due to the rigidity of the cadaveric tissue. The degree of coagulation necrosis in the tissue was also examined postexperimentally with Image $\mathrm{J}(n=12)$. A digital camera (9.1M DSC-H50, Sony Corp, Japan) was used to take images of pre-, intra-, and postoperation to show a sequence of photocoagulation. A two-tailed Student’s $t$-test was also used for statistical analysis and $p<0.05$ meant statistically significant.

3.1.

Optical Simulation

Figure 3 shows the spatial distribution of photons from optical simulation comparing diffusing and capped diffusing fibers at various distances of 1, 5, and 10 mm. At 1 mm, both fibers created similar photon distributions with generation of high irradiance along the fiber due to their close proximity to a planar detector. However, as the distance from the detector increased up to 10 mm, the distribution became wider due to light diffusion from scattering segments on the fiber surface. The diffusing fiber presented elongated shape (along $z$ axis) with relatively lower irradiance whereas the capped diffusing fiber created relatively circular distribution with higher irradiance, resulting from additional beam diffraction (along $x$ axis) through the glass cap [Fig. 3(a)].

Fig. 3

Simulated photon distribution from newly designed diffuser tip at various distances (1, 5, and 10 mm): (a) 2-D spatial distribution of irradiance for diffusing fiber (top) and capped diffusing fiber (bottom) and (b) longitudinal (along $z$-axis at $x=0\mathrm{mm}$) and horizontal (along $x$-axis at $z=10\mathrm{mm}$) irradiance distribution measured at 10 mm distance. Note that the distal end of each fiber was placed at $x=0$ and $z=25\mathrm{mm}$ and a $40\times 50\mathrm{mm}$ planar detector was used.

At a distance of 10 mm, longitudinal and horizontal distributions of the incident photons were compared in Fig. 3(b). Both directions demonstrated that the capped diffusing fiber yielded approximately 40% higher irradiance than the diffusing fiber (i.e., $\text{peak intensity}=10.8\mathrm{W}/{\mathrm{cm}}^2$ for the capped diffusing versus $7.7\mathrm{W}/{\mathrm{cm}}^2$ for the diffusing fiber). Based upon the longitudinal position, the width of the irradiance distribution was comparable between the two cases [Fig. 3(b)]. However, the capped diffusing fiber created around 30% wider horizontal distribution of the irradiance (i.e., $\text{Full-width-half-maximum}=15.2\mathrm{mm}$ for the capped diffusing versus 11.5 mm for the diffusing fiber). According to the simulation results, an additional layer from the glass cap circumscribed the longitudinal photon distribution and contributed to uniformly distribute the incident photons along $x$-axis.

3.2.

In Vitro Results

Figure 4 demonstrates the progression of laser-induced coagulation on tissue as a function of irradiation time for three fiber conditions: diffusing, capped diffusing, and capped diffusing fiber with PUR. The total applied energy (i.e., 120-W input power times irradiation time) was 840, 3600, 7200, and 14,400 J at 7, 30, 60, and 120 s, respectively. Overall, the degree of photocoagulation on the tissue surface developed gradually with the irradiation time. Coagulation initially developed in a vertical direction (i.e., perpendicular to the fiber axis) and later expanded horizontally (i.e., along the fiber axis) with the irradiation time. The shape of the coagulated area was almost rectangular for the three conditions, and at 120 s the area length (i.e., perpendicular to the fiber axis) and width (i.e., along the fiber axis) were measured to be 3.1 and 2.5 cm, respectively. In other words, the area length was equivalent to the length of the half arc length of the diffusing light 1 cm away from the fiber (i.e., $\pi \times 1\mathrm{cm}$ distance), and the area width became equivalent to the length of the diffusing fiber tip ($\sim 25\mathrm{mm}$). Compared to the diffusing and capped diffusing fibers, the last condition (capped diffusing fiber with PUR) overtly yielded more rapid and wider tissue coagulation (i.e., $9.6{\mathrm{cm}}^2$ for the last condition versus $0{\mathrm{cm}}^2$ for diffusing fiber and $5.2{\mathrm{cm}}^2$ for capped diffusing fiber at 7 s after irradiation). In fact, the coagulated area for the last condition became rapidly saturated after 30 s, compared to the two other conditions.

Fig. 4

Series of images of photocoagulated tissues at various irradiation times: diffusing fiber (top), capped diffusing fiber (middle), and capped diffusing fiber with polyurethane (PUR) layer (bottom). Note that each number in the left corner (${\mathrm{cm}}^2$) represents coagulated area (dotted in tissue), respectively.

Figure 5(a) shows the quantitative evaluation of coagulation depth in a radial direction (into the tissue) as a function of irradiation time. For a diffusing fiber, coagulation threshold time was around 7 s, whereas coagulation for two other conditions was initiated around 4 s after irradiation ($\text{coagulation thickness}=100$ to 200 μm). Similar to Fig. 4, a capped diffusing fiber with PUR increased the coagulation depth more rapidly than the other conditions. In the case of 1-min irradiation, the capped fiber with PUR created the coagulation necrosis of $3.5\pm 0.3\mathrm{mm}$, which was 5-fold and 1.5-fold thicker than the diffusing ($0.7\pm 0.2\mathrm{mm}$) and capped diffusing fibers ($2.5\pm 0.3\mathrm{mm}$), respectively [$p<0.001$; Fig. 5(a)].

Fig. 5

Quantitative evaluation of coagulated tissue as function of irradiation time: (a) radial depth (mm) of coagulation and (b) area (${\mathrm{cm}}^2$) of coagulation with three experimental conditions (diffusing, capped diffusing, and capped diffusing fiber along with polyurethane layer).

Figure 5(b) demonstrates the progression of coagulation area on the tissue surface that indicated lateral thermal expansion. The diffusing fiber presented that coagulation area increased almost linearly with the irradiation time as coagulation depth did in Fig. 5(a). On the other hand, for both capped diffusing fiber and capped diffusing fiber with PUR, the coagulation area initially increased but became saturated approximately 1 min after irradiation, whereas the overall tendency of the coagulation depth almost linearly increased with time for both cases. At 1-min irradiation, the capped diffusing fiber with PUR condition yielded 3.8-fold and 1.6-fold larger coagulation areas than the diffusing and capped diffusing fibers, respectively (i.e., $18.9{\mathrm{cm}}^2$ for capped diffusing fiber with PUR versus $5.0{\mathrm{cm}}^2$ for diffusing and $11.7{\mathrm{cm}}^2$ for capped diffusing fibers).

3.3.

Prototype and In Vivo Results

After in vitro validation of various diffusing fiber conditions, the design for the capped diffusing fiber was finalized, and the prototype optical device was made and incorporated with a balloon catheter for in vivo and cadaver studies as shown in Fig. 6(a). The capped diffusing fiber was placed in the center of an 8 cm long (1 cm in outer diameter) customized inflating tube [inset of Fig. 6(a)], of which the proximal end was connected to a customized inflating pump (variable pressure levels from 1 to 7 psi) to regulate the input pressure for saline supply. The distal end of the fiber tip was insecurely positioned (i.e., free end) inside the balloon. A 4 cm long PUR-based balloon catheter was tightly attached to the distal end of the inflating tube, and the size of the balloon was adjustable, depending upon the geometry of uterus and pump pressure levels. In order to distend the balloon catheter, saline was pumped through the tube to fill out the balloon until the target uterus was securely fixed for laser surgery. Prior to in vivo tests, the prototype was validated to ensure the tight sealing of the catheter connection at the distal end of the tube.

Fig. 6

Balloon catheter-combined diffusing fiber for endometrial coagulation: (a) prototype design consisting of capped diffusing fiber balloon catheter, inflating tube, and pump and (b) caprine uterine horn tissue coagulated in vivo with prototype device.

Figure 6(b) shows the acute thermal response of a caprine uterine horn tissue 2 h after a 30 s coagulation with the prototype device ($\text{applied energy}=3600\mathrm{J}$ and $\text{irradiance}=3.2\mathrm{W}/{\mathrm{cm}}^2$). After 30 s irradiation, photocoagulation entailed the uniform coagulation necrosis on the treated tissue, which appeared as the shape of the distended (i.e., 3 cm wide and 4 cm long due to the size of caprine uterus) balloon catheter [inset of Fig. 6(b)]. The overall thickness of coagulation necrosis was measured to be $2.8\pm 1.2\mathrm{mm}$ ($n=18$), which was slightly thinner than the estimated thickness (3 mm) of the distended uterine wall ($p=0.53$). No hemorrhage occurred in both endometrium and myometrium after the laser treatment. Figure 7 presents H&E-stained histology images of the laser-treated uterine tissue. Treatment sites were readily identified by the presence of endometrial gland changes, edema, endometrial connective tissue changes, and vacuolation of smooth muscle cells in myometrium.²⁹ The balloon-contacted tissue was merely coagulated without any thermal injury to adjacent tissue [Fig. 7(a)], in that endometrial glands, as well as the superficial epithelial lining on the luminal surface, were markedly obliterated along with sloughed epithelium. The coagulation process was also evidenced by protein coagulum created on the endometrium surface [Fig. 7(b)], smeared or atypical appearing epithelial cells [Fig. 7(c)], and amphophilic connective tissue surrounding vessels [Fig. 7(c)]. It was confirmed in Fig. 7(d) that the myometrial smooth muscle adjacent to the treatment site yielded minimal changes, such as cellular vacuolation (solid line) and slight discoloration (dotted line), representing no overt thermal damage or necrosis to the myometrium.

Fig. 7

H&E-stained histological images of caprine uterine tissue 2 h after in vivo coagulation testing: (a) uterine horn ($20\times$, L: lumen, C: coagulated endometrium, N: normal endometrial layer, M: myometrium), (b) focal site of protein coagulum ($100\times$; $\mathrm{bar}=10\mathrm{mm}$), (c) coagulated endometrial glands ($200\times$; $\mathrm{bar}=10\mathrm{mm}$), and (d) myometrium adjacent to treated lesion ($100\times$, E: endometrium and M:myometrium; $\mathrm{bar}=10\mathrm{mm}$).

3.4.

Cadaver Study

Followed by in vivo studies, a cadaveric human uterus was tested with the prototype device ($\text{applied energy}=2400\mathrm{J}$ and $\text{irradiance}=4.2\mathrm{W}/{\mathrm{cm}}^2$; Fig. 8). Unlike the in vivo caprine studies, a slightly longer balloon catheter (i.e., 5 cm long versus 4 cm long for in vivo) was used for the larger size of the human uterus [Fig. 8(a)]. However, due to the rigidity of the cadaveric tissue, the balloon was distended only up to 1.8 cm in diameter; accordingly, the irradiation time was limited to approximately 20 s (i.e., 30% shorter than the irradiation time for in vivo study, 30 s) because of the closer irradiation distance between the fiber and endometrial luminal surface (i.e., 0.9 cm for cadaver study versus 1.5 cm for in vivo study) and thereby, its 30% higher irradiance (i.e., $4.2\mathrm{W}/{\mathrm{cm}}^2$ for cadaver study versus $3.2\mathrm{W}/{\mathrm{cm}}^2$ for in vivo study). Figure 8(b) shows a sequence of endometrial coagulation with the prototype balloon catheter-assisted diffusing device. Prior to the treatment, the device securely held the uterus [far left image in Fig. 8(b)], and during the treatment (middle image), the scattered photons (532 nm) were visualized through the tissue, indicating the ongoing intra-operation. According to posttreatment (far right image), the overall thickness of coagulation necrosis was found to be $2.6\pm 0.6\mathrm{mm}$ ($n=12$). The coagulation thickness slightly thinner than the in vivo results ($p=0.50$) indirectly evidenced that the adjusted dosimetry for the cadaver study (i.e., $4.2\mathrm{W}/{\mathrm{cm}}^2$ and 20 s irradiation) was able to induce photocoagulation effects equivalent to those for the in vivo studies (i.e., $3.2\mathrm{W}/{\mathrm{cm}}^2$ and 30 s irradiation).

Fig. 8

Cadaver study with prototype diffusing optical device: (a) human uterus harvested from 59-year-old patient after radical hysterectomy and (b) sequence of endometrial coagulation with diffusing device (pre-, intra-, and posttreatment).