Design of Laser Treatment Protocols for Bacterial Disinfection in Root Canals Using Theoretical Modeling and MicroCT Imaging

Theoretical simulations of temperature elevations in root dentin are performed to evaluate how heating protocols affect the efficacy of using Er,Cr;YSGG pulsed lasers for bacterial disinfection during root canal treatments. The theoretical models are generated based on microCT scans of extracted human teeth. Heat transfer simulations are performed using the Pennes bioheat equation to determine temperature distributions in tooth roots and surrounding tissue during 500 mW pulsed Er,Cr;YSGG laser irradiation on the root canal for eradicating bacteria. The study not only determines the heat penetration within the deep dentin, but also assesses potential thermal damage to the surrounding tissues. Treatment protocols are identified for three representative tooth root sizes that are capable of maintaining elevated temperatures in deep dentin necessary to eradicate bacteria, while minimizing potential for collateral thermal tissue damage at the outer root surfaces. We believe that the study not only provides realistic laser heating protocols for various tooth root geometries, but also demonstrates utility of theoretical simulations for designing individualized treatments in the future. Introduction More than 15 million root canal treatment procedures are performed annually in the United States [ADA Survey of Dental Procedures Rendered 2007]. Root canal bacteria, particularly Gram-positive facultative anaerobes, seem to be remarkably resistant to local antimicrobial agents often used in root canal therapy. In fact, persistent cultivable bacteria are found in more than 40% cases [Card et al., 2002; Jha et al., 2006; Perez et al., 1993; Sjogren et al., 1997]. Moreover, the success rate of treatment in cases with pre-operative infections is significantly lower than in cases without infections, necessitating research on more effective root canal disinfecting protocols [Alley et al., 2004; Blicher et al., 2008; De Chevigny et al., 2008; Ng et al., 2011]. In addition to traditional approaches, intra-canal heating using a laser catheter has the potential to eradicate residual bacteria in deep root dentin. However, a carefully designed heating protocol is needed to achieve a desired thermal dosage that preserves sensitive surrounding tissues of the tooth root. Typically, the root canal system is irrigated with 0.5-5% sodium hypochlorite solution, and much of the infected hard tissue lining the canal is eliminated by mechanical instrumentation. Once this process is complete, a canal medicament such as calcium hydroxide is applied directly to the canal walls. The medicament mostly eradicates bacteria with which it comes in direct contact. The effectiveness of calcium hydroxide, even when mixed with chlorhexidine has been called into question. Another potential reason for failure of root canal disinfection with irrigants is the difficulty in distributing the irrigants in the extremely narrow or curved regions of root canals in sufficient quantities based on instrumental and visual limitations [Penesis et al., 2008]. Again, the irrigant must make direct physical contact with the bacteria in sufficient concentration for eradication. Bacterial colonies remaining in the root canal can reproduce, and may communicate with the periapical tissues despite root fillings, thus, leading to persistent infections and failure of the treatment [Shovelton 1964; Siqueira et al., 1996]. It has been reported that bacteria penetrate up to 800 μm from the root canal surface in root dentin [Drake et al., 1994; Haapasalo and Orstavik 1987; Love 2001; Perez et al., 1993]. One study of post-operative patients concluded that root canal procedures were successful in 68-80% of patients one year after treatment [Ng et al., 2007]. The largest factor affecting success rates is the inability to reach the apical infection with irrigant paste [Na et al., 2008]. Due to the limitations of conventional antibacterial methods, the application of heat has been proposed for bacterial disinfection of the root dentin [Gordon et al., 2007, Moritz et al., 2000]. Unlike traditional disinfection methods, heat delivered by irradiation of the root canal is capable of fully penetrating the dentin tubules via conduction. The Er,Cr;YSGG laser [Ishizaki et al., 2004; Matsuoka et al., 2005; Radatti et al., 2006; Yamazaki et al., 2001; Zhu et al., 2009] is of particular interest because the majority of its absorption is superficial. The absorption coefficient of the Er,Cr;YSGG laser is equal to approximately 10 cm [Aoki et al., 2008; Dederich and Bushick 2004], which means that the light penetrates only 1 μm into a dental material, making it best utilized for its thermal conductive properties and allowing it to be modeled as a heat flux for theoretical purposes. When used at a frequency of 20 Hz with pulse duration of 200 μs, this laser has been shown to be particularly capable of bacterial elimination, as it promises a reduction in bacterial re-growth by an order of magnitude when compared with traditional treatment techniques [Gordon et al., 2007]. Studies have demonstrated a 0.2% bacterial regrowth rate after laser treatment, compared to the 2.0% re-growth rate associated with conventional irrigant methods [Gordon et al., 2007]. One reason contributing to this difference is the use of heat. In addition to these benefits, the optical fiber tips of the lasers have diameters as small as 200 μm, which makes them capable of reaching the curved and narrow spaces close to the root apex, after preparing the canal to a size 20 file. If significant levels of disinfection could be achieved with lasers, this modality would be useful in future dental disinfection protocols. In order for laser irradiation to be clinically accepted as a root canal disinfection method, a number of parameters must be optimized, including the level and duration of heating required to eradicate bacteria without damage to the adjacent tissues. A prior theoretical study [Zhu et al., 2009] attempted to evaluate the resulting temperature field in tooth dentin under various heating protocols used in clinical studies. Though that study identified a heating duration to avoid collateral thermal tissue damage, it was based on a very simplified geometric model previously used in an in vitro clinical experiment [Gordon et al., 2007]. The simplified model incorporated assumptions in the tooth and canal geometry that limit the accuracy of the previous theoretical analyses. To our knowledge, there has been no theoretical study on the dentin temperature distribution arising from thermal treatment based on realistic tooth root geometry, including an assessment of temperature elevations at the cementum, periodontal ligament, and the surrounding bone. In this study, we focus on how tooth geometry affects temperature elevations in root dentin. Heat transfer simulations are performed using the Pennes bioheat equation [Pennes 1948] to understand the heat distribution in root dentin during 500 mW Er,Cr;YSGG laser irradiation for bacterial disinfection. Theoretical models are generated based on microCT scans of extracted human teeth. The study not only determines the heat penetration within the deep dentin, but also assesses potential thermal damage to the surrounding tissue structure. Realistic laser treatment protocols are identified that are capable of maintaining the required elevated temperature within deep dentin necessary to eradicate bacteria, while simultaneously minimizing collateral thermal tissue damage. The balance between these two factors is evaluated in the study. Methods MicroCT imaging of extracted human teeth Seven extracted single-root human teeth were provided by dentists in the School of Dentistry at the University of Maryland, Baltimore according to a protocol approved by the Institutional Review Board from the University of Maryland Baltimore County. An access preparation was performed, then the canals were prepared using a standard K-files. The teeth were all kept in individual test tubes, submerged in distilled water and stored in a refrigerator in the laboratory. Prior to scanning using a microCT imaging system (Skyscan 1172, Microphotonics, PA), each tooth was thoroughly rinsed with distilled water to remove debris. These teeth were then dried with compressed air and mounted inside a low density support structure to ensure maximum stability during the scan. The pixel size in each image was selected as 17.2 μm (medium scan) and each image file contains 2000 x 1048 pixels. Since the teeth have an aspect ratio of 3 or more, we used a set up called “oversized scan”, which combines multiple scans of individual segments. The rotation rate was set to be 0.4 at six increments. Lastly, the filter was chosen as an aluminum copper filter to produce an image with the best contrast. The scans took an average of six hours for each tooth. Following completion of the scans, the NRecon software package provided by the company was used in conjunction with six primary computer servers to reconstruct each specimen. Using the CTAn analysis software package, the reconstruction files allow generations of 3-D tooth models or cross-sectional images of the teeth. Although it would be ideal to import the 3-D tooth model to a commercial software for heat transfer simulation, it was not feasible due to limitations of the computational memory and a large number of irregularities in the geometry not admissible to the commercial software. In this study, we model the tooth structure as a two-dimensional axis-symmetric model as a compromise. The 2-D model captures the important geometrical variations pertaining to the tooth dentin structure without detrimental loss of detail. Figure 1 gives the reconstructed cross sections of the seven teeth. Based on the figure, it is clear that Tooth A is the longest of the seven provided specimens and Tooth D is the shortest. The variable brightness shown in the longitudinal-sectional images is indicative of varying densities, and thereby attenuation coefficients. The bright area surrounding the crown of the tooth is the enamel, shown most prevalent in Tooth A and Tooth C. Tooth E and Tooth D have a notably brighter material residing on the enamel surface; this is a composite filling material. Prior to obtaining the seven teeth from our collaborators at the Dental School of the University of Maryland, Baltimore, some teeth underwent a typical root canal procedure. Therefore, filling material may remain inside the root canal and it shows as a brighter region within the root canal in the microCT images of the teeth. When these longitudinal-sectional images were used to model the geometry, these regions were excluded from the root canal models because they are not a portion of the natural tooth geometry. The region comprised of dentin in each tooth appears as a homogeneous material. The dentin tubules, with diameters of 1-2 μm, are not visible since the pixel size was set to 17 μm. In this study, heat transfer simulations are performed on Tooth A (the largest), Tooth F (the middle), and Tooth D (the smallest). Mathematical formulation of the heat transfer model The roots of the three teeth of interest, the largest, the middle, and the smallest, are modeled as two-dimensional axis-symmetric geometries (Figure 2). Each geometry is imported into Comsol as an axis-symmetric model and embedded in a section of tissue to mimic realistic conditions inside the mouth. Both the tissue and the tooth model are transformed into solid entities using the “coerce to solid” function in Comsol. The longitudinal-sectional plane of the model is shown in Figure 3, where the tooth root is embedded in the tissue block. The tissue block is selected large enough so that the prescribed 37°C boundary condition on the tissue surface far away from the root canal surface is reasonable. We use the Pennes bioheat equation [Pennes 1948] to simulate the heat transport in the tissue as a result of laser treatment. Neglecting metabolic heat generation, one can write the governing equation for temperature T as ) 37 ( , , 2 , , tissue root b b tissue root tissue root tissue root T C T k t T C − + ∇ = ∂ ∂ ωρ ρ (1) where k is thermal conductivity, ρ is density, C is specific heat, and ω is the local blood perfusion rate. The thermal effect of the blood perfusion in the surrounding tissue is modeled as a heat source term with a strength proportional to the local blood perfusion rate, and the temperature difference between the body temperature (37°C) and the local tissue temperature. Table 1 lists the thermal properties used in the simulation based on previous studies [Craig and Peyton, 1961; Diao et al., 2003; Jakubinek and Samarasekera 2006]. In root dentin the blood perfusion rate is zero, while in the surrounding bony tissue it is 1.8 ml/100g/min [Diao et al., 2003]. The initial condition is 37°C in the entire domain. Boundary conditions are indicated in