Deep learning for tooth identification and numbering on dental radiography: a systematic review and meta-analysis.

OBJECTIVES: Improved tools based on deep learning can be used to accurately number and identify teeth. This study aims to review the use of deep learning in tooth numbering and identification. METHODS: An electronic search was performed through October 2023 on PubMed, Scopus, Cochrane, Google Scholar, IEEE, arXiv, and medRxiv. Studies that used deep learning models with segmentation, object detection, or classification tasks for teeth identification and numbering of human dental radiographs were included. For risk of bias assessment, included studies were critically analysed using quality assessment of diagnostic accuracy studies (QUADAS-2). To generate plots for meta-analysis, MetaDiSc and STATA 17 (StataCorp LP, College Station, TX, USA) were used. Pooled outcome diagnostic odds ratios (DORs) were determined through calculation. RESULTS: The initial search yielded 1618 studies, of which 29 were eligible based on the inclusion criteria. Five studies were found to have low bias across all domains of the QUADAS-2 tool. Deep learning has been reported to have an accuracy range of 81.8%-99% in tooth identification and numbering and a precision range of 84.5%-99.94%. Furthermore, sensitivity was reported as 82.7%-98% and F1-scores ranged from 87% to 98%. Sensitivity was 75.5%-98% and specificity was 79.9%-99%. Only 6 studies found the deep learning model to be less than 90% accurate. The average DOR of the pooled data set was 1612, the sensitivity was 89%, the specificity was 99%, and the area under the curve was 96%. CONCLUSION: Deep learning models successfully can detect, identify, and number teeth on dental radiographs. Deep learning-powered tooth numbering systems can enhance complex automated processes, such as accurately reporting which teeth have caries, thus aiding clinicians in making informed decisions during clinical practice.

Image-Guided Ultrasound Characterization of Volatile Sub-Micron Phase-Shift Droplets in the 20-40 MHz Frequency Range.

Phase-shift perfluorocarbon droplets are designed to convert from the liquid to the gas state by the external application of acoustic or optical energy. Although droplet vaporization has been investigated extensively at ultrasonic frequencies between 1 and 10 MHz, few studies have characterized performance at the higher frequencies commonly used in small animal imaging. In this study, we use standard B-mode imaging sequences on a pre-clinical ultrasound platform to both image and activate sub-micron decafluorobutane droplet populations in vitro and in vivo at center frequencies in the range of 20-40 MHz. Results show that droplets remain stable against vaporization at low imaging pressures but are vaporized at peak negative pressures near 3.5 MPa at the three frequencies tested. This study also found that a small number of size outliers present in the distribution can greatly influence droplet performance. Removal of these outliers results in a more accurate assessment of the vaporization threshold and produces free-flowing microbubbles upon vaporization in the mouse kidney.

Numerical study of a novel induced-charge electrokinetic micro-mixer.

A novel micro-mixer based on the induced-charge electrokinetic motion of an electrically conducting particle is proposed and numerically demonstrated in this paper. For most microfluidic applications, it is desired to mix different streams of solutions rapidly in a continuous flow mode. Therefore, in this work, we consider a mixing chamber containing an electrically conducting particle and the mixing chamber is located in the middle of a microchannel. Vortices are generated around the electrically conducting particle in an aqueous solution due to the interaction of the applied electric field and the induced surface charge on the particle. These vortices will enhance significantly the mixing of different solutions around the particle. The effectiveness of mixing the two streams entering the mixing chamber is numerically studied as functions of the applied electric field. Excellent mixing can be achieved in this system under two perpendicularly applied electric fields. The proposed micro-mixer is simple and easy to be fabricated for lab-on-a-chip applications.

Control of flow rate and concentration in microchannel branches by induced-charge electrokinetic flow.

This paper presents a numerical study of controlling the flow rate and the concentration in a microchannel network by utilizing induced-charge electrokinetic flow (ICEKF). ICEKF over an electrically conducting surface in a microchannel will generate vortices, which can be used to adjust the flow rates and the concentrations in different microchannel branches. The flow field and concentration field were studied under different applied electric fields and with different sizes of the conducting surfaces. The results show that, by using appropriate size of the conducting surfaces in appropriate locations, the microfluidic system can generate not only streams of the same flow rate or linearly decreased flow rates in different channels, but also different, uniform concentrations within a short mixing length quickly.

Micro-valve using induced-charge electrokinetic motion of Janus particle.

A new micro-valve using the electrokinetic motion of a Janus particle is introduced in this paper. A Janus particle with a conducting hemisphere and a non-conducting hemisphere is placed in a junction of several microchannels. Under an applied electric field, the induced-charge electrokinetic flow around the conducting side of the Janus particle forms vortices. The vortices push the particle moving forwards to block the entrance of a microchannel. By switching the direction of the applied electric field, the motion of the Janus particle can be changed to block different microchannels. This paper develops a theoretical model and conducts numerical simulations of the three-dimensional transient motion of the Janus particle. The results show that this Janus particle-based micro-valve is feasible for switching and controlling the flow rate in a microfluidic chip. This method is simple in comparison with other types of micro-valve methods. It is easy for fabrication, for operation control, and has a fast response time. To better understand the micro-valve functions, comparisons with a non-conducting particle and a fully conducting particle were made. Results proved that only a Janus particle can fulfill the requirements of such a micro-valve.

Concentrating molecules in a simple microchannel.

A simple method is proposed and tested to concentrate sample molecules from a dilute solution in a microchannel by electrokinetic means. The microfluidic chip has a straight microchannel connecting two wells and three electrodes. This method uses electrokinetic trapping and flow control simultaneously to concentrate a charged species of interest. A numerical model of the sample concentration process is presented in this paper. Using a fluorescent dye as the sample molecules, experimental investigation into the concentration process was performed. The 90 times of the concentration increase was achieved in 110 s. The numerical simulations of the concentrating and the subsequent dispensing processes agree well with the experimental results.

Numerical studies of electrokinetic control of DNA concentration in a closed-end microchannel.

A major challenge in lab-on-a-chip devices is how to concentrate sample molecules from a dilute solution, which is critical to the effectiveness and the detection limit of on-chip bio-chemical reactions. A numerical study of sample concentration control by electrokinetic microfluidic means in a closed-end microchannel is presented in this paper. The present method provides a simple and efficient way of concentration control by using electrokinetic trapping of a charged species of interest, controlling liquid flow and separating different sample molecules in the microchannel. The electrokinetic-concentration process and the controlled transport of the sample molecules are numerically studied. In this system, in addition to the electroosmotic flow and the electrophoresis, the closed-end of the chamber causes velocity variation at both ends of the channel and induces a pressure gradient and the associated fluid movement in the channel. The combined effects determine the final concentration field of the sample molecules. The influences of a number of parameters such as the channel dimensions, electrode size and the applied electric field are investigated.

Eccentric electrophoretic motion of a rectangular particle in a rectangular microchannel.

Understanding of the effects of the boundary - the channel walls - on the electrophoretic motion of particles in microchannels is very important. This paper developed an analytical solution of the electrophoretic mobility for eccentric motion of a rectangular particle in a rectangular microchannel. The simple geometry of the system does not limit the generality of the qualitative prediction of the model and the analytical solution. Several special cases are studied, and the effects of the degree of the eccentricity, the particle's size relative to the channel's size, and the relative zeta potentials on the particle's mobility are discussed. For the case where the particle's cross-section area is close to the cross-section area of the microchannel, the model's predictions are compared with the published experimental results and good agreement was found.