The rapid continuing introduction of the digital world within the health sector after the spread of the COVID-19 pandemic in early 2020, has hastened a shift towards digitisation and further developments of dental applications, tele-dentistry, and many others.1,2 These developments include the use of digital information and personal communication technologies and cloud computing to access and manage a variety of healthcare services.3,4 This advance in technology facilitates offsite-based consultations, prescriptions, patient monitoring, the uploading of records, and audio/visual consultations with assigned healthcare professionals.5
Machine learning and artificial intelligence have been recently introduced in orthodontics to assist the collection and analysis of a large amount of data to learn about a patient’s behaviour patterns.6,7 The recent advancements build on previous systems and structures to advance orthodontic research and develop new models for orthodontic treatment.8,9 Data are fuelling this progress in the healthcare sector and global researchers rely on it to provide services to patients.10 However, collecting data remains a challenge primarily due to the limited capabilities of the monitoring-based systems.11
Rigid adherence to prescribed retention protocols ensures that post-orthodontic treatment goals for long-term stable outcomes are in place.12 Monitoring a patient’s compliance is considered essential to ensure optimal adherence to the prescribed appliance wear protocols.13 Recently, individualised approaches have been developed to overcome the discrepancies between actual clinical progress, and subjective self-reported compliance to different removable retainer wear regimens. These modifications include systems such as microsensors, electronic reminder systems, and mobile device applications.14,15
The capabilities of these adjuncts have facilitated the use of artificial intelligence to digitally track data monitoring systems. However, the accuracy of current systems is often affected by challenges within the platform, the longevity of tracking, and the interaction between the monitoring systems and the experimental conditions (e.g., material composition, environmental conditions, and sustained long-term maintenance).16,17
New thermal monitoring microsensor systems have enhanced the clinical decision-making process and allowed future predictive algorithms to improve clinical practice. The current research group aimed, through this comparative
A commercially available TheraMon® microsensor (S version) was chosen as the most popular microsensor reported in the orthodontic literature.19 The microsensor has relatively small dimensions (13 mm × 9 mm × 4 mm) and a light weight of approximately 0.4 grams. The physical properties allow the integration of the microsensor into different removable orthodontic retainers which would make them less prone to corrosion and less susceptible to recording failure. The microsensor uses an application-specific integrated circuit with 16 kilobytes of Electrically Erasable Programmable Read-Only Memory (EEPROM), which facilitates its multiple reprogramming.20 The microsensor records appliance wear within the temperature range of 33.5°C to 38.5°C (attributable to changes in temperature caused by eating or drinking in the oral environment) at 15 min-intervals. Data may be stored for up to 100 days and then transferred to an associated software cloud in a process that could be repeated for a maximum of 15 months.
An
VFM and HR appliances with embedded thermal microsensor.
The retainer wires were adapted to the teeth and the microsensor was immersed in the acrylic when it was soft. A periodontal probe was used to assess the thickness of the covering acrylic layer from the mesial and occlusal surfaces while the material was still soft. After the setting of the acrylic, the palatal surface was assessed and adjusted using an air driven low-speed handpiece (Dentsply® Sirona, Ballaigues, Switzerland). The microsensor was embedded in the 1 mm thickness Essix plastic sheet of the VFR using a MiniSTAR S® pressure forming machine. Subsequently, the VFR sheet was trimmed around the microsensor leaving a border of 1 to 2 mm. A thin layer of plaster was added on the posterior buccal area of the construction cast, to flatten the surface for easy placement of the microsensor. A second sheet of material was pressure-formed on the cast and trimmed. The microsensor sheet was then bonded to the cast model sheet using auto-polymerising dental acrylic resin (Palapress®; Kulzer GmbH, Mitsui Chemicals, Hanau, Germany).
Following the retainer fabrication steps, the microsensors were activated by the software reader after which all retainers were immersed in a thermostatic water bath (Zhejiang Jinbo Electronic Co®, Ltd, China) filled with 8 litres of distilled water in a closed-box environment, at a controlled temperature of 35°C, which corresponds to the average intra-oral temperature.21 Prior to data collection, the accuracy of the water bath was examined for two consecutive days by measuring water temperature (35°C) to ensure consistent and stable performance. The microsensors were left for 1 week without any intermittent pauses and the experimental conditions were designated to replicate the oral environment and full-time wear of the orthodontic retainers.17 Water temperature and amount were checked twice daily. If volume decreased, water from a preheated water bath (35°C) was added to maintain constancy. The timestamp of the added water was recorded on a file sheet, to note any changes in temperature. The microsensor’s data were evaluated at the end of the week (Figure 2). The data were transferred from the microsensors to the TheraMon® client software (version 1.3.0.4, MC Technology GmbH, Hargelsberg, Austria) using a pen reader through an onboard antenna to a laptop running Windows 10 (64 bit). Data on the software was displayed as a graphical water bath time over a period of seven consecutive days. This totalled 168 hr representing full-time wear of the appliances for 1 week. Subsequently, data were displayed using the tracking graphical analyses, which were recorded by the microsensor every 15 min (Figure 3). The TheraMon® software has the capability to present a graph for short periods of time (days), which aligned with the 1-week trial. A generated analysis and temperature graphs presented precise information of the temperature changes during the experimental period.
Characteristics of included groups and flowchart of steps.
Detailed analysis of the temperature graph for one microsensor in the thick Hawley retainer group, with a temperature range between 34.54°C and 34.70°C.
Statistical analysis was conducted using the R statistical software (v4.1.2; R Core Team 2021). A simple descriptive analysis was conducted to calculate time and temperature discrepancies between actual time in the water bath and microsensor-recorded time were calculated, to find the mean values. An ANOVA and Tukey’s HSD test were applied to compare temperature means between the three groups and the results were deemed significant at
No technical errors occurred to any microsensor during the experimental period. The daily analysis revealed that all utilised microsensors were capable of providing a substantial performance function of 100% and able to record data for a continuous 1-week period. However, the detailed temperature graphs demonstrated small but significant differences in temperature recording between the groups.
The results revealed that the thin HR had a mean temperature reading of 34.81 + 0.04°C, followed by VFR at 34.77 + 0.09°C, and finally the thick HR with a mean of 34.73 + 0.05°C. The statistical analysis (ANOVA) yielded a
Temperature microsensors are commonly used to monitor body temperature changes. They consist of thermistor-based temperature gauge which records heat pressure at consistent time intervals which are then saved to memory.22 Fundamentally, the accuracy of a microsensor is determined by the closeness of the recorded readings to the actual value being monitored, taking into account all possible error sources.23 The accuracy of the current study is defined by the proximity of the microsensor’s readings to the true temperature of the water bath of 35°C. However, the precision of the readings is assessed by the distribution of readings per group, as measured by the standard deviation (SD). Therefore, the present aim was to assess the accuracy and precision of the TheraMon® microsensor in different thicknesses of HR (thin and thick) in comparison with a different material (VFR).
The overall results demonstrated that all microsensors were capable of detecting temperatures between 33.5°C and 38.5°C across the entire 1-week experimental period, thereby replicating the expected clinical conditions of retainer wear at an average intraoral temperature. This indicated excellent usefulness of the investigated microsensors despite the associated retainer material or thickness. Previous
The detailed temperature graphs demonstrated differences in the mean temperature recordings across the groups. The differences were generally small related to accuracy and precision. A statistically significant difference was shown between the two HR acrylic thicknesses (thick and thin). However, the lack of significant difference between the VFR and both HR groups may indicate less influence of the type of assessed material. The findings could be also attributable to the overlapping distributions between the two material types, as the VFR group had approximately twice the SD as the HR groups (Figure 4). This varying precision, as evidenced by a wider distribution of readings, was evident despite the uniform thickness of the VFR material surrounding all microsensors in that group. The results illustrate a potential need to control the thickness of acrylic coverage so that it is minimised.
Distribution of microsensor temperature readings within each group.
The detailed temperature results of each microsensor within the groups showed only small discrepancies, indicating a consistent utility for thermosensor use within different retainers. The water bath was an acceptable replica of the oral environment; however, continuous monitoring across the experiment period was mandatory to ensure temperature stability. However, recent advancements in microsensor-based technology have enabled the microsensors to be covered by a thin layer of polyurethane which has been shown to often be insufficient to protect the microsensors from the possibility of corrosion.14 Therefore, caution should be taken to ensure complete coverage of the microsensors during laboratory manufacture. The successful placement of the microsensors proved to be easier within the HR group, in comparison to the VFR group which is likely prone to breakage at the attachment area between the joined sheets.
Clinically, these differences could alter the detailed temperature detection, and subsequently affect the comprehensive recording of mean temperature. It would be advisable to control the acrylic thickness to 1 mm and attention given to full coverage of the microsensors to prevent corrosion from oral fluids. The laboratory manufacture and placement of the microsensor in the HR group was easier and less prone to failure in comparison to VFR. Nonetheless, patient preferences may require clear retainers, and so caution is advised in fabrication.
In recent years, modifications in microsensor technology have significantly improved accuracy. In turn, accurate monitoring of orthodontic patients in clinical settings is now possible to determine patient compliance with removable appliances over the short- and long-term. Sub-optimal adherence to retainer wear may be easily identified.19 Understanding the factors that affect the detection and performance abilities of microsensors would provide areas for further improvement. Such understanding should not only be confined to the laboratory steps but should also extend to the physiological and environmental factors that may influence the performance of thermal microsensors. Consequently, proper consideration of the thermal properties of retainer material could enhance accuracy and precision of detailed thermal detection in future applications of microsensors.
Although this
The accuracy of temperature detection of the TheraMon® microsensors was affected by the material thickness and composition of different removable retention appliances. However, the differences were small and did not appreciably impact on their clinical utility. For detailed temperature monitoring, it may be advisable to apply a thin layer of acrylic coverage in HRs within limits to control the thermal detection properties of the intra-oral device.