Paul Reaney discusses how artificial intelligence, alongside human input, is helping to treat obstructive sleep apnea (OSA).
Today in dentistry, there are so many technological advances that make diagnosis and treatment easier and better.
Patients do well and providers are less stressed when these tools are utilised. Digital radiography, CBCT imaging, intraoral scanning, and dental software are but a few examples.
Recently, dental sleep medicine (DSM) providers have been advantaged by artificial intelligence (AI) applications, robotic manufacturing, and complex algorithms that also make diagnosis and treatment of sleep related breathing disorders (SRBD) like obstructive sleep apnea (OSA) more predictable.
However, just as the clinician’s differential analysis and procedural/material decisions have always guided good treatment recommendations, so too is the case for DSM providers.
AI, enhanced by human intelligence (HI), is a powerful force in good clinical treatment and patient satisfaction. Great maths often needs the human touch.
A patient presented in clinic with the common risk factor for OSA. The STOP-BANG (Chung et al, 2008) questionnaire is one of the most trusted screening tools for sleep apnea (snoring, tired, observed stop breathing, high blood pressure, BMI, age 50+, neck 15 female/17 male or larger, gender male).
Positive answers to three of the eight factors indicate moderate risk. And therefore four or more moves the level to high risk.
The patient was scheduled for a home sleep test (HST) using the Night Owl (as per NICE guidelines).
Results indicated an apnea hypopnea index (AHI) of 17 events per hour. That means there were 17 reductions in breathing flow of at least 10%, that lasted at least 10 seconds and resulted in a decrease in the blood oxygen saturation of 3%!
Normal sleep will have 0 to five events per hour. Mild sleep apnea is from five to 15. Moderate 15-30 and severe is more than 30. We have seen patients with over 60 events per hour.
That means the sympathetic nervous system wakes the patient up enough to take a fuller breath, once every minute! Typically, this unbalance of the para and sympathetic systems during sleep causes sudden heart rate surges, cardiovascular challenges, and endocrine stress.
Readers should appreciate a fuller understanding of the pathogenicity and comorbidities of OSA. The graph below shows how many other conditions are worse in this subset of patients.
Precision HST with technology like the Night Owl, are excellent ways to assess many of the patient’s critical diagnostic and treatment confirmation metrics.
This particular test allows for up to 12 nights over a two-year period. Many clinicians prefer a two-night diagnostic test because up to 33% of mild and moderate patients can be misdiagnosed with a single night recording. Multi-night testing reduces failure rates and false negatives to 3%.
Today, state-of-the-art in precision inputs for DSM require scanning the upper and lower arches. Along with a digital protrusive bite record at the prescribed starting position.
Casretream Dental has long been involved in certifying digital treatment protocols and pathways. When the CS 3500 was introduced, they partnered with Prosomnus Sleep Technologies to establish the first qualified all digital pathway for inputting and manufacturing oral appliances.
The newer CS 3700 and CS 3800 scanners, continue to provide the precision representations of the mandible, maxilla and inter-arch relationship that sets you up for success.
The input requirements from manufacturers like Prosomnus have raised the bar in providing precision medical solutions that reduce side effect risk, lower dose, and improve efficacy and patient comfort/adherence.
Artisanal, handmade MADs using cold and hot pot/pressure processed PMMA in combination with advancement parts and pieces borrowed from orthodontic applications (Herbst arms, jackscrews, and straps) have long been the standard until the advent of CAD/CAM. And now artificial intelligent design and robotic manufacturing.
Stronger milled PMMA and MG6 technology (EVO material) have allowed devices to become smaller. These include more tongues space, improved comfort and better compliance.
The lower dose required by these advanced materials also result in continued improvement of efficacy and fewer side effects. Tooth movement is nonexistent, devices are easier to clean and patients prefer them over predicates and PAP solutions.
AI allows Prosomnus to replicate patient’s actual surface anatomy. They deliver the device within 1mm of the prescribed starting position.
Iterative advancement is easier for patients, maintains the mandibular position even with a relaxed mandible and is bilaterally symmetrical.
These levels of precision were not possible until recently. The technology, material composition and manufacturing advanced to meet the needs of DSM.
Even with all the technology, AI and robotic manufacturing, there is still room for and a need for ‘the human touch’.
Software struggles to interpret our feelings and concerns. The patient shown here was concerned with how the device would look at night to their spouse.
Embarrassed at not having maxillary incisors, the clinician and Prosomnus collaborated on manufacturing a device with receptacles for composite to simulate the look that the upper partial denture provided.
This humanisation was a key factor in gaining case acceptance from both the patient and indirectly his spouse. The software was able to record the tooth anatomy from the over-scan of the partial denture and add them to the edentulous space.
In a perfect world, patients would improve their AHI to below five. As well as have improved QOL scores, no side effects and excellent adherence.
Reducing the AHI from 20 to three events per hour for this patient, is a credible success indeed. Add to that, the patient reported sleeping better, was less tired, no longer snored and had more energy.
The smaller appliance and lower dose also resulted in excellent compliance. The patient reports seven hours per night wear, seven nights per week. No reported side effects complete the matrix for successful outcomes.
Chung F, Subramanyam R, Liao P, Sasaki E, Shapiro C and Sun Y (2008) High STOP-Bang score indicates a high probability of obstructive sleep apnoea. Br J Anaesth 108 (5): 768-75