Introduction

A sleep disorder that is highly prevalent in the global population is obstructive sleep apnea (OSA), and it is characterized by the collapse of the upper airway (UA), which can be partial or total, during sleep due to recurrent episodes of apnea or hypopnea [1]. Respiratory effort is associated with recovery from regular breathing after the aforementioned respiratory episodes, leading to daytime hypersomnolence, sleep fragmentation, and decreased sleep quality [2].

There is a wide variation in the epidemiology of OSA, ranging from 14% to 84% in men and 5% to 61% in women. Nonetheless, it is believed that these numbers underestimate the true prevalence, with many cases going undiagnosed [2–4].

International guidelines recommend diagnosing OSA through sleep studies such as polysomnography (PSG) or Home Sleep Apnea testing (HSAT) [2,5]. When there is a 90% reduction in airflow for at least 10 seconds, it is defined as apnea. In contrast, when there is at least a 50% reduction in airflow and a 3% reduction in oxygen saturation for at least 10 seconds, it is considered hypopnea. This latter respiratory event in adults can also be classified as an oxygen desaturation of 4% or more from the pre-event baseline in some accredited sleep centers. The apnea-hypopnea index (AHI) determines the severity of OSA: it is considered that there is no OSA when the AHI is below 5, mild when it is between 5–15, moderate when it is between 15–30, and severe when it is above 30. Asymptomatic individuals are only diagnosed with OSA if their AHI is > 15 [2,6].

Regarding the severity of OSA, some authors have found that relying solely on the AHI is insufficient, as it fails to capture the heterogeneity of the disease. To predict the long-term impact factors of OSA and given the absence of a validated scale, the international consensus suggests considering the following criteria: AHI, the duration of oxyhemoglobin desaturation below 90% as an indicator of hypoxemia; daytime sleepiness; the degree of obesity measured by body mass index (BMI); and the presence of comorbidities linked with OSA, such as hypertension (especially if it is resistant to treatment or exhibits a non-dipper pattern), type 2 diabetes mellitus, dyslipidemia, coronary artery disease, stroke, heart failure, or atrial fibrillation [7].

OSA is frequently associated with comorbidities, such as cardiovascular and metabolic diseases, and is observed in half of patients. One prevalent example is persistent secondary arterial hypertension and non-dipping nocturnal hypertension profiles. The prevalence of OSA is high in patients with type 1 diabetes mellitus. Additionally, OSA is linked to cerebrovascular diseases, with a prevalence of 62% in moderate OSA and 30% in severe OSA [2,8,9].

Continuous positive airway pressure (CPAP) has been the primary treatment for OSA since its publication in 1981. It is recommended for symptomatic OSA, patients with concurrent health conditions, or those diagnosed with moderate-to-severe disease through PSG [1]. CPAP therapy improves the quality of life and reduces the risk of accidents, cardiovascular events, and mortality associated with OSA. However, long-term adherence rates range from 30% to 50% [2,10,11].

As not all patients tolerate CPAP well, other treatment options are needed. Mandibular advancement devices (MAD) are the initial non-invasive option for non-adherence to CPAP, and the initial treatment approach applies to mild and moderate OSA cases that do not involve additional health conditions and lack significant oxygen level impacts. MADs position the mandible forward to manage upper airway openness during sleep, effectively reducing AHI. Although MADs may not achieve the same level of efficacy as CPAP in managing airway obstruction, they exhibit superior clinical and scientific adherence [2,11]. Despite their documented lower effectiveness compared to CPAP, MADs provide an alternative for patients who cannot tolerate CPAP therapy. An increasing number of articles indicate that using MADs for treating OSA is effective in improving PSG indices and objective and subjective measures of sleepiness, blood pressure, neuropsychological functioning, and quality of life. Research suggests that MADs reduce AHI by ≥50% in approximately 60–70% of patients, with AHI dropping below five events per hour in approximately 35–40% of cases [12].

These two complementary diagnostic tests, computed tomography (CT) and Drug-Induced Sleep Endoscopy (DISE), are not frequently used in OSA diagnosis and can be associated with significant expenses [13].

MADs can be categorized as either standardized (without the need for dental arch impressions) or customized (requiring impressions of the dental arches). Non-titratable MADs maintain the mandible in a fixed protruding position throughout the treatment without the ability to adjust it. In contrast, customized MADs are titratable, featuring mechanisms that allow incremental adjustments in mandibular protrusion based on the patient’s response to therapy. The increase in mandibular protrusion in titratable MADs is akin to the titration process used in CPAP therapy [14].

Factors related to anatomy and neuromuscular physiology are pivotal in the pathophysiology of OSA. Imaging diagnostic modalities can offer valuable insights into OSA evaluation and identify patients who are likely to benefit from MAD [2,15]. Among these methods, CT provides detailed images of the upper airway bone and soft tissues from the nasopharynx to the larynx. Axial plane images with high resolution (1–2 mm thickness) enable accurate measurements of the length and cross-sectional dimensions [2,16].

Although CT provides a fixed, planar depiction of complex, dynamic anatomical structures of the head and neck, it is considered a valuable predictive tool. Substantial variation in these measurements has been observed between asymptomatic individuals and those diagnosed with OSA. Moreover, cephalometry has been utilized to evaluate how changes in body position affect upper airway anatomy and function in individuals with OSA. Morphological predictors, such as retrognathia, micrognathia, long face, inferior hyoid bone positioning, accentuated mandibular plane, upper airway, soft palate, and tongue narrowing, have been identified. CT-based cephalometric analysis aids in the diagnosis and preoperative assessment of surgical OSA therapy [2,17].

Various surgical procedures and positional therapies have been introduced to enhance therapy adherence and efficacy when used alone or in combination [18]. Personalized treatment is crucial to achieve optimal long-term outcomes. To this end, tools such as DISE have been presented. Since its introduction in 1991, DISE has enabled the real-time evaluation of areas prone to vibration and collapse using flexible nasopharyngoscopy under sedation [2,19].

Research indicates that approximately 50% of surgical strategies are adjusted after DISE, in contrast to evaluations conducted while the patients are awake. Certain DISE observations correlate with the therapeutic approach outcomes [2,18]. However, DISE has indications and contraindications. Absolute contraindications included ASA 4 status, pregnancy, and allergies to DISE sedatives. Severe obesity is generally considered a relative contraindication, although specific UA features may warrant DISE consideration, even in these cases [2,18].

To enhance the accuracy of predicting MAD therapeutic success, adjustable intraoral devices and techniques such as chin lift and jaw thrust (Esmarch) can be used during DISE. These interventions should be carefully assessed to minimize patient discomfort and the associated micro-arousals [2,19].

This systematic review aimed to determine the most effective method for predicting the efficacy of MADs in OSA therapy. To date, no systematic review has directly compared the predictive values of CT and DISE for MAD therapy outcomes in patients with OSA. By addressing this gap, our study provides original insights into the diagnostic workup and individualized treatment planning for non-CPAP therapies. Given the imperative to individualize treatment and ensure safety in MAD prescriptions, this topic is of paramount importance.

Materials and methods

Results and discussion

Literature search

A total of one thousand eight hundred and nine references (n = 1809) were obtained through the extensive search of the 3 databases mentioned above (Embase = 952; PubMed = 508; Web of Science = 349). A total of 546 references were deleted at the start because they were duplicated. Of the 1263 references that remained after the removal of duplicates, 1179 were excluded after reviewing the titles and abstracts because they were not appropriate for this systematic review. The reasons remained: sampling with children or adolescents, Sampling with animals, therapeutic approaches that were not MAD, case reports and reviews, and patients without OSA. The full texts of the remaining 84 references were retrieved. Based on the inclusion criteria, mainly because of the lack of direct comparison between CT and DISE for the prediction of MAD success, 2 articles were selected for this systematic review (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flowchart.

https://doi.org/10.1371/journal.pone.0327974.g001

Craniofacial features, cephalometric evaluations, and location and type of UA obstruction

DISE.

Two articles used DISE to assess the UA and its collapse by introducing flexible nasofibroscopy through the nose [15–21]. Saleh et al used a propofol dosage of 1 mg/kg, followed by 20 mg doses every 2 minutes until the onset of the snoring-apnea cycle. The target-controlled infusion of propofol used was 1.5 μg/mL [21]. Van den Bossche et al. did not solely use propofol; instead, it began with endovenous administration of midazolam (1.5 mg), followed by propofol administered via target-controlled infusion (2.0–3.0 μg/ml) follow [15].

Both categorized the collapses of the UA in different ways: the location of the obstruction (soft palate, tonsils, tonsil base, and epiglottis); the mechanism of the obstruction (anteriorposterior, concentric, and posterior-lateral); the intensity of the obstruction (vibration, collapse, and collapse and vibration); and the duration of obstruction in the respiratory cycle (inhalation and exhalation) are classified [21]. According to Van den Bossche et al., the location of obstruction (palate, oropharynx, tongue base, pharyngeal lateral walls, and epiglottis), degree of collapse (complete, partial, or none), and direction of collapse (anteroposterior, laterolateral, or concentric) were classified [15].

The patients were divided into three groups: “complete response” (AHI < 5/h), “partial response,” and “no response” when the MAD was used to treat OSA, in Saleh et al. article. They found that the complete responders did not exhibit epiglottis collapse in the DISE (n = 6), and that they differed significantly from the partial responders (p = 0.029) [21].

In Van den Bossche et al. article, the patients were divided into four groups: “response,” “no response,” “deterioration,” and “no deterioration” when the MAD was used to treat OSA. The term ‘responses’ refers to a 50% decrease in the AHI, while ‘deterioration’ is primarily defined as an AHI increase of more than 10% from the baseline measurement. During the assessment, a notable contrast emerged between the individuals who showed a positive response and those who did not. Specifically, there was a marked variation in the occurrence of tongue base collapse, with a higher prevalence observed among those who responded positively than among those who did not (statistical significance, p = 0.0404). Additionally, in terms of palatal collapse, the disparity between responders and non-responders was evident, with a greater incidence noted among non-responders (statistical significance, p-value = 0.0240). At the initial assessment, notable differences were observed between patients who experienced deterioration and those who did not. Specifically, there were significant variances in baseline characteristics related to certain DISE parameters. These differences were particularly evident in the complete concentric palatal collapse (CCCp) measurement (statistical significance, p-value = 0.0494) and the occurrence of complete laterolateral oropharyngeal collapse (statistical significance, p-value = 0.0364) [15] (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. PICO—Computed tomography versus sleep endoscopy to predict the effectiveness of mandibular advancement devices in adult patients with obstructive sleep apnea.

https://doi.org/10.1371/journal.pone.0327974.t001

According to the odds ratios (OR), the two largest effects observed were the collapse of the responders’ tongue base (OR 3.29, 95% CI 1.02–10.64; p = 0.0464) and CCCp degradation (OR 28.88, 95% CI 1.18–704.35) [15]. (Table 1). Parte inferior do formulário

Computed tomography.

Regarding computed tomography, each of the two articles examined the potential collapse of the UA differently [15–21]. According to Saleh et al., the patients underwent Cone-Beam Computerized Tomography (CBCT) and were categorized as “more than 110 mm²” is the normal airway cross-sectional area; “52–110 mm²” is the mild to moderate range, and “less than 52 mm²” is the severe range [21]. In contrast, Van den Bossche et al. used computational fluid dynamics (CFD) to perform awake baseline low-radiation-dose CT scanning. From these data, the effective UA volume, excluding mouth leakage, was calculated. Measurements were taken for the velopharynx, oropharynx, hypopharynx, and distinct regions of the pharynx, as well as for the overall pharyngeal volume. Additional anatomical parameters, such as the minimal cross-sectional area and UA resistance were also assessed [15].

Patients were examined using CBCT both before and after receiving MAD. They showed a significant improvement in the following airway parameters in the patients after receiving MAD treatment: total volume of air (P < 0.001) and minimum and maximum cross-sectional areas (p = 0.047 and 0.038, respectively) [21] (Table 1).

In contrast, Van den Bossche et al. found no significant differences in UA CT parameters between patients who responded to the MAD and those who did not. Additionally, they suggested that there were no significant differences in the UA parameters on CT between patients who were deteriorating and those who were not [15] (Table 1).

DISE versus computed tomography.

The authors of the two articles included in this systematic review provided differing conclusions regarding the differences between CT and DISE [15–21]. According to Saleh et al., both DISE and CT can be valuable tools for predicting and analyzing the effectiveness of MAD treatment for OSA [21]. In contrast, Van den Bossche et al. stated that DISE is a more thorough examination for analyzing the likelihood of MAD success in the treatment of OSA [15] (Table 1).

Heart rate and BIS.

None of the articles addressed any of these variables during DISE [15–21].

Mean values of AHI.

The primary outcome variable was AHI in 2 studies [15–21]. The average AHI values were used to summarize the outcomes in the study by Saleh et al. article [21]. However, Van den Bossche et al. reported median AHI values [15].

Both studies showed a significant decline in AHI when the MAD was used [15–21], and one study found that the mean AHI was 31.4 ± 15.6/h before the MAD and 13.8 ± 15.2/h (P = 0.001) after MAD [21]. Conversely, in the study by Van den Bossche et al., the AHI’s median (interquartile range [IQR]) improved from 15.6 (10.4–23.5) to 9.0 (4.3–16.0) events/h (p < 0.0001) [15] (Table 1).

Average SaO₂.

Concerning this result, only Van den Bossche et al. have references on the Average SaO2. This remained unchanged both before and after the use of MAD, as seen in the following example: initial = 95.3% (94.1%–96.1%) and final = 95.3% (94.2%–96.0%) [15].

Time Under 90% SaO₂ (T90).

None of the studies included in our systematic review used the Time Under 90% SaO2 (T90) variable [15–21].

Enhancement in ESS scores and/or in PSQI scores.

This secondary outcome, measured using a questionnaire, was found in both studies [15–21]. However, only ESS, which analyzes excessive daytime sleepiness, was used in both cases [15–21]. In the study by Saleh et al., the initial median score of the ESS was 11.4 ± 5.3; however, the value after applying the MAD was not mentioned. Mention just that The ESS score significantly improved (P = 0.004) for the participants in the study following their adaptation to MAD [21]. In Van den Bossche et al., study, the use of the MAD resulted in a fall in the ESS from a median (IQR) score of 9/24 (5–12) to 6/24 (3–10) (p < 0.0001) [15].

After using the MAD, patients’ sleep quality improved according to Saleh et al., who used the PSQI. Specifically, prior to receiving MAD, the patient’s mean PSQI value was 17.5 ± 3.1; however, following therapy, it dropped to 6.5 ± 3.6 (p = 0.001) [21].

Literature quality (bias) assessment (QUADAS-2)

Risk of bias.

Fig 2 presents the assessment outcomes of study quality for both the included studies. We considered Van den Bossche et al. to be at low risk across all four risks of bias domains: patient selection, index test, reference standard, and flow and timing [15]. Two studies were at low risk of bias for three domains (patient selection, index test, and reference standard), but at unclear risk of flow and timing [21].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Results of QUADAS-2 assessment.

https://doi.org/10.1371/journal.pone.0327974.g002

The included studies enrolled consecutive patients [15–21]. Finally, it is unclear in Saleh et al. ’s study whether there was a suitable time gap between the index test and the reference standard [21].

Applicability concerns.

There were no applicability concerns for the two studies across the three applicability domains of patient selection, index test, and reference standard [15–21].

Discussion

Our study synthesizes current research findings on the diagnostic efficacy of neck CT and DISE in predicting the therapeutic success of MAD for managing OSA. This review involved a rigorous and transparent systematic literature search. Of the 1809 references searched in Embase, PubMed, and Web of Science, only two analyzed and compared CT and DISE in predicting MAD success for OSA. This demonstrates that this is an underexplored topic and that the study that our research group will produce shortly, taking into account all of the knowledge absorbed through the systematic review, will be relevant and lead to scientific advancements.

Consensus is lacking regarding the standards for the success of MAD therapy; nonetheless, recent scientific evidence indicates that factors such as low body mass index (BMI), supine-dominant OSA, neck circumference less than 40 cm, and sex may impact treatment outcomes with MAD [19].

The effectiveness of MAD in patients with OSA may differ based on the severity of the condition. Variables such as enlarged palatine tonsils or prominent pharyngeal pillars could contribute to less favorable outcomes, potentially resulting from partial collapse or compression of the base of the tongue. Given these considerations, targeting interventions that focus on the anterior positioning of the soft palate or base of the tongue before treatment could enhance the efficacy of MAD therapy [22]. Regardless of the location of UA obstruction, it is crucial to remember that multiple-level obstruction is the most common level of obstruction, whether assessed by CT or DISE [21].

Recent advancements in CT technology have enabled volumetric measurements using specialized software such as Sicat Air ® [17], generating three-dimensional images of craniofacial structures, mandible, hyoid bone, spine, and airways. Another aspect explored for its potential to predict MAD success through CT evaluation included assessing the angle of the skull base and measuring the distance between the sella turcica and the deepest point in the posterior cranial fossa [22]. MAD can also be predicted by factors such as total air volume and minimum and maximum cross-sectional area [21]. However, Bossche et al. (2022) stated that isolated CT analysis is not a reliable predictor of MAD or deterioration [15].

While CT scans are beneficial in specific scenarios, their routine use is not recommended because of associated expenses, restricted availability, and comparatively high radiation exposure [2,12,23].

An additional parameter integrated into DISE evaluates and strategizes the optimal therapeutic intervention for OSA [2,24–25]. This parameter involves personalized bite registration conducted by a sleep dentist before DISE to determine the patient’s Maximum Comfortable Protrusive Position (MCP). Consequently, the predictive value of DISE, combined with MCP bite registration, was assessed for MAD treatment outcome [26]. It is suggested that patients showing confident responses to MAD often exhibit a substantial increase in complete UA volume, indicating that MAD efficacy correlates with UA volume augmentation. Conversely, the absence of velopharyngeal volume increment appears to be associated with treatment deterioration [2,15,27].

Additionally, evidence suggests a link between the response to MAD treatment and the overall volume of the UA, especially focusing on the velopharyngeal region [14,28–29]. Another anatomical predictor that was evaluated was epiglottic collapse. When there is no collapse, the MAD have a higher chance of success [21].

Tailoring diagnostic and treatment methods for OSA is crucial for enhancing patient outcomes. Adopting personalized medicine for OSA represents a novel strategy, as traditional methods still need to adequately address the critical diagnostic and therapeutic aspects of the condition. A personalized therapeutic approach aims to meet each patient’s specific needs, thereby ensuring appropriate and optimal treatment.

One of the key clinical takeaways of this review is the importance of aligning diagnostic tools with patient-specific characteristics to optimize MAD therapy outcomes. DISE stands out for its ability to simulate sleep-related airway behavior, providing functional information that can directly inform treatment decisions. Identifying unfavorable collapse patterns—particularly tongue base collapse and complete concentric palatal collapse (CCCp)—can help clinicians avoid prescribing MADs to patients unlikely to respond [15,21,26]. This selective approach not only increases treatment efficiency, but also minimizes unnecessary costs and delays in care.

In contrast, CT offers a static but valuable anatomical overview that may be especially useful in treatment planning when craniofacial morphology plays a central role or when DISE is unavailable. Its ability to quantify airway volume and assess skeletal structures can support clinical decisions in a complementary way, though it lacks the dynamic insight of DISE [13,17,23].

From a practical standpoint, these modalities should not be considered as mutually exclusive. Rather, combining anatomical and functional assessments may represent the most effective path toward personalized therapy in OSA, particularly in cases where CPAP intolerance necessitates a tailored, non-invasive alternative, such as MAD.

Although there have been considerable improvements in identifying and treating this condition in recent years, implementing a tailored and holistic approach promises to enhance the management of OSA patients [2,30].

Both Neck CT and DISE allow three-dimensional evaluation of the UA. While neck CT is performed with the patient conscious and DISE requires sleep induced by medication, there is some convergence between the two methods [31]. CT imaging provides insight into obstructions related to the lateral walls of the oropharynx, as demonstrated by Zhang et al. in 2014. This suggests a potential partial replacement for DISE, given its requirements for a trained team, specialized facilities, time, expense, and the need for pharmacological sleep induction [2,30].

Although relatively few studies have directly compared DISE with Neck CT, both examinations are valuable for evaluating patients with OSA, each providing unique information [30]. As discussed earlier, DISE offers dynamic insights, whereas Neck CT provides more static information [2,18].

Studies comparing CT and DISE yield different results based on our knowledge and research [15,21]. According to one study, there are no significant differences or benefits of using DISE instead of CT for predicting MAD success [21]. In addition, Bossche et al. (2022), DISE is the only test that can predict the success of MAD for treating OSA [15].

However, the conformation of the UA, as assessed using DISE and/or Neck CT, fails to fully capture the intricate pathophysiology of OSA. This condition is multifaceted and cannot exclusively be attributed to anatomical factors. Variables such as arousal threshold, respiratory control stability, and genioglossus muscle responsiveness play crucial roles in OSA and must be considered as they can significantly differ between sleeping and awake states [2,32]. Some articles suggest that a potentially less invasive method for predicting treatment success and identifying patients responsive to MADs is through standard or temporary MADs. These possibilities are often more cost-effective and frequently demonstrate efficacy in managing OSA [2,33].

It is also important to note that while we can predict who will respond best to a MAD, it is equally crucial to understand adherence. However, success cannot be achieved without good treatment adherence. Therefore, several factors beyond anatomical considerations are important for analysis. Factors that may positively impact adherence to MAD include good communication skills of the practitioner, the use of a custom-made MAD, effective therapy, early detection of adverse effects, gradual adjustment of the MAD dosage, and favorable initial experience with the MAD. Conversely, elements that might deter commitment to MAD therapy include experiencing side effects such as discomfort or dental pain, opting for a thermoplastic MAD instead of a custom-made device, personality traits such as type D personality, MAD efficacy, dental procedures during MAD therapy, and an unsatisfactory initial experience due to inadequate guidance from healthcare providers. The literature commonly cites influences on adherence as side effects during MAD therapy, treatment efficacy, and the type of MAD employed [34].

Some limitations identified in this review include conducting a qualitative systematic review because only two studies met the inclusion criteria, heterogeneity in the methods of study evaluation, and lack of randomized clinical trials on the subject. These constraints can affect the synthesis, application, and generalizability of the reviewed information. Another potential challenge is the information bias stemming from our focus on studies published exclusively in English and involving adult populations. Another limitation is the diagnostic metrics used in the included studies. Both studies [15,21] adopted the Apnea–Hypopnea Index (AHI) as the primary outcome measure derived from full-night PSG, without reporting the Respiratory Disturbance Index (RDI). Respiratory effort-related arousals (RERAs) were excluded from the analysis. This may underestimate the total burden of sleep-disordered breathing events, particularly in patients with significant upper airway resistance without apneas or hypopneas. Future studies incorporating RDI could offer a more comprehensive assessment of the effectiveness of MAD therapies. Additionally, it should be noted that only one of the included studies [15] specified the version of the AASM scoring guidelines used (2007), whereas the other [21] did not. This lack of standardization in scoring criteria may introduce minor discrepancies in the definition of the apnea–hypopnea index (AHI), which could affect the comparability of results across studies. Nevertheless, our study is timely, particularly considering the increasing use of MADs. This study aimed to identify areas for improvement and advantages in using Neck CT and DISE to predict therapy outcomes in OSA, highlighting the importance of employing diverse tools for personalized and evidence-based therapeutic approaches.

The findings of this systematic review contribute to the broader field of sleep medicine by reinforcing the need for individualized diagnostic approaches in the management of OSA [15,21,30]. Given the phenotypic heterogeneity of OSA, no single diagnostic modality can reliably predict therapeutic success in mandibular MADs [2,15]. CT offers valuable anatomical insights into upper airway morphology [21], whereas DISE provides a dynamic and functional assessment that more closely reflects sleep-related airway behavior [15,26]. Within the broader framework of sleep medicine, these tools should be viewed as complementary elements of a multimodal evaluation strategy in line with the principles of precision medicine [26,30]. Their integration into clinical practice may facilitate the identification of patients most likely to benefit from non-CPAP therapies, such as MADs, particularly considering the well-documented challenges with long-term CPAP adherence [30–34]. Therefore, our findings underscore the importance of combining advanced diagnostic techniques with clinical, physiological, and patient-centered factors to optimize treatment outcomes in OSA [15,21,27].

Conclusions

Our systematic review compared the predictive value of CT and DISE in evaluating the efficacy of MAD in adult individuals with OSA. Despite the limited number of studies included in our review, our findings provide valuable insights into the diagnostic utility of these modalities.

Our systematic review identified and analyzed two studies. The first study reported no significant differences in the ability of CT and DISE to predict the effectiveness of MAD therapy in OSA patients. Conversely, the second study suggested that only DISE was capable of accurately predicting MAD efficacy, indicating the potential superiority of DISE over CT.

The conflicting findings between the two studies underscore the complexity of OSA management and the challenges associated with predicting treatment outcomes using different diagnostic modalities. While CT offers detailed anatomical information regarding airway morphology, DISE provides dynamic visualization of upper airway collapse during sleep, which may better simulate real-life conditions and the MAD response.

Supporting information

Acknowledgments

We would like to express our sincere gratitude to the Egas Moniz Center for Interdisciplinary Research (CiiEM) and Egas Moniz School of Health & Science in Almada, Portugal, and CUF Tejo Hospital in Lisboa, Portugal for their support and contributions to this study.

We extend special thanks to the study coordinators at CUF Tejo Hospital for their invaluable assistance and dedication throughout this research project. Their efforts were instrumental in completing the systematic review.

Finally, we express our heartfelt thanks to our families and friends for their unwavering support and encouragement throughout this research.