Despite the relevance of pre-surgical identification of AAA for SFE and other procedures involving the lateral wall of the maxillary sinus, such as Le Fort I osteotomies or Caldwell-Luc surgeries18, very few isolated studies have focused on the determinants of radiologic AAA detection. In this sense, and regardless of the extra-osseous position of the artery, the low detection rates showed by CBCT and CT (62.0%; 95%CI: 46.33–77.71)3,12 when compared to anatomical findings (100%)6,7,8,9,10 seem to be explained only by the small diameter of the vessel lumen, the technique (CBCT vs CT), and the skills and experience of the observers. Age and gender have also been related with radiological identification of AAA, although this relationship is based upon poorly consistent results11,15.
Two papers on AAA detection by CBCT reported important disparities in their chances to identify the vessel (52.8%12 to 90%18). As both samples report on patients with the same ethnicity, this wide variation seems to be related to factors linked to the CBCT device and to the observer12,18. Our results (51.1%) are close to the lower limit of this range in a sample including 57.5% females.
The usual course of AAA is mostly intraosseous within the lateral wall of the sinus19,20,21,22,23, followed in our study by the intrasinusal location (between the sinus membrane and the lateral wall). The anatomically described absence of a bony layer between the AAA and the Schneiderian membrane may influence the surgical handling of AAA, particularly during the procedure of membrane detachment8,21,24. A third possible situation of the vessel, just underneath the periosteum with a radiologically visible indentation in the lateral bony wall, was the least frequent one (3.3%) in our study, and less prevalent than reported in the literature20,21,22. The AAA can also describe a fully extra-osseous course2, but in these cases the artery lies within the flap, and the risk for haemorrhage would be associated to the incision, but not to the actual antrostomy21,25.
Information dealing with AAA diameter is usually reported as a categorical variable because the surgical relevance of damaging the vessel is size-dependent3. Hence, damage to a small AAA (< 1 mm) (19.5% in our series, 13.9–55.3% in the literature14,15 has a negligible surgical impact11 but damage to larger arteries may hamper visualization and hinder the surgical procedure13. The prevalence of AAA diameters between 1 and 2 mm is reported to range between 22.1 and 64.9%11,15. Vessels with diameters exceeding 2 mm, if damaged, are likely to result in bleeding important enough to interfere with the placement of the bony graft, which is a real surgical complication14. In our sample, 12.9% fell within this category (4.3 to 21.3% in other case series3. Larger vessels (> 3 mm) should be avoided or ligated to prevent severe iatrogenia21,25,26.
The relative position of AAA with reference to the alveolar ridge also influences osteotomy. There are wide variations reported in the literature (from 11.214 to 18.1 mm19) probably due to variations in the height of the residual crest. Again, our results rank between the reported values (15.2 mm for an average residual ridge height of about 7.7 mm). Thus, pre-surgical awareness of these variables is paramount to adequately design the osteotomy for SFE procedures.
The current study is the first report detailing explanatory models for AAA detection controlling for potential risk factors. The large size of the sample analysed—which is consistent with previous reports both in terms of age and gender distribution and artery diameter and position-, increases the external validity of our investigation. In addition, the fact that all participants in the study were being explored for dental implant treatments makes a potential selection bias unlikely. However, there are certain limitations for this kind of studies related to inter-observer variability (differences in visual observation and interpretation of images) and their experience. In this sense it has been suggested to include a higher number of observers27,28 and also that more experienced observers show higher detection rates29. In our study, two experienced surgeons scored a high concordance in AAA detection.
Another potential source of variability is the actual CBCT device28,30 and, therefore, extrapolation of our results to different technical equipment should be made with caution. In this vein, and when using CBCT for preoperative diagnosis, higher accuracy than the range of half a millimetre cannot be expected30, so clinicians should consider this fact when planning their surgeries to prevent overestimating the precision of CBCT examinations30. General surgical recommendations for SFE procedures include CBCT evaluation and a careful preparation of the bony window21 keeping a safety distance > 1 mm from AAA to avoid arterial damage.
The influence of age in AAA diameter is equivocal3,11. Although some groups have described a positive correlation11,31, other reports could not link age with the radiological detection or diameter of AAA15,19,22 as occurred in the current study.
On the other hand, gender resulted to be an explanatory variable in the multivariate model influencing artery detection. Larger diameters and higher detection rates were found amongst males15,18,19, which implies higher chances for intra-operative bleeding for this group15,16,17,18,19,20. The thickness of the sinus lateral bony wall also behaved as an explanatory variable, as larger AAAs were identified in thicker walls2,19. Thus, when facing a thick wall, the risk for bleeding should be anticipated3,27. However, the use of flexible models showed that this association is not linear, but chances for detection rapidly increase up to 6 mm thickness to decrease beyond this value. The width of the maxillary sinus also influences the surgical difficulty of SFE32 and has an effect on the probability for detecting AAA by CBCT. These chances significantly diminish when the distance between the lateral and medial walls of the sinus increase, particularly among women. Although our sample was mostly made of narrow sinuses (< 14 mm width), the AAA was nearly undetectable in wide sinuses (> 20 mm). The probability for detecting AAA is higher in patients with well-rounded ridges of an adequate height. The chances also increase with height for this particular ridge shape.
Bearing in mind the AAA detection rate by CBCT, particularly for vessels > 0.5 mm with potential to produce relevant bleeding if damaged with conventional rotary instruments, a careful surgical planning and the use of piezoelectric instruments are strongly recommended 26,33.