Low-Dose Dental Scans May Match Standard Imaging for Wisdom Tooth Surgery Planning

Every year, millions of people around the world have their mandibular third molars — the wisdom teeth — removed, and for many of them surgeons order a cone-beam computed tomography scan before the operation. The reason is simple: the lower wisdom teeth often sit perilously close to the mandibular canal, the bony tunnel through the jaw that carries the inferior alveolar nerve. If the nerve is bruised or severed during extraction, the patient can be left with lasting numbness of the lower lip and chin. A new study from the University of Hong Kong and the Chinese University of Hong Kong, published in BMC Medical Imaging, now suggests that clinicians may be able to slash the radiation dose of these pre-surgical scans by as much as 93 percent without losing the information that matters most for surgical planning.

The research team, led by Yongxin Zhang, Zekai Liu and senior authors Kuo Feng Hung and Qi Yong Hemis Ai, set out to answer a question that has quietly troubled oral and maxillofacial radiology for years. Standard-dose CBCT protocols deliver crisp, high-resolution images of the jaw, but they also expose patients to radiation that — while modest compared with medical CT — is not negligible, particularly for young patients who make up a large share of wisdom tooth extraction candidates. Radiation protection principles such as ALARA, as low as reasonably achievable, and its newer iteration ALADAIP, as low as diagnostically acceptable, indication-oriented and patient-specific, have long pushed clinicians toward the lowest dose that still answers the clinical question. What has been missing is hard evidence about how far the dose can fall before the anatomy that surgeons care about fades from view.

To find that limit, the investigators took an unusually rigorous approach. Rather than scanning patients repeatedly and multiplying their exposure, they worked with 20 dried human mandibles that still contained third molars in place. Each specimen was scanned on the same CBCT machine five times using different protocols: one standard-dose, normal-resolution protocol with a voxel size of 0.2 millimeters and a dose-area product of 322 mGy×cm², and four low-dose protocols with voxel sizes ranging from 0.15 to 0.4 millimeters and dose-area products falling to between 131 and a mere 23 mGy×cm². This ex vivo design, approved by the Institutional Review Board of the University of Hong Kong and Hospital Authority Hong Kong West Cluster, allowed the team to image identical anatomy repeatedly under different settings, isolating the effect of dose and resolution from biological variability between patients.

The central anatomical structure under scrutiny was the mandibular canal, a radiolucent channel winding through the body and ramus of the lower jaw. The team rated the canal’s visibility at three separate mesial-distal levels relative to the third molar — the distal, middle and mesial thirds — using a four-point scale, with higher scores indicating clearer delineation of the canal walls. For every scan they also assessed the spatial relationship between the canal and the apex, or root tip, of the third molar in two complementary ways. The qualitative assessment classified the relationship as non-contact, contact, or penetration, meaning the tooth root actually invading the canal. The quantitative assessment measured the actual three-dimensional distance between the canal and the root apex, the number surgeons use most directly when weighing the risk of nerve injury against the difficulty of the extraction.

The results delivered a message that will resonate with clinicians trying to balance image quality against patient dose. At every anatomical level examined, canal visibility scores remained high — 3.4 or above on the four-point scale — for all protocols except one: the combination of the lowest dose and the coarsest resolution, using a 0.4 millimeter voxel and a dose-area product of just 23 mGy×cm². That ultra-low protocol produced significantly poorer canal visibility compared with the standard-dose reference, with p values ranging from 0.006 to 0.026 across the three levels. Every other low-dose configuration, including protocols that cut the dose-area product by well over half, held its statistical ground against the standard scan.

Even more striking were the findings on the spatial relationship between canal and tooth. When the researchers compared the qualitative classifications — non-contact, contact, or penetration — and the measured canal-to-apex distances between the standard-dose protocol and each of the four low-dose protocols, they found no statistically significant differences for any pairing, as judged by the Wilcoxon signed-rank test for paired comparisons. In practical terms, a surgeon looking at a low-dose scan from most of the tested protocols would reach the same conclusions about nerve proximity as one looking at the full-dose image. The reliability of these readings was underpinned by the study’s statistical design, which used intraclass correlation coefficients to gauge agreement between assessors rating the same images.

The technical explanation for this resilience lies in the geometry and contrast of the anatomy itself. The mandibular canal is a relatively large structure defined by a radiolucent lumen bounded by corticated, radio-opaque walls, and distinguishing contact from penetration of a root apex does not demand the finest voxel size so long as the cortical boundary remains detectable. Voxel size and dose interact in CBCT: lowering milliampere settings reduces dose but increases image noise, while enlarging voxels smooths away fine detail. The study’s findings suggest that for the specific task of localizing the canal relative to a third molar apex, moderate reductions in both parameters stay within the diagnostic margin — but pushing both dose and voxel size to their extremes, as in the 0.4 millimeter, 23 mGy×cm² protocol, erodes the canal boundary beyond the point where confident reading is possible.

The authors, however, are careful not to overclaim, and their caution is worth emphasizing. They note explicitly that the absence of a statistically significant difference should not be interpreted as proof of equivalence between the protocols. Their assessments were subjective image ratings without an independent reference standard — there was no gold-standard micro-CT or histological ground truth against which each reading could be validated. Statistical non-significance in a limited sample of 20 mandibles is a different thing from demonstrated diagnostic equivalence, and the dry-mandible model, while elegant, lacks the soft tissue, scatter and patient motion that complicate real clinical imaging. These caveats place the findings in the category of promising evidence for dose reduction rather than a license to abandon standard protocols wholesale.

What the study does support is a shift in how CBCT protocols are chosen for wisdom tooth imaging. The authors conclude that low-dose CBCT may reasonably be considered in selected cases, but that protocol selection should remain indication-oriented, patient-specific and device-specific. A young patient with a panoramic radiograph already suggesting the canal runs far from the tooth roots may be an ideal candidate for a reduced-dose scan, while a case where the roots appear to hug or invade the canal — or where the anatomy is unusually dense or the patient cannot remain still — may still justify the full standard protocol. The findings also underscore that the coarsest, cheapest settings have a real cost in image quality, and that the middle ground of low-dose, moderate-resolution protocols represents the practical sweet spot.

Funded by the Oral Health Research and Innovation Fund of the University of Hong Kong’s Faculty of Dentistry, and published open access under a Creative Commons Attribution 4.0 license, the study arrives as dental radiology faces growing scrutiny over cumulative population radiation exposure. The dose-area product comparisons at the heart of the paper — 322 mGy×cm² for the standard protocol against as little as 23 mGy×cm² for the lowest setting — translate into a concrete arithmetic that clinicians can apply at the console. If further clinical studies with patient populations and reference standards confirm what this ex vivo work suggests, the routine pre-extraction CBCT could join the broader movement in medical imaging toward smarter, leaner scanning: not more pictures, but precisely enough picture to keep the inferior alveolar nerve safe and the surgeon informed.

Subject of Research: Comparing standard-dose and low-dose CBCT protocols for mandibular canal visibility and third molar surgical risk assessment

Article Title: Standard- versus low-dose CBCT in evaluating mandibular canal visibility and its impact on assessing the spatial relationship with mandibular third molars: a comparative analysis

Article References: Zhang, Y., Liu, Z., Shan, Z., Yang, Y., Ai, Q. Y. H., & Hung, K. F. (2026). Standard- versus low-dose CBCT in evaluating mandibular canal visibility and its impact on assessing the spatial relationship with mandibular third molars: a comparative analysis. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02730-4

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02730-4

Keywords: cone-beam computed tomography, radiation dose reduction, mandibular canal, mandibular third molar, inferior alveolar nerve, dental imaging, oral and maxillofacial radiology, ALADAIP, voxel size, wisdom tooth extraction, image quality, Hong Kong