Lack of testing and adherence to optimized low-dose CBCT protocols for children

Abstract

It takes a lot of effort and resources to conduct experiments to compare and determine the best low-dose CBCT protocol(s) for children with an optimized balance between the diagnostic value (image quality) and radiation safety (radiation dose) according to the ALADAIP (As Low as Diagnostically Acceptable being Indication-oriented and Patient-specific) principle.^{1, 2} It would be highly beneficial if the academic and clinical communities actively examined and applied the original studies that established these low-dose CBCT protocols. On February 22, 2024, we searched for the literature through the Web of Science Core Collection with the following search string: Topic—(pediatric OR paediatric OR child*) AND CBCT AND (low-dose OR “low dose” OR optimi*). The search string identified papers that mentioned these words and their derivatives in the title, abstract, and keywords. The search yielded 77 original articles. Studies were included if they tested multiple scanning protocols, recruited human subjects or used phantom heads, and evaluated the image quality with either subjective or objective measurements. Studies were excluded if they were not written in English. After screening their titles and abstracts, we identified five studies that established low-dose CBCT protocols for various indications in paediatric patients: Hidalgo Rivas et al. (2015), EzEldeen et al. (2017), Oenning et al. (2019), Brasil et al. (2019), and Ito et al. (2023).^3-7 An additional study by Lemberger et al. (2023)⁸ was identified by hand searching. No study was removed due to language issues (i.e., non-English).

The protocols advocated by these studies were surely very different in terms of the parameter settings, due to the different CBCT units as well as the different diagnostic indications tested, but they share common findings: The current and exposure time (mAs) can be reduced while maintaining sufficient image quality for clinical task in children. In this era of evidence-based dentistry, paediatric patients would benefit much more if such protocols were repeatedly tested and validated (and fine-tuned) by different research teams. Then, these protocols could be readily translated into daily clinical practice with steadfast evidence instead of being preliminary recommendations. In addition, the literature indicates that imaging performance and radiation dose outcomes in CBCTs cannot be directly extrapolated across different CBCT models due to significant variations in technical characteristics and clinical diagnostic efficacy.⁹ There is an urgent need for additional studies that use consistent methodologies to test new machines and protocols. In particular, only Lemberger et al. (2023) was a clinical study, whereas the other five were phantom head studies. More clinical studies would be beneficial, because the subjective image quality assessment based on real patient imaging data is more clinically relevant than objective assessments with IQ phantom such as sharpness level. Images from real patient scans are affected by patient movement, so the actual spatial resolution is usually reduced. This is different from ex vivo studies that can obtain data with a “nominal spatial resolution”¹⁰ as a phantom head does not move. Meanwhile, ex vivo experiments, particularly Monte Carlo simulations, can provide a more controlled environment to robustly calculate radiation doses. Because of ethical considerations, a patient may not undergo repeated scans with different protocols. This makes it challenging to compare protocols within the same set of patients. It is therefore pragmatic to develop specific image quality models that closely reproduce clinical conditions of paediatric patients, such as with soft tissue substitutes to account for scattering. Hence, both clinical and nonclinical studies are complementary to each other and should be encouraged.

On February 22, 2024, we collected the citation data of the abovementioned six studies from Dimensions, and their altmetrics data (e.g., mentions from social media platforms, online news, policy documents, and patent documents) from Altmetric. The citation statements were then manually retrieved from the citing papers. It was found that Hidalgo Rivas et al. (2015) was cited the most (60 times), followed by Oenning et al. (2019) (46 times), EzEldeen et al. (2017) (40 times), Brasil et al. (2019) (9 times), Lemberger et al. (2023) (3 times), and Ito et al. (2023) (0 time). In terms of altmetrics, only Oenning et al. (2019) was mentioned 16 times on X (previously known as Twitter), whereas the other five studies had zero altmetrics scores.

For Hidalgo Rivas et al. (2015) and Brasil et al. (2019), none of the citations were made to test or follow the recommended protocols. Many citations were made to support the position of using low-dose or optimized dose for CBCT scans, whereas some citations were made to justify/follow the use of static images (selected CBCT images) for subjective image quality assessment. Similarly, for Lemberger et al. (2023), the citations were made to support the position of using low-dose or optimized dose for CBCT scans without testing or following any established protocols.

For EzEldeen et al. (2017), only one study¹¹ cited it for reproducing the recommended protocol without testing it. The citing paper had the same first author and last author as the cited reference. We could not verify whether the exact protocol was followed or not, because the scanning parameters could not be retrieved from the publisher's website (claimed to be listed in its appendix). In fact, EzEldeen et al. was mainly cited to describe the use of low-dose CBCT images as the first step in the digital workflow for 3D printing or tooth segmentation.

Oenning et al. (2019) was cited by one study¹² that tested their recommended protocol against a standard protocol. Nevertheless, this time, the citing paper changed the FOV to 5 cm × 5 cm instead of 8 cm × 8 cm and evaluated vertical root fracture in an adult phantom instead of a paediatric one. The results showed that the diagnostic value of the low-dose protocol did not differ from the standard protocol. The first author and last author of the cited reference were co-authors of the citing paper, a situation similar to EzEldeen et al. (2017). In addition, Oenning et al. (2019) was mainly cited to justify the use of Mix-D (paraffin wax + chemicals) to cover the skull phantom to simulate soft tissue, and the use of static images (selected CBCT images) for subjective image quality assessment. As for X (ex-Twitter) mentions, the original post was by Oenning herself to promote the study, which was retweeted (re-posted) 15 times.

Collectively, the citation and altmetrics data seemed to suggest that there was a lack of interest from the scientific community to test, validate, or apply the recommended low-dose or optimized CBCT protocols for children introduced by existing studies. It may be that novelty is emphasized by research grant application reviewers as well as by journal editors and peer reviewers so that testing or replication of existing studies becomes discouraged. Alternatively, it may be that such technical studies in oral and maxillofacial radiology are not as “eye-catching” as clinical research works. The low-dose protocols devised from the abovementioned studies should be tested, validated, and promoted by the academia so that they have a better chance of being translated into clinical practice to maximize the impact of research. Hopefully, the development of artificial intelligence applied to optimization will spark renewed interest. Regardless of the CBCT units available, clinicians should be proactive in seeking optimized alternatives rather than uninformedly using or replicating the default protocols, especially when treating children.