CranioCatch Offers a New-Generation Assessment Standard in Dental Education
Scientific Evidence: CranioCatch Sets a New Standard for Assessment in Dental Education
A study with 266 students reveals the role of AI-supported performance-based assessment in oral radiology education
In dental education, “how much you know” and “how well you can make decisions in a clinical setting” are not always the same thing. A new study demonstrating this gap has been published in the journal Diagnostics (MDPI, August 2026): “Artificial Intelligence-Supported Performance-Based Assessment in Oral Radiology Education: A Constructive Alignment Perspective”.
Conducted by researchers from Lokman Hekim University, Ankara University, Başkent University, Gazi University and Ankara Medipol University (Coşkun Albayrak et al.), the study examines the contribution of the CranioCatch AI platform to dental education using objective data.
How Was the Study Conducted?
266 fourth-year dental students from Lokman Hekim University took part. Using CranioCatch’s AI-supported lesion detection platform, the students identified, delineated and diagnosed pathologies present on five different panoramic radiographs (fibrous dysplasia, dentigerous cyst, compound odontoma, radicular cyst and ameloblastoma).
Student performance was evaluated within a confusion matrix framework by comparing their annotations against CranioCatch’s AI detections, which had been validated by expert clinicians (two board-certified oral and maxillofacial radiologists). Sensitivity and precision values were calculated to generate a composite performance score for each student.
Key Findings
- It measures a different competency: The study found no statistically significant correlation between CranioCatch scores and students’ end-of-year grades or final exam results (ρ = 0.06–0.11, p > 0.05). According to the researchers, this is not a shortcoming but rather an important finding: unlike the knowledge-recall level (Knows / Knows How) measured by traditional written exams, AI-supported assessment captures applied diagnostic reasoning at the “Shows How” level of Miller’s Pyramid.
- Over-annotation is associated with lower performance: ROC analysis showed that as the number of regions marked by students increased (above a threshold of 8.5 regions), performance scores dropped significantly (p < 0.001). This indicates that uncertain students tend to adopt a “random marking” strategy, which reduces precision.
- Immediate feedback supports reflective learning: After the assessment, students could instantly compare their own annotations with expert-approved AI results — creating a learning loop that is not possible in conventional exams.
What Does This Mean in Terms of Constructive Alignment?
According to the principle of constructive alignment, developed by educational scientist John Biggs, learning outcomes, teaching activities and assessment methods must be consistent with one another. The researchers emphasise that traditional written exams cannot adequately measure applied skills such as radiographic interpretation, whereas platforms like CranioCatch fill this gap by serving as authentic assessment tools that reflect real clinical scenarios.
CranioCatch’s Role in Education
The study provides scientific support for the following aspects of CranioCatch:
- Objective assessment based on a standardised, expert-validated reference standard
- Measurable performance tracking through confusion matrix-based precision/sensitivity scoring
- Immediate feedback that lets students compare their own results with the AI
- Instructor flexibility: faculty members can incorporate their own expert judgement alongside the AI metrics
- An assessment dimension that complements rather than replaces traditional exams
The authors also highlight an important balance: while AI-supported tools offer great value in education, they should be used with a supervised and critical approach to mitigate the risk of students over-relying on AI outputs (automation bias). The way CranioCatch was used in this study was designed precisely in line with that principle — with expert validation and instructor oversight.
Conclusion
This multicentre academic collaboration involving 266 students shows that CranioCatch is not only a clinical diagnostic support system, but also a scientifically validated, competency-focused assessment tool for dental education.
Source: Coşkun Albayrak, S.; Gürel, F.S.; Kubat, G.; Coşkun, Ö.; Budakoğlu, I.İ.; Orhan, K. Artificial Intelligence-Supported Performance-Based Assessment in Oral Radiology Education: A Constructive Alignment Perspective. Diagnostics 2026, 16, 2542. https://doi.org/10.3390/diagnostics16162542 (Open access, CC BY 4.0)
Contact us to learn more about CranioCatch’s solutions for educational institutions.
