Bone density: when CBCT is measured against CT Hounsfield units

Evaluating bone density: can CBCT match medical CT scans?

In implantology and oral surgery, the precise measurement of bone density determines therapeutic success, from planning to primary implant stability. While Multi-Slice Computed Tomography (MSCT) serves as the gold standard thanks to Hounsfield Unit (HU) calibration, its constraints — high radiation dose and cost — limit its systematic use in the practice. Cone-Beam Computed Tomography (CBCT), which is more accessible, generates non-standardised Gray Values (GV), influenced by the field of view or reconstruction artefacts, often making their quantitative interpretation uncertain.

This observational study, conducted with 60 patients (34 men, 26 women; mean age 39.5 years), specifically aims to evaluate the correlation between MSCT HU and CBCT GV within four strategic maxillofacial regions: the anterior and posterior maxillary and mandibular sectors. The researchers tested the hypothesis that, despite the lack of absolute calibration, CBCT grey values could accurately reflect the relative bone density trends observed in MSCT. The clinical stakes are major for the practitioner: determining whether CBCT can serve as a reliable quantitative substitute for daily bone mass assessment.

Comparative evaluation methodology

This observational study, conducted between September 2024 and June 2025 at Baqubah Teaching Hospital, involved a cohort of 60 patients (34 men and 26 women) with a mean age of 39.5 ± 7.3 years. Inclusion exclusively concerned individuals requiring radiological examinations for dental or maxillofacial traumatic lesions, without underlying bone pathology at the analysed sites.

The protocol consisted of comparing bone density measurements via two distinct imaging systems:

• MSCT: Siemens Somatom 128-slice scanner, operated at 120 kVp and 200 mAs, with a slice thickness of 1 mm.
• CBCT: Planmeca ProMax 3D scanner, configured at 90 kVp and 12 mA, with a voxel size of 0.2 mm covering the entire maxillofacial region.

Four anatomical sites were selected for measurements: the anterior maxilla, the posterior maxilla, the anterior mandible, and the posterior mandible. For each site, Hounsfield units (HU) from MSCT and gray values (GV) obtained by CBCT were recorded using standardized circular regions of interest (ROI). Statistical processing utilized Pearson's correlation coefficient, paired t-tests for mean differences, as well as Bland-Altman analysis and linear regression to evaluate the predictive relationship between the two modalities.

Comparative analysis of densities: MSCT vs CBCT

The study, involving 60 patients (34 men, 26 women; mean age 39.5 ± 7.3 years), shows that the density values obtained by MSCT (Hounsfield Units - HU) are consistently higher than the grey values (GV) from CBCT across all anatomical sites evaluated. A statistically significant difference was observed for each region (p < 0.001).

Measurements performed in the trabecular bone (circular ROIs of 2 to 3 mm) highlight variable discrepancies depending on the location:

• Anterior mandible: This is where the average gap is most pronounced (72.6).
• Posterior maxilla: This area shows the lowest average difference (46.7).

Correlations and limits of interchangeability

Although both imaging modalities follow similar trends, statistical analysis reveals a strong positive correlation between HU and GV, particularly in the mandible.

However, the authors emphasize that a high correlation does not imply interchangeability of measurements. The Bland-Altman analysis shows very wide limits of agreement, ranging from -145.2 to +108.0. This result confirms that CBCT cannot provide absolute density measurements without rigorous prior calibration against MSCT.

Reliability of measurements

La reproductibilité des données est jugée excellente. L'analyse de fiabilité intra-observateur affiche des coefficients de corrélation intraclasse (CCI) supérieurs à 0,94 pour toutes les régions, avec un coefficient de variation (CV) inférieur à 4 %, garantissant la cohérence des mesures répétées au cabinet ou en centre hospitalier.

The illusion of absolute measurement in CBCT

The results of this study confirm a strong positive correlation (r = 0.85-0.91) between the Hounsfield units (HU) of medical CT and the gray values (GV) of CBCT. However, the Bland-Altman analysis reveals particularly wide limits of agreement, ranging from -145.2 to +108.0. This dispersion proves that, despite a common linear trend, the two modalities are not interchangeable for strict quantitative measurement.

The lack of CBCT standardisation remains the main obstacle. Unlike MSCT scanning, grey values are influenced by the type of device, exposure parameters, field of view (FOV), and beam hardening artefacts. These variables introduce "noise" that prevents the interpretation of GVs as an absolute bone density value.

For the clinician, CBCT remains a powerful tool for evaluating bone morphology and identifying relative density trends, particularly during implant planning or evaluation of the temporomandibular joint. However, the study highlights that without a specific calibration protocol, relying solely on CBCT values to dictate a drilling protocol or immediate loading remains risky.

Conclusion

This study demonstrates that while CBCT accurately reflects the distribution of bone density, it fails to provide a universal quantitative measurement without prior calibration.

Summary of results

This observational study conducted on 60 patients demonstrates a strong positive correlation (r = 0.85 to 0.91) between the Hounsfield units (HU) of the MSCT scanner and the gray values (GV) of the CBCT. However, the Bland-Altman analysis reveals wide limits of agreement (-145.2 to +108.0), proving that CBCT gray values systematically underestimate actual bone density compared to the MSCT standard.

In practical terms, for the practitioner:

• Do not confuse trends with absolute measurements: Use CBCT to assess relative bone density between different sites in the same patient, but never consider gray levels as interchangeable Hounsfield units for your loading protocols.
• Caution on primary stability: For complex cases where implant success depends on a precise density threshold, MSCT remains the reference, as beam artifacts and CBCT settings bias the actual quantitative value of the trabecular bone.
• Standardize your acquisitions: If you use grey levels for clinical follow-up, rigorously maintain the same field of view (FOV) and the same exposure parameters, as any variation artificially modifies the displayed grey value.

Technical Lexicon of the Study

Hounsfield Units (HU): A true gold standard in densitometry, this calibrated scale (using air and water as references) allows for a standardized quantification of bone density in MSCT, essential for rigorous implant planning.

Gray Values (GV): Gray values derived from CBCT. Unlike HU, these raw data are not standardized and depend on exogenous variables (field of view, positioning), making them a relative trend indicator rather than an absolute measurement.

Multislice Computed Tomography (MSCT): Multislice medical scanner used here as the reference modality. Although it involves higher radiation, it remains the gold standard for the quantitative evaluation of bone mass thanks to the reproducibility of its measurements.

Bland-Altman analysis: A crucial statistical tool for testing the interchangeability of two measurement methods. In this study, it demonstrates that despite a strong correlation, CBCT and MSCT cannot be substituted for one another without prior calibration.

Region of Interest (ROI): Data collection area selected by the examiner. The protocol here requires circular ROIs limited to the trabecular bone in order to isolate the cancellous density from interferences related to the cortical bone.

Voxel: Volume unit in 3D imaging (equivalent to the pixel in 2D). The study specifies a voxel size of 0.2 mm for CBCT scans, a resolution parameter that directly impacts the precision of bone analysis.

Information intended for healthcare professionals. This content may contain errors or truncated summaries. We recommend always verifying with the original source article. Delynov disclaims all responsibility regarding the use of this information. This document is not intended for patients or the general public.