Abstract
Purpose:
This study investigated the relationships among water-equivalent diameter (Dw), volume computed tomography dose index (CTDIvol), size-specific dose estimate (SSDE), and image noise in pediatric head computed tomography (CT) under active and inactive tube current modulation (TCM).
Subjects and Methods:
Fifty pediatric patients (0–14 years) undergoing noncontrast head CT were retrospectively analyzed. Scans were performed on a 128-slice CT scanner at 120 kVp. Twenty-five patients were examined using active TCM, and 25 with fixed tube current. CTDIvol and SSDE were calculated using IndoseCT software following American Association of Physicists in Medicine TG-293 recommendations. Image noise was quantified using the global noise index. Relationships among Dw, CTDIvol, SSDE, and noise were evaluated using linear regression.
Results:
Active TCM reduced CTDIvol from 38.5 ± 7.1 mGy to 32.8 ± 6.4 mGy (15.4%, P = 0.021) and SSDE from 39.2 ± 6.1 mGy to 29.6 ± 5.3 mGy (24.5%, P < 0.001). Image noise decreased from 10.2 ± 2.3 HU to 8.6 ± 1.9 HU (P = 0.037).
Conclusion:
Active TCM improves dose–noise efficiency in pediatric head CT by reducing radiation dose while maintaining stable image quality. Integration of Dw-based SSDE with automated noise analysis provides a quantitative framework for pediatric CT dose auditing and may support the development of size-specific diagnostic reference levels.
Keywords: Computed tomography, pediatric head computed tomography, size-specific dose estimate, tube current modulation, water-equivalent diameter
INTRODUCTION
Computed tomography (CT) is widely used in pediatric neuroimaging due to its rapid acquisition and high diagnostic accuracy. However, children are more radiosensitive than adults, and cumulative radiation exposure remains a significant concern. Multicenter surveys have demonstrated substantial inter-institutional variability in pediatric head CT doses, emphasizing the importance of systematic dose monitoring and protocol harmonization.[1,2,3]
The conventional scanner output metric, volume CT dose index (CTDIvol), does not account for patient size and tissue attenuation differences. Size-specific dose estimate (SSDE) improves dose interpretation by applying a size-dependent correction factor to CTDIvol according to American Association of Physicists in Medicine (AAPM) TG-293 recommendations.[4,5,6] SSDE is calculated as the product of CTDIvol and a size-correction factor derived from patient dimensions. Water-equivalent diameter (Dw) represents a combined geometric and attenuation-based parameter used to determine the correction factor in SSDE calculation. Because SSDE is computed from CTDIvol using a Dw-dependent correction factor, accurate Dw estimation is essential for reliable size-adjusted dose assessment in pediatric CT.[7,8,9]
Tube current modulation (TCM) automatically adjusts tube current according to patient attenuation, aiming to maintain image quality while minimizing unnecessary radiation exposure. Prior pediatric CT studies[8,10,11] have reported dose reductions with active TCM compared with fixed-current protocols.
On the other hand, image noise is an objective indicator of CT image quality. Automated noise analysis has been validated in pediatric CT imaging.[12,13,14] However, quantitative relationships among Dw, CTDIvol, SSDE, and image noise under active versus inactive TCM remain incompletely characterized. An integrated evaluation of size metrics, dose indices, and image noise is necessary for comprehensive pediatric dose optimization. This study quantitatively analyzes the relationships among Dw, CTDIvol, SSDE, and image noise in pediatric head CT under active and inactive TCM. To our knowledge, this study represents the first integrated quantitative evaluation of Dw, CTDIvol, SSDE, and automated noise metrics under active and inactive TCM in pediatric head CT.
SUBJECTS AND METHODS
Study design and population
This retrospective cross-sectional study included 50 pediatric patients (0–14 years) who underwent noncontrast head CT between January 2021 and December 2023. Ethical approval was obtained from the institutional review board (No. 1358/EC/KEPK-RSDK/2022).
Inclusion criteria were pediatric noncontrast head CT examinations performed for clinical indications with complete dose reports and artifact-free axial DICOM datasets. Exclusion criteria included incomplete metadata, improper positioning affecting region-of-interest segmentation, and metallic artifacts within the analyzed region. Patients were categorized into active and inactive TCM groups.
Computed tomography acquisition protocol
All examinations were performed using a 128-slice CT scanner (SOMATOM Definition AS, Siemens Healthcare, Erlangen, Germany). The acquisition parameters included tube voltage, tube current, rotation time, pitch, slice thickness, reconstruction kernel, and matrix size.[8,10] CT acquisition parameters are summarized in Table 1.
Table 1.
Computed tomography acquisition parameters for active and inactive tube current modulation protocols
| Parameter | Active TCM | Inactive TCM |
|---|---|---|
| Tube voltage (kVp) | 120 | 120 |
| Tube current (mA) | 100–350 (auto) | 250 (fixed) |
| Rotation time (s) | 0.5 | 0.5 |
| Pitch | 1 | 1 |
| Slice thickness (mm) | 1 | 1 |
| Reconstruction kernel | Standard (H30s) | Standard (H30s) |
| Matrix size | 512×512 | 512×512 |
TCM: Tube current modulation
Water-equivalent diameter and size-specific dose estimate calculation
Dw was automatically calculated at the mid-orbital slice using IndoseCT v20.b software (ITB, Bandung, West Java, Indonesia) implementing the AAPM TG-293 methodology.[4] SSDE was calculated using Equation (1):
SSDE = CTDIvol×fsize(Dw) (1)
Where fsize (Dw) represents the size-correction factor derived from the Dw according to the TG-293 formalism. Both Dw and SSDE values were computed automatically by the software. Manual verification of Dw and SSDE calculations was performed on a subset of datasets to confirm agreement with the software outputs. Previous studies have demonstrated strong relationships between SSDE and patient size metrics in pediatric CT imaging, supporting the reliability of Dw-based SSDE estimation for size-adjusted dose assessment.[6,7]
Image noise measurement
Image noise was quantified using the global noise index and measured using IndoQCT v25c (Diponegoro University, Semarang, Central Java, Indonesia). Automated noise quantification approaches have previously been validated for pediatric CT image-quality evaluation and allow reproducible measurement of noise distributions across image datasets.[12,13,14] The automated analysis workflow, including axial CT images, standard deviation (SD) maps, and histogram distributions, is illustrated in Figure 1. Additional technical details regarding Dw calculation, SSDE estimation, and automated noise analysis are available in the Supplementary Material.
Figure 1.
Workflow of automated image-noise quantification using IndoQCT: (a) Axial computed tomography image, (b) Standard-deviation map, and (c) Histogram used to determine the global noise index
Statistical analysis
Statistical analysis was performed using SPSS v26.0 (IBM Corp., Armonk, NY, USA). Data normality was assessed using the Shapiro–Wilk test. Quantitative variables are presented as mean ± SD. Independent-samples t-tests were used for group comparisons. Levene’s test was applied to evaluate the homogeneity of variance. Given the moderate sample size (n = 25 per group), residual plots were additionally inspected to verify homoscedasticity, and Welch correction was applied where appropriate. Levene’s test remains appropriate for moderate sample sizes, and the additional inspection of residual plots together with the Welch correction was applied to ensure robust variance comparison. Correlation strength was categorized as weak (r2 < 0.40), moderate (0.40–0.70), and strong (>0.70).
RESULTS
Water-equivalent diameter and volume computed tomography dose index
CTDIvol was 32.8 ± 6.4 mGy (active TCM) and 38.5 ± 7.1 mGy (inactive TCM), corresponding to a 15.4% reduction (P = 0.021). The relationship between Dw and CTDIvol is shown in Figure 2. A moderate association was observed under active TCM (r2 = 0.6324), whereas a weak-to-moderate association was observed under inactive TCM (r2 = 0.3788).
Figure 2.
Relationship between water-equivalent diameter and volume computed tomography dose index for active and inactive TCM protocols. Dw: Water-equivalent diameter, TCM: Tube current modulation, SSDE: Size-specific dose estimate
Water-equivalent diameter and size-specific dose estimate
SSDE values were 29.6 ± 5.3 mGy (active TCM) and 39.2 ± 6.1 mGy (inactive TCM), corresponding to a 24.5% reduction (P < 0.001). The Dw–SSDE relationship is illustrated in Figure 3. Under inactive TCM, a strong determination coefficient was observed (r2 = 0.9944), while under active TCM, the association was weak (r2 = 0.3242). Because SSDE is mathematically derived from CTDIvol using Dw-based correction factors, the strong association under fixed-current conditions reflects amplified size dependence.
Figure 3.
Relationship between water-equivalent diameter and size-specific dose estimate in pediatric head computed tomography for active and inactive TCM protocols. Dw: Water-equivalent diameter, TCM: Tube current modulation, SSDE: Size-specific dose estimate
Water-equivalent diameter and image noise
Image noise was 8.6 ± 1.9 HU (active) and 10.2 ± 2.3 HU (inactive), representing a 15.7% reduction (P = 0.037). The regression between Dw and noise is presented in Figure 4, demonstrating reduced size dependence under active TCM.
Figure 4.
Correlation between water-equivalent diameter and image noise under active and inactive tube current modulation conditions. Dw: Water-equivalent diameter, TCM: Tube current modulation
Comparison of all parameters
A comparative summary of CTDIvol, SSDE, and image noise parameters for active and inactive TCM protocols is presented in Table 2. The distribution of CTDIvol, SSDE, and image noise values for both protocols is illustrated in Figure 5, which demonstrates the reduced variability observed under active TCM.
Table 2.
Comparison of dose and noise parameters between active and inactive tube current modulation protocols
| Parameter | Active TCM (mean±SD) | Inactive TCM (mean±SD) | Difference (%) | P | R2 (active) | R2 (inactive) |
|---|---|---|---|---|---|---|
| CTDIvol (mGy) | 32.8±6.4 | 38.5±7.1 | ↓ 15.4 | 0.021 | 0.63 | 0.38 |
| SSDE (mGy) | 29.6±5.3 | 39.2±6.1 | ↓ 24.5 | <0.001 | 0.32 | 0.99 |
| Image noise (HU) | 8.6±1.9 | 10.2±2.3 | ↓ 15.7 | 0.037 | 0.39 | 0.58 |
↓decrease, TCM: Tube current modulation, SD: Standard deviation, CTDIvol: Computed tomography dose index, SSDE: Size-specific dose estimate
Figure 5.
Boxplots of volume computed tomography (CT) dose index, size-specific dose estimate, and noise values in pediatric head CT examinations: (a) Inactive tube current modulation (TCM) group; (b) active TCM group. SSDE: Size-specific dose estimate, CTDIvol: Volume computed tomography dose index
DISCUSSION
This study demonstrates that active TCM significantly improves dose–noise efficiency in pediatric head CT. Compared with fixed-current acquisition, active TCM reduced CTDIvol by 15.4% and SSDE by 24.5%, while simultaneously decreasing image noise by 15.7%. The concurrent reduction in both radiation dose and image noise indicates improved photon utilization efficiency rather than a simple trade-off between radiation output and image quality degradation. Similar benefits of attenuation-based TCM have been reported in previous CT optimization studies evaluating automatic exposure control strategies.[8,10,15]
From a radiophysical perspective, SSDE incorporates a Dw-dependent correction factor that intrinsically links dose estimates to patient size. Previous investigations have demonstrated strong correlations between SSDE and patient size metrics in pediatric CT imaging, supporting the reliability of Dw as a surrogate for patient attenuation and dose scaling.[4,6,7] Under fixed-current (inactive TCM) scanning conditions, this mathematical dependence can amplify size-related variability in SSDE values. Consequently, the very strong Dw–SSDE determination coefficient observed under inactive TCM (r2 = 0.9944) likely reflects the inherent dependence of SSDE on patient size rather than a direct relationship between scanner output and attenuation characteristics.
When TCM was activated, the scanner dynamically adjusted tube current according to attenuation differences across pediatric head sizes. This adaptive response reduced the apparent size-driven variability in SSDE and lowered the Dw–SSDE determination coefficient to 0.3242. This reduction does not contradict the mathematical dependence of SSDE on Dw; instead, it reflects the stabilizing effect of modulation on radiation output relative to patient attenuation. Similar modulation-induced reductions in inter-patient dose variability have been reported in studies evaluating automatic exposure control systems in CT imaging.[8,10,15]
The relationship between Dw and CTDIvol also provides insight into scanner output behavior under different acquisition modes. The moderate Dw–CTDIvol association observed under active TCM (r2 = 0.6324) indicates that TCM maintains proportionality between patient attenuation and radiation output while preventing excessive dose escalation in larger head diameters. In contrast, the weaker Dw–CTDIvol association under inactive TCM (r2 = 0.3788) suggests less predictable adaptation to patient size when tube current remains fixed. Similar patterns have been reported in previous studies investigating attenuation-based exposure control mechanisms in CT imaging.[8,10]
An important observation in this study is that the reduction in SSDE (24.5%) exceeded the reduction in CTDIvol (15.4%). This finding highlights the importance of size-adjusted dose metrics in pediatric CT. While CTDIvol reflects scanner output, SSDE incorporates patient size and attenuation, providing a more accurate estimate of the absorbed dose in pediatric patients. Previous investigations have emphasized the clinical relevance of SSDE for patient-specific dose assessment and pediatric CT dose auditing.[4,6,16]
Image noise behavior further supports the advantages of TCM. Under inactive TCM conditions, the Dw–noise association was moderate (r2 = 0.5841), indicating increased noise variability with head size. When TCM was enabled, this relationship decreased (r2 = 0.3911), suggesting more stable photon flux adaptation across pediatric head sizes. These findings are consistent with those reported by Greffier et al.,[10] who demonstrated improved dose–noise equilibrium under TCM, and by Tangsiwong et al.,[11] who showed that fixed-current scanning protocols may produce size-dependent noise amplification.
Compared with previous pediatric CT studies that examined CTDIvol or SSDE independently, the present study integrates patient size metrics (Dw), dose indices (CTDIvol and SSDE), and quantitative noise analysis within a unified analytical framework.[4,5,6,7] This multidimensional evaluation provides a more comprehensive understanding of dose–image quality relationships in pediatric CT and highlights the importance of incorporating patient size metrics into dose optimization strategies. A comparison between the present study and selected previous investigations on pediatric head CT dose optimization is summarized in Table 3.
Table 3.
Comparison with selected previous studies on pediatric head computed tomography dose optimization
| Study | Sample/population | Focus | Key findings | Comparison with present study |
|---|---|---|---|---|
| Markart et al. (2022)[10] | Pediatric head CT | Tube current modulation | TCM reduces inter-patient variability | Present study quantifies Dw–dose relationships under TCM |
| Kanal et al. (2022)[1] | Multicenter pediatric CT (USA) | DRLs | Established national DRLs and dose variability ranges | Supports need for standardized DRLs; present study adds Dw-based SSDE correlation |
| Zhang et al. (2024)[7] | Pediatric CT | Dw-based SSDE | Developed conversion factors for Dw-based SSDE | Present study extends by integrating Dw with noise analysis |
| Khan et al. (2025)[5] | Pediatric CT patients | SSDE | SSDE correlates strongly with patient size | Present study confirms SSDE–Dw relationship under TCM conditions |
| Zancopè et al. (2025)[6] | Clinical CT validation study | SSDE validation | SSDE correlates with effective dose | Supports validity of SSDE framework used in this study |
| Kellenberger et al. (2025)[8] | Pediatric head CT | Dose optimization | Automated modulation reduces dose variability | Confirms role of TCM in dose stabilization |
| Bos et al. (2025)[2] | International pediatric CT registry | DRLs (indication-based) | Demonstrated variability across institutions | Present study complements with quantitative Dw–dose relationships |
| Present study (2025) | 50 pediatric patients, Indonesia | Dw, CTDIvol, SSDE, noise | Dw influences dose and noise; TCM improves stability | Provides integrated quantitative evaluation under active vs inactive TCM |
Finally, regional evidence on Dw-based SSDE and noise behavior under TCM remains limited in Southeast Asia. Multicenter surveys have highlighted considerable variability in pediatric CT radiation doses across institutions, emphasizing the need for standardized dose monitoring and protocol optimization.[1,2,3] Therefore, the present investigation contributes important regional data supporting the implementation of size-aware dose monitoring strategies and the development of pediatric diagnostic reference levels tailored to local clinical practice.[16]
Several limitations should be acknowledged. First, the study was conducted at a single institution with a moderate sample size (n = 50), which may limit the generalizability of the findings. Second, all examinations were performed using a single CT platform at a fixed tube voltage of 120 kVp, and dose–size relationships may vary across scanner models, vendors, or acquisition settings. Future multicenter investigations involving larger pediatric cohorts and multivendor CT systems would allow more comprehensive characterization of size-specific dose behavior. Incorporation of task-based image quality metrics and detectability analysis may further refine pediatric CT protocol optimization.
CONCLUSION
Active TCM significantly enhanced dose–noise efficiency in pediatric head CT, reducing CTDIvol by 15.4%, SSDE by 24.5%, and image noise by 15.7% compared with fixed-current scanning. The strong Dw–SSDE association observed under inactive TCM reflects amplified size dependence inherent to the SSDE calculation, whereas active TCM mitigates apparent size-driven dose variability. These findings demonstrate that adaptive TCM not only reduces radiation output but also stabilizes image quality across variable pediatric head sizes. Integration of Dw-based SSDE and automated noise quantification provides a quantitative, reproducible framework for pediatric CT dose auditing and supports the development of size-aware diagnostic reference levels tailored to regional clinical practice.
Conflicts of interest
There are no conflicts of interest.
Acknowledgments
This study was supported by the Institute for Research and Community Service (LPPM), Diponegoro University, under research contract No. 222-506/UN.7.D2/PP/IV/2025. The authors thank the Department of Radiology, Dr. Kariadi Central General Hospital, and the Department of Physics, Diponegoro University, for their technical support. Ethical approval was obtained from the Institutional Review Board of Dr. Kariadi Central General Hospital (No. 1358/EC/KEPK-RSDK/2022).
Funding Statement
Nil.
SUPPLEMENTARY MATERIAL
Supplementary Material 1 : Detailed methodology for water-equivalent diameter (Dw), size-specific dose estimate (SSDE), and image noise analysis
Computed tomography image acquisition and data preparation
Pediatric non-contrast head CT examinations were performed using a 128-slice CT scanner (SOMATOM Definition AS, Siemens Healthcare, Erlangen, Germany). All scans were acquired at a tube voltage of 120 kVp with either active tube current modulation (TCM) or fixed tube current settings.
The reconstructed axial images were exported in DICOM format and imported into the IndoseCT v2.1 software for dose analysis and IndoQCT software for image noise analysis. For each examination, the mid-orbital slice was selected for patient size estimation and subsequent SSDE calculation, consistent with previously recommended pediatric CT dose assessment practices [1,5].
Water-equivalent diameter calculation
The water-equivalent diameter (Dw) was calculated following the AAPM TG-293 formalism [4,7]. Dw represents a patient-size metric that accounts for both geometric dimensions and attenuation properties of the scanned object.
The Dw value was computed using the following equation:
where:
AROI is cross-sectional area of the patient within the region of interest
HUROI is mean CT number of the segmented region
The segmentation of the patient cross-section was automatically performed using threshold-based body contour detection implemented in the IndoseCT software. The resulting Dw value was used to determine the size-dependent correction factor required for SSDE calculation.
Size-specific dose estimate
The size-specific dose estimate (SSDE) was calculated according to the American Association of Physicists in Medicine (AAPM) TG-293 recommendations. SSDE accounts for patient size by applying a correction factor to the scanner-reported CTDIvol.
SSDE was calculated using:
SSDE = CTDIvol × fsize(Dw)
where:
CTDIvol is scanner-reported volume CT dose index
fsize (Dw) is size-correction factor derived from the water-equivalent diameter
The size-correction factor was automatically determined based on Dw values using lookup tables implemented in IndoseCT software.
To verify the reliability of the automated calculations, manual SSDE estimation was performed for a subset of cases using the TG-293 correction factors. The manually derived values demonstrated close agreement with the software-generated results [4,6].
Automated image noise quantification
Image noise was quantified using the IndoQCT software, which implements an automated noise analysis based on the Christianson method [12-14].
The workflow consisted of the following steps:
Import axial CT images.
Generate pixel-wise standard deviation (SD) maps within soft-tissue regions.
Construct histogram distributions of SD values.
Determine the Global Noise Index (GNI) defined as the SD value at the maximum frequency of the histogram.
The automated approach reduces operator-dependent variability and provides reproducible quantitative assessment of image noise [12-14].
Validation of software-based measurements
The accuracy of IndoseCT for Dw-based SSDE estimation and IndoQCT for automated noise quantification has been previously validated in phantom and clinical studies [6,12,14]. These studies reported excellent agreement between automated software outputs and reference measurements, with intraclass correlation coefficients exceeding 0.95.
The consistency between manual calculations and software outputs observed in the present study further confirms the reliability of the computational workflow.
Statistical analysis
All statistical analyses were performed using SPSS v26.0 (IBM Corp., Armonk, NY, USA). Normality of the data distribution was evaluated using the Shapiro-Wilk test.
Group comparisons between active and inactive TCM protocols were performed using independent- samples t-tests. Homogeneity of variance was assessed using Levene's test, and Welch correction was applied where appropriate.
Correlation analyses between Dw, CTDIvol, SSDE, and image noise were performed using linear regression. The strength of correlation was categorized as:
Weak: R2 < 0.4 0
Moderate: 0.40 ≤ R2 ≤ 0.7 0
Strong: R2 > 0.7 0
A p-value < 0.05 was considered statistically significant.
REFERENCES
- 1.Kanal KM, Butler PF, Chatfield MB, Wells J, Samei E, Simanowith M, et al. U.S. diagnostic reference levels and achievable doses for 10 pediatric CT examinations. Radiology. 2022;302:164–74. doi: 10.1148/radiol.2021211241. [DOI] [PubMed] [Google Scholar]
- 2.Bos D, Wang Y, Stewart C, Chu PW, Luong J, Haubold J, et al. Diagnostic reference levels for indication-based CT categories in pediatric CT: Data from an international registry. Eur Radiol. 2025;35:7985–99. doi: 10.1007/s00330-025-11724-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bouchareb Y, Al Kharusi M, Al Maqbali A, Al Maimani A, Al Maskari H, Sirasanagandla SR, et al. Establishing diagnostic reference levels for paediatric CT imaging: A multi-centre study. Healthcare (Basel) 2025;13:1728. doi: 10.3390/healthcare13141728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.American Association of Physicists in Medicine Size-specific dose estimates (SSDE) for head CT: Report of AAPM Task Group 293. AAPM Report No. 293. 2019 [Google Scholar]
- 5.Khan R, Shah MA, Wahab S, Khan RA. Size-specific dose estimate and effective dose for pediatric computed tomography. J Med Phys. 2025;50:173–83. doi: 10.4103/jmp.jmp_55_24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zancopè N, De Monte F, Simeone E, Giannone A, Lombardi R, Mele A, et al. Validation of size-specific dose estimate calculation in a modern CT scanner and correlation with effective dose. Sci Rep. 2025;15:6091. doi: 10.1038/s41598-025-90509-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhang J, Zhang X. Size-specific dose estimation in pediatric CT: Preliminary study and conversion factors. Radiat Prot Dosimetry. 2024;200:677–86. doi: 10.1093/rpd/ncae104. [DOI] [PubMed] [Google Scholar]
- 8.Kellenberger CJ, Benz DC, Eberhard M, Wildi K, Goetti R, Hickethier T, et al. Radiation dose optimisation in paediatric head CT using attenuation-based auto prescription. Sci Rep. 2025;15:33276. doi: 10.1038/s41598-025-18097-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Poosiri S, Chuboonlap K, Kaewlaied N. Age-based size-specific dose estimate for pediatric computed tomography head examinations. Appl Sci. 2024;14:7848. [Google Scholar]
- 10.Markart S, Fischer TS, Wildermuth S, Dietrich TJ, Alkadhi H, Leschka S, et al. Organ-based tube current modulation and bismuth eye shielding in pediatric head computed tomography. Pediatr Radiol. 2022;52:2584–94. doi: 10.1007/s00247-022-05410-x. [DOI] [PubMed] [Google Scholar]
- 11.Chen XF, Zhang YC, Ding N, Liu L, Ji YD, Zhang C, et al. Radiation dose reduction and image quality in pediatric paranasal sinus CT with automatic tube current modulation and iterative reconstruction technique. Radiat Prot Dosimetry. 2024 doi: 10.1093/rpd/ncae192. [DOI] [PubMed] [Google Scholar]
- 12.Inkinen SI, Mäkelä T, Kaasalainen T, Peltonen J, Kangasniemi M, Kortesniemi M. Automatic head computed tomography image noise quantification with deep learning. Phys Med. 2022;99:102–12. doi: 10.1016/j.ejmp.2022.05.011. [DOI] [PubMed] [Google Scholar]
- 13.Ketola JHJ, Inkinen SI, Mäkelä T, Kaasalainen T, Peltonen JI, Kangasniemi M, et al. Automatic chest computed tomography image noise quantification using deep learning. Phys Med. 2024;117:103186. doi: 10.1016/j.ejmp.2023.103186. [DOI] [PubMed] [Google Scholar]
- 14.Ahmad M, Jacobsen MC, Thomas MA, Chen HS, Layman RR, Jones AK. A benchmark for automatic noise measurement in clinical computed tomography. Med Phys. 2021;48:640–7. doi: 10.1002/mp.14635. [DOI] [PubMed] [Google Scholar]
- 15.Pondard S, Desport C, Munier M, Kien N, Rousseau H, Merignac O, et al. Multicentric characterization of organ-based tube current modulation in head computed tomography: A dosimetric and image quality study. Phys Med. 2024;121:103389. doi: 10.1016/j.ejmp.2024.103389. [DOI] [PubMed] [Google Scholar]
- 16.Monteiro VC, Schelin HR, Legnani A, Filipov D, Freitas MB, Malthez ALMC. A study of SSDE, CTDI, and DLP dose indexes for establishing diagnostic reference levels in pediatric CT exams. Radiat Phys Chem. 2025;235:112813. [Google Scholar]