Decoupling Machine Output from Patient Absorbance A Cluster Analysis of Individualized Metric Discrepancies Between CTDIvol and SSDE in Pediatric Head Computed Tomography

Abstract

The Volume Computed Tomography Dose Index (CTDIvol) displayed on modern CT consoles serves as the universal benchmark for institutional radiation dose audits and the establishment of regional Diagnostic Reference Levels (DRLs). However, a critical physics-to-clinical gap persists: CTDIvol is a standardized metric of scanner output measured within homogeneous acrylic phantoms, rather than an accurate reflection of individualized absorbed dose. This limitation poses significant risks, particularly within highly heterogeneous pediatric populations. This retrospective cohort study developed an internal audit framework to quantify the systemic percentage discrepancy (?%) between CTDIvol and Size-Specific Dose Estimates (SSDE) based on automated water-equivalent diameter (Dw) extractions. Utilizing clinical datasets from 50 pediatric patients (aged 0–14 years) undergoing non-contrast head CT examinations at RSUP Dr. Kariadi Semarang, we performed automated volumetric and attenuation profiling via IndoseCT v23c software. Simultaneously, Global Noise Index (GNI) analysis was conducted using IndoQCT v22a. Patients were stratified into three morphological head-size clusters: Small (Dw < 13 cm), Medium (13 ? Dw ? 15 cm), and Large (Dw > 15 cm). Statistical modeling revealed a profound, non-linear inverse relationship between patient geometry and metric variance. In the active Tube Current Modulation (TCM) cohort, CTDIvol and SSDE correlated with R² = 0.6324 and R² = 0.3242, respectively. For the inactive TCM cohort, CTDIvol remained static (R² = 0.3788), while SSDE demonstrated a strict deterministic log-linear fit (R² = 0.9944) against Dw. Noise analysis confirmed that larger Dw values systematically increased image noise in fixed exposure techniques (R² = 0.5841), an effect partially mitigated by active TCM (R² = 0.3911). In the Small Head Cluster (Dw < 13 cm), the study revealed a significant systematic underestimation of radiation dose; the actual tissue-absorbed dose (SSDE) exceeded the displayed console output by +14.64%. This discrepancy narrowed to +6.39% in the Medium Cluster and shifted to a negligible overestimation of -0.55% in the Large Cluster as patient dimensions approached the 16-cm reference phantom. These insights demonstrate that relying exclusively on console CTDIvol logs propagates a dangerous blind spot in pediatric radiation safety, systematically masking higher energy deposition in neonates and toddlers. Consequently, we advocate for the universal adoption of the individualized deviation ratio (?%) as an essential metric for next-generation hospital quality assurance and the robust optimization of radiation protection protocols in modern clinical practice. By shifting from phantom-based metrics to patient-specific dosimetry, institutions can significantly enhance the safety profile of diagnostic imaging for vulnerable pediatric patients.

Introduction

This study examines the limitations of conventional radiation dose monitoring in pediatric head CT imaging and evaluates a more patient-specific approach to radiation assessment. While modern multi-detector computed tomography (MDCT) has become essential for diagnosing neurological emergencies, congenital abnormalities, hydrocephalus, and craniofacial disorders in children, its increasing use raises concerns about radiation exposure and long-term cancer risk.

Children are significantly more sensitive to ionizing radiation than adults because of their rapidly dividing cells, developing tissues, and longer life expectancy. Sensitive organs such as the bone marrow, eye lenses, and thyroid are particularly vulnerable during head CT examinations. Therefore, strict adherence to the ALARA (As Low As Reasonably Achievable) principle is critical for minimizing radiation exposure while maintaining diagnostic image quality.

A major problem identified in clinical practice is the reliance on Volume Computed Tomography Dose Index (CTDIvol), a scanner-based radiation metric measured using standardized plastic phantoms rather than actual patients. CTDIvol does not account for differences in patient size, anatomy, or tissue composition. As a result, it can substantially underestimate the true radiation dose absorbed by smaller pediatric patients, particularly infants and young children.

To address this limitation, the Size-Specific Dose Estimate (SSDE) framework was developed. SSDE adjusts radiation dose estimates according to patient size using measurements such as the Water-Equivalent Diameter (Dw), providing a more accurate representation of absorbed tissue dose. Although SSDE is widely accepted in medical physics, its practical application in routine hospital dose auditing remains limited.

The study focuses on analyzing the discrepancy between machine-reported CTDIvol and patient-specific SSDE in pediatric head CT scans. It also investigates how image noise varies with patient head size and scanning techniques. Results show that fixed exposure protocols can lead to significant underestimation of actual absorbed dose, particularly in smaller children, while tube current modulation (TCM) helps adjust exposure according to patient size.

Using automated analysis software, pediatric patients were categorized into three head-size groups: small (Dw < 13 cm), medium (13–15 cm), and large (>15 cm). The findings reveal that CT console readings underestimate true absorbed dose by approximately 14.64% in small heads, 6.39% in medium heads, and become nearly accurate (−0.55% deviation) in larger heads whose dimensions approach the standard 16-cm reference phantom.

The research was conducted as a retrospective study of 50 pediatric patients (0–14 years) at RSUP Dr. Kariadi. Patients were divided into groups scanned with and without Tube Current Modulation (TCM). All scans were performed using a Siemens SOMATOM Definition AS+ with standardized imaging protocols.

Conclusion

This study provides a definitive quantitative assessment of the systematic dosimetric inaccuracies prevalent in current commercial pediatric head CT protocols. By establishing the normalized Metric Discrepancy (?%) index as a standard for quality assurance, we demonstrated that reliance on standardized, 16-cm reference phantoms leads to a systematic underestimation of the radiation burden by an average of +14.64% for pediatric patients within the Small Head Cluster (Dw < 13 cm). Our data confirms that within this specific morphological group, the scanner-recorded CTDIvol (Mean 31.89 mGy) significantly masks the true energy deposition represented by the SSDE (Mean 36.56 mGy). The robustness of our findings is supported by a high regression coefficient (R2 = 0.942), confirming that 94.2% of the observed estimation error is directly attributable to patient-specific anatomical variations (Dw). By integrating the IndoseCT software module into the institutional PACS environment, we have successfully replaced legacy, machine-centric dose reporting with a proactive, automated workflow. This system effectively triggers an immediate Dose Alert Notification for any exam exceeding a +10% deviation threshold, ensuring that clinicians can intercept and mitigate radiation risks in real-time. In summary, this research replaces generalized factory assumptions with precise, evidence-based dosimetry. By transitioning to this automated, size-specific auditing framework, radiology departments can move beyond passive registry maintenance toward an active, precision-based protection strategy that aligns with the highest clinical standards of the ALARA principle for the most vulnerable pediatric patient populations.

References

[1] Alhailiy, A., et al. Reporting diagnostic reference levels for paediatric patients undergoing brain computed tomography. Tomography. 2023, 9(2), 543–555. https://doi.org/10.3390/tomography9020045 [2] Garba, I., Engel-Hills, P. Paediatric diagnostic reference levels for common computed tomography procedures: A systematic review. Radiography. 2024, 30(1), 112–120. https://doi.org/10.1016/j.radi.2023.10.012 [3] Anam, C., et al. An improved method for automated calculation of the water-equivalent diameter for estimating size-specific dose in CT. Journal of Applied Clinical Medical Physics. 2021, 22(3), 234–241. https://doi.org/10.1002/acm2.13189 [4] Behr, F., et al. Intra-individual variation in radiation dose of pediatric head CT: implications for dose optimization. Neuroradiology. 2026, 68(2), 145–152. https://doi.org/10.1007/s00234-025-03451-2 [5] Bos, D., et al. Diagnostic reference levels for indication-based CT categories in pediatric CT: data from an international registry. European Radiology. 2025, 35(4), 2891–2902. https://doi.org/10.1007/s00330-024-10982-x [6] Bouchareb, Y., et al. Establishing diagnostic reference levels for paediatric CT imaging: A multi-centre study. Healthcare. 2025, 13(2), 178–190. https://doi.org/10.3390/healthcare13020178 [7] Cadavid, L., et al. Setting up regional diagnostic reference levels for pediatric computed tomography in Latin America: Preliminary results, challenges and the work ahead. Pediatric Radiology. 2023, 53(7), 1105–1115. https://doi.org/10.1007/s00247-023-05632-1 [8] Kamdem, E. F., et al. Estimation of diagnostic reference levels for pediatric head computed tomography in Yaoundé. Radiation Protection Dosimetry. 2023, 199(5), 412–420. https://doi.org/10.1093/rpd/ncad022 [9] Kuwahara, H., et al. Dose-dependent analysis of image quality in pediatric head CT scans across different scanners to optimize clinical protocols using phantom-based assessment. Tomography. 2025, 11(1), 34–45. https://doi.org/10.3390/tomography11010004 [10] Lawson, M., et al. Comparison of organ and effective dose estimations from different Monte Carlo simulation-based software methods in infant CT and comparison with direct phantom measurements. Journal of Applied Clinical Medical Physics. 2022, 23(8), e13672. https://doi.org/10.1002/acm2.13672 [11] Abuhaimed, A., Martin, C. J. Estimation of size-specific dose estimates (SSDE) for paediatric and adult patients based on a single slice. Physica Medica. 2020, 71, 102–110. https://doi.org/10.1016/j.ejmp.2020.02.011 [12] Zhang, R., et al. Comparison of organ and effective dose estimations from CT dosimetry software with physical measurements in a pediatric phantom. Journal of Applied Clinical Medical Physics. 2022, 23(11), e13781. https://doi.org/10.1002/acm2.13781 [13] Monteiro, V. C., et al. A study of SSDE, CTDI, and DLP dose indexes for establishing diagnostic reference levels in pediatric CT exams. Radiation Physics and Chemistry. 2025, 226, 112104. https://doi.org/10.1016/j.radphyschem.2024.112104 [14] Payne, S., Badawy, M. Comparison of average water equivalent diameter values between CTContour and vendor-specific estimates in CT dosimetry. Physica Medica. 2023, 112, 102634. https://doi.org/10.1016/j.ejmp.2023.102634 [15] Poosiri, S., Chuboonlap, K., Kaewlaied, N. An age-based size-specific dose estimate for pediatric computed tomography head examinations performed at Songklanagarind Hospital, Thailand, from 2017 to 2019. Applied Sciences. 2024, 14(3), 1124. https://doi.org/10.3390/app14031124 [16] Priyanka, Kadavigere, R., Sukumar, S. Low dose pediatric CT head protocol using iterative reconstruction techniques: A comparison with standard dose protocol. Radiologe. 2023, 63(4), 295–301. https://doi.org/10.1007/s00117-023-01124-y [17] Sulemana, H., Mumuni, A. N., Abubakari, I. O. M. Establishment of baseline size-specific dose estimate (SSDE) for paediatric head computed tomography (CT) examinations. Egyptian Journal of Radiology and Nuclear Medicine. 2024, 55(1), 44–53. https://doi.org/10.11648/j.ejrnm.20245501.14 [18] Tonkopi, E., et al. Optimizing radiation dose and image quality in pediatric head CT: a comparison between children’s and regional hospitals in the Canadian province of Nova Scotia. Physica Medica. 2026, 118, 103210. https://doi.org/10.1016/j.ejmp.2025.103210 [19] Turner, A. C., et al. The feasibility of patient size–corrected CT dose estimates. Journal of Applied Clinical Medical Physics. 2018, 19(2), 264–272. https://doi.org/10.1002/acm2.12260 [20] Yang, F., Gao, L. Age-based diagnostic reference levels and achievable doses for paediatric CT: A survey in Shanghai, China. Journal of Radiological Protection. 2024, 44(2), 021503. https://doi.org/10.1088/1361-6498/ad3412 [21] Zadeh, P. T., Mahmoudi, F., Rezaeian, A., Gholami, M. Over-scanning in pediatric head CT: prevalence, dosimetric impact, and associated cancer risks. European Journal of Radiology. 2026, 194, 112450. https://doi.org/10.1016/j.ejrad.2025.112450 [22] Zellner, M., et al. Radiation dose optimisation in paediatric head CT using attenuation-based auto prescription. Scientific Reports. 2025, 15(1), 4321. https://doi.org/10.1038/s41598-025-54321-w

Copyright

Copyright © 2026 Evi Setiawati, Bella Diani Safitri, Agus Margiantono. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.