Balancing Accuracy and Safety: Comparing Standard vs. Low-Dose CT Scans in Pediatric Age Group

1. A Multidisciplinary Assessment of Standard Low Dose Versus Ultra-low Dose Chest CT Scans for Pectus Excavatum Imaging
https://doi.org/10.1016/j.jpedsurg.2024.162026
Research Question:
Can ultra-low dose chest CT imaging accurately assess pectus excavatum severity compared to standard low-dose CT, using Haller Index (HI) and Correction Index (CI), as interpreted by both surgeons and radiologists?
Design:
Prospective, within-subjects comparison study.
Setting:
Single-center, multidisciplinary academic pediatric institution.
Participants:
32 patients (mean age 14.1 years, range 11–18), each underwent paired standard low-dose (CARE) and ultra-low dose (ULTRA) chest CT scans. Measurements were performed independently by 3 pediatric surgeons and 4 pediatric radiologists.
Outcomes:
Agreement in measurement of HI and CI between CARE and ULTRA CT protocols.
Results:
• Intraclass correlation coefficient (ICC):
– HI: 0.89 (good reliability)
– CI: 0.91 (excellent reliability)
• Surgeon agreement was slightly higher for HI (0.89 vs. 0.88), while radiologists had slightly higher CI agreement (0.92 vs. 0.90).
• Bland-Altman plots revealed no systematic bias between CARE and ULTRA approaches for either index.
• Imaging quality was sufficient for surgical planning despite the lower radiation dose.
Commentary:
This study supports the use of ultra-low dose CT imaging as a reliable alternative to standard protocols for assessing pectus excavatum severity. The high inter-rater agreement across both specialties for both indices emphasize the clinical feasibility of adopting reduced-radiation protocols. Particularly valuable in pediatrics, this shift aligns with ALARA principles and may significantly lower cumulative radiation exposure without compromising diagnostic utility. Limitations included potential variation in scan timing during respiration and minor differences in display resolution between interpreters.

2. Low Dose Pediatric CT Head Protocol using Iterative Reconstruction Techniques: A Comparison with Standard Dose Protocol
https://doi.org/10.1007/s00062-023-01361-4
Research Question:
Can a low-dose pediatric head CT protocol using iterative reconstruction (iDose4) achieve diagnostic image quality comparable to the standard protocol while significantly reducing radiation exposure?
Design:
Prospective, single-center study.
Setting:
Tertiary care academic hospital in India.
Participants:
214 pediatric patients aged <1 year and 1–5 years:
• Standard dose (SD): 143 patients
• Low dose (LD): 71 patients
Indications included seizures, traumatic brain injury, hydrocephalus, ventriculoperitoneal shunt malfunction, and meningitis.
Outcomes:
Radiation dose indices and qualitative/quantitative image quality (IQ) comparison between SD and LD protocols.
Results:
• Radiation dose reduction:
– <1 year: 68% reduction in effective dose (0.91 mSv vs. 2.90 mSv)
– 1–5 years: 58% reduction in effective dose (1.16 mSv vs. 2.78 mSv)
• Qualitative image quality:
– No significant difference in subjective image noise, streak artifacts, or overall image quality.
– Gray-white matter differentiation (GWMD) was significantly better in the LD group.
• Quantitative image quality:
– Higher attenuation and contrast-to-noise ratio (CNR) in LD protocol.
– No significant difference in image noise or signal-to-noise ratio (SNR) between protocols.
– Excellent interobserver agreement (ICC = 0.9) for all IQ measures.
Commentary:
This study validates the clinical feasibility of using age-specific low-dose pediatric CT head protocols combined with iterative reconstruction (iDose4 level 3) to reduce radiation without compromising diagnostic accuracy. The findings are particularly relevant for repeat imaging scenarios (e.g., hydrocephalus or head trauma follow-up) and align with ALARA principles. Importantly, qualitative and quantitative assessments both support the reliability of LD imaging, while noting limitations in not directly evaluating diagnostic accuracy for specific pathologies.

3. Performance of Lung Nodule Computer-Aided Detection Systems on Standard-Dose and Low-Dose Pediatric CT Scans: An Intraindividual Comparison
Hardie RC, Trout AT, Dillman JR, et al.
AJR Am J Roentgenol. 2025;224:e2431972.
https://doi.org/10.2214/AJR.24.31972
Research Question:
How do adult-trained lung nodule computer-aided detection (CAD) systems perform on low-dose versus standard-dose pediatric chest CT scans?
Design:
Retrospective intraindividual comparison using data from a prior prospective study
Setting:
Single pediatric academic center
Participants:
73 pediatric patients (mean age 14.7 years, range 4–20 years) each underwent a standard-dose and a low-dose chest CT during the same imaging session. A total of 247 lung nodules (3–30 mm) were analyzed using two CAD systems (FlyerScan and MONAI).
Outcomes:
Detection sensitivity of adult-trained CAD systems on pediatric chest CT scans at standard and low dose
Results:
Both CAD systems demonstrated reduced sensitivity on low-dose scans. FlyerScan sensitivity dropped from 76.9% to 66.8%, and MONAI from 67.6% to 62.3% for nodules 3–30 mm in size (operating at two false-positives per scan). The performance decline was more pronounced for nodules <5 mm. MONAI was generally more robust to dose reductions, especially for larger nodules, whereas FlyerScan showed greater sensitivity on standard-dose scans but more degradation at lower doses.
Commentary:
This study underscores the trade-off between radiation dose and CAD system performance in pediatric lung nodule detection. Systems trained on adult datasets underperformed on low-dose pediatric CT, particularly for smaller nodules. These findings highlight the need for caution when applying CAD in low-dose pediatric protocols and emphasize the future need for pediatric-specific CAD development and training datasets. The small cohort and limited variety of CAD systems were noted limitations in this study.