Inter- and intra-individual radiation dose variability in oncologic chest and abdominal computed tomography

Abstract

Purpose

Computed tomography (CT) plays a central role in oncologic imaging, yet repeated examinations contribute substantially to cumulative radiation exposure. This study aimed to evaluate inter- and intra-individual radiation dose variability in chest and abdominal CT and the impact of CT device model and protocol standardization.

Materials and Methods

In this retrospective single-center study, 42441 CT scans from 4986 adult oncologic patients were analyzed. Dose metrics (CTDIvol, DLP, SSDE, effective dose) were extracted using automated dose monitoring. Inter- and intra-individual radiation dose variability was assessed across four CT device models and various protocol subtypes. Intra-individual radiation dose variability was calculated relative to the lowest dose per patient and compared across CT devices and protocol subtypes.

Results

Radiation dose varied substantially between devices, with CTDIvol differences of up to 2.4-fold in chest CT (2.98–7.26 mGy) and 1.7-fold in abdominal CT (5.26–8.77 mGy). The median intra-individual radiation dose variability was 93.7% (IQR 19.8–142.0%) in non-contrast chest CT, 66.3% (31.1–105.4%) in contrast-enhanced chest CT, 19.8% (11.8–32.3%) in non-contrast abdominal CT, and 28.2% (16.6–41.0%) in contrast-enhanced abdominal CT. When consecutive scans were performed on the same scanner, intra-individual radiation dose variability decreased to 14.7% (IQR 8.1–33.1%), 18.1% (9.5–31.3%), 11.7% (7.8–19.8%), and 15.3% (8.3–24.7%), respectively, indicating substantial device-specific effects.

Conclusion

Significant radiation dose variability persists in oncologic CT, both between and within patients, despite the use of standardized protocols. Device-adapted dose management and consistent device use may improve dose consistency, support optimization in oncologic imaging, and reduce radiation exposure.

Key Points

• Substantial radiation dose variability persists across CT devices despite protocol standardization.

• Intra-individual radiation dose variability is significant and highest in non-contrast chest CT.

• Consistent use of the same scanner reduces intra-individual radiation dose variability significantly.

• Internal diagnostic reference levels may improve radiation dose consistency and minimize exposure.

Citation Format

• Moradians AF, Rosok D, Serger R et al. Inter- and intra-individual radiation dose variability in oncologic chest and abdominal computed tomography. Rofo 2026; DOI 10.1055/a-2786-2534

Zusammenfassung

Ziel

Die Computertomografie (CT) spielt eine zentrale Rolle in der onkologischen Bildgebung, jedoch tragen wiederholte Untersuchungen wesentlich zu einer kumulativen Strahlenexposition bei. Ziel dieser Studie war es, die inter- und intraindividuelle Variabilität der Strahlendosis bei Thorax- und Abdomen-CTs zu untersuchen und den Einfluss des CT-Gerätemodells sowie der Standardisierung der Protokolle zu analysieren.

Material und Methoden

In dieser retrospektiven Einzelzentrumsstudie wurden 42441 CT-Scans von 4986 erwachsenen onkologischen Patient:innen analysiert. Der volumengewichtete CT-Dosisindex (CTDIvol), das Dosis-Längen-Produkt (DLP), das Size-Specific Dose Estimate (SSDE) und die effektive Dosis (ED) wurden mittels eines automatisierten Dosisüberwachungssystems extrahiert. Die inter- und intraindividuelle Strahlendosisvariabilität wurde zwischen vier CT-Gerätemodellen und verschiedenen Protokollsubtypen evaluiert. Die intraindividuelle Dosisvariabilität wurde jeweils im Verhältnis zur niedrigsten individuellen Dosis berechnet und geräte- sowie protokollübergreifend verglichen.

Ergebnisse

Die Strahlendosis variierte deutlich zwischen den Geräten, mit bis zu 2,4-fachen Unterschieden des CTDIvol bei Thorax-CTs (2,98–7,26 mGy) und bis zu 1,7-fachen Unterschieden bei Abdomen-CTs (5,26–8,77 mGy). Die mediane intraindividuelle Dosisvariabilität betrug 93,7% (IQR 19,8–142,0%) für native Thorax-CTs, 66,3% (31,1–105,4%) für Thorax-CTs mit Kontrastmittel, 19,8% (11,8–32,3%) für native Abdomen-CTs und 28,2% (16,6–41,0%) für Abdomen-CTs mit Kontrastmittel. Wurden die Folgeuntersuchungen jeweils am gleichen Scanner durchgeführt, verringerte sich die intraindividuelle Variabilität auf 14,7% (IQR 8,1–33,1%), 18,1% (9,5–31,3%), 11,7% (7,8–19,8%) bzw. 15,3% (8,3–24,7%), was auf substanzielle gerätespezifische Faktoren hindeutet.

Schlussfolgerung

Trotz standardisierter Untersuchungsprotokolle besteht bei der onkologischen Computertomografie eine signifikante Variabilität der Strahlendosis, sowohl zwischen als auch innerhalb einzelner Patient:innen. Ein gerätespezifisches Dosismanagement und die konsequente Nutzung gleicher CT-Geräte könnten die Dosiskonstanz verbessern, zur Optimierung der onkologischen Bildgebung beitragen und die Strahlenexposition reduzieren.

Kernaussagen

• Trotz standardisierter Protokolle besteht eine substanzielle Strahlendosisvariabilität zwischen unterschiedlichen CT-Geräten.

• Die intraindividuelle Strahlendosisvariabilität ist signifikant und am höchsten bei nativen Thorax-CTs.

• Eine konsequente Nutzung des gleichen CT-Geräts verringert die intraindividuelle Strahlendosisvariabilität deutlich.

• Institutionsinterne Referenzwerte können zur Verbesserung der Dosiskonstanz und Reduktion der Strahlenexposition beitragen.

Keywords

Computed Tomography - Oncologic Imaging - Dose Monitoring - Radiation Dose Variability - Diagnostic Reference Levels (DRLs) - CT Protocol Optimization

Introduction

The increasing prevalence of oncologic diseases has led to a growing need for accurate and rapid radiological imaging. In recent years, computed tomography (CT) has become indispensable in oncologic imaging due to its superimposition-free cross-sectional visualization with ever-increasing accuracy. Between 2007 and 2021, the number of CT scans in Germany increased by 40% in the outpatient sector and even doubled in the inpatient sector, where at least half of all examinations are still performed, highlighting the growing reliance on this technology [1] [2] [3]. Consequently, approximately 13 million CT examinations were conducted in Germany in 2020 alone [3].

While technological advances have reduced radiation doses, the overall population-wide exposure continues to rise due to the increasing number of CT scans [4] [5]. Therefore, the average effective dose (ED) from radiologic examinations in Germany reached 1.6 mSv per inhabitant in 2018, with markedly higher cumulative doses in patients undergoing repeated oncologic staging scans [1] [3] [6]. CT accounts for a disproportionate share of medical radiation exposure, especially when compared to non-ionizing modalities such as MRI [1] [3] [4] [7]. Repeated radiation exposure is associated with an increased risk of radiation-induced malignancies. According to the International Commission on Radiological Protection (ICRP), a nominal risk coefficient of 5% per Sievert is applied for stochastic effects including cancer mortality for radiation protection purposes [8]. Furthermore, based on the linear non-threshold (LNT) model, even doses of 100 mSv or less may carry a stochastic risk [8]. Oncologic patients are particularly vulnerable due to the frequent use of CT in staging and surveillance.

Beyond the rising number of scans, significant variability in radiation dose has been reported both between and within institutions, despite the existence of national dose reference levels (DRLs) [1] [4] [9] [10] [11] [12]. Inconsistent monitoring of radiation dose and image quality may result in diagnostic uncertainty and unnecessary radiation exposure, particularly in patients undergoing serial CT follow-ups. Consequently, systematic evaluation of protocol adherence and dose consistency is essential to ensure radiation dose minimization in line with the “as low as reasonably achievable” (ALARA) principle [13]. This is particularly relevant in oncologic imaging, where repeated examinations require consistent image quality over time.

This study aims to assess inter- and intra-individual radiation dose variability in oncologic chest and abdominal CT at a high-volume center, focusing on the impact of CT devices and protocol standardization. Furthermore, it examines the potential benefits of establishing standardized, institution-specific protocols and DRLs, contributing to optimized and consistent radiation dose application in the clinical routine.

Materials and Methods

Study Design and Patient Cohort

Ethical approval for this retrospective, single-center observational study was granted by the University Hospital Essen ethics committee (ID: 23-11510-BO) and the requirement for written informed consent was waived. All adult patients who underwent at least two CT scans of the chest or abdomen for oncologic staging or follow-up between January 2021 and August 2023 were identified using the radiology information system (RIS) database (Centricity version 5.0, GE Healthcare IT, Barrington, Illinois, USA). Patients under 18 years of age, low-dose chest CT scans, single examinations, and non-staging-related scans were excluded to ensure cohort consistency and reduce disproportionate variability. The cohort was stratified by scan region (chest or abdomen), protocol (non-contrast or contrast-enhanced; single- or multiphase), and CT device. No formal sample size calculation was performed due to the retrospective nature of the study. However, the large number of scans included over the study period ensured sufficient statistical power for subgroup analyses and robust descriptive evaluation.

CT Devices

CT scans were performed on four different CT device models in a single institution at a high-volume center: SOMATOM Definition AS+, SOMATOM Definition Flash, SOMATOM Force (n=2) and SOMATOM Definition Edge (all: Siemens Healthcare, Forchheim, Germany). Detailed chest and abdominal CT protocols are listed in [Table 1]. Device-specific protocol use was considered in the inter-individual radiation dose analysis.

Table 1 Technical and scan parameters for CT devices and protocols. * All scanners: Siemens Healthcare, Forchheim, Germany. ** Where multiple values are listed (rotation time, tube voltage peak, and tube current), they correspond to the CT protocols (in order): chest non-contrast, chest contrast, abdomen non-contrast, abdomen contrast.

Scan parameter
CT device*	SOMATOM Force	SOMATOM Definition Flash	SOMATOM Definition AS+	SOMATOM Definition Edge
Number of rows	192 (2×96)	128 (2×64)	64w/128 channels	64w/128 channels
Detector width (mm)	52.5	38.4	19.2	38.4
Collimation	0.6	0.6	0.6	0.6
Pitch	0.6	0.6	0.6	0.6
Reconstructed slice thickness (mm)	0.6	0.6	0.6	0.6
Rotation time (s)**	0.25/0.25/0.5/0.5	0.28/0.28/0.5/0.5	0.3/0.3/0.5/0.5	0.28/0.28/0.5/0.5
Tube voltage selection (ATVS)	on	on	on	on
Tube voltage peak (kVp)**	90–130/90–130/100–140/100–140	100–130/100–130/110–140/110–140	100–130/100–130/120–140/120–140	100–130/100–130/120–140/120–140
Tube current modulation (ATCM)	on	on	on	on
Minimum tube current (mAs)**	22.88/18.64/41.72/22.75	26.47/26.52/76.47/41.96	25.57/22.35/50.62/56.44	24.38/23.00/58.40/22.23
Median tube current (mAs)**	49.13/45.98/81.71/94.41	40.45/56.33/85.41/83.01	40.72/61.74/78.03/124.20	77.08/87.96/97.60/104.20
Maximum tube current (mAs)**	110.90/209.80/128.40/260.70	71.88/165.10/97.99/248.30	107.40/209.20/195.60/296.80	154.60/254.60/198.50/319.50

Radiation Dose Assessment

Radiation dose parameters were extracted using automated dose monitoring software (Radimetrics Enterprise Platform, Bayer Healthcare, Leverkusen, Germany), which captures patient demographics (sex, age, and BMI) and technical data from the DICOM header and structured dose report. Quantitative dose metrics included the volume-weighted CT dose index (CTDIvol), dose-length product (DLP), and size-specific dose estimate (SSDE). Effective dose (ED) and organ doses were calculated based on Monte Carlo simulations, based on weighting factors recommended in the 103rd ICRP publication [8] [14]. To evaluate radiation dose levels, DRLs for CT imaging as established by the German Federal Office for Radiation Protection (Bundesamt für Strahlenschutz, BfS) were used as benchmarks in inter-individual radiation dose analyses. These DRLs, defined as the 75^th percentile of representative dose distributions across institutions, provide a national reference for dose optimization [15].

Statistical Analysis

Statistical analysis was performed using GraphPad Prism (version 10.4.2., GraphPad Software, Boston, USA). Descriptive statistics, including means and standard deviations, medians and interquartile ranges (IQRs), ranges and coefficients of variation, were calculated for patient demographics and dose metrics. Radiation dose variability was assessed using IQRs and quartile ratios (QR, 75th-to-25th percentile ratio). All analyses were repeated across protocols and CT device subgroups. Data distribution was evaluated using D’Agostino and Pearson and Shapiro-Wilk normality tests. Demographic parameters were recorded for the whole study population. However, the statistical analysis did not include detailed dose values, device-specific dose distribution, or intra-individual dose variability of multiphase abdominal scans due to heterogeneity in acquisition schemes and clinical indications. Intra-individual radiation dose variability was evaluated by identifying the lowest CTDIvol, DLP, SSDE, and ED per patient for each protocol category. All subsequent CT scans using the same protocol were compared to this reference, and deviations were calculated. Analyses were repeated for each CT device model to identify potential dose optimization margins based on achievable minimum doses, in line with the ALARA principle. Group comparisons were performed using the Mann-Whitney U or Kruskal-Wallis test, followed by Dunn’s post hoc multiple comparison test where applicable. The Wilcoxon signed-rank test was used to compare non-independent subgroups. A two-sided p < 0.05 was considered statistically significant.

Results

Patient Characteristics

Out of 73513 scans from 22067 patients performed during the study period, 42441 scans from 4986 adult patients met the inclusion criteria. The study cohort comprised 47.2% (2349/4986) female patients. The median age at first examination was 63.4 years (IQR 55.1–71.4), ranging from 18.1 to 95.2 years. The median BMI was 25.5 kg/m² (IQR 22.7–29.1). Patients underwent a total of 2 to 16 CT scans, including examinations of the chest, the abdomen, or both. The median number of scans was 5 (3–8) for each region. The median follow-up interval spanned 400 days (IQR 173–708). All distributions deviated significantly from normality (all p < 0.0001).

Chest CT accounted for 38.2% (16231/42441) of all scans and was performed in 79.3% (3952/4986) of patients. Among these, 5.3% (866/16231) were non-contrast scans, performed in 7.8% (307/3952) of patients, and 94.7% (15365/16231) were contrast-enhanced scans, conducted in 96.5% (3814/3952) of patients. Abdominal CT represented 61.8% (26210/42441) of all scans. Of these, 54.1% (14177/26210) were single-phase examinations performed in 71.4% (3560/4986) of patients, while 45.9% (12033/26210) were multiphase examinations performed in 32.4% (1617/4986) of patients. Among the single-phase abdominal scans, 3.2% (448/14177) were non-contrast examinations performed in 5.3% (187/3560) of patients, while 96.8% (13729/14177) were contrast-enhanced examinations performed in 97.5% (3471/3560) of patients ([Table 2], [Fig. 1]).

Table 2 Patient characteristics and CT scan distribution. Data are given as absolute numbers and percentages or as median with interquartile range (IQR). * All scanners: Siemens Healthcare, Forchheim, Germany.

Parameter	n (%)
Number of CT scans
Total	42441
Chest
All CT devices*	16231 (38.2%)
SOMATOM Force	8229 (19.4%)
SOMATOM Definition Edge	6001 (14.1%)
SOMATOM Definition AS+	1605 (3.8%)
SOMATOM Definition Flash	396 (0.9%)
Abdomen (single-phase)
All CT devices*	14177 (33.4%)
SOMATOM Force	7570 (17.8%)
SOMATOM Definition Edge	4883 (11.5%)
SOMATOM Definition AS+	1406 (3.3%)
SOMATOM Definition Flash	318 (0.7%)
Abdomen (multiphase)
All CT devices*	12033 (28.4%)
SOMATOM Force	4376 (10.3%)
SOMATOM Definition Edge	6061 (14.3%)
SOMATOM Definition AS+	1147 (2.7%)
SOMATOM Definition Flash	449 (1.1%)
Number of Patients
Total	4986
Male	2637 (52.8%)
Female	2349 (47.2%)
Age at examination (years)	63.4 (55.1–71.4)
BMI (kg/m²)	25.5 (22.7–29.1)
CT scans per patient	5 (3–8)
Individual examination period (days)	400 (173–708)

Fig. 1 Flowchart illustrating the selection process and distribution of the study population.

The five most prevalent oncologic CT indications in the total cohort were non-small cell lung cancer (NSCLC) (21.7%, 9210/42441), malignant melanoma (11.6%, 4923/42441), breast cancer (7.3%, 3098/42441), uveal melanoma (3.8%, 1612/42441), and hepatocellular carcinoma (3.7%, 1570/42441). Distribution patterns varied by sex: in male patients, the most prevalent indications were NSCLC (21.8%, 4808/22106), malignant melanoma (13.1%, 2895/22106), hepatocellular carcinoma (5.6%, 1240/22106), uveal melanoma (4.3%, 950/22106), and renal cell carcinoma (4.2%, 934/22106). In female patients, leading indications were NSCLC (21.6%, 4389/20335), breast cancer (15.2%, 3084/20335), malignant melanoma (9.9%, 2022/20335), leiomyosarcoma (4.0%, 819/20335), and uveal melanoma (3.3%, 673/20335).

Radiation Dose Distribution Across CT Devices

In this section, all dose metrics are consistently reported in the following order, unless otherwise stated: CTDIvol (mGy), DLP (mGycm), and ED (mSv), with values given as median [IQR]. This structure applies to both chest and abdomen CT analyses.

Chest CT

Radiation doses in chest CT varied markedly by scanner and protocol. The overall values for CTDIvol, DLP, and ED in non-contrast chest scans were as follows: 5.15 mGy [3.24–7.69], 183 mGycm [112–270], and 3.56 mSv [1.93–4.80]. The corresponding scanner-specific doses ranged from 2.98 mGy [2.18–3.62], 95 mGycm [75–126], and 1.71 mSv [1.38–2.07] on the SOMATOM Force to 7.26 mGy [5.67–9.26], 249 mGycm [193–317], and 4.58 mSv [3.95–5.52] on the SOMATOM Definition Edge. Therefore, dose differences between scanners in terms of CTDIvol reached a maximum factor of 2.44, equivalent to an 144% increase (p < 0.0001). Radiation dose variability, quantified by the QR, ranged from 1.59 (SOMATOM Definition AS+) to 1.66 (SOMATOM Force), indicating moderate variability within scanners.

In contrast-enhanced chest scans, the overall doses were (CTDIvol, DLP, ED): 3.88 mGy [2.67–5.82], 139 mGycm [96–207], and 2.35 mSv [1.61–3.57]. Again, the lowest corresponding values were achieved with the SOMATOM Force (2.78 mGy [2.11–3.66], 100 mGycm [75–134], 1.64 mSv [1.38–1.98]), whereas the highest were recorded on the SOMATOM Definition Edge (6.33 mGy [4.91–8.77], 224 mGycm [173–314], 3.85 mSv [3.34–5.25]). The maximum inter-scanner differences in CTDIvol reached a factor of 2.28 (128% increase, p < 0.0001), while radiation dose variability ranged from a QR of 1.57 on the SOMATOM Definition Flash to 1.79 on the SOMATOM Definition Edge, indicating scanner-dependent variability ([Table 3]). [Fig. 2] illustrates the radiation doses (represented by CTDIvol values) across scanners and their relationship to the national DRL of 8 mGy for chest CT. The Kruskal-Wallis test confirmed significant differences in CTDIvol distribution between scanners (p < 0.05).

Table 3 Radiation dose metrics for chest and abdominal CT across CT devices and protocol types. Data are given as median with interquartile range (IQR). CTDIvol: Volume-weighted CT dose index; DLP: Dose-length product; SSDE: Size-specific dose estimate; ED: Effective dose; QR: Quartile ratio of the 75^th to the 25^th percentile; CT devices: Force: SOMATOM Force; Definition Flash: SOMATOM Definition Flash; Definition AS+: SOMATOM Definition AS+; Definition Edge: SOMATOM Definition Edge (all: Siemens Healthcare, Forchheim, Germany).

Protocol	CT device	CTDIvol (mGy)			DLP (mGycm)		SSDE (mGy)		ED (mSv)
		Median	IQR	QR	Median	IQR	Median SSDE	IQR	Median ED	IQR
Chest	Total	3.91	2.68–5.88	2.19	140	96–209	4.76	3.32–7.41	2.37	1.62–3.60
Non-contrast	Total	5.15	3.24–7.69	2.37	183	112–270	7.03	4.11–9.80	3.56	1.93–4.80
	Force	2.98	2.18–3.62	1.66	95	75–126	3.66	2.97–4.29	1.71	1.38–2.07
	Definition Flash	4.53	4.14–6.74	1.63	163	152–263	5.58	5.00–8.32	2.62	2.32–3.62
	Definition AS+	3.59	2.84–4.51	1.59	120	94–145	4.50	3.59–4.96	2.10	1.69–2.50
	Definition Edge	7.26	5.67–9.26	1.63	249	193–317	9.39	7.85–11.02	4.58	3.95–5.52
Contrast	Total	3.88	2.67–5.82	2.18	139	96–207	4.73	3.31–7.34	2.35	1.61–3.57
	Force	2.78	2.11–3.66	1.73	100	75–134	3.37	2.76–4.11	1.64	1.38–1.98
	Definition Flash	4.21	3.39–5.31	1.57	150	119–188	5.20	4.49–6.23	2.58	2.24–3.04
	Definition AS+	4.34	3.45–5.46	1.58	148	119–190	5.51	4.72–6.57	2.78	2.43–3.23
	Definition Edge	6.33	4.91–8.77	1.79	224	173–314	8.03	6.73–10.46	3.85	3.34–5.25
Abdomen	Total	5.75	4.41–7.64	1.73	291	216–400	7.33	5.94–9.10	4.13	3.26–5.24
Non-contrast (single-phase)	Total	7.58	5.86–10.17	1.74	517	389–691	10.65	8.74–13.95	7.88	6.35–10.34
	Force	6.80	5.66–9.26	1.64	472	373–625	9.60	8.17–12.22	7.08	6.08–9.28
	Definition Flash	5.26	4.67–5.54	1.19	357	328–364	6.04	5.67–6.61	4.64	4.03–5.84
	Definition Edge	8.77	7.28–10.99	1.51	589	475–749	13.21	11.28–15.66	9.43	7.93–11.21
	Definition AS+	6.69	5.49–9.05	1.65	476	376–656	8.98	7.22–10.41	6.75	5.62–8.75
Contrast (single-phase)	Total	5.70	4.41–7.67	1.74	286	213–390	7.25	5.89–8.96	4.09	3.24–5.12
	Force	5.05	3.95–6.56	1.66	259	193–346	6.48	5.40–7.84	3.58	2.93–4.39
	Definition Flash	5.58	4.56–7.02	1.54	285	221–373	6.95	6.02–7.98	4.07	3.55–4.85
	Definition Edge	6.20	4.83–8.18	1.69	307	233–425	7.85	6.56–9.57	4.46	3.75–5.45
	Definition AS+	8.11	6.30–11.11	1.76	404	303–574	9.89	8.37–12.60	5.94	5.00–7.62

Fig. 2 Device-specific volume-weighted CT dose index (CTDIvol) for chest CT (a) and single-phase abdominal CT (b) across CT devices, by protocol type. Total: all CT devices; Force: SOMATOM Force; Definition Flash: SOMATOM Definition Flash; Definition AS+: SOMATOM Definition AS+; Definition Edge: SOMATOM Definition Edge (all: Siemens Healthcare, Forchheim, Germany). The red line indicates the national German diagnostic reference level (DRL) of 8 mGy for chest CT (a) and 12 mGy for abdominal CT (b) in adult patients [15]. Whiskers are defined according to Tukey, extending to 1.5× the interquartile range. Asterisk indicates significant difference: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Abdomen CT

The overall doses in non-contrast abdominal scans were as follows (CTDIvol, DLP, ED): 7.58 mGy [5.86–10.17], 517 mGycm [389–691], and 7.88 mSv [6.35–10.34]. As with the findings on chest CT, these varied by scanner, ranging from 5.26 mGy [4.67–5.54], 357 mGycm [328–364], and 4.64 mSv [4.03–5.84] on the SOMATOM Definition Flash to 8.77 mGy [7.28–10.99], 589 mGycm [475–749], and 9.43 mSv [7.93–11.21] on the SOMATOM Definition Edge. Consequently, the maximum inter-scanner difference in CTDIvol reached a factor of 1.67, equivalent to a 67% increase. However, this difference was not statistically significant (p = 0.4017), likely due to the limited number of scans for this specific protocol and scanner combination. Radiation dose variability, as quantified by the QR, ranged from 1.19 (SOMATOM Definition Flash) to 1.65 (SOMATOM Definition AS+).

On contrast-enhanced abdominal scans, the overall doses were (CTDIvol, DLP, ED): 5.70 mGy [4.41–7.67], 286 mGycm [213–390], and 4.09 mSv [3.24–5.12]. The lowest corresponding values were recorded on the SOMATOM Force (5.05 mGy [3.95–6.56], 259 mGycm [193–346], 3.58 mSv [2.93–4.49]), while the SOMATOM Definition AS+ showed the highest (8.11 mGy [6.30–11.11], 404 mGycm [303–574], 5.94 mSv [5.00–7.62]). The maximum inter-scanner difference in CTDIvol was 1.61-fold (61% increase, p < 0.0001), with dose variability ranging from a QR of 1.54 (SOMATOM Definition Flash) to 1.76 (SOMATOM Definition AS+) ([Table 3]). Radiation doses across scanners and their relation to the national DRL of 12 mGy for abdominal CT are illustrated in [Fig. 2]. Significant differences in CTDIvol distribution between scanners were also confirmed for this modality (p < 0.05). A similar pattern of dose distribution was observed for DLP, SSDE, and ED for each region and protocol ([Fig. 3]).

Fig. 3 Radiation dose parameters for single-phase abdominal CT, stratified by CT device and protocol type. a CTDIvol: Volume-weighted CT dose index; b DLP: Dose-length product; c SSDE: Size-specific dose estimate; d ED: Effective dose; Total: all CT devices; Force: SOMATOM Force; Definition Flash: SOMATOM Definition Flash; Definition Edge: SOMATOM Definition Edge; Definition AS+: SOMATOM Definition AS+ (all: Siemens Healthcare, Forchheim, Germany). Whiskers are defined according to Tukey, extending to 1.5× the interquartile range. Asterisk indicates significant difference: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Intra-individual Radiation Dose Variability

Intra-individual radiation dose variability was assessed by comparing all CT scans performed with identical protocol categories to the lowest recorded CTDIvol for each patient. This approach allowed for a comprehensive quantification of dose variability across repeated examinations. Dose variability was assessed for all dose parameters (CTDIvol, DLP, SSDE and ED), while the following results focus on CTDIvol for clarity.

In non-contrast chest CT scans, data from 1.9% (95/4986) of the patient cohort with consecutive scans in this category were available for analysis. The intra-individual dose variability ranged from 17.9% [IQR 5.8–78.8] to 140.7% [22.7–188.4], with a median variability of 93.7% [19.8–142.0], indicating substantial dose variability. In contrast-enhanced chest CT scans, which represented a larger subgroup with 63.2% (3152/4986) of the cohort. The variability ranged from 15.8% [6.3–40.2] to 120.4% [41.8–188.3], with a median variability of 66.3% [31.1–105.4].

Abdominal CT showed considerably less intra-individual variability. For non-contrast abdominal CT scans, 2.0% (99/4986) of patients had consecutive scans available for comparison. The variability ranged from 10.6% [5.3–22.8] to 26.0% [14.5–41.4], with a median variability of 19.8% [11.8–32.3]. For contrast-enhanced abdominal CT scans, representing 61.0% (3043/4986) of patients, variability ranged from 11.2% [4.5–23.1] to 44.2% [23.6–70.4], with a median variability of 28.2% [16.6–41.0]. These findings highlight that dose variability between consecutive scans is more pronounced in chest CT – particularly in non-contrast protocols – compared to abdominal CT (p < 0.0001).

Device-Specific Radiation Dose Variability

When the analysis was restricted to scans performed on the same CT device, intra-individual radiation dose variability was markedly reduced across all regions and protocol categories. In chest CT, intra-individual dose variability decreased from 93.7% to 14.7% [IQR 8.1–33.1] in non-contrast scans and from 66.3% to 18.1% [9.5–31.3] in contrast-enhanced scans, corresponding to relative reductions of 84.3% and 72.7%, respectively (both p < 0.0001). Similarly, in abdominal CT, radiation dose variability decreased from 19.8% to 11.7% [7.8–19.8] in non-contrast scans and from 28.2% to 15.3% [8.3–24.7] in contrast-enhanced scans, equating to relative reductions of 40.9% and 45.7%, respectively (both p < 0.0001) ([Table 4], [Fig. 4]). These results indicate that scanner-specific factors – such as calibration, algorithmic dose modulation, or hardware sensitivity – substantially contribute to intra-individual dose variability, even when protocol settings are standardized. The data underscore the importance of consistent scanner usage in longitudinal oncologic imaging to enhance dose stability.

Table 4 Intra-individual radiation dose variability in repeated chest and abdominal CT scans across different CT devices versus on same CT device, measured by volume-weighted CT dose index (CTDIvol) and stratified by protocol type. Data are given as median with interquartile range (IQR).

Protocol	Minimum radiation dose variability (%)		Mean radiation dose variability (%)		Maximum radiation dose variability (%)
	Median	IQR	Median	IQR	Median	IQR
Scans across different CT devices
Chest
Non-contrast	17.9	5.8–78.8	93.7	19.8–142.0	140.7	22.7–188.4
Contrast-enhanced	15.8	6.3–40.2	66.3	31.1–105.4	120.4	41.8–188.3
Abdomen
Non-contrast	10.6	5.3–22.8	19.8	11.8–32.3	26.0	14.5–41.4
Contrast-enhanced	11.2	4.5–23.1	28.2	16.6–41.0	44.2	23.6–70.4
Scans on same CT device
Chest
Non-contrast	9.4	4.2–21.7	14.7	8.1–33.1	18.8	9.9–37.1
Contrast-enhanced	10.5	4.6–21.7	18.1	9.5–31.3	23.9	11.3–42.8
Abdomen
Non-contrast	9.4	5.4–17.5	11.7	7.8–19.8	16.4	9.4–25.3
Contrast-enhanced	9.2	3.9–17.8	15.3	8.3–24.7	20.0	9.9–33.5

Fig. 4 Intra-individual radiation dose variability in chest and abdominal CT. Whiskers are defined according to Tukey, extending to 1.5× the interquartile range. By limiting the analysis to the same CT device, the median intra-individual radiation dose variability was reduced from 93.7% to 14.7% in non-contrast chest CT, from 66.3% to 18.1% in contrast chest CT, from 19.8% to 11.7% in non-contrast abdominal CT, and from 28.2% to 15.3% in contrast-enhanced abdominal CT, corresponding to relative reductions of 84.3%, 72.7%, 40.9%, and 45.7%, respectively.

Discussion

Although CT accounts for only approximately 10% of all X-ray examinations, it contributes up to 68% of the total ED from medical imaging in Germany – a proportion that is expected to increase further in the coming years [1] [3] [4] [7]. Oncologic patients are particularly vulnerable due to the high frequency of repeated staging and follow-up CT scans. Ensuring consistent and minimized radiation exposure is therefore essential to reduce cumulative dose and maintain diagnostic image quality over time.

Advances in CT technology, such as automatic tube current modulation (ATCM), automatic tube voltage selection (ATVS), and iterative reconstruction techniques, have enabled meaningful dose reductions and decreased the mean ED for individual CT scans in Germany by 15–16% over the past decade compared to before 2010 [4] [5] [16] [17] [18] [19]. More recently, deep learning-based reconstruction methods have shown additional potential to reduce image noise and improve diagnostic quality at lower dose levels, achieving ED reductions of up to 45% in oncologic CT imaging while maintaining diagnostic quality [20] [21] [22]. Despite these innovations, population-wide radiation exposure continues to rise, primarily driven by the increasing number of CT scans performed [1] [3] [4] [6].

The first key finding of this study was that substantial inter-scanner differences in radiation dose persist, even within a single high-volume institution using standardized protocol settings. Across all CT protocols, maximum inter-scanner CTDIvol variability ranged from 1.6 to 2.4, depending on the scan region and protocol type. The differences were consistent across other metrics such as DLP, SSDE, and ED. Furthermore, variation in QR between scanners suggested that not only absolute dose levels but also the spread of delivered doses varied by CT device. These findings indicate that scanner-specific implementation of protocols can significantly influence radiation dose, even when standardization efforts are in place.

Secondly, we observed considerable intra-individual dose variability between consecutive scans using the same protocol. This variability was greatest in non-contrast chest CT, with a median deviation of nearly 100%, and lowest in non-contrast abdominal CT, with a median deviation below 20%. These region-specific differences may reflect the greater complexity of dose modulation in chest CT imaging, which is more sensitive to respiratory motion and differences in tissue attenuation along the z-axis [23]. In contrast, the more homogeneous anatomy of the abdomen, combined with reduced motion artifacts, likely contributes to more stable dose modulation. However, a lack of voltage selection and the use of multiphase protocols can still lead to increased variability. Recent registry analyses have shown that kV modulation by patient size is rarely used in practice, despite its potential to reduce dose, and multiphase scans remain common [24]. The observed differences between CT devices emphasize the crucial role of scanner-specific implementation in dose delivery. Previous studies and systematic reviews have identified substantial variations in DRLs for the same procedures, both between and within institutions, linked to differences in scanner model, protocol configurations, slice thickness and number, patient characteristics, and operator-dependent adjustments [4] [9] [10] [11] [23] [24] [25] [26] [27] [28] [29]. These factors include both automatic modifications, such as tube current modulation, and manual adjustments to tube voltage, scan range, and patient positioning. A large international cohort study involving 151 institutions found that most dose variability stems from local decisions on protocol settings rather than from scanner or patient characteristics [12]. In addition to these known contributors, the observations likely also reflect subtle distinctions in hardware performance, detector sensitivity, and algorithmic behavior that remain unaccounted for.

A third and crucial observation was that intra-individual dose variability decreased markedly when consecutive scans were performed on the same CT device. This effect was evident across all regions and protocols. For example, in contrast-enhanced chest CT, the dose variability was reduced from 66.3% to 18.1% when same-scanner comparisons were performed. These results emphasize that beyond intrinsic patient-related factors, such as body habitus, manual parameter entry, patient positioning, and automated exposure control adjustments, scanner-specific calibration, algorithmic behavior, and hardware characteristics have a measurable impact on radiation dose, even under standardized protocol conditions [26]. Our observations expand upon a prior study by Suntharalingam et al., who reported CTDIvol variability of 9.6% in chest CT and 10.1% in abdominal CT, even under consistent hardware and protocol conditions with automated exposure controls applied [23]. In contrast, our study reflects real-world clinical data from a large, heterogeneous cohort, incorporating longer follow-up intervals, more than two scan time points, and up to four different device models. This broader scope likely captures additional technical and biological influences, such as changes in body habitus over time due to disease progression (e.g., weight loss, muscle atrophy), which may affect automatic exposure modulation systems or trigger manual adjustments.

Although CTDIvol is not a patient-specific metric per se, such changes could affect dose output and influence automated tube current or voltage modulation systems, or lead to manual parameter adjustments by the staff. ATVS has been shown to improve image quality and dose consistency in an operator-independent manner, though small anatomical differences may still trigger different voltage selection [17] [23]. These factors could, in turn, contribute to increased intra-individual variability over longer follow-up periods.

Importantly, most observed radiation doses were below the national DRLs. Accordingly, systematic dose monitoring enables continuous assessment and adjustment of scan protocols for different scanners [13]. Our data support the implementation of internal, scanner-specific DRLs to better reflect real-world clinical performance and support ongoing dose monitoring and protocol adjustment. This is consistent with the heterogeneous reporting standards observed across institutions [9]. Similar conclusions have been drawn in international research, which found that most international variabilities in dose are attributable to local choices regarding technical parameters and protocol rather than to scanner models or patient characteristics [12]. These findings also underscore the need for robust institutional dose management strategies that go beyond basic protocol harmonization. Even within standardized frameworks, scanner-specific implementation and operator practices can introduce considerable variability. Routine dose tracking – particularly at the device level – can help identify inconsistencies, guide protocol refinement, and ensure adherence to the ALARA principle. In the long term, such strategies may reduce patient exposure and improve comparability of serial imaging, which is critical in oncology.

The study has several limitations. It was conducted at a single institution, which may limit the generalizability of scanner-specific effects to other centers. Although all scans were performed on Siemens CT devices, variations in device calibration, protocol implementation, and clinical use cases may differ across institutions. Furthermore, changes in patient-related factors (e.g., weight, disease status) were not formally controlled for and could contribute to dose variability. Finally, the retrospective design may introduce selection bias, and post hoc subgroup analyses carry the inherent risk of data-driven interpretation.

Future studies should investigate these effects in multicenter settings and across different scanner vendors to validate our findings and facilitate the development of harmonized protocol standards. In addition, integration of deep learning-based dose optimization tools may further enhance real-time personalization of dose delivery, particularly in longitudinal oncologic imaging.

Conclusion

This study revealed substantial inter- and intra-individual variability in radiation dose among oncologic patients undergoing repeated chest and abdominal CT examinations, despite the use of standardized protocol settings. Dose levels and variability differed significantly between CT devices, underscoring the impact of scanner-specific implementation on dose delivery. While intra-individual dose variability decreased markedly when consecutive scans were performed on the same scanner, residual variability persisted. These findings indicate that protocol standardization alone is insufficient to ensure dose consistency and emphasize the need for scanner-specific dose management strategies. Routine dose monitoring and the establishment of internal, device-adapted DRLs may improve consistency, minimize radiation exposure, and support the ALARA principle in clinical radiology.

Clinical relevance of the study

• Repeated CT imaging, particularly of oncologic patients, requires awareness of potentially avoidable cumulative radiation exposure.

• Substantial CT device-specific differences in radiation dose delivery are a concern, even with protocol standardization.

• Systematic radiation dose monitoring and consistent scanner use are essential to reduce radiation exposure and variability in follow-up CT imaging and to enhance patient safety.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 Bos D, Guberina N, Zensen S. et al. Radiation Exposure in Computed Tomography. Deutsches Ärzteblatt international 2023; 120: 135-141
• 2 Bundesamt für Strahlenschutz. Röntgendiagnostik: Häufigkeit und Strahlenexposition für die deutsche Bevölkerung (18.10.2024). Accessed May 18, 2025 at: https://www.bfs.de/DE/themen/ion/anwendung-medizin/diagnostik/roentgen/haeufigkeit-exposition.html
• 3 Bundesamt für Strahlenschutz (BfS), Bundesministerium für Umwelt, Naturschutz, Reaktorsicherheit und Verbraucherschutz. Umweltradioaktivität und Strahlenbelastung – Jahresbericht 2020. 2020 http://nbn-resolving.de/urn:nbn:de:0221–2023092039261
• 4 Bundesamt für Strahlenschutz. Computertomografie – Expositionspraxis in Deutschland und Möglichkeiten der Dosisreduktion (20.09.2022). Accessed May 18, 2025 at: https://www.bfs.de/DE/themen/ion/anwendung-medizin/diagnostik/roentgen/ct-untersuchungen.html
• 5 Schegerer AA, Nagel H-D, Stamm G. et al. Current CT practice in Germany: Results and implications of a nationwide survey. Eur J Radiol 2017; 90: 114-128
• 6 Nekolla EA, Schegerer AA, Griebel J. et al. Häufigkeit und Dosis diagnostischer und interventioneller Röntgenanwendungen: Trends zwischen 2007 und 2014. Radiologe 2017; 57: 555-562
• 7 Miglioretti DL, Smith-Bindman R. Overuse of computed tomography and associated risks. Am Fam Physician 2011; 83: 1252-1254
• 8 ICRP. The 2007 Recommendations of the International Commission on Radiological Protection. ICRP publication 103. Ann ICRP 2007; 37: 1-332
• 9 Garba I, Zarb F, McEntee MF. et al. Computed tomography diagnostic reference levels for adult brain, chest and abdominal examinations: A systematic review. Radiography (Lond) 2021; 27: 673-681
• 10 Kumsa MJ, Nguse TM, Ambessa HB. et al. Establishment of local diagnostic reference levels for common adult CT examinations: a multicenter survey in Addis Ababa. BMC Med Imaging 2023; 23: 6
• 11 Smith TB, Zhang S, Erkanli A. et al. Variability in image quality and radiation dose within and across 97 medical facilities. J Med Imaging (Bellingham) 2021; 8: 52105
• 12 Smith-Bindman R, Wang Y, Chu P. et al. International variation in radiation dose for computed tomography examinations: prospective cohort study. BMJ 2019; 364: k4931
• 13 Council of the European Union. Council Directive 2013/59/Euratom of 5 December 2013 laying down basic safety standards for protection against the dangers arising from exposure to ionising radiation, and repealing Directives 89/618/Euratom, 90/641/Euratom, 96/29/Euratom, 97/43/Euratom and 2003/122/Euratom. Official Journal of the European Union 2013. http://data.europa.eu/eli/dir/2013/59/2014–01–17
• 14 McNitt-Gray MF. AAPM/RSNA Physics Tutorial for Residents: Topics in CT. Radiation dose in CT. Radiographics 2002; 22: 1541-1553
• 15 Bundesamt für Strahlenschutz. Bekanntmachung der aktualisierten diagnostischen Referenzwerte für diagnostische und interventionelle Röntgenanwendungen [Vom 17. November 2022] (01.07.2024). Accessed June 16, 2025 at: https://www.bfs.de/SharedDocs/Downloads/BfS/DE/fachinfo/ion/drw-roentgen.pdf?__blob=publicationFile&v=1
• 16 Kordolaimi SD, Argentos S, Pantos I. et al. A new era in computed tomographic dose optimization: the impact of iterative reconstruction on image quality and radiation dose. J Comput Assist Tomogr 2013; 37: 924-931
• 17 Krazinski AW, Meinel FG, Schoepf UJ. et al. Reduced radiation dose and improved image quality at cardiovascular CT angiography by automated attenuation-based tube voltage selection: intra-individual comparison. Eur Radiol 2014; 24: 2677-2684
• 18 Lee S, Yoon S-W, Yoo S-M. et al. Comparison of image quality and radiation dose between combined automatic tube current modulation and fixed tube current technique in CT of abdomen and pelvis. Acta Radiol 2011; 52: 1101-1106
• 19 Mayer C, Meyer M, Fink C. et al. Potential for radiation dose savings in abdominal and chest CT using automatic tube voltage selection in combination with automatic tube current modulation. AJR Am J Roentgenol 2014; 203: 292-299
• 20 Kobayashi N, Nakaura T, Yoshida N. et al. Impact of deep learning reconstruction on radiation dose reduction and cancer risk in CT examinations: a real-world clinical analysis. Eur Radiol 2025; 35: 3499-3507
• 21 Paudyal R, Shah AD, Akin O. et al. Artificial Intelligence in CT and MR Imaging for Oncological Applications. Cancers (Basel) 2023; 15
• 22 Rusanov B, Hassan GM, Reynolds M. et al. Deep learning methods for enhancing cone-beam CT image quality toward adaptive radiation therapy: A systematic review. Med Phys 2022; 49: 6019-6054
• 23 Suntharalingam S, Stecker FF, Guberina N. et al. How Much Is the Dose Varying between Follow-Up CT-Examinations Performed on the Same Scanner with the Same Imaging Protocol?. PLoS One 2016; 11: e0152961
• 24 Smith-Bindman R, Kang T, Chu PW. et al. Large variation in radiation dose for routine abdomen CT: reasons for excess and easy tips for reduction. Eur Radiol 2024; 34: 2394-2404
• 25 Fathelrahman SA, Ahmed AT, Khalid NH. et al. Study the variations in radiation doses in different multi-slice CT scan machines. GSC Adv. Res. Rev 2023; 17: 221-228
• 26 Harri PA, Moreno CC, Nelson RC. et al. Variability of MDCT dose due to technologist performance: impact of posteroanterior versus anteroposterior localizer image and table height with use of automated tube current modulation. AJR Am J Roentgenol 2014; 203: 377-386
• 27 Kaasalainen T, Palmu K, Reijonen V. et al. Effect of patient centering on patient dose and image noise in chest CT. AJR Am J Roentgenol 2014; 203: 123-130
• 28 Lange I, Alikhani B, Wacker F. et al. Intraindividual variation of dose parameters in oncologic CT imaging. PLoS One 2021; 16: e0250490
• 29 Pantos I, Thalassinou S, Argentos S. et al. Adult patient radiation doses from non-cardiac CT examinations: a review of published results. Br J Radiol 2011; 84: 293-303

Publication History

Received: 04 August 2025

Accepted after revision: 04 January 2026

Article published online:
30 January 2026