Comparative image quality and dosimetric performance of two generations of dedicated breast CT systems

Abstract

Background

Dedicated breast computed tomography (bCT) systems offer detailed imaging for breast cancer diagnosis and treatment. As new bCT generations are developed, it is important to evaluate their imaging performance and dose efficiency to understand differences over previous models.

Purpose

To characterize the imaging performance and dose efficiency of a second‐generation (GEN2) bCT system and compare them to those of a first‐generation (GEN1) system.

Methods

The imaging performance was evaluated through key metrics: modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) in the projection domain. In the image domain, contrast‐to‐noise ratio (CNR), signal‐to‐noise ratio (SNR), and the visibility of calcifications were analyzed using a quality control (QC) phantom with masses and calcification clusters. Air kerma and tube output were measured and mean glandular dose (MGD) estimated for different phantom sizes for dosimetric characterization of the acquisition protocols set by the automatic exposure control (AEC).

Results

GEN2 outperformed GEN1 at higher spatial frequencies, with 57% of the MTF observed at 1 cycles/mm compared to 43% for GEN1. For a 2 mm diameter mass, GEN2 showed 60% higher CNR and 63% higher SNR. However, for larger masses, GEN1 outperformed GEN2, with CNR and SNR values higher by 12% to 44% and 14% to 43%, respectively. GEN2 also achieves higher DQE across the frequency spectrum, with 45% at 1 cycle/mm, compared to GEN1's 20%. Regarding calcifications in the QC phantom, the 320 µm calcifications resulted in distinct full‐width‐at‐half‐maxima (FWHM ± SD), with 897 ± 58 µm for GEN1 and 811 ± 127 µm for GEN2, with a p‐value of 0.19. For 290 µm calcifications, GEN1's FWHM was 866 ± 129 µm, while GEN2's was narrower at 665 ± 57 µm, with a p‐value of 0.01. The tube output was higher for GEN1 (45.2 mGy/mAs) compared to GEN2 (31.5 mGy/mAs). Additionally, GEN2 resulted in 8% lower MGD values compared to GEN1.

Conclusion

While GEN1 offers better CNR and SNR for larger masses, GEN2 provides superior resolution for calcifications, better MTF, improved DQE, and lower MGD at AEC‐determined settings.

1. INTRODUCTION

Breast cancer remains the most common cancer diagnosis among women and is a major cause of cancer‐related mortality worldwide. ¹ , ² Early detection and accurate diagnosis are critical for improving survival rates. ³ , ⁴ , ⁵ However, traditional breast imaging modalities often face challenges, particularly when imaging dense breast tissue. For example, mammography can struggle with tissue overlap, ultrasound may miss deeper or smaller lesions, and MRI can lead to higher false‐positive rates. These limitations can hinder the early detection of small or subtle lesions, potentially impacting treatment outcomes. ⁶ , ⁷ , ⁸ , ⁹

Dedicated breast computed tomography (bCT) systems have emerged as a promising technology that addresses many of these limitations. Unlike conventional imaging methods, bCT provides high‐resolution, three‐dimensional images without breast compression, enhancing both diagnostic accuracy and patient comfort. ¹⁰ , ¹¹ , ¹² Previous studies have characterized the imaging performance, dose efficiency, and clinical applications of earlier‐generation bCT systems, highlighting their potential advantages and setting the stage for ongoing innovations in this technology. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

As newer generations of bCT systems are developed, it becomes important to evaluate their imaging performance and dose efficiency to ensure they offer significant improvements over previous models. In this study, we aimed to characterize a new bCT system, the second‐generation ‘Vera’ (GEN2), and compare its imaging performance and dose efficiency to the first‐generation ‘CBCT1000’ (GEN1) system, both developed by Koning Corp. (Norcross, GA, USA).

2. MATERIAL AND METHODS

Our evaluation of image quality and dosimetry differences between the two systems focused on several metrics: modulation transfer function (MTF), noise power spectrum (NPS), detective quantum efficiency (DQE), contrast‐to‐noise ratio (CNR), signal‐to‐noise ratio (SNR), visibility of calcifications, air kerma, tube output and resulting mean glandular dose (MGD) at the automatic exposure control (AEC)‐determined settings.

For this analysis, we used two distinct types of phantoms. The first, a quality control (QC) breast phantom (Figure 1), consists of breast‐adipose equivalent material and incorporates various embedded targets, such as masses and calcifications. The second, a slab phantom (Figure 2), was designed to mimic different breast sizes.

FIGURE 1.

QC phantom used for estimation of CNR, SNR, and calcification visibility. (Left) A schematic representation showing modular sections: calcification speck (Module 1), contrast resolution (Module 2), uniformity (Module 3), and mounting section (Module 4), with dimensions specified. (Right) Detailed view of Module 2, highlighting the inserts for evaluating the contrast resolution, acquired with GEN2 and used for CNR calculations. WW/WL: 350/‐250. CNR, contrast‐to‐noise‐ratio; QC, quality control; SNR, signal‐to‐noise ratio.

FIGURE 2.

Three phantom sizes used to mimic different breast volumes for absorbed dose assessment: small (left), medium (center), and large (right). CWD refers to the chest‐wall diameter, and CND refers to the chest‐wall‐to‐nipple distance.

2.1. Technical specifications and operational details of GEN1 And GEN2 bCT systems

The GEN1 system, which was previously installed at our institute but has since been replaced by the recently installed GEN2 system, differs in several key technical specifications, including acquisition geometry (larger SDD but lower SID), detector type, pixel pitch, x‐ray tube voltage, and voxel size. These differences are summarized in Table 1, which provides a detailed comparison of the two systems.

TABLE 1.

Comparative overview of GEN1 and GEN2 bCT systems, with the specifications that are different between the two systems highlighted in bold.

	GEN1	GEN2
Detector model	Paxscan 4030 (Varian medical system, Palo Alto, CA, USA)	Xineos 3030 HS (Teledyne Dalsa, Waterloo, ON, Canada)
Detector dimensions	397 mm × 298 mm	295 mm × 295 mm
Pixel pitch	0.388 mm (0.194 mm binned 2×2)	0.152 mm
Source to detector distance (SDD)	923 mm	950 mm
Source to isocenter distance (SID)	650 mm	600 mm
X‐ray tube voltage	49 kV	49‐65 kV*
Focal spot (nominal)	0.3 mm	0.3 mm
Pulse width	8 ms	5 ms
Projections per revolution	300	225
Permanent additional filtration	1.5 mm Al	1.5 mm Al
First half value layer	1.39 mm Al	1.39 mm Al
Rotation time (per revolution)	10 s	10 s
Reconstruction algorithm	FDK‐based	FDK‐based
Reconstructed voxel size	0.273 mm	0.200 mm

*Note: 49 kV is the tube voltage used for standard clinical imaging.

Abbreviations: bCT, breast computed tomograph; GEN1, first‐generation; GEN2, second‐generation.

2.2. Evaluation of the MTF using the edge approach

The MTF for the two bCT systems were assessed in the projection domain using the edge spread function technique. ²¹ For comparability, GEN2's x‐ray tube was set to 49 kV with 1.5 mm Al filtration, since this is its standard mode of clinical operation. A tungsten‐based edge testing device (TX5, Ion Beam Applications, Louvain‐la‐Neuve, Belgium) was placed directly on the detector at an angle of 3 degrees relative to the pixel grid, ensuring alignment with the central x‐ray beam. ²² Using the COQ plug‐in in ImageJ, ²³ we determined the MTF values along the horizontal and vertical axes. An average of two MTF measurements provided the final bCT system MTF values in the projection domain.

2.3. Evaluation of the normalized NPS

For both bCT systems, NNPS were determined in the projection domain at air kerma at the isocenter levels of 5.0 mGy, 12.7 mGy, and 25.6 mGy for GEN1, and 4.5 mGy, 11.0 mGy, and 22.3 mGy for GEN2, with the slight mismatch in dose levels due to the limited selection of exposure settings that can be set for a scan. Identical spectral conditions (49 kV with a 1.5 mm Al filter) were used as those used for MTF analysis, along with additional 14 mm of Al added at the source to mimic the attenuation of x‐rays through breast tissue. The procedure for each system involved capturing 300 projections for GEN1 and 225 projections for GEN2. Each set of projections underwent a standard flat‐field correction process. For the NNPS calculation, the COQ plug‐in of ImageJ was again used, resulting in a one‐dimensional NNPS, representing the radial average of the two‐dimensional one. ²⁴

2.4. Evaluation of the DQE

The DQE for GEN1 and GEN2 systems at 12.7 mGy and 11.0 mGy, respectively, was calculated using the following equation, where $q_0$​ denotes the photon fluence incident on the detector surface.

$$\mathrm{DQE}=\frac{{\mathrm{MTF}}^2}{q_0\mathrm{NNPS}}$$ (1)

The x‐ray spectrum and photon fluence of each system were modeled, as described in Section 2.1, using the SpekPy software (version 2.0.8, October 2020). ²⁵ The implementation details for GEN2 were documented in our previous research, and the same methodology was applied to GEN1. ²⁶ Additionally, the air kerma was measured after passing the x‐ray beam output through various thicknesses of aluminum to validate the accuracy of the modeled spectra and calculate the first half‐value layer.

2.5. CNR and SNR calculation

The CNR was computed to assess the system's ability to differentiate tissue types or materials using five (2 mm, 3 mm, 4 mm, 6 mm, and 8 mm in diameter, Figure 1) of the seven spherical epoxy resin masses present in a QC breast phantom. Signal ROIs, small enough to avoid edge effects, and background ROIs were measured and averaged over ten slices along the z‐axis. The phantom was scanned using both the GEN1 and GEN2 systems, with air kerma levels set at 12.7 mGy and 11.0 mGy, respectively.

The CNR was then calculated as the average pixel intensity difference between object (${\overline{I}}_o$) and background (${\overline{I}}_{bg}$) ROIs, normalized by the background ROI's standard deviation (SD) (${\sigma }_{bg}$).

$$\mathrm{CNR}=\frac{{\overline{I}}_o-{\overline{I}}_{bg}}{{\sigma }_{bg}}$$ (2)

To account for the effect of mass size on the evaluation, the SNR (referred to as SNR_Rose) was calculated using the Rose signal definition. ¹³ , ²⁷ This metric takes into account the number of pixels ($N$) within the specified ROI that had the same physical size but different pixel pitch, given by:

$${\mathrm{SNR}}_{\mathrm{Rose}}=\mathrm{CNR}\times \sqrt{N}$$ (3)

2.6. Evaluation of calcification visibility

The QC phantom used in this study includes a calcification speck module (Module 1, Figure 1) containing six patterns of calcifications with diameters of 320 µm, 290 µm, 270 µm, 240 µm, 220 µm, and 160 µm. These patterns are uniformly distributed along a circle with a 40 mm radius, all located within the same phantom slice.

To assess the spatial resolution of both systems, 10 mm‐long profiles were positioned to pass through the centers of six specks in two calcification clusters (320 and 290 µm). Intensity profiles were extracted, and the full‐width‐at‐half‐maximum (FWHM) was calculated for each of the six calcifications of the same size in each cluster. The FWHM values were then averaged to obtain a mean FWHM for each cluster. The SD and 95% confidence intervals (CIs) of the sample mean were also calculated and reported. Additionally, the two sets of FWHM were compared for statistical significance using a paired t‐test, and the resulting p‐values were included.

2.7. Air kerma measurements

A pencil beam ion chamber (10 × 6‐3CT, Radcal Corp., Monrovia, CA, USA), linked to a dosimeter (Accu‐Gold+ Touch Pro control unit, Radcal Corp.), was positioned at the isocenter and aligned with the central ray of the x‐ray beam. Air kerma measurements were conducted for both systems across the full range of tube current settings: 40 mA, 50 mA, 64 mA, 80 mA, 100 mA, 125 mA, and 160 mA.

2.8. MGD estimation

We aimed to assess the MGD during standard bCT acquisition protocol for GEN1 and GEN2, as per their corresponding AECs. To achieve this, we first modeled the x‐ray spectrum and tube output of each system using SpekPy, considering the spectral and geometric characteristics established in Section 2.1.

Initially, we determined the number of photons per energy bin and the photon fluence (${\mathrm{\Phi }}_{\mathrm{E}}$) in photons/cm². We conducted a Monte Carlo (MC) simulation using the Geant4 toolkit (v.11.1.3, Nov 2023), based on a previously validated algorithm for the GEN1 system. ¹⁴ The validation was performed by comparing the simulation results with experimental measurements. The simulation tracked 10⁶ photons, and a pencil beam ion chamber (100 mm × 10 mm) was simulated by positioning the detector at the isocenter. The pixels corresponding to the chamber's size and shape were selected, ensuring alignment perpendicular to the central ray.

The simulation results were averaged to obtain the mean fluence (${\overline{\mathrm{\Phi }}}_{MC}$). The ${\mathrm{\Phi }}_{\mathrm{E}}$ was then normalized (${\mathrm{\Phi }}_E^{\prime }$), and the normalized values were then scaled by the average pixel intensity from the pencil beam simulation (${\overline{\mathrm{\Phi }}}_{MC}$), resulting in the adjusted energy fluence $({\mathrm{\Phi }}_N)$:

$${\mathrm{\Phi }}_N\left[i\right]={\mathrm{\Phi }}_E^{\prime }\left[i\right]\times {\overline{\mathrm{\Phi }}}_{MC}$$ (4)

Next, the normalized energy fluence (${\mathrm{\Phi }}_N$) was multiplied by the mass energy‐absorption coefficient for air (${\mu }_{en}$/ρ) obtained from NIST data tables, ²⁸ to calculate the air kerma at isocenter (AK _I ):

$${\mathrm{AK}}_I\left[i\right]={\mathrm{\Phi }}_N\left[i\right]\times \left(\frac{{\mu }_{en}}{\rho }\right)\left[i\right]\times 1.602\times {10}^{-16}\mathrm{J}/\mathrm{keV}$$ (5)

These values were validated by comparing them with empirical air kerma measurements taken at the isocenter for each system, as described in Section 2.7.

Then, the mean absolute error (MAE) in percentage was calculated as:

$$\mathrm{MAE}\left(\%\right)=\frac{100\%}{n}\sum _{i=1}^n\left|\frac{y_{i_{measured}}-y_{i_{simulated}}}{y_{i_{measured}}}\right|$$ (6)

where $y_{measured}$ and $y_{simulated}$ represent the empirical measurements and the calculated AK _I , respectively.

Finally, the normalized glandular dose (DgN) was calculated for breasts with a fixed volume density of 14.3%, ²⁹ and these results were compared to the values for the GEN1 system. The simulation was then adapted for the GEN2 system to similarly determine the DgN. Our findings were compared with the values for different breast sizes previously published by Sechopoulos et al., ¹⁴ and the mean difference (MD) in percentage was calculated:

$$\mathrm{MD}\left(\%\right)=\frac{100\%}{n}\sum _{i=1}^n\frac{y_{i_{published}}-y_{i_{present}}}{y_{i_{published}}}$$ (7)

Where $y_{published}$ and $y_{present}$ represent the previously published values and our findings.

2.9. Phantom analysis: MGD

After obtaining the DgN conversion factors, we calculated the MGD for three different breast phantom sizes: small, medium, and large, as depicted in Figure 2. The dimensions of the phantoms are listed in Figure 2. The three phantoms were placed at the isocenter, and the AEC of the bCT system determined the optimal tube current for each case. The conversion factor for each breast size was multiplied by the corresponding air kerma value to obtain the MGD.

3. RESULTS

3.1. Presampled MTF at the detector

Figure 3 shows the measured presampled MTF for the GEN1 and GEN2 systems. A slight difference was observed between the horizontal and vertical directions for both GEN1 and GEN2, but too small to be clearly distinguishable in the printed figure.

FIGURE 3.

GEN1 (blue‐dashed line) and GEN2 (orange line). Measured MTF in the projection domain plotted as a function of frequency. GEN1, first‐generation; GEN2, second‐generation.

The MTF at 1.0 cycles/mm is 57% for GEN2 and 43% for GEN1.

3.2. NNPS at the detector

Figure 4 shows the measured NNPS for the GEN1 and GEN2 systems.

FIGURE 4.

Shows the NNPS measured at the detector, in the projection domain, for both systems and plotted against frequency. The GEN1 is represented by dashed lines, while the GEN2 system is indicated by a solid line. NNPS, normalized noise power spectrum; GEN1, first‐generation; GEN2, second‐generation.

The NNPS at 1.0 cycles/mm is 1.9 × 10⁻⁴, 7.3 × 10⁻⁵ and 3.7 × 10⁻⁵ for GEN2 compared to 4.0 × 10⁻⁵, 1.3 × 10⁻⁵, and 6.94 × 10⁻⁶ for GEN1.

3.3. DQE

Figure 5 presents the DQE for GEN1 and GEN2. The number of incident photons per cm² ($q_0$) retrieved from the SpekPy simulation was 1.07 × 10⁴ and 1.06 × 10⁴ photons for GEN1 and GEN2, respectively.

FIGURE 5.

Depicts the DQE measured at the detector for both systems, plotted against frequency. The GEN1 system is represented by a blue dashed line, while the GEN2 system is indicated by an orange solid line. DQE, detective quantum efficiency; GEN1, first‐generation; GEN2, second‐generation.

GEN2 achieves higher DQE across the frequency range, with 60% DQE at 0.5 cycles/mm and 45% at 1 cycle/mm, compared to GEN1's 50% and 20%, respectively.

3.4. CNR and SNR

In Figure 6, CNR and SNR measurements for GEN1 and GEN2 systems are plotted against mass diameter.

FIGURE 6.

The left plot shows CNR values, while the right plot shows SNR values. Error bars indicate the SD of measurements. The GEN1 system is represented by blue diamond markers and dashed lines, while the GEN2 system is indicated by orange circle markers and solid lines. CNR, contrast‐to‐noise ratio; GEN1, first‐generation; GEN2, second‐generation; SNR, signal‐to‐noise ratio.

GEN2 exhibited a 60% higher CNR for the 2 mm mass compared to GEN1. However, for larger masses, GEN1 outperformed GEN2, with CNR differences of 12%, 16%, 21%, and 44% higher for the 3 mm, 4 mm, 6 mm, and 8 mm masses, respectively. Similarly, SNR results showed that GEN2 had a 63% higher SNR for the 2 mm mass. For larger masses, GEN1 again demonstrated superior performance, with SNR differences of 14%, 13%, 22%, and 43% higher for the 3 mm, 4 mm, 6 mm, and 8 mm masses, respectively.

3.5. Visibility of calcifications

Figure 7 depicts the analysis of two sets of calcifications, 320 and 290 µm in size. For the 320 µm calcifications, the FWHM values were 897 ± 58 µm (95% CI: [836, 959]) for GEN1 and 811 ± 127 µm (95% CI: [678, 944]) for GEN2, with a p‐value of 0.19 (not significant). For the 290 µm calcifications, the FWHM values were 866 ± 129 µm (95% CI: [730, 1002]) for GEN1 and 665 ± 57 µm (95% CI: [605, 725]) for GEN2, with a p‐value of 0.01 (significant).

FIGURE 7.

Comparison of calcification measurements for 320 and 290 µm clusters using GEN1 and GEN2 systems. Graphs show 10 mm long intensity profiles (depicted in yellow in images) for both systems, with FWHM values, SD, 95% CI, and statistical significance (p‐values) included. GEN1 is represented in blue, and GEN2 is represented in orange. The Gaussian fits in the graphs were derived from the calculated average FWHMs. WW/WL: 1420/130. CI, confidence interval; FWHM, full‐width‐at‐half‐maxima; GEN1, first‐generation; GEN2, second‐generation; SD, standard deviation.

3.6. Air kerma measurements

Figure 8 illustrates the linear relationship between tube current and air kerma for both the GEN1 and GEN2 systems. The linear fit equations are provided for both systems: Y = 0.25 X for GEN1 and Y = 0.23 X for GEN2 with R² = 0.99 in both cases.

FIGURE 8.

Air Kerma as a function of tube current for GEN1 (blue diamonds) and GEN2 (orange circles). The 95% CI were also plot for GEN1 and GEN2. CI, confidence interval; GEN1, first‐generation; GEN2, second‐generation.

3.7. MGD

The MAE (min; max) between empirical measurements and the simulated AK _I was computed as 3% (1%; 4%) for GEN1 and 3% (1%; 6%) for GEN2.

The conversion factors for GEN1 and GEN2 are derived from the calculations and listed in Table 2:

TABLE 2.

The table presents the number of photons at isocenter, air kerma per milliampere seconds at the isocenter for both GEN1 and GEN2 systems.

System	Num. of photons	Tube output at isocenter (mGy/mAs)
GEN1	1.44 × 1014	45.2
GEN2	9.57 × 1013	31.5

Abbreviations: GEN1, first‐generation; GEN2, second‐generation.

In Table 3, the DgN coefficients for breasts of varying sizes are presented:

TABLE 3.

DgN coefficients in mGy per mGy of air kerma for breasts of varying sizes, considering a glandularity of 14.3% by volume. Air kerma at the isocenter of the bCT system (49 kV/1.5 mm Al) was taken as reference.

Chest‐wall diameter (cm)	Chest wall‐to‐nipple distance (cm)	DgN (mGy/mGy) GEN1 previously published 14	DgN (mGy/mGy) GEN1 (this work)	Diff GEN1/GEN1‐published	DgN (mGy/mGy) GEN2
10	5	0.52	0.50	3.72%	0.50
	7.5	0.54	0.52	4.20%	0.54
	10	0.56	0.54	4.03%	0.55
12	6	0.47	0.46	3.44%	0.47
	9	0.50	0.48	3.27%	0.49
	12	0.51	0.49	3.55%	0.50
14	7	0.44	0.42	3.06%	0.43
	10.5	0.46	0.45	2.36%	0.45
	14	0.47	0.46	2.60%	0.47
16	8	0.40	0.39	3.33%	0.40
	12	0.42	0.42	2.19%	0.42
	16	0.43	0.43	2.05%	0.43
18	9	0.38	0.37	2.29%	0.37
	13.5	0.39	0.38	2.28%	0.39
	18	0.40	0.39	2.00%	0.40

Abbreviations: bCT, breast computed tomography; Dgn, normalized glandular dose GEN1, first‐generation; GEN2, second‐generation.

From Table 3, the MD (min; max) between the current GEN1 values and the previously published GEN1 values was 2.96% (2.00%; 4.20%).

3.8. Phantom analysis: MGD

Table 4 presents the MGD for small, medium, and large phantom sizes in both GEN1 and GEN2 systems. On average, GEN2 delivers 8% lower MGD than GEN1.

TABLE 4.

MGD for different phantom sizes in GEN1 and GEN2 systems. The AECdetermined the optimal tube current to be used for each case.

Phantom size (cm) (CWD × CND)	Tube current (mA)	Air Kerma at isocenter (mGy)		MGD (mGy)
		GEN1	GEN2	GEN1	GEN2
Small (14 × 6)	40	10.2	8.7	4.3	3.7
Medium (17 × 10)	64	16.0	15.1	5.9	5.7
Large (18 × 14)	80	20.4	18.5	7.9	7.3

Abbreviations: AEC, automatic exposure control; CWD, chest‐wall diameters; CND, chest‐to‐nipple distances; GEN1, first‐generation; GEN2, second‐generation; MGD, mean glandular dose.

4. DISCUSSION

When comparing the GEN2 with the older GEN1, several key performance improvements and trade‐offs become evident. GEN2 demonstrates a clear enhancement in spatial resolution (Figure 3). This increased resolution is largely due to GEN2's detector with smaller pixel pitch and no binning, although this also results in higher noise levels (Figure 4). In contrast, GEN1 benefits from a larger pixel size and 2 × 2 binning, which contribute to lower noise levels.

GEN2 also achieves higher DQE across the frequency spectrum, with 60% DQE at 0.5 cycles/mm and 45% at 1 cycle/mm, compared to GEN1's 50% and 20%, respectively (Figure 5). This enhanced photon detection efficiency contributes to the overall improved image quality in GEN2, despite the increased noise.

In terms of CNR and SNR, GEN1 performs better at larger mass diameters, indicating its effectiveness in detecting contrast differences in larger masses (Figure 6). However, GEN2 provides superior resolution and edge definition for smaller features, such as calcifications, as shown by its narrower FWHM values (Figure 7).

The air kerma analysis indicates a linear relationship with tube current for both systems. GEN1 has a slightly higher tube output rate due to its longer pulse width and more projections per revolution. In contrast, GEN2, with its shorter pulse width and fewer projections, reduces the overall dose while still maintaining sufficient image quality, likely due to the lower electronic noise of its CMOS detector. Although GEN1 produced a higher number of photons at the isocenter, GEN2 had a higher air kerma per mAs. Additionally, although GEN2 generally exhibits slightly higher DgN values, it maintains efficiency in dose management, delivering lower MGD across different phantom sizes. The glandular dose decreases with increasing breast diameter and decreasing chest‐wall‐to‐nipple distance, in alignment with previously published studies. ¹⁴

The MGD analysis in Table 4 highlights that at the AEC settings, GEN2 consistently provides lower MGD values across small, medium, and large phantom sizes, indicating its improved dose efficiency while preserving image quality. However, the suitability and comparability of the clinical performance of the images resulting from the two generations when using their AEC at the manufacturer‐determined settings, as tested here, remains to be evaluated.

5. CONCLUSION

Although GEN1 demonstrates higher CNR and SNR for larger masses, GEN2 provides superior spatial resolution for calcifications, better MTF, and improved DQE. While GEN2's smaller pixel pitch results in higher noise levels, it compensates with enhanced photon detection efficiency and sharper edge definition. Additionally, GEN2 delivers these improvements while consistently providing lower MGD at the AEC settings across different breast sizes, making it more efficient overall.

CONFLICT OF INTEREST STATEMENT

Ioannis Sechopoulos is a Scientific Advisory Board member of Koning Corp.

Pautasso JJ, Van Speybroeck CDE, Michielsen K, Sechopoulos I. Comparative image quality and dosimetric performance of two generations of dedicated breast CT systems. Med Phys. 2025;52:2191–2200. 10.1002/mp.17623