Christian Sommer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Christian Sommer has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Stephan Scheidegger
ZHAW School of Engineering
5 shared publications - Gerd Lutters
Kantonsspital Aarau
3 shared publications - Mathias Sebastian Weyland
ZHAW School of Engineering, Zurich University of Applied Sciences, Technikumstr. 9, CH-8401 Winterthur, Switzerland
3 shared publications - Ismail Özden
Kantonsspital Aarau
2 shared publications Patrik Eschle
Zurich University of Applied Science, ZHAW School of Engineering, Winterthur, Switzerland
2 shared publications- Carolina Duran
ZHAW School of Engineering, 8401 Winterthur, Switzerland
1 shared publication
Latest Publications
- Open Access
- 18 June 2026
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 71 (12) , 125006
Radiomics-informed CT protocol optimization through feature-level trade-off analysis in a phantom study
Robust radiomic features are critical for reliable quantitative CT analysis; however, their reproducibility is limited by variations in acquisition protocols. This study aims to (1) identify radiomic features predictive of low-contrast detectability, (2) quantify their sensitivity to acquisition parameters, and (3) integrate these factors through a feature-level trade-off framework for CT protocol optimization. A phantom containing 30 inserts was scanned across multiple CT protocols, and 107 radiomic features were extracted. Feature robustness was assessed using CV and ICC, followed by ComBat harmonization, which approximately doubled the proportion of reproducible features. A random forest model achieved strong performance (AUC = 0.94, accuracy = 0.90), with first-order features showing the highest predictive power. Sensitivity analysis using Spearman’s correlation revealed that texture features were more influenced by protocol variations ( ρ ≈ 0.15–0.17), while shape features were more associated with dose ( ρ ≈ 0.13–0.14). A consistent trade-off was observed between detectability and sensitivity (−0.07 to −0.17), indicating that maximizing detectability alone is insufficient for protocol design. The proposed framework enabled systematic protocol ranking, demonstrating that moderate acquisition settings provide a more balanced and robust performance. These findings establish a quantitative foundation for radiomics-informed CT protocol optimization, with potential for further validation in clinical settings.
- 18 June 2026
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 71 (12) , 125006
Open AccessRadiomics-informed CT protocol optimization through feature-level trade-off analysis in a phantom study
Robust radiomic features are critical for reliable quantitative CT analysis; however, their reproducibility is limited by variations in acquisition protocols. This study aims to (1) identify radiomic features predictive of low-contrast detectability, (2) quantify their sensitivity to acquisition parameters, and (3) integrate these factors through a feature-level trade-off framework for CT protocol optimization. A phantom containing 30 inserts was scanned across multiple CT protocols, and 107 radiomic features were extracted. Feature robustness was assessed using CV and ICC, followed by ComBat harmonization, which approximately doubled the proportion of reproducible features. A random forest model achieved strong performance (AUC = 0.94, accuracy = 0.90), with first-order features showing the highest predictive power. Sensitivity analysis using Spearman’s correlation revealed that texture features were more influenced by protocol variations ( ρ ≈ 0.15–0.17), while shape features were more associated with dose ( ρ ≈ 0.13–0.14). A consistent trade-off was observed between detectability and sensitivity (−0.07 to −0.17), indicating that maximizing detectability alone is insufficient for protocol design. The proposed framework enabled systematic protocol ranking, demonstrating that moderate acquisition settings provide a more balanced and robust performance. These findings establish a quantitative foundation for radiomics-informed CT protocol optimization, with potential for further validation in clinical settings.
- Open Access
- 21 May 2024
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 69 (11) , 115026
Probabilistic U-Net model observer for the DDC method in CT scan protocol optimization
Optimizing complex imaging procedures within Computed Tomography, considering both dose and image quality, presents significant challenges amidst rapid technological advancements and the adoption of machine learning (ML) methods. A crucial metric in this context is the Difference-Detailed Curve, which relies on human observer studies. However, these studies are labor-intensive and prone to both inter- and intra-observer variability. To tackle these issues, a ML-based model observer utilizing the U-Net architecture and a Bayesian methodology is proposed. In order to train a model observer unaffected by the spatial arrangement of low-contrast objects, the image preprocessing incorporates a Gaussian Process-based noise model. Additionally, gradient-weighted class activation mapping is utilized to gain insights into the model observer’s decision-making process. By training on data from a diverse group of observers, well-calibrated probabilistic predictions that quantify observer variability are achieved. Leveraging the principles of Beta regression, the Bayesian methodology is used to derive a model observer performance metric, effectively gauging the model observer’s strength in terms of an ‘effective number of observers’. Ultimately, this framework enables to predict the DDC distribution by applying thresholds to the inferred probabilities (Part of this work has been presented at: Stocker D, Sommer C, Gueng S, Stäuble J, Özden I, Griessinger J, Weyland M S, Lutters G, Scheidegger S (2023). Probabilistic U-Net Model Observer for the DDC Method in CT Scan Protocol Optimization. The 56th SSRMP Annual Meeting 2023, November 30. - December 1., 2023, Luzern, Switzerland).
- 21 May 2024
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 69 (11) , 115026
Open AccessProbabilistic U-Net model observer for the DDC method in CT scan protocol optimization
Optimizing complex imaging procedures within Computed Tomography, considering both dose and image quality, presents significant challenges amidst rapid technological advancements and the adoption of machine learning (ML) methods. A crucial metric in this context is the Difference-Detailed Curve, which relies on human observer studies. However, these studies are labor-intensive and prone to both inter- and intra-observer variability. To tackle these issues, a ML-based model observer utilizing the U-Net architecture and a Bayesian methodology is proposed. In order to train a model observer unaffected by the spatial arrangement of low-contrast objects, the image preprocessing incorporates a Gaussian Process-based noise model. Additionally, gradient-weighted class activation mapping is utilized to gain insights into the model observer’s decision-making process. By training on data from a diverse group of observers, well-calibrated probabilistic predictions that quantify observer variability are achieved. Leveraging the principles of Beta regression, the Bayesian methodology is used to derive a model observer performance metric, effectively gauging the model observer’s strength in terms of an ‘effective number of observers’. Ultimately, this framework enables to predict the DDC distribution by applying thresholds to the inferred probabilities (Part of this work has been presented at: Stocker D, Sommer C, Gueng S, Stäuble J, Özden I, Griessinger J, Weyland M S, Lutters G, Scheidegger S (2023). Probabilistic U-Net Model Observer for the DDC Method in CT Scan Protocol Optimization. The 56th SSRMP Annual Meeting 2023, November 30. - December 1., 2023, Luzern, Switzerland).
- Open Access
- 17 February 2022
- Published by Elsevier in Zeitschrift für Medizinische Physik
- Vol. 32 (2) , 209-217
Feasibility of a method for low contrast CT image quality assessment using difference detail curves for abdominal scans
- 17 February 2022
- Published by Elsevier in Zeitschrift für Medizinische Physik
- Vol. 32 (2) , 209-217
Open AccessFeasibility of a method for low contrast CT image quality assessment using difference detail curves for abdominal scans
- Open Access
- 28 June 2018
- Published by Springer Nature in European Radiology Experimental
- Vol. 2 (1) , 1-10
Effect of thyroid shielding during mammography: measurements on phantom and patient as well as estimation with Monte Carlo simulation
During mammography, the thyroid is exposed to scattered radiation from breast tissue and the device. This may increase the risk of radiation induced thyroid cancer. We investigated the scatter radiation exposition of the thyroid and the effect of a tailored thyroid protection in phantom and patient as well as by using Monte Carlo simulation (MCS). The protective effect of a modified thyroid protection, the relevance of the protective effect and acceptance by patients have been investigated. Phantom and patient measurements provided higher values for the surface dose at thyroid position than expected from MCS (phantom 0.32 mGy; patients 0.38 mGy; MCS 0.16 mGy). Phantom measurements indicated scatter contributions from both breast tissue and collimator/tube system. The value found in our patient study is within the range of the literature (0.22–0.39 mGy). The thyroid protection significantly reduced the surface dose but the dose (0.016 mGy) was higher than that expected from the lead equivalent value. However, the impact of the collar to the effective dose was small (< 4%). The collar was not visible on mammograms. Scatter from the collimator/tube system contributed with 50% to the thyroid dose. Due to the relative small fraction of dose deposited in the thyroid when compared to the mean glandular dose to the breast, a collar is not mandatory in general. Not being associated with the risk of obscuring parts of mammograms, such a collar may be used for young women considering their higher radio sensitivity.
- 28 June 2018
- Published by Springer Nature in European Radiology Experimental
- Vol. 2 (1) , 1-10
Open AccessEffect of thyroid shielding during mammography: measurements on phantom and patient as well as estimation with Monte Carlo simulation
During mammography, the thyroid is exposed to scattered radiation from breast tissue and the device. This may increase the risk of radiation induced thyroid cancer. We investigated the scatter radiation exposition of the thyroid and the effect of a tailored thyroid protection in phantom and patient as well as by using Monte Carlo simulation (MCS). The protective effect of a modified thyroid protection, the relevance of the protective effect and acceptance by patients have been investigated. Phantom and patient measurements provided higher values for the surface dose at thyroid position than expected from MCS (phantom 0.32 mGy; patients 0.38 mGy; MCS 0.16 mGy). Phantom measurements indicated scatter contributions from both breast tissue and collimator/tube system. The value found in our patient study is within the range of the literature (0.22–0.39 mGy). The thyroid protection significantly reduced the surface dose but the dose (0.016 mGy) was higher than that expected from the lead equivalent value. However, the impact of the collar to the effective dose was small (< 4%). The collar was not visible on mammograms. Scatter from the collimator/tube system contributed with 50% to the thyroid dose. Due to the relative small fraction of dose deposited in the thyroid when compared to the mean glandular dose to the breast, a collar is not mandatory in general. Not being associated with the risk of obscuring parts of mammograms, such a collar may be used for young women considering their higher radio sensitivity.
- Open Access
- 7 September 2017
- Published by Walter de Gruyter GmbH in Current Directions in Biomedical Engineering
- Vol. 3 (2) , 517-519
Evaluation of low contrast resolution and radiation dose in abdominal CT protocols by a difference detail curve (DDC) method
The use of optimised CT protocols regarding radiation exposure is a legal requirement. Since low contrast visibility is intrinsically varying within the CT slice, there is no adequate method for optimisation of dose and image quality. We developed a method to access image quality in a way that represents the situation closer to a real patient. This method is based on a novel difference detail curve (DDC) phantom with low contrast objects representing native tissue contrast and contrast media with different densities and diameters. The position of the contrast objects have been evaluated by a noise level analysis of CT slices of different manufactures. The dose – length – product can be measured within the phantom simultaneously. For all tested manu-factures and CT protocols, the noise analysis revealed a similar spatial variation of the signal -to-noise ratio (SNR). For the DDC method, contrast steps of 6 (4-8) Hounsfield Units (HU) are adequate. For the different CT units, comparable low contrast detectability is associated with remarkably varying dose levels (CTDI range from 8 to 18 mGy for native contrast and 9-16 mGy for contrast media). The novel DDC phantom is sensitive to protocol optimisations and therefore suitable for rating subtle effects caused by protocol optimisation.
- 7 September 2017
- Published by Walter de Gruyter GmbH in Current Directions in Biomedical Engineering
- Vol. 3 (2) , 517-519
Open AccessEvaluation of low contrast resolution and radiation dose in abdominal CT protocols by a difference detail curve (DDC) method
The use of optimised CT protocols regarding radiation exposure is a legal requirement. Since low contrast visibility is intrinsically varying within the CT slice, there is no adequate method for optimisation of dose and image quality. We developed a method to access image quality in a way that represents the situation closer to a real patient. This method is based on a novel difference detail curve (DDC) phantom with low contrast objects representing native tissue contrast and contrast media with different densities and diameters. The position of the contrast objects have been evaluated by a noise level analysis of CT slices of different manufactures. The dose – length – product can be measured within the phantom simultaneously. For all tested manu-factures and CT protocols, the noise analysis revealed a similar spatial variation of the signal -to-noise ratio (SNR). For the DDC method, contrast steps of 6 (4-8) Hounsfield Units (HU) are adequate. For the different CT units, comparable low contrast detectability is associated with remarkably varying dose levels (CTDI range from 8 to 18 mGy for native contrast and 9-16 mGy for contrast media). The novel DDC phantom is sensitive to protocol optimisations and therefore suitable for rating subtle effects caused by protocol optimisation.
- Open Access
- 24 May 2017
- Published by Elsevier in The Journal of Arthroplasty
- Vol. 32 (10) , 3213-3218
Determining 3D Kinematics of the Hip Using Video Fluoroscopy: Guidelines for Balancing Radiation Dose and Registration Accuracy
- 24 May 2017
- Published by Elsevier in The Journal of Arthroplasty
- Vol. 32 (10) , 3213-3218
Open AccessDetermining 3D Kinematics of the Hip Using Video Fluoroscopy: Guidelines for Balancing Radiation Dose and Registration Accuracy