{"id":{"repo_id":"cambridge","oai_identifier":"oai:www.repository.cam.ac.uk:1810/390598"},"canonical_url":"https://search.dev.ndltd.org/etd/cambridge/oai:www.repository.cam.ac.uk:1810/390598","repository":{"repo_id":"cambridge","name":"Cambridge University","base_url":"https://api.repository.cam.ac.uk/server/oai/request"},"display":{"title":"Improving Quality of Prostate MRI","abstract":"Prostate cancer (PCa) is the most common cancer among men in the UK and the second leading cause of cancer-related deaths in men worldwide, with its incidence expected to double by 2040. Chapter 1 of this thesis explored the anatomy of the prostate gland, epidemiology, classification, and grading of PCa. Given that prostate MRI has been established as a crucial tool in the PCa diagnostic pathway, the latter half of this chapter introduced key concepts of prostate MRI, including its evolving role and the widely adopted international guidelines, Prostate Imaging Reporting and Data System (PI-RADS). Furthermore, image quality is fundamental to prostate MRI, as it impacts nearly every subsequent step in patient management, including biopsy, treatment, and surveillance. Therefore, Chapter 1 concluded with an introduction to prostate MRI image quality, highlighting our collaborative work with the European Society of Radiology in updating Prostate Imaging Quality scoring system (PI-QUAL) to version 2. Some simple concepts may have the potential to enhance image quality. Chapter 2 investigated the potential benefits of repositioning patients from supine to prone during prostate MRI. Since diffusion-weighted imaging (DWI) is particularly susceptible to susceptibility artefacts caused by rectal air at the prostate-rectum interface, imaging patients in the prone position may shift rectal air away from the prostate, especially when appreciable rectal gas was present. This repositioning could have a positive impact on prostate MRI image quality, especially on DWI, offering a straightforward yet effective method to optimize imaging. However, image quality of T2-weighted imaging (T2WI) was reduced when imaging prone compared to supine because of motion artefacts. Only 20% of the patients preferred the prone position. Additionally, acquiring prone DWI added a mean scanning time for was 8 minutes and 18 seconds. Hence, the prone DWI could serve as a mitigation technique in selected cases with significant rectal air. With the growing demand for prostate MRI, there is increasing interest in integrating novel techniques into clinical practice to reduce scan time and potentially enhance image quality. One such innovation is artificial intelligence (AI). Chapter 3 reviewed the application of deep learning-based reconstruction (DLR) in prostate MRI, a technique designed to maintain—or even improve—image quality while also shortening scan duration. DLR allowed the acquisition time to be reduced by 33% for T2WI and 49% for DWI compared to standard-of-care sequences without compromising image quality or PI-RADS categorisation. Therefore, DLR could be a promising technique for enhancing image quality, and has been adopted in the clinical scanning sequence at Cambridge University Hospital. Since its initial release in 2020, PI-QUAL has gained significant attention and was subsequently updated to version 2 in 2024. Key changes in PI-QUAL v2 include reducing the emphasis on dynamic contrast-enhanced imaging (DCE) and expanding the evaluation criteria to accommodate both biparametric (bp) and multiparametric (mp) MRI. Chapter 4 presented a study that compares scoring distribution and inter-reader variability between PI-QUAL v1 and v2. When used to evaluate mpMRI, the inter-reader agreement for PI-QUAL v1 and v2 was comparable. Nevertheless, a notable shift from “optimal” to “acceptable” quality was demonstrated when moving from v1 to v2, with DCE tending improving quality from “inadequate” (bpMRI) to “acceptable” (mpMRI). Hence, DCE can potentially serves as a “safety net” when image quality of bpMRI is suboptimal. In Chapters 2, 3, and 4, experienced uroradiologists assessed prostate MRI image quality, however, their evaluations were not always consistent. Chapter 5 presented a systematic investigation of radiologists' perceptions of prostate MRI quality using controlled synthetic degradations based on PI-QUAL v2 criteria. The study provided insights into inter-reader agreement and how degradation affects quality domains such as SNR, structure delineation, and anatomical mismatch. A U-shaped trend in agreement was observed in the majority of the quality metrics, with greater variability at intermediate quality levels and stronger consensus at the extremes. Additionally, overlap between criteria, such as the impact of low SNR on structure delineation, was identified. These findings emphasize the complexity of image evaluation and the need for clearer guidelines and automated tools to standardize prostate MRI quality assessment. Chapter 6 presented the development of an automated Minimum Technical Standards of Prostate Imaging Reporting and Data System (PI-RADS MTS) checker capable of evaluating prostate MRI studies from a variety of scanner types. The tool assessed 29 criteria, categorized into DICOM tag-based, image-based, and combined types. Results showed high compliance for many criteria, but challenges remain—particularly with PI-RADS MTS spatial resolution and FOV standards. On the other hand, the American College of Radiology’s Prostate Cancer MRI Center Designation Technical Parameters Criteria (ACR TPC) proposed less stringent spatial resolution standards. The software was also capable of assessing compliance with ACR TPC. The differences in compliance between PI-RADS MTS and ACR TPC highlight the need for practical and clinically achievable standardization. Furthermore, the image-based criteria were evaluated by deep-learning models, which showed strong performance in assessing prostate coverage and fat suppression. Overall, the tool demonstrates promise for improving standardization and quality assurance in prostate MRI. Finally, Chapter 7 will conclude this PhD thesis and propose some potential future works.","abstract_html":"Prostate cancer (PCa) is the most common cancer among men in the UK and the second leading cause of cancer-related deaths in men worldwide, with its incidence expected to double by 2040. Chapter 1 of this thesis explored the anatomy of the prostate gland, epidemiology, classification, and grading of PCa. Given that prostate MRI has been established as a crucial tool in the PCa diagnostic pathway, the latter half of this chapter introduced key concepts of prostate MRI, including its evolving role and the widely adopted international guidelines, Prostate Imaging Reporting and Data System (PI-RADS). Furthermore, image quality is fundamental to prostate MRI, as it impacts nearly every subsequent step in patient management, including biopsy, treatment, and surveillance. Therefore, Chapter 1 concluded with an introduction to prostate MRI image quality, highlighting our collaborative work with the European Society of Radiology in updating Prostate Imaging Quality scoring system (PI-QUAL) to version 2. Some simple concepts may have the potential to enhance image quality. Chapter 2 investigated the potential benefits of repositioning patients from supine to prone during prostate MRI. Since diffusion-weighted imaging (DWI) is particularly susceptible to susceptibility artefacts caused by rectal air at the prostate-rectum interface, imaging patients in the prone position may shift rectal air away from the prostate, especially when appreciable rectal gas was present. This repositioning could have a positive impact on prostate MRI image quality, especially on DWI, offering a straightforward yet effective method to optimize imaging. However, image quality of T2-weighted imaging (T2WI) was reduced when imaging prone compared to supine because of motion artefacts. Only 20% of the patients preferred the prone position. Additionally, acquiring prone DWI added a mean scanning time for was 8 minutes and 18 seconds. Hence, the prone DWI could serve as a mitigation technique in selected cases with significant rectal air. With the growing demand for prostate MRI, there is increasing interest in integrating novel techniques into clinical practice to reduce scan time and potentially enhance image quality. One such innovation is artificial intelligence (AI). Chapter 3 reviewed the application of deep learning-based reconstruction (DLR) in prostate MRI, a technique designed to maintain—or even improve—image quality while also shortening scan duration. DLR allowed the acquisition time to be reduced by 33% for T2WI and 49% for DWI compared to standard-of-care sequences without compromising image quality or PI-RADS categorisation. Therefore, DLR could be a promising technique for enhancing image quality, and has been adopted in the clinical scanning sequence at Cambridge University Hospital. Since its initial release in 2020, PI-QUAL has gained significant attention and was subsequently updated to version 2 in 2024. Key changes in PI-QUAL v2 include reducing the emphasis on dynamic contrast-enhanced imaging (DCE) and expanding the evaluation criteria to accommodate both biparametric (bp) and multiparametric (mp) MRI. Chapter 4 presented a study that compares scoring distribution and inter-reader variability between PI-QUAL v1 and v2. When used to evaluate mpMRI, the inter-reader agreement for PI-QUAL v1 and v2 was comparable. Nevertheless, a notable shift from “optimal” to “acceptable” quality was demonstrated when moving from v1 to v2, with DCE tending improving quality from “inadequate” (bpMRI) to “acceptable” (mpMRI). Hence, DCE can potentially serves as a “safety net” when image quality of bpMRI is suboptimal. In Chapters 2, 3, and 4, experienced uroradiologists assessed prostate MRI image quality, however, their evaluations were not always consistent. Chapter 5 presented a systematic investigation of radiologists' perceptions of prostate MRI quality using controlled synthetic degradations based on PI-QUAL v2 criteria. The study provided insights into inter-reader agreement and how degradation affects quality domains such as SNR, structure delineation, and anatomical mismatch. A U-shaped trend in agreement was observed in the majority of the quality metrics, with greater variability at intermediate quality levels and stronger consensus at the extremes. Additionally, overlap between criteria, such as the impact of low SNR on structure delineation, was identified. These findings emphasize the complexity of image evaluation and the need for clearer guidelines and automated tools to standardize prostate MRI quality assessment. Chapter 6 presented the development of an automated Minimum Technical Standards of Prostate Imaging Reporting and Data System (PI-RADS MTS) checker capable of evaluating prostate MRI studies from a variety of scanner types. The tool assessed 29 criteria, categorized into DICOM tag-based, image-based, and combined types. Results showed high compliance for many criteria, but challenges remain—particularly with PI-RADS MTS spatial resolution and FOV standards. On the other hand, the American College of Radiology’s Prostate Cancer MRI Center Designation Technical Parameters Criteria (ACR TPC) proposed less stringent spatial resolution standards. The software was also capable of assessing compliance with ACR TPC. The differences in compliance between PI-RADS MTS and ACR TPC highlight the need for practical and clinically achievable standardization. Furthermore, the image-based criteria were evaluated by deep-learning models, which showed strong performance in assessing prostate coverage and fat suppression. Overall, the tool demonstrates promise for improving standardization and quality assurance in prostate MRI. Finally, Chapter 7 will conclude this PhD thesis and propose some potential future works.","abstract_has_math":false,"creators":["Lee, Kang-Lung"],"institution":"University of Cambridge","degree_name":"Doctor of Philosophy (PhD)","degree_level":"Doctoral","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Barrett, tristan"],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-05-02","date_published":"2025-05-02","updated_at":"2026-07-22T22:24:06Z","subjects":["Prostate MRI","Image quality"],"languages":["eng"],"rights":[],"rights_urls":["https://www.repository.cam.ac.uk/bitstreams/77bae2e0-305a-4983-8b06-d64beaf8e72d/download","http://purl.org/NET/rdflicense/allrightsreserved"],"identifier_entries":[]},"links":{"outbound_url":"https://doi.org/10.17863/CAM.122151","outbound_label":"DOI","outbound_source":"dc:identifier.doi"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Barrett, tristan"]},{"key":"dc:creator","label":"Author","values":["Lee, Kang-Lung"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.issued","label":"Date","values":["2025-05-02"]},{"key":"dc:publisher.institution","label":"Dc Publisher Institution","values":["University of Cambridge"]},{"key":"dc:relation.isreferencedby.uri","label":"Dc Relation Isreferencedby URI","values":["https://www.repository.cam.ac.uk/handle/1810/390598"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"dc:type.qualificationlevel","label":"Dc Type Qualificationlevel","values":["Doctoral"]},{"key":"dc:type.qualificationname","label":"Dc Type Qualificationname","values":["Doctor of Philosophy (PhD)"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Prostate MRI","Image quality"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["https://www.repository.cam.ac.uk/bitstreams/77bae2e0-305a-4983-8b06-d64beaf8e72d/download","http://purl.org/NET/rdflicense/allrightsreserved"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.17863/CAM.122151"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://www.repository.cam.ac.uk/bitstreams/4dce76ed-6f53-4d93-b4de-a6da4cfe88cc/download"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Prostate cancer (PCa) is the most common cancer among men in the UK and the second leading cause of cancer-related deaths in men worldwide, with its incidence expected to double by 2040. Chapter 1 of this thesis explored the anatomy of the prostate gland, epidemiology, classification, and grading of PCa. Given that prostate MRI has been established as a crucial tool in the PCa diagnostic pathway, the latter half of this chapter introduced key concepts of prostate MRI, including its evolving role and the widely adopted international guidelines, Prostate Imaging Reporting and Data System (PI-RADS). Furthermore, image quality is fundamental to prostate MRI, as it impacts nearly every subsequent step in patient management, including biopsy, treatment, and surveillance. Therefore, Chapter 1 concluded with an introduction to prostate MRI image quality, highlighting our collaborative work with the European Society of Radiology in updating Prostate Imaging Quality scoring system (PI-QUAL) to version 2. Some simple concepts may have the potential to enhance image quality. Chapter 2 investigated the potential benefits of repositioning patients from supine to prone during prostate MRI. Since diffusion-weighted imaging (DWI) is particularly susceptible to susceptibility artefacts caused by rectal air at the prostate-rectum interface, imaging patients in the prone position may shift rectal air away from the prostate, especially when appreciable rectal gas was present. This repositioning could have a positive impact on prostate MRI image quality, especially on DWI, offering a straightforward yet effective method to optimize imaging. However, image quality of T2-weighted imaging (T2WI) was reduced when imaging prone compared to supine because of motion artefacts. Only 20% of the patients preferred the prone position. Additionally, acquiring prone DWI added a mean scanning time for was 8 minutes and 18 seconds. Hence, the prone DWI could serve as a mitigation technique in selected cases with significant rectal air. With the growing demand for prostate MRI, there is increasing interest in integrating novel techniques into clinical practice to reduce scan time and potentially enhance image quality. One such innovation is artificial intelligence (AI). Chapter 3 reviewed the application of deep learning-based reconstruction (DLR) in prostate MRI, a technique designed to maintain—or even improve—image quality while also shortening scan duration. DLR allowed the acquisition time to be reduced by 33% for T2WI and 49% for DWI compared to standard-of-care sequences without compromising image quality or PI-RADS categorisation. Therefore, DLR could be a promising technique for enhancing image quality, and has been adopted in the clinical scanning sequence at Cambridge University Hospital. Since its initial release in 2020, PI-QUAL has gained significant attention and was subsequently updated to version 2 in 2024. Key changes in PI-QUAL v2 include reducing the emphasis on dynamic contrast-enhanced imaging (DCE) and expanding the evaluation criteria to accommodate both biparametric (bp) and multiparametric (mp) MRI. Chapter 4 presented a study that compares scoring distribution and inter-reader variability between PI-QUAL v1 and v2. When used to evaluate mpMRI, the inter-reader agreement for PI-QUAL v1 and v2 was comparable. Nevertheless, a notable shift from “optimal” to “acceptable” quality was demonstrated when moving from v1 to v2, with DCE tending improving quality from “inadequate” (bpMRI) to “acceptable” (mpMRI). Hence, DCE can potentially serves as a “safety net” when image quality of bpMRI is suboptimal. In Chapters 2, 3, and 4, experienced uroradiologists assessed prostate MRI image quality, however, their evaluations were not always consistent. Chapter 5 presented a systematic investigation of radiologists' perceptions of prostate MRI quality using controlled synthetic degradations based on PI-QUAL v2 criteria. The study provided insights into inter-reader agreement and how degradation affects quality domains such as SNR, structure delineation, and anatomical mismatch. A U-shaped trend in agreement was observed in the majority of the quality metrics, with greater variability at intermediate quality levels and stronger consensus at the extremes. Additionally, overlap between criteria, such as the impact of low SNR on structure delineation, was identified. These findings emphasize the complexity of image evaluation and the need for clearer guidelines and automated tools to standardize prostate MRI quality assessment. Chapter 6 presented the development of an automated Minimum Technical Standards of Prostate Imaging Reporting and Data System (PI-RADS MTS) checker capable of evaluating prostate MRI studies from a variety of scanner types. The tool assessed 29 criteria, categorized into DICOM tag-based, image-based, and combined types. Results showed high compliance for many criteria, but challenges remain—particularly with PI-RADS MTS spatial resolution and FOV standards. On the other hand, the American College of Radiology’s Prostate Cancer MRI Center Designation Technical Parameters Criteria (ACR TPC) proposed less stringent spatial resolution standards. The software was also capable of assessing compliance with ACR TPC. The differences in compliance between PI-RADS MTS and ACR TPC highlight the need for practical and clinically achievable standardization. Furthermore, the image-based criteria were evaluated by deep-learning models, which showed strong performance in assessing prostate coverage and fat suppression. Overall, the tool demonstrates promise for improving standardization and quality assurance in prostate MRI. Finally, Chapter 7 will conclude this PhD thesis and propose some potential future works."]},{"key":"dc:format.checksum.md5","label":"Dc Format Checksum Md5","values":["6de0338260360f91a23609c4a74b1889","87eda9de84448d1f82354d60eee3eb5f"]},{"key":"dc:title","label":"Title","values":["Improving Quality of Prostate MRI"]}]}],"canonical_facts":{"dc:contributor.advisor":["Barrett, tristan"],"dc:creator":["Lee, Kang-Lung"],"dc:date.issued":["2025-05-02"],"dc:description.abstract":["Prostate cancer (PCa) is the most common cancer among men in the UK and the second leading cause of cancer-related deaths in men worldwide, with its incidence expected to double by 2040. Chapter 1 of this thesis explored the anatomy of the prostate gland, epidemiology, classification, and grading of PCa. Given that prostate MRI has been established as a crucial tool in the PCa diagnostic pathway, the latter half of this chapter introduced key concepts of prostate MRI, including its evolving role and the widely adopted international guidelines, Prostate Imaging Reporting and Data System (PI-RADS). Furthermore, image quality is fundamental to prostate MRI, as it impacts nearly every subsequent step in patient management, including biopsy, treatment, and surveillance. Therefore, Chapter 1 concluded with an introduction to prostate MRI image quality, highlighting our collaborative work with the European Society of Radiology in updating Prostate Imaging Quality scoring system (PI-QUAL) to version 2. Some simple concepts may have the potential to enhance image quality. Chapter 2 investigated the potential benefits of repositioning patients from supine to prone during prostate MRI. Since diffusion-weighted imaging (DWI) is particularly susceptible to susceptibility artefacts caused by rectal air at the prostate-rectum interface, imaging patients in the prone position may shift rectal air away from the prostate, especially when appreciable rectal gas was present. This repositioning could have a positive impact on prostate MRI image quality, especially on DWI, offering a straightforward yet effective method to optimize imaging. However, image quality of T2-weighted imaging (T2WI) was reduced when imaging prone compared to supine because of motion artefacts. Only 20% of the patients preferred the prone position. Additionally, acquiring prone DWI added a mean scanning time for was 8 minutes and 18 seconds. Hence, the prone DWI could serve as a mitigation technique in selected cases with significant rectal air. With the growing demand for prostate MRI, there is increasing interest in integrating novel techniques into clinical practice to reduce scan time and potentially enhance image quality. One such innovation is artificial intelligence (AI). Chapter 3 reviewed the application of deep learning-based reconstruction (DLR) in prostate MRI, a technique designed to maintain—or even improve—image quality while also shortening scan duration. DLR allowed the acquisition time to be reduced by 33% for T2WI and 49% for DWI compared to standard-of-care sequences without compromising image quality or PI-RADS categorisation. Therefore, DLR could be a promising technique for enhancing image quality, and has been adopted in the clinical scanning sequence at Cambridge University Hospital. Since its initial release in 2020, PI-QUAL has gained significant attention and was subsequently updated to version 2 in 2024. Key changes in PI-QUAL v2 include reducing the emphasis on dynamic contrast-enhanced imaging (DCE) and expanding the evaluation criteria to accommodate both biparametric (bp) and multiparametric (mp) MRI. Chapter 4 presented a study that compares scoring distribution and inter-reader variability between PI-QUAL v1 and v2. When used to evaluate mpMRI, the inter-reader agreement for PI-QUAL v1 and v2 was comparable. Nevertheless, a notable shift from “optimal” to “acceptable” quality was demonstrated when moving from v1 to v2, with DCE tending improving quality from “inadequate” (bpMRI) to “acceptable” (mpMRI). Hence, DCE can potentially serves as a “safety net” when image quality of bpMRI is suboptimal. In Chapters 2, 3, and 4, experienced uroradiologists assessed prostate MRI image quality, however, their evaluations were not always consistent. Chapter 5 presented a systematic investigation of radiologists' perceptions of prostate MRI quality using controlled synthetic degradations based on PI-QUAL v2 criteria. The study provided insights into inter-reader agreement and how degradation affects quality domains such as SNR, structure delineation, and anatomical mismatch. A U-shaped trend in agreement was observed in the majority of the quality metrics, with greater variability at intermediate quality levels and stronger consensus at the extremes. Additionally, overlap between criteria, such as the impact of low SNR on structure delineation, was identified. These findings emphasize the complexity of image evaluation and the need for clearer guidelines and automated tools to standardize prostate MRI quality assessment. Chapter 6 presented the development of an automated Minimum Technical Standards of Prostate Imaging Reporting and Data System (PI-RADS MTS) checker capable of evaluating prostate MRI studies from a variety of scanner types. The tool assessed 29 criteria, categorized into DICOM tag-based, image-based, and combined types. Results showed high compliance for many criteria, but challenges remain—particularly with PI-RADS MTS spatial resolution and FOV standards. On the other hand, the American College of Radiology’s Prostate Cancer MRI Center Designation Technical Parameters Criteria (ACR TPC) proposed less stringent spatial resolution standards. The software was also capable of assessing compliance with ACR TPC. The differences in compliance between PI-RADS MTS and ACR TPC highlight the need for practical and clinically achievable standardization. Furthermore, the image-based criteria were evaluated by deep-learning models, which showed strong performance in assessing prostate coverage and fat suppression. Overall, the tool demonstrates promise for improving standardization and quality assurance in prostate MRI. Finally, Chapter 7 will conclude this PhD thesis and propose some potential future works."],"dc:format.checksum.md5":["6de0338260360f91a23609c4a74b1889","87eda9de84448d1f82354d60eee3eb5f"],"dc:identifier.doi":["https://doi.org/10.17863/CAM.122151"],"dc:identifier.uri":["https://www.repository.cam.ac.uk/bitstreams/4dce76ed-6f53-4d93-b4de-a6da4cfe88cc/download"],"dc:language":["eng"],"dc:publisher.institution":["University of Cambridge"],"dc:relation.isreferencedby.uri":["https://www.repository.cam.ac.uk/handle/1810/390598"],"dc:rights":["https://www.repository.cam.ac.uk/bitstreams/77bae2e0-305a-4983-8b06-d64beaf8e72d/download","http://purl.org/NET/rdflicense/allrightsreserved"],"dc:subject":["Prostate MRI","Image quality"],"dc:title":["Improving Quality of Prostate MRI"],"dc:type":["Thesis"],"dc:type.qualificationlevel":["Doctoral"],"dc:type.qualificationname":["Doctor of Philosophy (PhD)"]},"updated_at":"2026-07-22T22:24:06Z"}