Towards quantitative image analysis of respiratory imaging data for pulmonary disease assessment.
Lead Research Organisation:
University of Oxford
Department Name: Population Health
Abstract
The proposed research encompasses new generation of computational algorithms and applied research in cancer image analysis that will impact not only the medical imaging field, but primarily contribute toward a significant understanding of cancer, treatment planning and response assessment. While the main focus here is on lung tumours, many of the concepts being explored have applicability beyond (e.g. liver diseases). The developed computational algorithms will be capable of learning automatically, building on available data, but weak-supervision (either coming from our a priori knowledge or human interaction when necessary) will be also available to provide biomedically plausible outcomes. Therefore, it is expected that the new generation algorithms can also be applied in other fields of engineering and mathematics that make use of imaging for instance in computational modelling.
Technical Summary
This research project goes substantially beyond the state-of-the-art in medical image analysis in term of:
- a methodology by a novel paradigm extending current local or global formulation to a unified non-local counterpart considering all components of image registration (similarity measure, motion/deformation model, optimisation method) jointly with quantitative imaging;
- a hypothesis that this unified image analysis framework will be physiologically explicitly grounded in the analysed data (potential to link with advances in machine learning, and opening up to future opportunities for collaboration on big data), and be consequently more relevant to biomedical applications.
- challenge on improving the performance of motion estimation, the author believes that the presented methodological advances will eventually lead to the development of real-time (close to real-time) motion estimation for highly dimensional biomedical data. Such an advance will have immediate impact beyond personalised precise radiotherapy planning and delivery.
- a methodology by a novel paradigm extending current local or global formulation to a unified non-local counterpart considering all components of image registration (similarity measure, motion/deformation model, optimisation method) jointly with quantitative imaging;
- a hypothesis that this unified image analysis framework will be physiologically explicitly grounded in the analysed data (potential to link with advances in machine learning, and opening up to future opportunities for collaboration on big data), and be consequently more relevant to biomedical applications.
- challenge on improving the performance of motion estimation, the author believes that the presented methodological advances will eventually lead to the development of real-time (close to real-time) motion estimation for highly dimensional biomedical data. Such an advance will have immediate impact beyond personalised precise radiotherapy planning and delivery.
Organisations
Publications
Ana I.L. Namburete
(2023)
Normative spatiotemporal fetal brain maturation with satisfactory development at 2 years
Ana I.L. Namburete
(2023)
Normative spatiotemporal fetal brain maturation with satisfactory development at 2 years
Chen Y
(2024)
VertXNet: an ensemble method for vertebral body segmentation and identification from cervical and lumbar spinal X-rays
in Scientific Reports
Franklin JM
(2020)
Tumour subregion analysis of colorectal liver metastases using semi-automated clustering based on DCE-MRI: Comparison with histological subregions and impact on pharmacokinetic parameter analysis.
in European journal of radiology
Gonzales RA
(2021)
MOCOnet: Robust Motion Correction of Cardiovascular Magnetic Resonance T1 Mapping Using Convolutional Neural Networks.
in Frontiers in cardiovascular medicine
Heinrich M
(2020)
Handbook of Medical Image Computing and Computer Assisted Intervention
| Title | Heart and lung segmentations for MIMIC-CXR/MIMIC-CXR-JPG and Montgomery County TB databases |
| Description | Segmenting the heart and lungs from chest X-ray images is essential for accurate disease diagnosis, enabling the calculation of image-derived digital biomarkers for cardiopulmonary health assessment. Furthermore, creation of annotated data sets can enhance development of machine learning applications in disease detection, reduce image noise for better clarity, or promote standardization for tracking disease progression or treatment effects. Generation of manual segmentations is time intensive process and it also requires access to trained medical professionals to verify the quality and accuracy of the scans. Originally made to train heart and lung segmentation/detection models for cardiomegaly diagnosis, the manual segmentations in this data paper are published here in hope of aiding development of AI models relating to heart and lungs identification in chest X-rays. This database presents the heart and lung segmentations for 200 semi- randomly chosen MIMIC-CXR/MIMIC-CXR-JPG posterior-anterior chest X-rays for the purpose of training detection and segmentation networks. Additionally, it contains the heart segmentations for the 138 posterior-anterior chest X-rays in the Montgomery Country tuberculosis database. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2023 |
| Provided To Others? | Yes |
| Impact | When developing AI tools for medical purposes a large number of high quality ground truth segmentations are a fundamental requirement. This can be a major hurdle as each segmentation is time intensive and qualified medical personnel is needed validate the accuracy and consistency of the segmentations. This database of 200 manually segmented lungs and 338 manually segmented heart, was originally curated for [1], a study developing multimodal diagnostic tool for cardiomegaly. Specifically, this database was used to train Faster-RCNN and Mask-RCNN models to build an pipeline to automatically retrieve cardiothoracic ratio and cardiopulmonary area ratio biomarker values. We hope this database of segmentations can be used to train novel segmentation and detection networks and ease the burden of manual segmentations by increasing the pool of widely available scan/segmentation pairs. [1] Duvieusart, B., Krones, F., Parsons, G., Tarassenko, L., Papiez, B.W., Mahdi, A. (2022). Multimodal Cardiomegaly Classification with Image-Derived Digital Biomarkers. In: Yang, G., Aviles-Rivero, A., Roberts, M., Schönlieb, CB. (eds) Medical Image Understanding and Analysis. MIUA 2022. Lecture Notes in Computer Science, vol 13413. Springer, Cham. doi: 10.1007/978-3-031-12053-4_2 |
| URL | https://physionet.org/content/heart-lung-segmentations-data/1.0.0/ |
| Title | Image-derived cardiomegaly biomarker values for 96K chest X-rays in MIMIC-CXR/MIMIC-CXR-JPG |
| Description | Cardiomegaly is a condition characterized by an abnormal enlargement of the heart, its identification is of paramount importance as it associate with a wide range of cardiac conditions. It is primary identified via the cardiothoracic ratio (CTR), however this metric can be inaccurate as it is affect by external factors such as breathing and body position. Multimodal approaches could mitigate these limitations by integrating non-imaging data, however reliable and explainable integration of imaging and non-imaging data remains a significant challenge. While this database does not directly use multimodal data, it hopes to tackle this challenge by extracting cardiomegaly biomarkers (CTR and cardiopulmonary area ratio) from chest X-rays. Thus encapsulating the relevant imaging information into individual datapoints, allowing easy integration of 'imaging' data with non-imaging data for more reliable diagnostic tools. The values were extracted from over 93,000 posterior-anterior MIMIC-CXR scans using detection and segmentation neural networks, tuned for cardiac and pulmonary identification. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2024 |
| Provided To Others? | Yes |
| Impact | This database presents automatically extracted CTR and CPAR values for more than 93,000 PA chest X-rays in MIMIC-CXR / MIMIC-CXR-JPG [1]. It was originally generated in the context of [2], a study developing a multimodal approach to cardiomegaly diagnosis. Specifically, a subset of these CTR and CPAR values were used in a XGBoost model with other relevant cardiac data from MIMIC-CXR [1]. We hope this database will facilitate multimodal approaches on the cardiomegaly identification challenge by extracting the relevant information from the chest X-rays thus simplifying the integration of imaging and non-imaging data. Furthermore, we hope this database will incite continued research into the use of CTR and CPAR biomarkers by making them readily available. [1] Johnson, A. et al. (2019). MIMIC-CXR-JPG - chest radiographs with structured labels (version 2.0.0). PhysioNet. https://doi.org/10.13026/8360-t248. [2] Duvieusart, B. et al. (2022). Multimodal Cardiomegaly Classification with Image-Derived Digital Biomarkers. Medical Image Understanding and Analysis. MIUA 2022: Lecture Notes in Computer Science, vol. 13413, pp. 13-27. Springer, Cham. https://doi.org/10.1007/978-3-031-12053-4_2 |
| URL | https://physionet.org/content/cxr-cardiomegaly/1.0.0/ |
| Title | METHOD AND APPARATUS FOR USE WITH A SCANNING APPARATUS |
| Description | Embodiments of the present invention provide a computer-implemented method of determining displacement information, comprising receiving (520) image data (485) comprising first pixel data (610) corresponding to movement of a scanning apparatus (130), with respect to at least one object, in a first direction and a second pixel data (620) corresponding to movement of the scanning apparatus (130) in a second direction, and determining (520) displacement information indicative of a displacement of at least a portion of the second pixel data (620) with respect to the first pixel data (610) by minimising a cost function indicative of a similarity between the first and second pixel data. |
| IP Reference | US2021272242 |
| Protection | Patent / Patent application |
| Year Protection Granted | 2021 |
| Licensed | No |
| Description | 24th UK Conference on Medical Image Understanding and Analysis |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | I co-organised the 24th conference on Medical Image Understanding and Analysis (MIUA2020) in Oxford. MIUA is the principal UK forum for communicating research progress within the community interested in image analysis applied to medicine and related biological science. The meeting is designed for the dissemination and discussion of research in medical image understanding and analysis, and aims to encourage the growth and raise the profile of this multi-disciplinary field by bringing together the various communities |
| Year(s) Of Engagement Activity | 2020 |
| URL | https://miua2020.com/ |
| Description | HDR UK Summer School |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Postgraduate students |
| Results and Impact | Summer School on Heath Data Research - delivered a lecture on medical image analysis |
| Year(s) Of Engagement Activity | 2019 |
| Description | Research workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | I co-organised the international workshop on GRaphs in biomedicAl Image anaLysis (GRAIL 2018), organised as a satellite event of MICCAI 2018 in Granada, Spain. We aimed to highlight the potential of using graph-based models for biomedical image analysis. Our goal was to bring together scientists that use and develop graph-based models for the analysis of biomedical images and encourage the exploration of graph-based models for difficult clinical problems within a variety of biomedical imaging contexts. |
| Year(s) Of Engagement Activity | 2018 |
| URL | https://grail-miccai.github.io/ |
| Description | Third International Workshop on Graphs in Biomedical Image Analysis Held in Conjunction with MICCAI 2020 |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | I co-organised the international workshop on GRaphs in biomedicAl Image anaLysis (GRAIL 2020), organised as a satellite event of MICCAI 2020 in Lima, Peru (however due to pandemic, the meeting was held virtually). We aimed to highlight the potential of using graph-based models for biomedical image analysis. Our goal was to bring together scientists that use and develop graph-based models for the analysis of biomedical images and encourage the exploration of graph-based models for difficult clinical problems within a variety of biomedical imaging contexts. |
| Year(s) Of Engagement Activity | 2020 |
| URL | https://grail-miccai.github.io/ |