Index

Abnormality detection on musculoskeletal radiographs

Thesis Description

The primary objective of this thesis project is to develop an algorithm that can determine whether a musculoskeletal X-ray study is normal or abnormal. For this purpose, we only consider X-rays of the upper extremities including the shoulder, humerus, elbow, forearm, wrist, hand, and finger. By abnormalities we consider fractures, hardware, degenerative joint diseases, lesions, subluxations, and other deviations from the standard structural composition and morphology. Given an X-ray image as an input, the devised algorithm should output a labeled image which indicates the presence or absence of an abnormality.  Such a system could be used to enhance the confidence of the radiologist or prioritize subsequent analysis and treatment options.

The task to determine abnormality on musculoskeletal radiographs is particularly critical since more than 1.7 billion people around the globe are affected by musculoskeletal conditions [12]. Since a radiograph is the cheapest, best available and usually the first measure to detect musculoskeletal abnormalities, automatic detection and localization of such potential abnormalities enables a faster initial diagnosis, saves valuable time for physicians, and reduces the number of subsequent diagnostic treatments required on the patient.  This will also reduce the work pressure and fatigue of radiologists [10] which is caused by overwhelming number of X-ray studies they have to diagnose every day [11].

In this project we will use a large public data set called ‘MURA-v1.1’ published by Stanford Machine Learning Group of Stanford University [1]. The data set consists of 14,863 studies from 12,173 patients with a total of 40,561 multi-view radiographic images. Board-certified radiologists from Stanford Hospital manually labeled the radiographs as normal or abnormal. Out of 14,863 studies 9,045 are normal and 5,818 are abnormal.

The project is structured into three parts. First, a learning-based classification algorithm is used to predict whether a radiograph is normal or abnormal [1,2]. Second, anatomical information derived from the dataset’s annotation is incorporated to additionally predict the anatomical origin of the radiograph [3,4,6,7,8]. In a last step, the abnormality is localized and visualized by incorporating the results from the previous steps in combination with targeted feature space analysis. All components should then be combined to a framework capable to predict, localize and visualize musculoskeletal abnormality. Algorithmic development is based on recent advances in deep learning techniques building upon the DenseNet [9] and ResNet [13] neural network architecture. A main aspect of the work is the conception and implementation of an integration strategy of additional anatomical information. It shall also be analyzed to what extent this information can support and improve the classification of abnormal and normal radiographs. Prior work of multi-task/multi-label optimization is investigated and examined for applicability to this project’s task [3,4,5,6,7].  The project is fixed to a six-month period timeline and will be concluded by a detailed project report. Technical implementation of the prototype will be performed within the PyTorch environment for the Python programming language.

References

1. Rajpurkar P., Irvin J., Bagul A., Ding D., Duan T., Methta H., Yang B., Zhu K., Laird D., Ball R., Langlotz C., Shpanskaya K., Lungren M., Ng A. , “MURA: Large Dataset for Abnormality Detection in Musculoskeletal Radiographs” 1st Conference on Medical Imaging with Deep Learning (MIDL 2018)
2. Guendel S., Grbic S., Gerogescu B., Zhou K., Ludwig R., Meier A., “Learning to recognize abnormalities in chest x-rays with location aware dense networks.” arxiv preprint arXiv:1803.04565 ,2018
3. Guendel S., Ghesu F., Grbic S., Gibson E., Gerogescu B., Maier A.,“ Multi-task Learning for Chest X-ray Abnormality Classification on Noisy Labels “ arxiv preprint arXiv:1905.06362 ,2019
4. Yang, X., Zeng, Z., Yeo, S.Y., Tan, C., Tey, H.L., Su, Y., “A novel multi-task deep learning model for skin lesion segmentation and classification.” arxiv preprint arXiv:1703.01025 ,2017
5. Vesal S., Ravikumar N., Maier A., ‘‘A Multi-task Framework for Skin Lesion Detection and Segmentation’’ arxiv preprint arXiv:1808.01676 ,2018
6. Kendall A., Gal Y., Cipolla R.,”Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics” arxiv preprint arXiv:1705.07115 ,2017
7. Vandenhende S., Brandere B., Gool Luc., “Branched Multi-Task Networks: Deciding What Layers To Share“ arxiv preprint arXiv:1904.02920 ,2019
8. Berlin L., ”Liability of interpreting too many radiographs.” American Journal of Roentgenology, 175(1):17–22, 2000
9. Huang G., Liu Z., Weinberger K.Q., and van der Maaten, Laurens, “Densely connected convolutional networks.” arXiv preprint arXiv:1608.06993, 2016.
10. Lu Y., Zhao S., Chu P.W., and Arenson R.L., “An update survey of academic radiologists’ clinical productivity.” Journal of the American College of Radiology, 5(7):817–826, 2008.
11. Nakajima Y., Yamada K., Imamura K., and Kobayashi K.. ,”Radiologist supply and workload: international comparison.” Radiation medicine, 26(8):455–465, 2008.
12. URL http://www.boneandjointburden.org/2014-report.
13. He K., Zhang X., Ren S., Sun J., “Deep Residual Learning for Image Recognition” arxiv preprint arXiv:1512.03385

Localization and Standard Plane Regression of Vertebral Bodies in Intra-Operative CBCT Volumes

Automatic detection of standard planes in surgical FD-CT volumes

Thesis Description

Intra-articular fractures are commonly treated by open reduction and internal fixxation (ORIF). This procedure comprises first, reorient the bone fracture into the normal position and secondly fix it using metal screws, plates or rods. Malreduction of the fracture, intra-articular position of the screws, remaining gaps, or steps offs may lead to malunion or post-traumatic osteoarthritis. The use of mobile C-arms to acquire 3D images during intervention has become a standard since it enables the evaluation of fractures of complex anatomical regions. Two dimensional images often lack information about fracture reduction and implant position in non-planar joints [1]. This includes fractures of the tibial head as well as the calcaneus, ankle injuries involving the syndesmosis or spinal injuries among others. After fracture treatment, if the surgeon is not satisfied with the result, a correction can be made within the frame of the intervention and can avoid the patient a revision surgery in the future. Several studies show intraoperative revision rates depending on the anatomical region up to 40% [2].

Acquisition of standard planes that contains key anatomical structures is decisive for the assessment of intervention results. Multiplanar reconstruction (MPR) is the standard method for reconstruction of the 3D image which allows the generation of slices from arbitrary viewpoint and orientation. Absence of information about position between patient and the C-arm device results in the need for adjustment of the standard planes at a workstation in the operating room. Till now, surgeons must manually find standard planes orientation and position which takes from 46 to 55 second depending
on the experience level of the surgeon and can thus be considered a time-consuming and complicated task [3].

No methods for the fully automatic adjustment of standard planes of mobile C-arms have been described. However, it is possible to find several works in other modalities as ultrasound. In [4] a CNN is used to detect transventricular and transcerebrall standard planes in fetal brain ultrasound. The network learns the mapping between a 2D plane, and the transformation required to move the plane towards the standard plane in the volume. Another approach used in [5] is based on reinforcement learning approach to automatically localize transthalamic and transcerebellar standard planes in 3D fetal ultrasound.

This thesis aims to design a framework for the automatic adjustment of standard planes in different anatomical joint regions using deep learning algorithms. The thesis will comprise the following work items:

• Literature overview of state-of-the-art automatic standard plane adjustment
• Characterization of standard planes for different anatomical regions
• Design and formalization of the to be developed method
• Overview and explanation of the algorithms used
• Implementation of the plane detection framework
• Evaluation of results

References

[1] Paul Alfred Grützner. Rontgenhelfer 3D: Handbuch intraoperative 3D-Bildgebung mit mobilen C-Bögen. Bengelsdorf & Schimmel, 2004.

[2] Jochen Franke, Klaus Wendl, Arnold J Suda, Thomas Giese, Paul Alfred Grützner, and Jan von Recum. Intraoperative three-dimensional imaging in the treatment of calcaneal fractures. JBJS, 96(9):e72, 2014.

[3] Michael Brehler, Joseph Gorres, Jochen Franke, Karl Barth, Sven Y Vetter, Paul A Grützner, Hans-Peter Meinzer, Ivo Wolf, and Diana Nabers. Intra-operative adjustment of standard planes in C-arm CT image data. International journal of computer assisted radiology and surgery, 11(3):495-
504, 2016.

[4] Yuanwei Li, Bishesh Khanal, Benjamin Hou, Amir Alansary, Juan J Cerrolaza, Matthew Sinclair, Jacqueline Matthew, Chandni Gupta, Caroline Knight, Bernhard Kainz, et al. Standard plane detection in 3D fetal ultrasound using an iterative transformation network. In International Conference on Medical Image Computing and Computer-Assisted Intervention, p. 392-400. Springer, 2018.

[5] Haoran Dou, Xin Yang, Jikuan Qian, Wufeng Xue, Hao Qin, Xu Wang, Lequan Yu, Shujun Wang, Yi Xiong, Pheng-Ann Heng, et al. Agent with warm start and active termination for plane localization in 3D ultrasound, 2019.

Micro CT Denoising Using Low Parameter Models

Computed Tomography (CT) is widely used as a diagnostic tool due to its ability to acquire three-dimensional structures while preserving great bone-soft tissue contrast. Motivated by these contrast properties, it is instructive to use high-resolution CT imaging (Micro CT) in preclinical osteoporosis research to resolve bone structures in mice. Especially in vivo Micro CT imaging of mouse tibia bones is interesting for understanding osteoporosis and developing a medication [1]. However, radiation dose and image quality are strongly connected. A significant amount of radiation must be deposited in the imaged object to acquire a desired contrast. When scanning a living animal, the deposited energy will harm the tissue and increase the risk of cancer and other diseases. Therefore, minimizing the dose is crucial, which is usually connected to degraded image quality.
Using denoising algorithms can leverage image quality. Here, iterative reconstruction algorithms have been successfully applied in the past. While their algorithms are usually based on reasonable statistical assumptions, these methods are computationally costly and limited in their denoising performance. In recent years, deep learning approaches have shown promising results in terms of image quality.

The goal of this master thesis is to use the deep learning-based joint bilateral filtering (JBFnet) [2] to denoise Micro CT data of mouse tibia bones. The JBFnet is a promising approach for denoising Micro CT data as it requires only a few trainable parameters while achieving state-of-the-art denoising performance. Hence, the integrity of the denoised structures can be claimed which is crucial considering the tiny bone features that shall be restored. After achieving reasonable denoising results, multiple modifications of the JBFnet are planned to adapt the filtering better to the respective noise characteristics of the data. In the last part of the thesis, an extensive performance evaluation of the network and its modifications will be performed.

[1] A. Grüneboom, L. Kling, S. Christiansen, L. Mill, A. Maier, K. Engelke, H. H. Quick, G. Schett, and M. Gunzer, “Next-generation imaging of the skeletal system and its blood supply,” Nature Reviews Rheumatology, vol. 15, no. 9, pp. 533–549, 2019.

[2] M. Patwari, R. Gutjahr, R. Raupach, and A. Maier, “Jbfnet-low dose ct denoising by trainable joint bilateral filtering,” in International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 506–515, Springer, 2020.

Multi-scale Tissue Segmentation on Canine Cutaneous Tumors

Cutaneous tumors, i.e., tumors originating from skin cells, are one of the most common tumor types in dogs [1]. As cost efficiency is an important driver in animal care, it would be strongly beneficial to support veterinary pathologists in the diagnosis of those tumors and their respective subtypes. Besides the use in a decision-support system, computerized segmentation and classification of tumors can potentially increase precision for therapeutic options and – through quantitative evaluation – provide new insights into tumor development with inter-species relevance, which includes humans.

This thesis aims to perform tissue segmentation from a data set comprising of the nine most common canine cutaneous tumor types of whole slide images. A particularly challenging component is the combination of predictions performed for differing magnifications: while some tissue types can be spotted with higher accuracy using lower magnifications, for others human experts will utilize higher magnifications, especially for fine-grained differentiation versus neighboring tissue segments.

The thesis comprises the following items:

• Literature review concerning detection on multi-scale images
• Training of state-of-the-art segmentation networks on multiple scales of microscopy images
• Analysis of detection results on single scales
• Development of a multi-scale fusion system, achieving high precision at multiple image scales
• Documentation and presentation of the findings, documentation of code

[1] Murphy, S. (2006). Skin neoplasia in small animals 3. Common canine tumours. In practice, 28(7), 398-402.

Real-Time Prospective Respiratory Triggering for Free-Breathing Lung Computed Tomography

Thesis Description

Respiratory diseases are among the leading causes of death, according to the World Health Organization. With more than 3 million deaths in 2016, chronic obstructive pulmonary disease is the third leading cause of death worldwide [1]. An important tool for diagnosing respiratory diseases is computed tomography of the lungs [2]. The current state of the art approach for this is breath-hold CT, which requires patients to follow breathing cues and hold their breath on command [3]. This procedure is not appropriate for certain groups of patients who are unable to follow instructions. Some of these patients are unconscious, mentally impaired, or are infants and young children [4]. In these cases, medication must be administered to stop the breathing so that sufficient scans can be acquired. This not only carries risks for the patient, but also requires additional clinical staff [4, 5, 6].

To improve these issues, free-breathing computed tomography has been proposed. In this approach, scans are performed while the patient continues to breathe [4, 5, 7]. Because the lungs move during the scan, this method produces images with more artifacts, compared to the breath-hold approach. To optimize image quality and for comparability between scans, it would be beneficial to scan at times with little lung movement, such as during inhalation and exhalation. One way to achieve comparability is retrospective respiratory triggering, which uses the respiratory waveform to select the correct phase after CT images are acquired. However, this is not ideal for clinical use because of the high radiation dose involved. To achieve lower radiation exposure, prospective respiratory triggering utilizes a shorter duration scan triggered by a respiratory gating device. [8, 9]

To address the challenges associated with the continuous lung movement, two recording modes are further investigated.

• Sequential mode acquires images without CT couch motion, with image size limited by collimator width.  Multiple images are stacked to cover a larger area. This is expected to work better for children, who tend to have a high respiratory rate and a small lung area to scan [10].
• The flash spiral mode utilizes a technique previously used for cardiac imaging with a high-pitch spiral CT [11]. This makes it possible to scan the entire lung in one pass in less than one second. The advantages of this high pitch mode for lung imaging have already been demonstrated [4, 5, 8]. Promising results were achieved with regularly breathing patients, although the limitations of triggering with respect to irregular breathing were noted by Goo et al. [8]

For both of these approaches, triggering algorithms expand on phase space based respiratory triggering as presented by Werner et al.. A set of criteria is used to define target regions in the phase space representation of the respiratory signal data in order to emit a trigger depending on amplitude and velocity [12]. The goal is to achieve the best possible robustness on patient data with a focus on especially challenging breathing patterns.

This thesis deals with the following work items:

1. Customization of respiratory triggering algorithms to both acquisition modes
   1. Sequential respiratory triggering
   2. Flash spiral respiratory triggering
2. Evaluation of robustness against other approaches
   1. Commercial respiratory gaiting
   2. Possibly: Reinforcement learning based triggering
3. Image based validation of the reconstruction result
   1. With simulated breathing signals
   2. With breathing signals from real patients

References

[1] World Health Organization (WHO) et al. Global health estimates 2016: estimated deaths by age, sex and cause. Geneva: WHO, 2018.

[2] Darel E Heitkamp, Matthias M Albin, Jonathan H Chung, Traves P Crabtree, Mark D Iannettoni, Geoffrey B Johnson, Clinton Jokerst, Barbara L McComb, Anthony G Saleh, Rakesh D Shah, et al. Acr appropriateness criteria® acute respiratory illness in immunocompromised patients. Journal of thoracic imaging, 30(3):W2–W5, 2015.

[3] Toshizo Katsuda, Shigeru Eiho, Chikazumi Kuroda, and Tsutomu Hashimoto. Analysis of breath holding for lung ct imaging. Radiography, 11(4):235–241, 2005.

[4] Michael M Lell, Michael Scharf, Achim Eller, Wolfgang Wuest, Thomas Allmendinger, Florian Fuchs, Stephan Achenbach, Michael Uder, and Matthias S May. Feasibility of respiratory-gated high-pitch spiral
ct:: Free-breathing inspiratory image quality. Academic radiology, 23(4):406–412, 2016.

[5] Ilias Tsiflikas, Christoph Thomas, Dominik Ketelsen, Guido Seitz, Steven Warmann, Claus Claussen, and Juergen Schaefer. High-pitch computed tomography of the lung in pediatric patients: an intraindividual comparison of image quality and radiation dose to conventional 64-mdct. RoFo, 186(6):585–590, 2014.

[6] Mike Sury, Ian Bullock, Silvia Rabar, and Kathleen DeMott. Sedation for diagnostic and therapeutic procedures in children and young people: summary of nice guidance. Bmj, 341, 2010.

[7] Hyun Woo Goo. Combined prospectively electrocardiography-and respiratory-triggered sequential cardiac computed tomography in free-breathing children: success rate and image quality. Pediatric radiology, 48(7):923–931, 2018.

[8] Hyun Woo Goo and Thomas Allmendinger. Combined electrocardiography-and respiratory-triggered ct of the lung to reduce respiratory misregistration artifacts between imaging slabs in free-breathing children: initial experience. Korean journal of radiology, 18(5):860, 2017

[9] Cyrus Behzadi, Michael Groth, Frank Oliver Henes, Dorothee Schwarz, Andr´e Deibele, Philipp GC Begemann, Gerhard Adam, and Marc Regier. Intraindividual comparison of image quality using retrospective
and prospective respiratory gating for the acquisition of thin sliced four dimensional multidetector ct of the thorax in a porcine model. Experimental lung research, 41(9):489–498, 2015.

[10] Edmond A Hooker, Daniel F Danzl, Mary Brueggmeyer, and Edith Harper. Respiratory rates in pediatric emergency patients. The Journal of emergency medicine, 10(4):407–410, 1992.

[11] Stephan Achenbach, Mohamed Marwan, Tiziano Schepis, Tobias Pflederer, Herbert Bruder, Thomas Allmendinger, Martin Petersilka, Katharina Anders, Michael Lell, Axel Kuettner, et al. High-pitch spiral
acquisition: a new scan mode for coronary ct angiography. Journal of cardiovascular computed tomography, 3(2):117–121, 2009.[12] Ren´e Werner, Thilo Sentker, Frederic Madesta, Tobias Gauer, and Christian Hofmann. Intelligent 4d ct sequence scanning (i4dct): concept and performance evaluation. Medical physics, 46(8):3462–3474, 2019.

Deep-learning-based behaviour prediction of rear-end road users when changing lane as a system design reference for highly automated driving

Motivation

Until fully automated vehicles reach full market saturation, a mixed operation between human-driven and highly automated vehicles will characterize traffic activities. Mutual understanding of driving intentions is therefore crucial for harmonizing road traffic.

The prediction of driving intentions of other road users is a subject of numerous scientific researches and is the link between environmental perception and maneuver planning. The driving environment is determined by kinematic vehicle parameters and their temporal history and as well as the context of the traffic situation. Based on these, predictions about future trajectories of the surrounding vehicles are made possible and one’s own target behavior can be derived in the form of a target trajectory.

However, the influence of one’s own driving behaviour on other road users is only part of very few investigations yet. For the acceptance of highly automated driving functions, it is not only essential that the driving behaviour is perceived safe and comfortable by passengers of a highly automated vehicle, but also predictable by other road users.

Approach

The thesis aims to gain specific insights on interactions between road users. The core target of this thesis is to train a model, which describes how a lane change influences the behaviour of rear-end road users on highways. The highD dataset will be used to build a ‘Deep Learning’ model, which learns the dependencies between lane changes and the reactions caused by it.

The goals of this thesis are:

• Defining the scenario
• Identifying the relevant input and output parameters for the deep learning module
• Creating and training of a suitable model
• Using the developed model to provide a reference for mutual interactions between road users and to derive possible behavioural patterns
• Assessing the impact on rear road users when changing lane for highly automated driving

Predictive Maintenance for SINAMICs Frequency Converter

Thesis Description

Maintaining machines causes large expenses in the modern industrial world. For example, in the product manufactoring sector, maintenance makes up to 15%-60% of their costs [1]. To reduce these costs, more efficient methods of maintenance were invented. In general, Maintenance can be divided into three different categories [2]:

1. Corrective: The machine’s failure has already occured.
2. Preventive: The maintanence is done on a regular basis to decrease the likelihood of a failure.
3. Predictive: Keep track of the machine’s condition and prediction of its failure occurance.

Predictive Maintenance uses industrial data modeling and analysis to perform equipment fault diagnosis through real-time monitoring. By direct application of domain knowledge onto given problems, it is possible to predict that status and its trend with respect to possible failures. With this, maintenance plans become more accurate. This reduces unplanned downtime as well as cost of operation [1]. One common problem in the field of predictive maintenance is the detection of anomalies. As the name implies, the classifier’s task is to detect anomalies within data by comparing it to expected data. These anomalies are used as indicators for problems and basis for the failure prediction [2]. Anomalies can be found in any kind of data. In this thesis the focus is on machines with a SINAMICS Low
Voltage Converter [3]. In general, a converter adapts the constant AC to AC with various frequency and voltage [4]. The SINAMICS Low Voltage Converter produces sensor data in the high frequency spectrum based on a low voltage [3]. The data is gathered and computed on an Industrial Edge [5]. Edge Computing is defined as any computing done at the edge of the network that produced the data. Compared to Cloud Computing which bottleneck is the massive data transfer, Edge Computing saves a lot of bandwith since the data is already processed and only the relevant information is sent to the cloud [6].
Anomaly detection is already used in different security aspects like fraud, intrusion detection but also regarding medical health anomalies. As a result, there exist various methods for different kind of data (continous, discret, labeled or unlabeled etc.). Since in this thesis not all anomalies are known, the given data is unlabeled and there is a time element, the focus in this thesis will be on methods for unsupervised time series data, for example, Hidden Markov Models, Clustering, Neural Network and Nearest neighbour, etc. [7, 8, 9]. The goal of this thesis is to apply different methods for the detection and classification of signal anomalies resulting in a comparison between them regarding resource usage (e.g. RAM), runtime and accuracy for a specific type of converter. To be able to accurately predict any failures at the machine, the anomalies will be ranked regarding their severness in respect to the machine. This thesis aims to find a robust, exible and unsupervised model for anomaly detection in high frequency, low voltage sensor data as a basis for Predictive Maintenance. The data is computed directly on an Industrial Edge device (227E, 427E) attached to the SINAMICS Low Voltage Converter. In summary, the thesis deals with the following points:

1.  Define anomalies and their consequences/meanings for the SINAMICS Low Voltage Converter:
   (a) Detection and identification of anomalies
   (b) Ranking of found anomalies based on their inuence
2. Development of different models for anomaly detection and classification
3. Evaluation based on the different models considering the following points:
   (a) Accuracy
   (b) Flexibility
   (c) Resource Usage
   (d) Robustness
   (e) Runtime

References
[1] Bei-ke Zhang Cheng Cheng and Dong Gao. A predictive maintenance solution for bearing production line
based on edge-cloud cooperation. In 2019 Chinese Automation Congress (CAC), pages 5885{5889, 2019.
[2] Masaru Kurihara Pushe Zhao, Shigeyoshi Chikuma Junichi Tanaka, Tojiro Noda, and Tadashi Suzuki. Advanced
correlation-based anomaly detection method for predictive maintenance. In 2017 IEEE International
Conference on Prognostics and Health Management (ICPHM), pages 78{83. IEEE, 2017.
[3] Siemens Global Website. Siemens Digital Industries. https://new.siemens.com/global/en/products/
drives/sinamics/low-voltage-converter.html, 1996-2021. [Online; accessed 23-02-2021].
[4] Pawel Szczesniak Zbigniew Fedyczak and Marius Klytta. Ac drives with buck-boost voltage switched frequency
converters without dc storage. In 2018 International Symposium on Power Electronics, Electrical
Drives, Automation and Motion (SPEEDAM), pages 507{512, 2018.
[5] Siemens Global Website. Industrial Edge. https://new.siemens.com/global/en/products/automation/
topic-areas/industrial-edge/simatic-edge.html, 1996-2021. [Online; accessed 18-02-2021].
[6] Jie Cao Weisong Shi, Youhuizi Li Quan Zhang, and Lanyu Xu. Edge computing: Vision and challenges.
IEEE Internet of Things Journal, 3(5):637{646, 2016.
[7] Charu C. Aggarwal Manish Gupta, Jing Gao and Jiawei Han. Outlier detection for temporal data: A survey.
IEEE Transactions on Knowledge and Data Engineering, 26(9):2250{2267, 2014.
[8] Lei Clifton Marco A.F. Pimentel, David A. Clifton and Lionel Tarassenko. A review of novelty detection.
Signal Processing, 99:215{249, 2014.
[9] Arindam Banerjee Varun Chandola and Vipin Kumar. Anomaly detection: A survey. ACM Computing
Surveys, 41:1{58, 2009.

Optimization of the fat-water separation for muscle imaging at 7 T with application in quantitative Na23/K39 MRI

Magnetic Resonance Imaging (MRI) is a noninvasive imaging method, without ionizing radiation, which
has the ability to particularly resolve soft tissue such as muscle, fat, and connective tissue. Numerous
studies have shown the ability of MRI to detect alterations in skeletal muscle structure and composition,
for example, patients with Duchenne Muscular Dystrophy (DMD) [1]. The muscle destruction observed
in muscle dystrophy patients is associated with ion homeostasis dysregulation and chronic in
ammation,
which leads ultimately to bro-fatty replacement of muscle tissue [2]. An increase in total sodium
concentration (TSC) has been observed in these patients in addition to elevated intra-cellular weighted
sodium signal (ICwS) based on an inversion recovery (IR) method. [3] Na23/K39 MRI detects a muscular
Na+ overload in these patients and thus it could depict early changes in the ion homeostasis in skeletal
muscle tissue of these patients. [2]
In this thesis, we aim to optimize the Na23/K39 quantication in muscle tissue by combining Na23/K39
MRI data with fat fractions (the proportion of the acquired signal derived from fat protons) obtained
by H1 MRI. As sodium and particularly potassium concentrations are strongly reduced in fat tissue
compared to muscle tissue, the Na23/K39 quantication in fat inltrated muscles is generally distorted.
Thus, a fat correction needs to be applied to obtain the real ion concentrations of muscle tissue. This
can be achieved by applying the Multiple point Dixon technique for Water/Fat decomposition on the
H1 MRI Images acquired at 7T. The original fat quantication is based on the fact that fat and water
possess dierent resonance frequencies in MRI, which is called the Chemical Shift. Using a gradient
echo-based Dixon acquisition with very closely spaced echo times, the fat and water from these images
can be successfully separated using Dixon’s fat water separation algorithm [4]. It is based on the
principle that the signal in the image acquired when the water and fat have the same phase, interfere
constructively whereas the images acquired when water and fat are in opposed phase they interfere
destructively. Thus fat inltration in the muscle tissue of a muscle dystrophy patient could be studied
using the Fat fraction [5].
The goal of the thesis is to create an image processing protocol at 7T correctly maps the fat and
water in the calf muscles and generate a fat-fraction which can further be used on the Na23 images where
both the 1H and Na23 images are acquired at a 7T MRI. Currently, no such protocols exist at 7T, and
adapting them from 3T comes with various challenges. This thesis aims to eectively facilitate easy and
early diagnosis for various muscular dystrophies.
1
The following points will be covered in this thesis:
1. Design and production of a measurement phantom
(a) Prerequisites: Multiple compartments (e.g. cylinders) containing dierent fat-water fractions,
size tting into 1H knee coil @ 7T (and other typical 1H coils, e.g. @3T)
(b) Research on typical fat-fractions in muscle tissue/materials suitable for achieving dierent
fat fractions/phantoms used in other publications on fat-water imaging; and thereby design
the phantom.
2. Optimization of measurement protocol for fat-water imaging at 7T
(a) Compare and evaluate existing H1 MRI acquisition protocols for fat-water separation at 3T,
choose the most suitable .rotocol to be translated to 7T.
(b) Optimize multiple-echo acquisition protocols by calculating optimized TEs for fat-water separation
at 7T.
3. Optimization of post-processing for fat-water imaging
(a) Comparison of dierent algorithms for fat-water separation using the ISMRM fat-water toolbox.
(b) Study eects of B0 inhomogeneities using dierent B0 shimming protocols to resolve phase
wrapping artifacts.
4. Application of fat-water imaging in vivo
Application of optimized fat-water imaging protocol to healthy subjects (approx. 5); Quantication
of fat fraction in healthy muscle tissue and comparison of results to literature; Maybe (towards
the end of thesis): application to patients with muscular dystrophies
5. Comparison of fat-water imaging at dierent eld strengths
Optimize protocol (or use existing protocols from literature) for dierent eld strengths, e.g.3T
and 7T; Compare resulting fat fraction values in phantom and in vivo measurements (healthy
muscle)

References
[1] Erika L. Finanger, Barry Russman, Sean C. Forbes, William D. Rooney, Glenn A.Walter, and Krista
Vandenborne. Use of skeletal muscle mri in diagnosis and monitoring disease progression in duchenne
muscular dystrophy. Physical medicine and rehabilitation clinics of North America, 23(1):1{ix, 2012.
[2] Teresa Gerhalter, Lena V. Gast, Benjamin Marty, Jan Martin, Regina Trollmann, Stephanie
Schussler, Frank Roemer, Frederik B. Laun, Michael Uder, Rolf Schroder, Pierre G. Carlier, and
Armin M. Nagel. 23na mri depicts early changes in ion homeostasis in skeletal muscle tissue of patients
with duchenne muscular dystrophy. Journal of Magnetic Resonance Imaging, 50(4):1103{1113,
2019.
[3] Armin M. Nagel, Marc-Andre Weber, Arijitt Borthakur, and Ravinder Reddy. Skeletal Muscle MR
Imaging Beyond Protons: With a Focus on Sodium MRI in Musculoskeletal Applications, pages
115{133. Springer Berlin Heidelberg, Berlin, Heidelberg, 2014.
[4] W. T. Dixon. Simple proton spectroscopic imaging. Radiology, 153(1):189{94, 1984.
[5] T. J. Bray, M. D. Chouhan, S. Punwani, A. Bainbridge, and M. A. Hall-Craggs. Fat fraction mapping
using magnetic resonance imaging: insight into pathophysiology. Br J Radiol, 91(1089):20170344,
2018.
3

Incorporating GAN-Translated Tomosynthesis Images for improved automatic Lesion Detection in Mammography Images

Computer Aided Diagnosis (CAD) systems assist physicians in the interpretation of medical images. One of the most common application for CAD systems is in the field of mammography where they help in the detection and analysis of lesions e.g. masses, microcalcifications and tumors. X-ray mammography is the most used method for breast cancer screening and CAD systems for this modality are well established. CAD systems do also exist for other modalities like ultrasound and magnetic resonance imaging. A further modality is digital breast tomosynthesis (DBT) which uses image recordings from different angles to produce a 3D-representation of the breast. These 3D representations can overcome some shortcomings of conventional X-ray mammography like overlapping breast tissue due to the necessary breast compression during an examination. Using digital breast tomosynthesis has shown to increase accuracy and leading to a reduced false positive rate in breast cancer screening.

CAD systems are usually trained for a single modality and therefore often not applicable to other modalities. They rely upon state of the art machine learning approaches like deep learning. In general deep learning needs high amounts of data for training which for specific problems is often not available or hard to come by. Especially labelling and annotation of images is a labour-intensive and expensive task. Being able to come by the issue of limited modality usage and the problem of few data by combining different data sets would be beneficial for an automatic lesion detection method. However, combining data from different sources usually makes further model adaption necessary.

For this thesis two data sets of labeled mammography images are available. The first data set contains X-ray mammograms of 237 patients. The second consists of digital breast tomosynthesis images of 42 patients. The task of this thesis is to incorporate tomosynthesis images into a method for lesion detection in mammograms. This incorporation should improve the methods ability to detect lesions in mammograms and expand the method’s usability to a second modality.

This task can be considered as a problem of Domain Adaption [1] with two source and one target distribution. The source distributions are mammograms and tomosynthesis images whereas the target distribution are mammograms. One method of Domain Adaption appropriate for this task is adversarial-based Domain Adaptation with generative models. In the medical context Generative Adversarial Networks (GANs) have been applied for e.g. the translation of images between modalities, denoising of images and artifact correction [2]. In this thesis GANs will be used to generate synthetic mammograms out of tomosynthesis data which will be used to enhance the data set used for the training of a deep convolutional neural network (CNN). The hypothesis is that using such an enhanced data set for model training leads to an improved performance in lesion detection compared to training with a smaller data set solely consisting of genuine mammograms.

The master thesis contains several milestones:

1. Creation of a baseline deep CNN model trained and evaluated for lesion detection in mammograms.
2. Development and optimization of a Cycle-GAN [3] to translate tomosynthesis images into mammograms.
3. Evaluation of the classification performance for lesion detection of a deep CNN model trained on mammograms and Cycle-GAN generated mammograms.
4. Application of the model obtained in 2 to verify the hypothesis on another publicly available mammogram data set.

[1] Wang M. and Deng W. (2018). Deep visual domain adaptation: A survey . Neurocomputing, 312,135–153. https://doi.org/10.1016/j.neucom.2018.05.083
[2] Armanious K., Jiang C., Fischer M., Küstner T., Hepp T., Nikolaou K., Yang B. (2020). MedGAN: Medical image translation using GANs. Computerized Medical Imaging and Graphics,79, 1–16. https://doi.org/10.1016/j.compmedimag.2019.101684
[3] Zhu J. Y., Park T., Isola P. and Efros A. A. (2017). Unpaired Image-to-Image Translation Us-ing Cycle-Consistent Adversarial Networks. Proceedings of the IEEE International Conference on Computer Vision, 2017-October, 2242–2251. https://doi.org/10.1109/ICCV.2017.244