Master degree

FOR

2019

Anders Johannessen

Faculty of Information Technology and Electrical Engineering
Department of Engineering Cybernetics
Norwegian University of Science and Technology
Master’s thesis in Cybernetics and Robotics

Affine Alignment of Ultrasound
Volumes Using Deep Learning
This thesis explores the possibilities of deep learning based image registration of 3D ultrasound volumes. Two models for affine transformation of a moving image in reference to a target image is being compared. The first model is inspired by the affine transformation framework proposed by de Voss et al. (2019). This model passes the full ultrasound image directly through a deep learning network to predict the optimal affine transformation matrix that maximizes the Normalized Cross-Correlation (NCC) similarity measure. The second model proposed in this thesis is inspired by the framework for Large Deformation Diffeomorphic Metric Mapping (LDDMM) developed by Yang et al. (2017). This model divides the 3D image into multiple smaller patches of uniform size before passing them through a deep learning network for transformation matrix prediction. The change in NCC between the moving and target image before and after transformation, is being used as a metric for performance of the two models.

Supervisors: Gabriel Kiss and Hans Torp
June 2019

Torjus Haukom

Faculty of Information Technology and Electrical Engineering
Department of Electronic Systems
Norwegian University of Science and Technology
Master’s thesis in Electronics Systems Design and Innovation

Basal Strain Estimation in Transesophageal Echocardiography using Unsupervised Deep Learning
This thesis aims to contribute towards the full automation of perioperative echocardiographic monitoring through investigating the feasibility of fast, automatic longitudinal strain estimation in the basal segments from unselected 4-chamber, 2-chamber, and longaxis TEE images using unsupervised deep learning methods. A strain estimation pipeline is proposed, composed of two major components: myocardial landmark detection and frame-to-frame displacement estimation. Using the estimated displacements, the detected landmarks can be tracked through the cardiac cycle and used to estimate strain. The landmark detection algorithm assumes known mitral annulus location and employs a series of filtering operations to highlight a suitable landmark in the myocardial segment below it. The displacements are estimated using a fully convolutional neural network (CNN) and cubic B-spline interpolation, inspired by recent work in image registration. The CNN is trained in an unsupervised manner, removing the need for manual annotation of the ground truth, and estimates a low-resolution displacement field. This low-resolution field is then interpolated to produce a dense displacement field describing the motion of each individual pixel between two consecutive frames.

Supervisors: Gabriel Hanssen Kiss and Ilangko Balasingham
June 2019

Trym Nordal

Faculty of Information Technology and Electrical Engineering
Department of Engineering Cybernetics
Norwegian University of Science and Technology

Master’s thesis in Cybernetics and Robotics
Automatic Detection of Mitral Annular Plane Systolic Excursion from Transesophageal Echocardiography Using Deep Learning
The study presented in this thesis is based on the hypothesis that perioperative monitoring and assessment of cardiac function from TEE can be automated by utilizing the power of deep learning. A pipeline for automatic detection of the global systolic functional parameter MAPSE (mitral annular plane systolic excursion) has been proposed, implemented and tested against clinical measurements from St. Olavs University hospital, Trondheim, Norway. The pipeline consists of a convolutional neural network (CNN) for detection of a set of mitral landmarks in two-dimensional TEE B-mode recordings of the left ventricle, and necessary post processing components in order to get a final MAPSE estimate. The CNN landmark detector has been trained with 131 two-chamber and four-chamber recordings.

Supervisor: Lasse Løvstakken and Gabriel Hanssen Kiss
June 2019


Christoffer Pedersen

Department of Energy and Process Engineering

Studies have shown that it is possible to get less than 10 bacteria per day. cubic meters of air (referred to as ultra-clean requirements (10CFU / m3)) in operating rooms equipped with mixing ventilation. However, variations in measured CFU level exist between operations, leading to uncertainty about how consistent blend ventilation manages to meet the ultra-clean requirement. Studies have suggested that the activity level of the surgical staff is an important factor affecting the bacterial level in the air, but few studies have quantified this hypothesis.
Autumn and spring semester 2019 I wrote a project and a master's thesis in collaboration with the operating room of the Future. The aim of the thesis was to analyze the conditions that must be present in an operating room equipped with mixing ventilation in order for the ultra-clean requirement to be met, and in our task the focus was on activity level. To find out, we arranged standardized play operations in the FOR living room at AHL, which is equipped with mixing ventilation, and performed bacterial measurements of the air at different activity levels on the operating staff.
The measurements from 5 play operations showed an average bacterial level of 12.9 CFU / m3, with a standard variation of 6, when the operating personnel were active. When the operating staff was at rest, the bacterial level was 2.5CFU / m3.
The results indicate that activity level is an important factor in keeping a low bacterial level in the air during surgery. An operating room equipped with mixing ventilation can fairly consistently meet the ultra-clean requirement if the right conditions are present, but single operations can differ quite a bit from the requirement, even if the right conditions are present.


Minchao Fan

Department of Energy and Process Engineering

A suitable ventilation system in operating theater should use relatively less investment and energy consumption to create an environment which perfectly meets the standard requirements of clean room, at the same time ensure the staffs’ and patients’ comfort. In the evaluation framework, evaluation indices were determined from the aspects of ventilation effectiveness (sub-indicators: bacteria and particle concentration, temperature and relative humidity), energy consumption as well as users’ satisfaction (sub-indicators were thermal comfort and noise level) and the major aspects were given weights 0.49, 0.35 and 0.16, respectively based on expert survey and analytic hierarchy process method. Through the combination of weights and the corresponding index values from the measurements during mock surgeries, the fuzzy comprehensive evaluation method was used to obtain the final comments of the degree of membership of the object to be evaluated as “unsuitable, tolerable and suitable”. This evaluation framework is flexible since more evaluation indicators can be added, such as lifecycle cost and environmental influence; the evaluation criteria of the indicators can also be adjusted according to the applied countries or regions.

Two operating rooms with laminar air flow and mixing ventilation at St. Olavs Hospital were evaluated respectively. The evaluation results showed that both operating rooms could provide clean and temperature-appropriate operating environment, while operating room with mixing ventilation had more dissatisfaction vote in thermal comfort and the electricity consumption was at the level a little bit above median value of all the operating rooms. Therefore, the case operating room with mixing ventilation was evaluated less suitable than the one with laminar air flow system. From the evaluation results and energy calculation, we may suggest that design and selection of the ventilation system for operating rooms should focus on appropriate air volume on the basis of meeting the demand of safety, then make rational use of the heat recovery technology as well as setback strategy. Regularly test the air cleanliness, temperature, humidity and noise in the operating theaters and keep tracking the feelings of the operating team members to provide the basis for possible system adjustment.

Supervisor: Prof. Guangyu Cao

Helena Kuivjõgi

Department of Energy and Process Engineering

In healthcare facilities and hospital environment, it is essential to ensure a certain thermal comfort level for all occupants. Unsatisfactory thermal conditions in operating rooms (ORs) will influence the performance of surgical staff and increase the risk of developing surgical site infections (SSIs) possibility of patient.

During my exchange semester spring 2019 in Norwegian University of Science and Technology (NTNU) I completed my master thesis in collaboration with the Operating Room of the Future with the support of Liv-Inger Stenstad and Jan-Gunnar Skogås. My supervisor was professor Guangyu Cao from the Department of Energy and Process Engineering at NTNU.

The purpose was to investigate the thermal comfort level of patient and surgical staff in operating rooms at St. Olavs hospital and give the suggestions for improvement. The main work of this master thesis has been performed in two ORs with LAF ventilation ("Stue 8" at "Bevegelsessenteret") and in mixing ventilation ("Stue 1" at "AHL-senteret") solution.

This study includes measurements and survey in both operating rooms. Measurements shows, that in the mixing ventilation OR, the air temperature and air velocity are comfortable for patient, but uncomfortable for surgeon and assistant nurse thermal comfort (warm). In contrast, in LAF OR, the air temperature and air velocity are comfortable for surgical staff, but may be uncomfortable for patient (higher velocity feels cold).

Based on survey results there is an acceptable thermal comfort level on surgical staff in the operating room with LAF, but in the OR with mixing ventilation the occupants feel in some cases cold (clean zone nurse) and in some cases extremely warm (surgeons). The results show that the thermal comfort level varies greatly from surgeons to patient in different operating room, therefore this should get more attention in hospitals.

2018

Geir Andre Pedersen

Master of Biomedicine at the Department of Natural Sciences, Faculty of Health Sciences, OsloMet

In cooperation with St. Olav's hospital.

«Quality assurance and technical aspects of embolization of uterine myomas»

Radiological intervention procedures are constantly growing as a result of new and more modern devices and examination methods. Due to the fact that the largest human-induced radiation dose to the population comes from radiological procedures, it is important to provide a knowledge base on the quality and effectiveness of such procedures. It is important to meet national quality requirements and that the treatment the patients receive is effective and safe.

The purpose was to carry out an internal quality assurance study of the procedure UAE at St Olav's hospital as part of systematic quality work. This quality assurance study was conducted to elucidate whether the patients receive the treatment they are entitled to, and whether the treatment is in line with current recommendations. It was therefore chosen to look at a well-established method for embolization of uterine myomas (UAE) to assess this. Both efficiency and patient safety were in this quality study compared to international and national findings.

Supervisor: Knut Haakon Stensæth (St. Olav's Hospital),
Supervisors: Berit J. Brattheim (NTNU) and Siv-Marit Lamøy (St. Olav's hospital).
Posted: 15.12.2018.

Martin Henrik Hassel

Master of Science Department of Electronic Systems, Faculty of Information Technology and Electrical Engineering, NTNU

"Perioperative monitoring of cardiac function based on transesophageal echocardiographic data"

Major cardiac interventions are known to have a negative impact on heart function. Patients that undergo cardiac surgery have their heart carefully monitored and evaluated perioperatively in order to determine whether the heart restores and maintains the desired functionality. Currently, cardiac monitoring remains largely qualitative. Quantitative measures for cardiac function has been an active field of research for years and with technological advancement being made to transesophageal echocardiography, it is expected that automated functional monitoring can be facilitated. TEE provides acquisitions of high- quality and frame rate, making automated computation of both global and regional cardiac function feasible. MAPSE, mitral annular peak velocities and strain are well established cardiac parameters that describe motion, rate of motion and deformation of the myocardium, and they are reported to relate well with cardiac function. The main aim of this thesis is to compute cardiac functional parameters automatically by a combining speckle tracking and velocity estimation from tissue Doppler measurements.

Supervisor: Gabriel Hanssen Kiss, FOR-NorMIT, NTNU
Date: June 27th 2018

Anders Mostrøm Nilssen

Master of Science Department of Energy and Process Engineering, Faculty of  Engineering, NTNU

«Characterization of the airflow distribution in close proximity to the patient in an operating room"

Supervisor: Guangyu Cao, NTNU
Co-Supervisors: Liv-Inger Stenstad og Jan Gunnar Skogås, FOR

Guri Bråthen

Master of Science Department of Energy and Process Engineering, Faculty of  Engineering, NTNU

«Simulation and experimental study of the indoor enviroment of a Norwegian hospital»

Supervisor: Guangyu Cao
Co-Supervisor: Liv- Inger Stenstad

2017

Kristine Rise Fry

Master of Science in Industrial Design, NTNU

"Design of interactive telemedical interface between primary and secondary health care units"

Supervisor: Thomas Porathe
Co-supervisor: Håvard Sjøvoll, Alexander Vanvik EGGS design
Date: June 2017

Marta Havåg Ranestad

Master of Science in Electronics, Department of Electronic Systems, NTNU

"Fast Non-Rigid Registration Using Polynomial Expansion with Diffeomorphic Constraints Applied to Cardiac Ultrasound Volumes"

This thesis attempted to develop a fast, non-rigid, diffeomorphic image registration implementation and was written as a continuation of a project that was completed during the fall of 2016 [1] which developed and implemented a rigid image registration code. Non-rigid diffeomorphic image registration is already used in many diagnostic and research related medical situations, however due to the computational demands it is only used as a post-processing tool. This thesis focused mostly on how it could be used with 3D echocardiography to estimate a vector field representation of the movement of the heart, preferably in real-time.

Supervisor: Ilangko Balasingham, IES, NTNU 
Co-supervisor: Gabriel Hanssen Kiss, ISB, NTNU
Date: June 2017


Madeleine Charlotte Aviles Storås

Master of Science in Energy and Enviromental Engineering, NTNU

“Characterization of the airflow distribution in close proximity to a patient in operating rooms with laminar airflow at St. Olavs hospital”

The objective of this study is to characterize the airflow distribution in close proximity to a patient in one operating room with a laminar airflow system at St. Olavs hospital.

Supervisor: Guangyu Cao, NTNU
Co-supervisor: Liv-Inger Stenstad and Jan Gunnar Skogås, NTNU
Date: August 7th 2017

2016

Liv-Inger Stenstad

"Information to patients prior to radiological examinations"   


Lise Hagen

"An observational study at St. Olavs Hospital, Department Røros Hospital"


Marthe Siren Anvik, Mikkel Treu Os and Jon Erik Medhus

"Operation planning at the Department of Orthopedic Surgery"


Jacob Nyman

"Distribution of operating rooms, sequence of operations and time for start of operations with different operating time when surgeons are working in more than one OR"

2014

Marit Furre Amundsen

«Quality of life after surgery for chronic rhinosinusitis»


Ida Leirheim Fagerli and Cecilie Liseth

”Safety in the operating room”

2012

Siri Fenstad Ragde

Master in molecular medicine
"Characterization of surgical staff's exposure to surgical smoke at St. Olavs Hospital".


Cathrine Soleglad, Torstein Høie and Mari Thoresen

Master in media communication and information technology "Innovation in the health sector - A qualitative investigation of innovation at St. Olavs Hospital".


Marit Rødevand and Ragnhild Byrkjeland

"Visulaisation of medical images on smart-phones, laptops, tablet computers".

2011

Anne Karin Wik

Master in health science    

"Quality of life and degree of leak in patients with anal incontinence before and after treatment with sacral nerve stimulation"

2010

Jenny Kristin Aasland

"Vascular Surgery in Norway - An investigation based on the Norwegian Vascular Register  (NorKar)"

2007

L.Nilsson

“Accuracy of electromagnetic tracked freehand laparoscopic 3D ultrasound”


Lars Eirik Bø

Ultrasound simulator

2006

Camilla Berge

"Infrarenal abdominal aortic aneurysm repair. Time-trends and results during a 20-year period"


Magnus Strømmen

"Obesity and health related quality of life study"

​​​Did you find what you were looking for?