Computational Surgineering (NEW)

Overview

The course is meant to let students dive deep into the hospital daily routine, as well as, “get their hands dirty” developing solutions for the clinical routine. The students attending this course will prototype the concepts developed in the lecture IN2286 or new ideas resulting in a computer-aided solution tailored for clinical needs of our partners in several hospitals within Munich.
The course is divided in three parts (a) lectures, (b) clinical hospitation (“job shadowing”), (c) project development including presentations.

  1. Introduction - week 1
  2. OR training (for students who missed that training in IN2286, together with IGS students) - week 1
  3. Tools and methods for software development for medical image processing and computer-assisted interventions (for students who missed that training in IN2106/IN4136, together with PMSD) - week 2
  4. Hospitation at the clinical partner’s department for at least one week - starting week 3
  5. Introduction to regulatory aspects of medical software development (for students who missed that training in IN2286, together with IGS students) - week 3
  6. Presentation Requirements specification - presented to doctors - week 4
  7. Weekly reviews (consultation, help with bug fixing, administrative stuff, connection to experts) - starting in week 5, every week
  8. Prototype demonstration - presented to the doctors - end of the semester

Prerequisites and Registration

Experience in programming, in particular Python and/or C/C++ are needed. Basics of Medical Imaging (e.g., IN2021, IN2022) and Computer Vision (e.g., ) are highly recommended.

Registration through the TUM Matching System is mandatory. Your chances to be assigned to the course increase if you give the course a higher rank in your choices. If you already have a potential project, notify the course tutors via e-mail as soon as possible. This increases your chances to be assigned to the course, but you have to register through the matching system in any case. For further details about how the matching system works and its schedule please check its documentation.

Objectives

Students completing this practical course successfully will be able to:

  • Use a subset of common software development tools for medical image processing and computer-assisted interventions
  • Consider regulatory constraints needed to be taken into account when developing medical software
  • Understand the daily clinical routine within one specialty/clinical department
  • Refine a clinical solution for an unmet clinical need to be able to generate a prototype
  • Analyze one clinical application in order to generate a requirement specification for the chosen clinical challenge
  • Develop a clinical prototype (mainly software, if applicable also involving hardware) and demonstrate it on phantoms, ex-vivo, or using retrospective data
  • Present their work in front of an audience of medical technologists and clinicians

Teaching and Learning Method

In total, 18 to 21 students will be recruited, with students that attended IN2286 having priority. They will be divided into groups of 3 students and form a team.
Students will be presented with all concepts that former students have proposed since 2020. They will then select their priorities and 6 to 7 groups will be assigned. If a group of 3 students brings a new concept and a clinical partner, we will also accept this.
Students who did not attend IN2286 will get an introduction on how to behave in the OR. Furthermore, two classes will be given on software development and present typical libraries and tools used in medical image processing and computer-assisted interventions (e.g., Git, Qt, Jupyter Notebooks, PyTorch, OpenGL, OpenCV, Unity, ITK/VTK, 3D Slicer, ImFusion SDK, ROS, RViz, Gazebo). Additionally, a class on regulatory aspects for medical software development will be offered.
At this point, students will be assigned to our clinical partners to be in the hospital for a full week attending all possible surgeries our clinical partners perform as if the students were medical “Famulanten”. Here students will not only attend the surgery they will develop a solution for, but also other surgeries in the same clinical department to get a broader view of the routine in the hospital and their requirements and needs.
The students will then start preparing a software requirement specification in close contact with the clinical partners, and will start with the implementation of a software prototype/proof-of-concept. We will make sure students use tools of machine learning, computer vision, augmented reality, and medical robotics. Within the first 3-4 weeks of the course, the students will present their requirement specifications to the clinical partners to make sure they modeled their needs properly.
Each student group will be assigned a tutor and will be supervised weekly by him/her, to make sure the software development is running.

Master Practical Course - Computational Surgineering (IN2106)

Lecturer (assistant)
Number0000001367
TypePractical course
Duration6 SWS
TermWintersemester 2022/23
Language of instructionEnglish
Position within curriculaSee TUMonline
DatesSee TUMonline

Dates

  • 26.07.2022 13:00-14:00 Online: Videokonferenz / Zoom etc., Link: https://tum-conf.zoom.us/j/65609363155 Meeting ID: 656 0936 3155 Passcode: 076093

Admission information

See TUMonline
Note: Preliminary meeting: Topic: Computational Surgineering WS23 - Preliminary Meeting Time: Jul 26, 2022 01:00 PM Amsterdam, Berlin, Rome, Stockholm, Vienna Link: https://tum-conf.zoom.us/j/65609363155 Meeting ID: 656 0936 3155 Passcode: 076093 Students who have attended IN2286 will be prioritized. Also, BMC students during later semesters.

Objectives

Students completing this practical course successfully will be able to: - Use a subset of common software development tools for medical image processing and computer-assisted interventions - Consider regulatory constraints needed to be taken into account when developing medical software - Understand the daily clinical routine within one specialty/clinical department - Refine a clinical solution for an unmet clinical need to be able to generate a prototype - Analyze one clinical application in order to generate a requirement specification for the chosen clinical challenge - Develop a clinical prototype (mainly software, if applicable also involving hardware) and demonstrate it on phantoms, ex-vivo, or using retrospective data - Present their work in front of an audience of medical technologists and clinicians

Description

The course is meant to let students dive deep into the hospital daily routine, as well as, “get their hands dirty” developing solutions for the clinical routine. The students attending this course will prototype the concepts developed in the lecture IN2286 or new ideas resulting in a computer-aided solution tailored for clinical needs of our partners in several hospitals within Munich. The course is divided in three parts (a) lectures, (b) clinical hospitation (“job shadowing”), (c) project development including presentations. 1. Introduction - week 1 2. OR training (for students who missed that training in IN2286, together with IGS students) - week 1 3. Tools and methods for software development for medical image processing and computer-assisted interventions (for students who missed that training in IN2106/IN4136, together with PMSD) - week 2 4. Hospitation at the clinical partner’s department for at least one week - starting week 3 5. Introduction to regulatory aspects of medical software development (for students who missed that training in IN2286, together with IGS students) - week 3 6. Presentation Requirements specification - presented to doctors - week 4 7. Weekly reviews (consultation, help with bug fixing, administrative stuff, connection to experts) - starting in week 5, every week 8. Prototype demonstration - presented to the doctors - end of the semester

Prerequisites

Experience in programming, in particular Python and/or C/C++ are needed. Basics of Medical Imaging (e.g., IN2021, IN2022) and Computer Vision (e.g., ) are highly recommended.

Teaching and learning methods

In total, 18 to 21 students will be recruited, with students that attended IN2286 having priority. They will be divided into groups of 3 students and form a team. Students will be presented with all concepts that former students have proposed since 2020. They will then select their priorities and 6 to 7 groups will be assigned. If a group of 3 students brings a new concept and a clinical partner, we will also accept this. Students who did not attend IN2286 will get an introduction on how to behave in the OR. Furthermore, for students who did not attend IN2106/IN4136, two classes will be given on software development and present typical libraries and tools (e.g., Git, Qt, Jupyter Notebooks, PyTorch, OpenGL, OpenCV, Unity, ITK/VTK, 3D Slicer, ImFusion SDK, ROS, RViz, Gazebo). Additionally, a class on regulatory aspects for medical software development will be offered for students who did not attend IN2286. At this point, students will be assigned to our clinical partners to be in the hospital for a full week attending all possible surgeries our clinical partners perform as if the students were medical “Famulanten”. Here students will not only attend the surgery they will develop a solution for, but also other surgeries in the same clinical department to get a broader view of the routine in the hospital and their requirements and needs. The students will then start preparing a software requirement specification in close contact with the clinical partners, and will start with the implementation of a software prototype/proof-of-concept. We will make sure students use tools of machine learning, computer vision, augmented reality, and medical robotics. Within the first 3-4 weeks of the course, the students will present their requirement specifications to the clinical partners to make sure they modeled their needs properly. Each student group will be assigned a tutor and will be supervised weekly by him/her, to make sure the software development is running.

Examination

At the end of the semester, the students will present their work to the clinical partners and the tutors in the form of a presentation with a demo. The grading will be based on the requirement specification (25%), the presentation (25%), and the implementation including code (50%).

Recommended literature

- Esteban, J., Simson, W., Witzig, S. R., Rienmüller, A., Virga, S., Frisch, B., ... & Hennersperger, C. (2018). Robotic ultrasound-guided facet joint insertion. International journal of computer assisted radiology and surgery, 13(6), 895-904. - Roodaki, H., Filippatos, K., Eslami, A., & Navab, N. (2015, September). Introducing augmented reality to optical coherence tomography in ophthalmic microsurgery. In 2015 IEEE International Symposium on Mixed and Augmented Reality (pp. 1-6). IEEE. - Matinfar, S., Nasseri, M. A., Eck, U., Kowalsky, M., Roodaki, H., Navab, N., ... & Navab, N. (2018). Surgical soundtracks: automatic acoustic augmentation of surgical procedures. International journal of computer-assisted radiology and surgery, 13(9), 1345-1355. - Esteban, J., Grimm, M., Unberath, M., Zahnd, G., & Navab, N. (2019, October). Towards fully automatic X-ray to CT registration. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 631-639). Springer, Cham. - Nasseri, M. A., Eder, M., Nair, S., Dean, E. C., Maier, M., Zapp, D., ... & Knoll, A. (2013, July). The introduction of a new robot for assistance in ophthalmic surgery. In 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (pp. 5682-5685). IEEE. - Navab, N., Heining, S. M., & Traub, J. (2009). Camera augmented mobile C-arm (CAMC): calibration, accuracy study, and clinical applications. IEEE transactions on medical imaging, 29(7), 1412-1423. - Padoy, N., Blum, T., Ahmadi, S. A., Feussner, H., Berger, M. O., & Navab, N. (2012). Statistical modeling and recognition of surgical workflow. Medical image analysis, 16(3), 632-641. - Twinanda, A. P., Shehata, S., Mutter, D., Marescaux, J., De Mathelin, M., & Padoy, N. (2016). Endonet: A deep architecture for recognition tasks on laparoscopic videos. IEEE transactions on medical imaging, 36(1), 86-97.

Links