Machine Learning Seminar: Introduction to Machine Learning and the Kaggle-club

The first meeting will be held at HVL, Wednesday February 21st, 14.15-16.00, in room D113 (Auditorium 8). The address is Kronstad, Inndalsveien 28. See this mazemap link for room location:

If you are interested in attending, please fill out the doodle-link:

Tentative agenda for the first meeting:

  • Welcome and general introduction
  • Short introduction of the participants (name, background, interests)
  • Two short talks (10-15 min each)
    • Sean Meling Murray: A Gentle Introduction to Machine Learning and Neural Networks
    • Jens Wahl: Information on the kaggle-club at the Department of Mathematics at UiB
  • Time for coffee and socializing.

MedViz Seminar: Radiological and metabolic patterns related to endometrial cancer risk and prognosis

We are happy to announce that this spring’s second MedViz seminar will take place on Friday 23rd February at 12.00.

Location: The library (H112), Department of Radiology, Haukeland University Hospital

Time: 12:00-13:00

Title: Radiological and metabolic patterns related to endometrial cancer risk and prognosis

Speakers: Elin Strand (MSc, PhD) and Kristine E. Fasmer (Medical physicist, PhD student), Bergen Gynecologic Cancer Research Group, University of Bergen


Elin Strand: ‘Is metabolomic profiling a useful tool in endometrial cancer?’

Endometrial cancer has a high prevalence among post-menopausal women in industrialized countries. The aim of the current study is to explore whether certain metabolic
patterns are related to clinical phenotype using samples from the Molecular Markers in Treatment of Endometrial Cancer (MoMaTec) patient cohort. Serum metabolites are measured on a multiplex system including 183 metabolites, which are subsequently determined by liquid chromatography-mass spectrometry. Both supervised and unsupervised statistical methods are used to search for metabolic patterns related to prognosis. Identifying metabolic signatures in serum and tumor may represent a promising approach to further evaluate genetic traits and metabolic pathways relevant for tumorgenesis and progression in endometrial cancer. In the future, this may also serve as a useful platform for identifying targets for treatment and disease prevention. Preliminary results and thoughts and ideas for future projects will be summarized.

Kristine E. Fasmer: ‘Does it matter whether your abdominal fat is stored inside or outside in endometrial cancer? High BMI is an established risk factor for endometrial cancer (EC) development, but the underlying
mechanisms linking obesity to cancer development are complex and only partly understood. In a recent work by our group (Mauland et al, Oncotarget 2017), diagnostic CT scans of the abdomen were evaluated for
quantification of abdominal fat volumes in 227 patients with histologically confirmed endometrial carcinomas. High content of visceral fat compared to the total abdominal fat content was shown to independently
predict reduced EC survival. In an ongoing study of 120 EC patients, we further investigate the links between abdominal fat distribution, tumour characteristics and steroid levels measured in blood serum from patients
with EC. This talk will both give a summary of the fat segmentation method, the recent published work by the group and also some preliminary results from ongoing studies.


New Seminar Series: Machine Learning organized by HUS – HVL – UIB

We hereby wish to invite you to a new series of seminar-based meetings on machine learning, to be held at the three organising institutions, Western Norway University of Applied Sciences (HVL), University of Bergen (UiB) and Haukeland University Hospital (HUS). The idea is to create an informal platform where people interested in machine learning can meet and discuss across institutions and disciplines. Experienced users, newcomers, and everyone in between are welcome to participate. In this way, we can share experiences, learn from each other and maybe develop a foundation for future collaboration.

Planned meetings for spring 2018:

  • Wednesday February 21st, 14.15-16.00, HVL
  • Wednesday March 21rst, 14.15-16.00, UiB
  • Wednesday April 18th, 14.15-16.00, Haukeland
  • Wednesday May 23rd, 14.15-16.00, HVL
  • Wednesday June 20th, 14.15-16.00, UiB

RSNA 2017: Explore. Invent. Transform.

The 103rd scientific assembly and annual meeting of the Radiological Society of North America (RSNA) was held November 26th – December 1st, 2017 in Chicago. Among more than 11,000 submitted abstracts, PhD candidate Sigmund Ytre-Hauge was selected to present a recent study conducted at Haukeland University Hospital and the University of Bergen: Preoperative Tumor Texture Analysis from MRI Predicts High-Risk Status and Reduced Survival in Endometrial Carcinomas.

Radiologist Sigmund Ytre-Hauge has been a PhD candidate since 2015, affiliated to Bergen Abdominal Imaging Group and Bergen Gynecologic Cancer Research Group, with his project Functional imaging for individualized treatment of uterine cancer.

– A great experience! Just being there, at the world’s #1 radiology assembly among 50,000+ other attendees, is awesome. To get this unique opportunity to present our research to the international community is a great acknowledgement, Sigmund says enthusiastically shortly after returning to Norway.

The presented study is recently also submitted as an original article to a scientific journal. The study is co-authored by Julie A. Dybvik, Arvid Lundervold, Øyvind O. Salvesen, Camilla Krakstad, Kristine E. Fasmer, Henrica M. Werner, Balaji Ganeshan, Erling Høivik, Line Bjørge, Jone Trovik and Ingfrid S. Haldorsen.