PhD Student in Machine learning for radiological precision medicine

Do you want to contribute to top quality medical research?

To be a doctoral student means to devote oneself to a research project under supervision of experienced researchers and following an individual study plan. For a doctoral degree, the equivalent of four years of full-time doctoral education is required.

The research group

The research group is led by Fredrik Strand, MSc MD PhD, who has a mixed background of medicine, engineering and economics, and is a breast radiologist at the Karolinska University Hospital. The group consists of around ten full-time members: research specialist, post doc, PhD students, and research engineers. Our research is internationally recognized for the development and evaluation of AI for radiological breast cancer diagnostics with recent publications in Lancet Digital Health, JAMA Oncology, Radiology, MICCAI, among others. Our group has expertise in radiology, biostatistics and machine learning. We have a strong collaboration with research groups at KTH, at KI, and internationally within two EU projects and a new collaboration with a group at the center for Computational Precision Health at University of California Berkeley and San Francisco. The announced doctoral student position will form part of that collaboration.

The doctoral student project and the duties of the doctoral student

Our research initially focused on AI for early detection and screening mammography but in recent years it has expanded to the diagnostic pipeline and disease prognostic factors. Developing, and evaluating, new ground-breaking machine-learning approaches for radiological precision medicine for breast cancer will be the focus for the doctoral student. The project aims to develop AI in breast imaging for therapy response prediction to push precision medicine and personalize treatment planning – especially for the cases in current breast cancer care for which established predictors fail. Our overall aim is to revolutionize personalized treatment planning in breast cancer care through AI and multimodal imaging. To achieve this, we aim to address three main research questions:

1. Prognostic image biomarkers for Neoadjuvant Therapy Efficacy: Can AI algorithms, utilizing breast imaging modalities, improve the prognostic accuracy of response to neoadjuvant therapy?
2. Predictive image biomarkers for Optimal Treatment Identification: If neoadjuvant therapy is likely to succeed, can AI-driven analysis determine the most effective treatment for individual patients?
3. Dynamic Treatment Evaluation: Can we develop AI models that dynamically evaluate the treatment efficacy during the neoadjuvant therapy using imaging acquired to monitor the disease status (regression/progression/stable)?

The doctoral student will be responsible for driving the technical innovations necessary to address the posed research questions. Specifically, the doctoral student’s role will include large-scale and multi-modal medical data curation (imaging and non-imaging, clinical data entries), understanding and development of deep learning model (both convolutional networks and vision transformers), and specifying along with the supervisory team which clinical factors and/or radiological extracted features can impact the AI models’ development and clinical translation. Technical development will be carried out together with our collaborators at UC Berkeley, University of Barcelona and KTH in Stockholm.

What do we offer?

A creative and inspiring environment full of expertise and curiosity. Karolinska Institutet is one of the world's leading medical universities. Our vision is to pursue the development of knowledge about life and to promote a better health for all. At Karolinska Institutet, we conduct successful medical research and hold the largest range of medical education in Sweden. As a doctoral student you are offered an individual research project, a well-educated supervisor, a vast range of elective courses and the opportunity to work in a leading research group. Karolinska Institutet collaborates with prominent universities from all around the world, which ensures opportunities for international exchanges. You will be employed on a doctoral studentship which means that you receive a contractual salary. Employees also have access to our modern gym for free and receive reimbursements for medical care.

Eligibility requirements for doctoral education

In order to participate in the selection for a doctoral position, you must meet the following general (A) and specific (B) eligibility requirements at latest by the application deadline.

It is your responsibility to certify eligibility by following the instructions on the web page Entry requirements (eligibility) for doctoral education.

A) General eligibility requirement

You meet the general eligibility requirement for doctoral/third-cycle/PhD education if you:

1. have been awarded a second-cycle/advanced/master qualification (i.e. master degree), or
2. have satisfied the requirements for courses comprising at least 240 credits of which at least 60 credits were awarded in the advanced/second-cycle/master level, or
3. have acquired substantially equivalent knowledge in some other way in Sweden or abroad.*

Follow the instructions on the web page Entry requirements (eligibility) for doctoral education.

*If you claim equivalent knowledge, follow the instructions on the web page Assessing equivalent knowledge for general eligibility for doctoral education.

B) Specific eligibility requirement

You meet the specific eligibility requirement for doctoral/third-cycle/PhD education if you:

- Show proficiency in English equivalent to the course English B/English 6 at Swedish upper secondary school.

Follow the instructions on the web page English language requirements for doctoral education.

Verification of your documents Karolinska Institutet checks the authenticity of your documents. Karolinska Institutet reserves the right to revoke admission if supporting documents are discovered to be fraudulent. Submission of false documents is a violation of Swedish law and is considered grounds for legal action.

(A) and (B) can only be certified by the documentation requirement for doctoral education.

Skills and personal qualities

Necessary:

• Experience with developing, evaluating and explaining state-of-the-art machine-learning algorithms, with a strong focus on deep learning including vision transformers, for computer vision tasks.
• Previous experience in image analysis/processing.
• Bachelor’s and/or MSc degree in machine learning, computer science, electrical and computer engineering or a related field with a strong focus on machine learning
• Excellent programming skills (strong focus in python and machine learning libraries).
• Self-driving qualities, with a strong interest in the research advancements in the biomedical image analysis field.
• A strong personal interest to improve cancer care
• Very good communication skills in both writing and speaking, in English

Desirable/advantageous:

• You have a high professional ambition level combined with a kind and collaborative attitude
• Previous scientific publication as first author
• Experience with DICOM image analysis
• Experience with developing state-of-the-art ML algorithms for radiology images
• Experience with full-stack production-grade Python code, containerization, version control
• Experience with statistical evaluations within the cancer, and specifically breast cancer
• Biological understanding of the development and treatment of cancer
• Effective at adapting to new frameworks, querying databases, and debugging code
• You have a history of helping others to succeed in general, to succeed in their research and development work
• You have experience with any frontend framework such as Svelte, React

Terms and conditions

The doctoral student will be employed on a doctoral studentship maximum 4 years full-time.

Application process

Submit your application and supporting documents through the Varbi recruitment system. Use the button in the top right corner and follow the instructions. We prefer that your application is written in English, but you can also apply in Swedish.

Your application must contain the following documents:

- A personal letter and a curriculum vitae
- Degree projects and previous publications, if any
- Any other documentation showing the desirable skills and personal qualities described above
- Documents certifying your general eligibility (see A above)
- Documents certifying your specific eligibility (see B above)

Selection

A selection will be made among eligible applicants on the basis of the ability to benefit from doctoral education. The qualifications of the applicants will be evaluated on an overall basis.

Karolinska Institutet uses the following bases of assessment:

- Documented subject knowledge of relevance to the area of research
- Analytical skill
- Other documented knowledge or experience that may be relevant to doctoral studies in the subject.

All applicants will be informed when the recruitment is completed.

https://ki.se/en/research/groups/computational-breast-imaging-fredrik-strands-group

Want to make a difference? Join us and contribute to better health for all

1. Fredrik Strand, fredrik.strand@ki.se

1. Per Hydbring, SACO/Nick Tobin, SACO, per.hydbring@ki.se/nick.tobin@ki.se
2. Henry Wölling, SEKO, henry.wolling@ki.se
3. Helen Eriksson, OFR, helen.eriksson@ki.se