Master Thesis - Optimal Edi Catheter Placement in Neonates Using Multivariate Time-Series ML

With a passion for life

Join our diverse teams of passionate people and a career that allows you to develop both personally and professionally. At Getinge we exist to make life-saving technology accessible for more people. To make a true difference for our customers – and to save more lives, we need team players, forward thinkers, and game changers. 

Are you looking for an inspiring career? You just found it.

Master Thesis project regarding Optimal Edi Catheter Placement in Neonates Using Multivariate Time-Series Machine Learning in the Innovation & Technology team at Getinge, Solna Sweden.

At Getinge, our passion is to secure that every person and community have access to the best possible care, offering hospitals and life science institutions products and solutions that aim to improve clinical results and optimize workflows. Every day we collaborate to make a true difference for our customers – and to save more lives.

Are you looking for an inspiring environment for your thesis project? You just found it. We are now looking for you that wants to execute your thesis project in our ACT-TR function at our Solna site. ACT-TR is a department working with innovation, clinical and technical research within the product area of Acute Care Therapies. 

Introduction

Electrical Activity of the Diaphragm (Edi) catheter is a special nasogastric tube equipped with embedded electrodes that measure the electrical activation of the diaphragm in real time. It’s both a feeding tube and a sensor — it goes through the nose, down the esophagus, and stops just behind the diaphragm with the tip in the stomach. Tiny electrodes near the tip pick up the diaphragm’s electrical activity, which reflects how much the patient is trying to breathe. In neonatal intensive care units (NICUs), precise placement of the Edi catheter is essential for the safe and effective operation of neurally adjusted ventilatory assist (NAVA) systems. Misplacement can impair ventilatory efficiency and jeopardize patient outcomes. Clinically, catheter positioning is currently determined through the users interpretation of electromyographic (EMG) and electrocardiographic (ECG) signals acquired via a 9-lead electrode array. However, due to weak or absent EMG signals and frequent electrode displacement due to patient motion, this process demands continuous expert oversight.

Previous work at Getinge demonstrated the use of MiniROCKET for classifying adult Edi catheter positions using multichannel ECG recordings, resulting in a significant computational advantage over traditional deep CNNs like ResNet. Building on this foundation, our project aims to extend these methods to neonatal populations.

This study will integrate EMG-based signal quality assessments across all 8 leads to derive a high-level classification of catheter positioning, distinguishing among Well-Positioned, Too Far In, and Too Far Out categories. Additionally, ECG-based models—particularly MiniROCKET and related variants—will be evaluated to assess and enhance prediction robustness, computational efficiency, and real-time integration potential.

The assignment

To develop a Python-based, server-ready pipeline for classifying Edi catheter positioning in neonates using multichannel ECG data. The system should output actionable recommendations (e.g., “adjust up”, “adjust down”, “well-positioned”) by aggregating signal quality measures and machine learning predictions across the 8-lead array. Integration readiness for deployment into the Getinge demonstration system is a key deliverable. In addition, the project will investigate the potential of transfer learning by training models on adult and neonatal datasets and evaluating their cross-domain performance. This will help assess the feasibility of developing a generalized model that can accurately support both adult and neonatal applications. 

Number of students: 1

Start date: According to agreement during start of VT26
30 HP 

Your qualifications:

• Courses in machine learning and AI
• Programming skills in Matlab
• Able to express yourself well in speech and writing in English
• Relevant courses in signal processing are appreciated but not necessary 

Why should you apply?

Getinge is a leading medtech company on an exciting transformation journey constantly looking for new ways to innovate together with our customers to meet the healthcare challenges of the future. This is an exciting opportunity for you who in the future want to work in medtech and with Critical Care products in an environment that is characterized by great care for the end customer, high pace, continuous learning and continuous improvement.

Why should you apply? 

Getinge is a leading MedTech company on an exciting transformation journey constantly looking for new ways to innovate together with our customers to meet the healthcare challenges of the future.  

This is an exciting opportunity for you who aspire to work in MedTech in the future and in an environment that is characterized by great care for the end customer, high pace, continuous learning and improvement. 

Welcome with your application no later than 2025-12-07  

For questions, please contact Fredrik Jalde at Fredrik.jalde@getinge.com

About us 

With a firm belief that every person and community should have access to the best possible care, Getinge provides hospitals and life science institutions with products and solutions aiming to improve clinical results and optimize workflows. The offering includes products and solutions for intensive care, cardiovascular procedures, operating rooms, sterile reprocessing and life science. Getinge employs over 12,000 people worldwide and the products are sold in more than 135 countries.

#LI-SE1