Hi,

The Joint Applied Mathematics and Statistics Seminar continues.

On Thursday, Dec. 10th., Ilmari Ahonen will give a talk starting at 10:15 at room M1 (Quantum).

Title:
On the high-throughput analysis of patient-derived high-content cancer micro tissue cultures

Abstract:

The recent technological advances in 3D micro tissue models have enabled  the study of cancer in ever more realistic, yet artificial conditions.  More specifically, direct culturing of patient-derived samples allows  for personalized drug screens in a controlled environment with high  clinical relevance. Compared to traditional cell lines, the patient  derived cultures introduce a new level of complexity that the related  image analysis methodology needs to address. In this paper, we propose an analysis pipeline that encompasses all key steps from the image
processing to final statistical modelling of the treatment effects. Efficiency, robustness to varying image quality and high-throughput are prioritized in each step. The analysis consists of three main steps: 1) separation of the image foreground from the background, 2) supervised
classification between malignant and benign tissue texture and 3) statistical modelling of the total area covered by each tissue type. The proposed pipeline is applied to an imaging data set containing data from multiple patients, cultivation methods and treatments. The results obtained with the proposed methods both support the current understanding of tumor-stroma interaction and bring new insight into the analysis of patient derived 3D micro tissues.

All interested are warmly welcome!

Hi,

The Joint Applied Mathematics and Statistics Seminar continues.

On Thursday, Dec. 10th., Ilmari Ahonen will give a talk starting at 10:15 at room M1 (Quantum).

Title:
On the high-throughput analysis of patient-derived high-content cancer micro tissue cultures

Abstract:

The recent technological advances in 3D micro tissue models have enabled  the study of cancer in ever more realistic, yet artificial conditions.  More specifically, direct culturing of patient-derived samples allows  for personalized drug screens in a controlled environment with high  clinical relevance. Compared to traditional cell lines, the patient  derived cultures introduce a new level of complexity that the related  image analysis methodology needs to address. In this paper, we propose an analysis pipeline that encompasses all key steps from the image
processing to final statistical modelling of the treatment effects. Efficiency, robustness to varying image quality and high-throughput are prioritized in each step. The analysis consists of three main steps: 1) separation of the image foreground from the background, 2) supervised
classification between malignant and benign tissue texture and 3) statistical modelling of the total area covered by each tissue type. The proposed pipeline is applied to an imaging data set containing data from multiple patients, cultivation methods and treatments. The results obtained with the proposed methods both support the current understanding of tumor-stroma interaction and bring new insight into the analysis of patient derived 3D micro tissues.

All interested are warmly welcome!