Bildanalys för pålitlig och kostnadseffektiv cancerdetektion

Tidsperiod: 2016-01-01 till 2018-12-31

Projektledare: Joakim Lindblad

Finansiär: Vetenskapsrådet

Bidragstyp: Internationellt samarbete

Budget: 900 000 SEK

A crucial aspect for reliable automated analysis of cell samples for cancer detection is accurate and robust quantification of very subtle changes in the nuclear chromatin distribution through image analysis. This requires high resolution, perfect focus for each cell nucleus and robustness against staining variations. This is very hard and expensive to achieve with conventional optics.Recently (2013) a new way of creating images in a light microscope was invented, Fourier ptychographic microscopy, FPM. This technique holds promise to address all three of the crucial problems in chromatin distribution quantification. However, the technique needs improvements in terms of computational efficiency and it needs to be adapted and optimized for the needs of cancer detection. We plan to do that within this project.We have since 2008 had a collaboration between India and Uppsala University on developing a new cost effective system for cervical cancer screening. The FPM technique holds promise for increasing the performance of the developed system. We also propose to use the same technique for addressing two other common and increasing cancers: oral cancer and thyroid cancer. The Indian project partners have a strong background in cancer research, the Swedish team combines long-standing expertise in the field of digital cytometry with expertise in robust and precise image analysis.