Automated cytology

Computer-assisted quantitative cytology for improved early detection of cervical cancer

Cervical cancer is the second most common cancer in women worldwide with about 80% of cases occurring in low-income countries*. The Papanicolaou or Pap test is the primary screening test for the disease involving the microscopic examination of cells sampled from in and around the cervix. Precancerous changes can be treated before they have a chance to progress to cancer. Cancer can also be successfully treated provided it is detected early.

In developed countries widespread screening and treatment programmes have significantly reduced the incidence of invasive cervical cancer. Nevertheless the Pap test is not a perfect test with one in every 10 to 20 positive cases missed in routine screening. MedTech West is a partner in developing computerised cell image analysis methods to both improve screening accuracy and to facilitate the production of a low-cost and robust screening solution for the developing world.

Our approach

The problem of missed positives can be attributed primarily to interpretation error and sampling error (wherein diagnostic cells do not make it onto the microscope slide in the first place). Our approach addresses both of these errors using a combination of automation with a computer and robotic microscope, and novel quantitative cell analysis techniques based on malignancy-associated change (MAC) analysis. MACs are subtle sub-visual changes in the appearance of normal-looking cells on an abnormal Pap slide. MAC analysis obviates the need to find diagnostic cells and also to perform an exhaustive review of all of the cellular material. Thus it not only addresses sampling error but also reduces the complexity and cost of automated screening. In the future our approach promises more accurate and cost-effective screening for cervical cancer and concomitantly reduced mortality from the diseases.

Research team

MedTech West Partner

Dr. Andrew Mehnert, Department of Signal Processing and Biomedical Engineering, S2, Chalmers University of Technology.

Technical research partners

Prof. Mikael Persson, Department of Signal Processing and Biomedical Engineering, S2, Chalmers University of Technology.

Prof. Ewert Bengtsson, Centre for Image Analysis, Uppsala University


For more information on this project, contact Henrik Mindedal at