New pathology software improves cancer diagnosis

14 August 2014

 New software developed jointly by the MedUni Vienna and the Vienna-based firm Tissuegnostics helps pathologists to identify cancerous tissue with greater accuracy and remove some of the variability in current methods based on visual analysis of tissue samples.

 The software was tested by the Clinical Institute of Pathology, at the Medical University of Vienna, the Ludwig Boltzmann Institute for Cancer Research (LBI-CR) and the VetMedUni Vienna.

In the study, published in the journal PlosOne, scientists analysed 30 liver cell carcinomas and were able to classify them into categories ranging from 'negative' to 'strongly positive' using the software .

This study measured the expression of the proteins STAT5AB and JUNB in an aggressive T-cell lymphoma. STAT5 plays an important role in the development of leukaemia and liver cancer. The JUNB gene is involved in the development of tumours in lymph gland tissue. The software uses certain algorithms and highly sensitive digital photography, enabling it to represent the matrix of cells and the cell nucleus better than under the microscope.

Researcher Lukas Kenner said, “The new program of course does not make pathologists redundant, however it is a supplementary method that considerably increases diagnostic certainty." The MedUni Vienna expert also anticipates that the new technology will contribute to the changes in cancer cells, which are currently classified into four categories, being specified in much more detail in the future. It may in future be possible to create much more detailed categories, giving clinicians a further tool with which to choose the correct and tailored therapy option.

“Cancer therapies are expensive. This new software will also help us to assess more effectively where expensive therapy is justified, but also which cases do not need it, thereby also sparing the patient," added Kenner.


Schlederer M, et al. Reliable quantification of protein expression and cellular localization in histological sections. PlosOne


To top