New Technology Can More Accurately Diagnose Ovarian Cancer

A new method to differentiate between two gynecologic tumors, malignant epithelial ovarian tumors (MEOTs) and borderline epithelial ovarian tumors (BEOTs), has been created by Chinese scientists from the Suzhou Institute of Biomedical Engineering and Technology.

This new methodology uses Modality-based Attention and Contextual (MAC)-Net.


MEOTs account for 90% of all patients diagnosed with ovarian cancer. It is considered the most lethal type of gynecologic malignancy. MEOTs have a 5 year survival rate of 35%. Most patients require surgery as well as chemotherapy.

BEOTs have a low potential for malignancy and a 5-year survival rate of 92%. They don’t have stromal invasion. For BEOTs, conservative treatments are often used in effort to maintain function of the ovaries and preserve fertility for patients.

Clearly, these two forms of diagnosis are very different. Differentiating between the two is necessary for understanding prognosis, the best form of treatment, and ensuring patients receive the best, most personalized care.

Unfortunately, as things stand, these conditions are differentiated mainly through MRIs. These results are often subjective, the tests are time consuming, and the accuracy is low (just 74-89%).

This is where MAC-Net can come into play. This technological essentially creates an automatic diagnosis, simply by determining the top and the bottom of the tumor. It doesn’t need an accurate boundary of the tumor.


MAC-Net technology can be used for MRIs, PET scans, CTs and more. It may assist in the diagnosis not simply of ovarian cancer but also lung cancer, bowel cancer, prostate cancer, liver cancer, and breast cancer.

If patients can be given an automatic diagnosis, their long-term outcomes may be vastly different.

The technology is a MICNN which uses MA and C-MPL. MA helps by assessing the importance of MRI modalities. C-MPL uses tumor distribution knowledge to make an accurate prediction.

You can read more about this new technology here.

Share this post

Share on facebook
Share on twitter
Share on linkedin
Share on pinterest
Share on print
Share on email