After a MRI or CT scan of a patient suffering from seizures and severe headaches, an invasive biopsy needs to be taken to determine with certainty that it concerns a glioma. The biopsy tissue is used to run tests to determine the genetic characteristics of the glioma, one of which could be 1p/19q codeletion. This genetic mutation consists of the loss of the short arm of chromosome 1 (1p) and of the long arm of chromosome 19 (19q). If a patient tests positive to this 1p/19q codeletion, this establishes the diagnosis of oligodendroglioma and indicates a better prognosis and better response to treatment. Imagine, however, being able to diagnose this patient with oligodendroglioma without the need for a biopsy.
At Quantib, we are working on a future where diagnoses like oligodendroglioma can be made without the invasive procedure of a biopsy. We work on smart software that can support and guide the radiologist when diagnosing a patient. This might sound like a distant prospect but in fact, it may become a reality in the near future.
In a study led by Dr. Marion Smits funded by the Dutch Cancer Society, the Departments of Radiology and Medical Informatics at Erasmus MC developed an algorithm based on approximately 80 patients diagnosed with gliomas. Available data within this set included different MRI sequences and results from genetic tests (Next Generation Sequencing). Researchers trained the software to recognize the visual difference between glioma with 1p/19q codeletion and those without. Testing the performance of the software against the knowledge of an experienced neuroradiologist resulted in a higher score for the algorithm.
As Quantib is a spinoff of Erasmus MC, we have a strong collaboration with their radiology department and imaging research group and are actively engaged in the most recent developments in this field. We have a very strong quality focused team with a track record in certifying machine learning algorithms for CE and FDA. We have a neurodegenerative product already on the market and many algorithms in the pipeline.