The four-minute passage of a patient in the tunnel of this new “machine” provides as much information to his attending physician as if he had passed “30 scans in the 2000s”, calculates Professor Frédéric Courbon, head of the imaging department at Toulouse Cancer University Institute (IUCT). The machine in question which, for a layman, has all the appearance of an ordinary scanner, in whiter and more futuristic, is a technological platform called Omni Legend and developed by the French company GE Healthcare.
The prototype, tested as a world first and used for several months at the Oncopole in Toulouse, is the only one to be in operation in France to date. The cancer professor Jean-Pierre Delord does not hesitate to call it “revolutionary”. So what does this “TEP Scan” have more than the others? First, therefore, this speed of data acquisition which reduces the duration of this agonizing examination, especially for agitated or painful patients. With the corollary, the possibility of giving access to more patients per day. “All the patients who need it will be able to use this machine”, assures Jean-Pierre Delord. Finally, this speed makes it possible to reduce the dose of radioactive product injected and the duration of exposure to the rays.
The contribution of artificial intelligence
The Omni Legend’s second superpower: its resolution. It is able to detect cancerous lesions even “of a millimeter”. Practitioners can thus know if the cancer is really localized, or not, and adapt the treatment early on accordingly. To be convinced of this, the IUCT team conducted a clinical study including 45 patients. The latter were kind enough to lie down in the tunnel of the Omni Legend but also in those of two other TEP Scans of previous generations and from the same manufacturer. Then, three specialists compared “blindly” the anonymized images of the three machines. “We wanted the quality to be at least equivalent, it turns out to be 75% higher,” assures Nicolas Durand-Schaeffer, director of imaging at GE Healthcare.
The new technological platform also uses artificial intelligence to “reconstruct the data”. This aspect is still the subject of a second clinical study, which consists of ensuring that the AI does not produce “false positives”, that it does not see tumors, even tiny ones, where it does not there are none with the risk of causing “overtreatment”.
The ambition of the IUCT is to make its new tool accessible to around 4,500 patients per year, referred by clinicians from all over Occitania. And to use it to refine their diagnoses on a broad spectrum of cancers, ranging from leukemias to melanomas, passing from affections of the prostate, breast, ENT or digestive system.