How would machine learning help cure cancer

Can cancer be cured in 15 years?

09/30/2019 - Can cancer be cured in 15 years? Four experts were confronted with this question at the 18th European Health Congress in Munich. Cancer researcher Prof. Dirk Jäger in particular found a clear answer to this.

“I'm skeptical of such slogans,” said moderator Prof. Herbert Rebscher right at the beginning of the event in the Hilton Park Hotel, “but good, we want to underline this sensational title with a little substance today.” From a medical point of view, hardly anyone is likely to do this better suited than Prof. Dirk Jäger, Head of the Medical Oncology Department at Heidelberg University Hospital. He too does not like promises of salvation, especially since “we cannot cure many benign diseases either” - such as a harmless cold. But, and that was the crucial point in Jäger's keynote speech, "there are areas at the moment in which cancer research is exploding."

For example, in drug therapy with “targeted substances”, so-called immunotherapeutics. “The immune system,” says Jäger, “is playing an increasingly important role in cancer research.” In contrast, the age of “standard therapies” is gradually coming to an end: “We are realizing more and more that it is not right when we make our therapy decisions the basis of shockingly little information, which is essentially based on histology and imaging. ”For example, patients with colon carcinoma are generally recommended chemotherapy,“ although we know that only 17 percent benefit from it. ”83 percent do not benefit either because the cancer can be treated without chemotherapy or because metastases develop anyway.

Use the power of the immune system

Instead of a therapy based on the watering can principle, Jäger is convinced that the future belongs to personalized medicine: “Ultimately, we will design an individual therapy for each tumor disease. In doing so, we will use the power of the immune system specifically against tumors. ”Every day, hundreds of tumor cells arise in each of us, which our immune system recognizes and kills - tumors could only develop if the immune system was undermined.

Certain cancers are already being fought with so-called checkpoint inhibitors. These antibodies do not act against cancer cells directly; they intervene in the control of the immune response against tumors - at the "immune checkpoints". They are used very successfully in the treatment of black skin cancer. Every fifth skin cancer patient responds to this therapy and achieves a survival time of more than ten years.

“This is a small revolution in oncology,” says Dirk Jäger. In addition, “artificial immune cells” would become increasingly important in the future, which are created in a very complex but effective process: “To do this, you have to take immune cells from the patient, place them in a test tube and modify them genetically, for example by installing an artificial receptor. Then these modified immune cells are given back to the patient as a single infusion. In this way we can cure 80 percent of childhood leukemia cases today, and over 50 percent of adults. "

According to Dirk Jäger, the future of oncology could and should look like this:

  • Less standard therapies
  • Individualized therapies for more and more cancer patients
  • Medicines that are individually "designed" for each individual patient
  • Immune cells are increasingly used in cancer therapy
  • There will be complex diagnostic platforms
  • Therapies are simulated with the help of artificial intelligence before the patient receives them

The arduous journeys of the present

Dr. Johannes Bruns, Secretary General of the German Cancer Society, returned in his lecture to the present, in which every cancer patient “has to go an incredibly long way - for example from general practitioner to specialist to the nearest specialist, at some point perhaps to the oncology center.” Such centers should be funded , because: "What Professor Jäger has described cannot be done everywhere." And: "The performance law must adapt to the development of innovations."

With Dr. Djork-Arné Clevert heard a scientist who works as a Senior Research Scientist in Bayer's research department. He reported on algorithms that can help drug research become faster and more efficient, for example in the search for new active ingredients. “What previously took six months,” says Clevert, “we can now do in four days.” With his team, Clevert “has developed a new language in chemistry - we use concepts from linguistics to convert one chemical structure into another chemical one Translate structure. In this way we can optimize molecules and, for example, transform them into a more soluble structure in a short time. ”Machine learning helps to develop molecules much faster and, according to Clevert,“ usher in a new era in drug development. ”

What progress costs

But how should such a new era in cancer therapy be financed? Dr. Sabine Richard from the AOK Federal Association did not answer. However, she pointed out: “We have to think about it: What are we willing to pay for the progress?” Herbert Rebscher added that in the cancer therapy of the future, not so much products but treatment processes would have to be financed. Dirk Jäger is “not worried about reimbursement, because there is also huge potential for savings. We have an extremely large number of ineffective chemotherapies today. If we become more precise and effective, new forms of treatment can be a bit more expensive - in the end we still save money with them. "

So will cancer be curable in 15 years? At this question, Prof. Jäger rolls his eyes: “First of all, we need greater freedom to bring innovations to patients.” Only when this “political problem” has been resolved can predictions about the status of cancer therapy in 15 Years.