Cancer treatment is controlled by a multi-disciplinary team of doctors, surgeons, oncologists, radiologists, biologists and pathologists. And the latter play a little known but important role. By observing the tumor under a microscope, they determine the cell characteristics and limitations of the cancerous tissue. This ranking is crucial when it comes to choosing which treatments to use.
At the moment, we are seeing accelerated changes in this particular role. Pathologists, who up to now have spent their lives bent over their microscope, observing cells, are faced with the onslaught of genetic data on tumors.
The first international congress on the metamorphosis of pathology was recently held in Paris, at the initiative of Professor Jean-Christophe Fournet. The tendency to classify tumors depending on their genetic profile rather than their appearance under the microscope is increasing with the more widespread access to DNA sequencing. For example, breast cancers are already subdivided into tens of different groups according to their genetic signature, which allows their treatment to be adapted accordingly. Three French scientists who have been particularly involved in this personalized approach to cancer are Pierre Laurent-Puig (tumors of the colon), Olivier Cussenot (prostate cancer) and Fabrice André (breast cancer).
Nowadays, the leading cancer treatment centers are sequencing all three billion chemical bases in the DNA of each patient’s tumor, thanks to the fact that the cost of sequencing is halving roughly every five months. In the near future, a genetic identity card will synthesize the thousands of mutations in tumor DNA to choose a custom cocktail of drugs that will target the specific characteristics of the cancer, so each patient will receive completely personalized treatment.
This revolution is the result of the intersection between genomics and information technology. The explosion in computing power always follows the law articulated in 1965 by Gordon Moore, one of the founders of Intel, that the power of microprocessors doubles at constant cost every eighteen months. So, Xeon Phi, the new Intel chip, is capable of more than a trillion operations a second! When used to support genomics, this increased power allows for an increasingly complete reading of the DNA involved.
Pathologists are faced with a real ‘digital storm’: they need to interpret trillions of items of information to classify each tumor, although so far all patient data has been held on a single Excel spreadsheet. This torrent of data is closer to what astrophysicists or specialists in nuclear physics have to deal with than traditional pathologists. How can the profession adapt itself to such a brutal mutation? Will pathologists and oncologists have to outsource their brains to so-called Cloud computing: in other words, hand over decision-making to huge databases located on remote servers? In any case, it seems impossible that a doctor could learn by heart the 100,000 genetic mutations discovered every day, which feed the databases in a constant stream. So it is becoming urgent for cancer specialists to observe how astrophysicists manage the Exabytes (billions of billions of data items) they produce. Otherwise, the power in medicine might just change hands.
Laurent Alexandre is urological surgeon, president of DNAVision, author of “The Death of Death” (Editions JC Lattes) – l.alexandre at dnavision.be
Source: From an article originally published in LE MONDE SCIENCE ET TECHNO, 24 January 2012.