IBM has unveiled a clinical-analytics platform to provide doctors with insight into patients' conditions using natural-language processing and machine-learning capabilities similar to Watson.
IBM is moving
on from "Jeopardy" to finding treatments using genetic data. The
company announced it has developed a data-analytics platform called Clinical
Genomics that uses algorithms and analytics similar to that of Big Blue's
Watson supercomputer to find treatments for conditions based on a patient's
use Clinical Analytics to analyze patients' similarities, predict outcomes,
evaluate risk benefits and view treatment options. In addition to genetic data,
the software also takes into account a patient's age, gender, symptoms and past
platform combines knowledge of patient histories with automatic statistical
analysis and machine learning, said Dr. Haim Nelken, manager of IBM Research's
health care and privacy solutions division in Haifa, Israel.
a standard-based generic framework that can be used in the context of various
diseases and for a range of decision-support tasks, such as patients'
similarity assessment, prospective outcome prediction, risk-benefits analysis,
and presentation of the treatment options most likely to succeed," Nelken
wrote in an email to eWEEK.
can access data through a standard Web browser on tablets or PCs.
Genomics fits into the growing trend in health care of using big data to develop personalized medicine
is the ability to use a patient's personal genetic characteristics to prescribe
medical treatment for conditions, such as cancer, hypertension and AIDS.
Genomics enables doctors to aggregate medical statistics and develop recommendations
and weighted predictions, Nelken wrote. The software analyzes medical
guidelines and the knowledge clinicians provide, and correlates it with patient
data to provide doctors with personalized treatment options.
platforms such as Clinical Genomics, analyzing a patient's genetic profile
along with other data such as family history required large-scale data sources,
IRCCS Istituto Nazionale dei Tumori (INT)
, a research and treatment
cancer center in Italy focused on preclinical and clinical oncology, is
conducting a pilot project using Clinical Genomics, which IBM announced March
decisions in today's complex environment requires computerized methods that can
analyze the vast amounts of patient information available to ease clinical
decision making," Dr. Marco A. Pierotti, scientific director at INT, said
in a statement. "By providing our physicians with vital input on what
worked best for patients with similar clinical characteristics, we can help
improve treatment effectiveness and the final patient outcome."
In the first
stages of the trial at INT, researchers looked at data from patients suffering
from Soft Tissue Sarcoma, a cancer of soft tissue in muscles, tendons, fat and
blood vessels, said Nelken.
additional hospital systems will integrate Clinical Genomics into their
personalized medicine programs. The platform will keep data anonymous and
identify cases by search criteria such as age, gender, symptoms and diagnosis,
genomics can improve treatment and limit the number of unnecessary procedures,
said Nelken. (A recent survey
in the March 2012 issue of the journal Health Affairs
found that using
computerized systems led doctors to order more lab tests.)
Genomics uses natural-language processing and machine-learning capabilities
that are similar to Watson, IBM reported. The new platform's algorithms and
analytics capabilities also complement the deep question and answering features
of Watson, according to the company. On March 1, IBM announced a new advisory
board of experts for Watson that will research how the supercomputing
technology can help clinician workflows.
development in personalized medicine, Dell announced on Nov. 10 that it would donate its cloud
to the Translational Genomics Research
Institute (TGen) to house data for a personalized medicine trial being
conducted to find treatments for pediatric cancer. TGen is a nonprofit
organization focused on using genomics to develop medical treatment.