Download National Cancer Institute`s New Tool Puts Cancer Risk in Context

2 of 8 | JNCI J Natl Cancer Inst, 2015, Vol. 107, No. 12
news
choosing their dose and fractionation
schemes. What this study does is take us
to the next level in that approach, allowing us to further personalize care.”
Max Diehn, M.D., Ph.D., an assistant professor of radiation oncology at
Stanford University School of Medicine
in Palo Alto, Calif., noted that “although
the larger study does show that a gene
expression signature correlated with
radiation resistance in previous studies is more highly expressed in metastases than primary tumors,” the paired
analysis did not. “This suggests expression of the signature may be an intrinsic
property of a given patient’s tumor,” said
Diehn. “However, it’s also possible that
tumors with the more resistant gene
expression profile may be more likely
to metastasize to certain organs.” If that
can be confirmed, Diehn said, it would
provide support for factoring where the
metastasis is when choosing radiation
doses. Still, he emphasized, “the findings
are preliminary and would need to be
validated.”
Torres-Roca agreed that the RSI could
be identifying something inherent to the
organ. “Or could it be that metastases
that go to the lung are by chance more
radiosensitive?” he said. “Or is it that the
metastases become more sensitive when
they go to the lung? It’s one of those
things that make you want to dig more
to find out why this is.”
The RSI has been validated in retrospective studies. Testing samples from
a prospective phase III clinical trial that
will randomize patients to receive or not
receive radiation is needed next. TorresRoca is trying to obtain funding to support this analysis. “Billions of dollars are
going into developing targeted therapies,”
he said, “but radiation is already readily
available, highly effective and we’ve been
using it for a long time. By using biological information from the tumor, we could
learn how to do it better.”
Haas-Kogan warned that the effort
to personalize radiation oncology might
be more challenging than using targeted agents in patients with specific
genetic mutations. “The targets of radiation are somewhat diffuse and varied,”
she said. “It’s not like medical oncology,
where you can look at a single target to
see if an agent is effective. Radiation is
less specific in its scope and anticancer
effect. It affects multiple pathways in
the tumor, which is why it is so effective.” Even so, “this type of information
could guide us to identifying who might
need more aggressive treatment to control tumor tissue,” Haas-Kogan continued. “Or it could possibly be expanded
beyond radiation sensitivity to tell us
who might experience more risk and
side effects or who might be more prone
to radiation-induced second malignancies. It is an exciting path to pursue.”
Torres-Roca and Steven A. Eshrich, Ph.D.,
another coauthor, both hold stock and leadership positions in Cvergenx Inc., and hold
patents on and receive royalties from the RSI.
© Oxford University Press 2015.
DOI:10.1093/jnci/djv397
First published online December 11, 2015
National Cancer Institute’s New Tool Puts Cancer Risk in Context
By Caroline McNeil
Type “cancer risk assessment” into
Google, and you’ll come up with a list of
assessment tools for particular cancers,
most with a strong focus on personal risk
factors related to lifestyle, exposures,
and medical and family history.
Would it help also to get a broader
view of cancer risk? The National Cancer
Institute thinks so. NCI has teamed with
Dartmouth researchers Steven Woloshin,
M.D., and Lisa Schwartz, M.D., to create
“Know Your Chances” (http://knowyourchances.cancer.gov/), a website that aims
to put cancer risk in perspective.
“We want to get people to understand
the risks of dying from specific cancers
in comparison to each other and to noncancer causes of death, and at different
times in their lives,” said Eric J. (Rocky)
Feuer, Ph.D., chief of the Statistical
Research and Applications Branch in
NCI’s Surveillance Research Program and
the senior investigator in this project.
Without this larger picture, it’s hard to
make sense of cancer statistics, according to Woloshin and Schwartz, both professors at the Dartmouth Institute for
Health Policy and Clinical Practice and
codirectors of its Medicine in the Media
Program.
“It’s difficult to read a newspaper or
magazine, watch television, or surf the
Internet without hearing about cancer,”
they say in the introduction to the website. “Unfortunately, these messages are
often missing basic facts needed for people to understand their chance of cancer:
the magnitude of the chance and how
it compares with the chance of other
diseases.”
The Charts
“Know Your Chances” has four sections:
1. Big-picture charts give the chance of
dying over 10-year periods, by age,
sex, and race for different cancers;
other major causes of death; and all
causes combined.
2. Custom charts allow users to generate charts by age, sex, and race for
different causes of death and over
different time frames.
3. Your chances let users see leading
causes of death by sex, race, and
exact age.
4. Special cancer tables show the risk of
diagnosis, as well as death, for particular cancers and across cancer
sites.
The data and estimates for the new
charts come from NCI’s DevCan statistical algorithms and database
(http://surveillance.cancer.gov/devcan/). DevCan calculates probabilities
of developing and dying from specific cancers by using incidence
data from the agency’s Surveillance,
Epidemiology, and End Results (SEER)
program; mortality counts from the
National Center for Health Statistics;
and U.S. census data.
The idea of comparing probabilities and using other ways to provide a
broader context builds on earlier work
by the Dartmouth researchers, who have
published widely on the communication
of medical and statistical information,
especially in relation to risk.
“We saw an opportunity to marry the
kind of calculation we do to the kind of
risk communication messages they do,”
said NCI’s Feuer.
Accurate risk communication messages are based on several principles,
Schwartz and Woloshin said. One is to
put the number of deaths in numerical
context—e.g., 3 of 100 people, or 3%—
rather than the absolute number of people expected to die of a disease. Give the
denominator, in other words, as well as
the numerator.
Another is to avoid making statements of relative risk without giving
actual figures. A 50% risk reduction
sounds large but may not mean much if
risk is reduced from 4% to 2%. But if risk
is reduced from 60% to 30%, a 50% reduction can mean a lot.
A guiding principle is to put risks in
accurate perspective. “We wanted to put
risk in context,” Schwartz said, “by using
denominators, by comparing each specific cancer to other cancers and other
major causes of death, and by using
standardized time frames and formats.”
The 10-year time frame is arbitrary
but makes sense, they said, because it
avoids the exaggerated risk that comes
with time frames that are too long as
well as those that are too short. “Over-alifetime risk distorts the picture, but so
can too short a time frame,” Woloshin
said. That can cause you to underestimate your risk.”
The 10-year time frame also allows
people time to do things such as make
lifestyle changes or consider proven
screening tests, they note on the website.
Setting Priorities
The charts present risk by age, sex, and
race, but so far not by factors related
to lifestyle or heredity. However, NCI
plans to add smoking, the single most
important risk factor for a variety of cancers and other causes of death.
In an earlier version of the charts
(J. Natl. Cancer Inst. 2008;100:845–53),
the Dartmouth authors included how
smoking affects risk. Data and methods
are now being updated using National
Health Interview Surveys that include
smoking status and follow-up data on
mortality and cause of death.
As for family history, lifestyle, and
other risk factors, finding comparable
data sources may be more problematic.
“We see it as providing
breadth, giving the broad
landscape. It could help
a person make decisions
about priorities—diabetes,
lung cancer, heart attack,
for instance—before
making decisions about
lifestyle changes or
screening.”
“We have to consider feasibility,”
Schwartz said. “Right now we don’t have
good data for all of them. But even with
extra risk factors we might not know that
much more.”
In any case, “Know Your Chances”
is not designed to do exactly the same
thing as other risk assessment tools,
which may provide estimates for a single disease as a function of several risk
factors. Instead, it could complement
them, Feuer said. “We see it as providing breadth, giving the broad landscape.
It could help a person make decisions
about priorities—diabetes, lung cancer,
heart attack, for instance—before making decisions about lifestyle changes or
screening.
“Both facts and values go into decision making,” Woloshin noted. “These
charts give facts. The values are individual, how you process the facts. Research
has shown people process the same facts
differently.”
That finding suggests that the charts’
facts could be the basis for doctor–
patient discussions of risks and what to
do about them.
In fact, “the charts could be the first
step in a huge educational process,” said
Otis Brawley, M.D., chief medical officer
of the American Cancer Society. He said
physicians could
use the charts
to help patients
avoid the common pitfall of
exaggerating
the chance of
dying from one
disease, such as
prostate cancer,
while minimizEric J. (Rocky) Feuer,
ing the greater
Ph.D.
chance of dying
from others. For instance, a 55-yearold white man has a 1 in 1,000 chance
of dying from prostate cancer within
10 years, compared with 2 in 1,000 for
high blood pressure, 3 in 1,000 for stroke,
and 10 in 1,000 for lung cancer, according
to the charts.
“One of the great problems we deal
with in our society is that people don’t
understand or they misperceive risk,”
Brawley said. “These charts could reduce a
lot of mental anguish and a lot of concern.”
© Oxford University Press 2015.
DOI:10.1093/jnci/djv398
First published online December 11, 2015
Targeted Therapies Improve Outlook for Chronic Lymphocytic
Leukemia
By Vicki Brower
In a pilot trial of an autologous cell
therapy for chronic lymphocytic leukemia (CLL), four of 14 patients achieved
complete responses and four had partial responses, some of which are ongoing 4.5 years after treatment. University
of Pennsylvania researchers recently
reported mature data from the pilot
trial of the treatment, CTL-019, begun
in 2010, in which patients received
an infusion of their own T cells that
were genetically modified to express
a chimeric antigen receptor (CAR) (Sci.
Transl. Med. 2015;7:303;doi10.1126/scitranslmed.aac5415;
doi:10.1126/scitranslmed). CAR T cells target the CD19
protein found on the surface of B cells
that characterize CLL.
news
NEWS | 3 of 8