CHICAGOMuch of cancer treatment rightly relies on evidence from
randomized clinical trials. However, definitive clinical trials have
not been done on some problem diagnoses, such as ductal carcinoma in
Many major clinical investigations also fail to gather information
about such aspects of treatment as trade-offs, for example, between
potentially improved survival and breast preservation for women with
DCIS, or the degree to which patients are willing to tolerate
chemotherapy toxicities in exchange for possibly improved survival,
for example, in ovarian carcinoma.
Outcomes researchers consequently are developing tools or study
approaches to fill in these knowledge gaps. Two such researchers
described their efforts at a symposium on quality of life and
outcomes sponsored by Northwestern University and Evanston
Bruce E. Hillner, MD, professor of medicine, Virginia Commonwealth
University, Richmond, has been applying decision modeling using the
Markov process to compare treatment strategies for DCIS. The disease
is a relatively new enigma for oncology, he said, because it is being
diagnosed more frequently (now accounting for between 10% and 15% of
all new cases of breast cancer), and its natural history is poorly understood.
Women are not necessarily offered a choice of treatment, however. Nor
do they have solid clinical data on which to base an informed
decision. No randomized clinical trial has been conducted to evaluate
mastectomy vs breast-conserving surgery in this setting, and there
has been only one trial comparing breast-conserving surgery alone and
breast-conserving surgery plus radiotherapy, he said.
Although two ongoing trials are investigating breast-conserving
surgery in combination with radiotherapy or with tamoxifen
(NSABP-24), results from these trials are at least a year away.
Dr. Hillner used decision modeling to compare actuarial, average, and
discounted survival for mastectomy, breast-conserving surgery, and
breast-conserving surgery plus radiotherapy. He then factored in the
quality of life for women who had both breasts preserved.
According to this modeling plan, women who opted for mastectomy would
not die from breast cancer at 10 and 20 years after treatment for
DCIS. (Women could die from other causes.) Although actuarial
survival at 20 years dropped off significantly for the two other
treatment alternatives, average survival was similar in all groups:
17.94 years for mastectomy, 17.73 years for breast-conserving surgery
plus radiotherapy, and 17.38 years for breast-conserving surgery alone.
In terms of discounted survival, which takes into account the value
of being alive today vs survival sometime in the future, the
differences were minimal. When compared with women who had
mastectomy, women who had breast-conserving surgery plus radiotherapy
would sacrifice only 38 days of survival, and women who had
breast-conserving surgery alone would sacrifice 103 days. Thus, for a
compromise of only about 1 to 3 months in survival, women could
retain both breasts for 50% to 66% of the 20-year interval.
Although this decision-modeling plan does not provide treatment
answers, it does lay the groundwork for making informed
choices, he said.
Elizabeth Calhoun, PhD, assistant professor, Northwestern University
Medical School, has been developing health state utilities that
consider the effect of chemotherapy toxicity on quality of life. She
has used utility assessments to assess the preferences or attitudes
of women with ovarian cancer, women at risk of the disease (such as
the daughters of women under treatment for ovarian cancer), and
healthy women and physicians.
These utilities evaluate a set of hypothetical scenarios related to
the severity of ototoxic, nephrotoxic, and neurotoxic side effects of
chemotherapy on a scale of 0 (death) to 1 (perfect health). The
objective was to determine the lowest amount of full-quality life a
woman would accept as being equal to living one full year in the
defined health state.
Women in essence answered the question: How many months of life would
you give up to avoid a mild, moderate, or severe side effect?
The utilities were tested in 39 women with ovarian cancer, 15 women
at risk of developing the disease, 30 healthy women at baseline risk,
and 11 physicians.
Overall, healthy women were the most willing to give up more time in
a poorer state of health for less time in perfect health. Women with
ovarian cancer and women at high risk for the disease did not differ
on any of the health states while physicians underestimated their
patients willingness to sacrifice quantity of time for quality
It is important to assess how health state utilities or other
instruments that assess patients preferences can be used in
making treatment decisions, she said. She added that these and other
types of instruments are needed to learn how far patients are willing
to go with treatments that affect overall quality of life.