Prediction Tool Helps Determine Whether Older Breast Cancer Patients Will Benefit from Radiation
By Amelia Scholtz
statistical tool is now available to predict the likely individual
benefit of radiation therapy for older women who have undergone
breast-conserving surgery for breast cancer.
The prediction tool is based on a nomogram that uses patient-specific
factors to calculate a patient’s risk of requiring subsequent
mastectomy—the most common treatment for local breast cancer
recurrence—within 5 and 10 years after breast-conserving surgery alone
and after breast-conserving surgery plus adjuvant radiation therapy.
“The question of when it is okay to treat older women with breast
cancer using lumpectomy without radiation has been an ongoing source of
controversy,” said Benjamin Smith, M.D., an assistant professor in the
Department of Radiation Oncology at The University of Texas MD Anderson
Cancer Center. He and colleagues from MD Anderson and the University of
Chicago conducted the study that produced the nomogram. “Our concern
was that the National Comprehensive Cancer Network guideline statement
was not sufficiently nuanced to truly capture the breadth of types of
patients for whom radiation therapy can be omitted,” Dr. Smith said.
The decision-making tool
The decision-making tool, which is available on MD Anderson’s Web site,
first asks the user to designate the patient’s age and race, the
tumor’s size and estrogen receptor status, and the lymph node status.
The tool then provides the described patient’s risks of needing a
mastectomy by 5 and 10 years after surgery with and without radiation.
For example, a physician might select an age of 66–69 years, white
race, tumor size of 2.0 cm or less, negative or borderline-positive
estrogen receptor status, and negative nodal status that was assessed
clinically but not pathologically. For a patient with those
characteristics, the nomogram calculates the risks of needing
mastectomy by 5 and 10 years to be 9% and 21%, respectively, if the
patient does not receive radiation therapy; however, if the patient
does receive radiation therapy, the risks of needing mastectomy by 5
and 10 years are 2% and 5%, respectively.
The online tool offers advantages for both clinicians and patients.
“For physicians, risk prediction tools that allow you to take the
experience from the literature and then apply it to your patient can
really improve decision making,” Dr. Smith said. The tool also
facilitates patient autonomy in decision-making by showing the likely
benefit of radiation therapy in easily understood terms.
The research behind the tool
The process of creating the nomogram began with Dr. Smith and his
colleagues using the Surveillance, Epidemiology, and End
Results–Medicare database to identify women who were diagnosed with
breast cancer at ages 66–79 years between 1992 and 2002 and who had no
cancer history. Women 80 years and older were excluded because the
team’s earlier research suggested that those patients were unlikely to
benefit significantly from radiation therapy. Also excluded were
patients diagnosed with ductal carcinoma in situ and those with missing
data. The final cohort included approximately 16,000 patients, 89% of
whom received radiation therapy after their initial surgery.
The next step was to identify risk factors for subsequent mastectomy.
Younger age, larger tumor size, and black race were found to be
independent risk factors for mastectomy in the patient cohort. The
researchers also studied which patients received the greatest benefit
from adjuvant radiation therapy. Estrogen receptor positivity was
associated with lower tumor grade, which itself was associated with a
smaller benefit from radiation. Patients who had pathologically
confirmed disease in one or more lymph nodes or who had node-negative
disease that had been diagnosed only clinically were the most likely to
benefit from radiation therapy.
Yu Shen, Ph.D., a professor in the Department of Biostatistics, and
Diane Liu, M.S., a statistical analyst in that department, used these
data to generate the nomogram in consultation with
physician-scientists. They then internally validated the nomogram, and
after some refinement, its accuracy was similar to that of predictive
tools used for other purposes in clinical practice.
The observational data used to produce the nomogram had several
limitations worth noting. The cohort included relatively few black
women, making it difficult for researchers to determine why black race
appeared to be a risk factor for subsequent mastectomy. Additionally,
the use of mastectomy as a proxy for local disease recurrence does not
account for the minority of recurrences that are treated with
lumpectomy. Therefore, the absolute risk of recurrence is almost
certainly higher than the risk of mastectomy recorded in the study.
Finally, data for additional possible risk factors such as surgical
margin status and lymphovascular space invasion were not available.
Despite these limitations, the results produced by the nomogram are
consistent with the findings of a 10-year, prospective, randomized
trial reported by Hughes and colleagues at the annual American Society
of Clinical Oncology meeting in 2010. Moreover, Dr. Smith is hopeful
that his team will ultimately be able to externally validate the data
on which the nomogram is based, possibly using data from other
population-based tumor registries.
Physicians and patients can access the prediction tool and other
clinical calculators at
information, call Dr. Benjamin Smith at 713-563-2380.
articles in OncoLog, June 2013 issue:
Home/Current Issue | Previous Issues
| Articles by
Topic | Patient
About Oncolog | Contact OncoLog | Sign
Up for E-mail Alerts
©2014 The University of
Texas MD Anderson Cancer Center
1515 Holcombe Blvd., Houston, TX 77030
1-877-MDA-6789 (USA) / 1-713-792-3245
Referral Legal Statements Privacy