SAN ANTONIO – A new prognostic tool that uses four clinical and pathological variables may help to guide decisions about extending adjuvant endocrine therapy for postmenopausal women with estrogen receptor–positive (ER+) breast cancer, according to a study reported at the San Antonio Breast Cancer Symposium.
ER+ breast cancer is well known for recurring long after endocrine therapy stops, but the risk varies widely, ranging from 10% to 40% (N Engl J Med. 2017;377:1836-46), noted lead investigator Ivana Sestak, MS, PhD, a lecturer in medical statistics at the Queen Mary University of London. “A few trials have shown that extended endocrine therapy can reduce the risk of recurrence, but careful assessment of potential side effects and actual risk of developing a late distant recurrence is essential,” she said.
Courtesy of Dr. Sestak
Dr. Ivana Sestak
The investigators used the CTS5 to stratify patients into a low-risk group (risk of late distant recurrence less than 5%), an intermediate-risk group (risk between 5% and 10%), and a high-risk group (risk more than 10%). The observed rates of distant recurrence between years 5 and 10 were about 3% for the low-risk group, 7% for the intermediate-risk group, and 19% for the high risk-group. In addition, the CTS5 outperformed the original Clinical Treatment Score (CTS0), which was developed to predict recurrence between 0 and 10 years (J Clin Oncol. 2011;29:4273-8).
“We have developed a simple prognostic tool for the prediction of late distant recurrences which will help clinicians and their patients in the decision-making process about extended endocrine therapy,” Dr. Sestak commented. “The CTS5 was highly prognostic for the prediction of late distant recurrences and identified a large proportion of women, 42%, as low risk, where the value of extended endocrine therapy is limited. The CTS5 was also more prognostic than the already published CTS0 and should be used in this context for the prediction of late distant recurrence.”
“We aim to make the CTS5 algorithm and risk curve, with a read-out table, available to clinicians, and it will also be published in our manuscript,” she added.
Session attendee Frankie Ann Holmes, MD, of the Texas Oncology/US Oncology Network in Houston commented, “Just identifying high risk doesn’t necessarily translate into benefit, which is what we see with the Breast Cancer Index: You get the high risk, but then you learn if there is actually benefit to the extended therapy. Does your assay have a benefit portion to it?”
“No, we can’t look at the predictive benefit [with the CTS5]. This assay is purely a prognostic tool to predict late distant recurrences,” Dr. Sestak replied. “In these two trials, we do not have information on how many patients actually went on to extended endocrine therapy. You have to remember, these are old trials – they finished in about 2007-2008 – so not many women would have been given extended endocrine therapy at that time point.”
Session attendee Laura J. van’t Veer, PhD, of the University of California, San Francisco, asked, “How do you feel this will translate for risk up to 20 years, for which the question of extended endocrine therapy might also be very relevant?”
“For the purpose of this analysis, we only looked at out to 10 years. But I agree, it’s also important if we could apply a prognostic tool out to 20 years,” Dr. Sestak replied. “We have longer follow-up on some of the ATAC women, and we might look into that to see if we see any benefit of using a prognostic tool in the prediction of late distant recurrences.”
Study details
The investigators developed and trained the new tool using data from 4,735 women from the ATAC trial. They then validated the tool using data from 6,711 women from the BIG 1-98 trial.
The final CTS5 model contained four clinical variables, Dr. Sestak reported: number of involved nodes, size of the tumor, grade of the tumor, and age of the patient.
In the ATAC population, the CTS5 model did a better job than the original CTS0 model of predicting late distant recurrence. CTS5 improved the prediction of late distant recurrence by a factor of 2.47, whereas CTS0 improved the predictive value by a factor of 2.04. The CTS5 model performed similarly well regardless of whether patients had received chemotherapy.
In the BIG 1-98 population, the findings were much the same: The CTS5 model improved prediction of late distant recurrence by 2.07, while the CTS0 model improved prediction of late distant recurrence by 1.84. Performance of the CTS5 model was again similarly good regardless of whether patients had received chemotherapy.
Observed rates of distant recurrence between years 5 and 10 were similar in the ATAC and BIG 1-98 populations for the CTS5-defined low-risk group (2.5% and 3.0%, respectively), intermediate-risk group (7.7% and 6.9%), and the high-risk group (20.3% and 17.3%).
When the two trials’ populations were combined, the observed rate was 3.0% in the CTS5-defined low-risk group, 7.3% in the intermediate-risk group, and 18.9% in the high-risk group.
In addition, the main results held up among all node-negative women combined and among all women who had between one and three positive nodes combined. “For women with four or more positive lymph nodes, the CTS5 was not informative and categorized virtually all women into the high-risk group,” Dr. Sestak noted.
The investigators did not look at whether local or regional recurrences modulated the risk of late distant recurrence, she said. However, women who had experienced isolated local recurrence during the first 5 years would have been included in analysis.
“A strength of our study is that we used clinicopathological parameters that are measured in all breast cancer patients, and there is no need for further testing,” noted Dr. Sestak, who disclosed that she has received fees for advisory boards and lectures from Myriad Genetics.
On the other hand, it is unclear how the CTS5 would perform among premenopausal women and among women with HER2-positive disease given that two trials took place before routine HER2 testing and HER2-directed therapy were used.
SOURCE: Sestak I et al. SABCS 2017 Abstract GS6-01.