CHICAGO — A year ago, results from TAILORx made a big splash by showing that use of the 21-gene signature (Oncotype DX, Genomic Health) could identify women with breast cancer who do not need chemotherapy.
That finding — which was the primary endpoint of TAILORx — showed that after 9 years of follow-up, postmenopausal women with hormone receptor-positive (HR+), human epidermal growth factor receptor 2-positive (HER 2+), node-negative breast cancer with intermediate genomic risk did not benefit from the addition of chemotherapy (CT) to endocrine therapy.
Now, a year later, a preplanned analysis of a secondary endpoint shows that "greater precision" may be attained when clinical risk features of the disease are also considered, and gives new results for premenopausal women.
"Integration of genomic and clinical risk offers the potential for greater precision in prognosis and ultimately guiding the use of adjuvant chemotherapy," said lead author Joseph A. Sparano, MD, from the Montefiore Medical Center in New York City.
"With this new analysis, it is clear that women ages 50 or younger [premenopausal] with a Recurrence Score result between 16 and 20 and low clinical risk do not need chemotherapy," Sparano commented in a press release from Genomic Health. "Furthermore, the Oncotype DX test in combination with clinical risk factors could identify premenopausal women with higher clinical risk who may benefit from ovarian function suppression and more aggressive anti-estrogen therapy," he added.
In the Highlights of the Day session, Nancy Davidson, MD, of the Fred Hutchinson Cancer Research Center, University of Washington, Seattle, described Sparano's presentation as "data rich" and the TAILORx trial as "the gift that keeps on giving."
In this new analysis, the investigators refined a model that used genomic tests and added classical clinical information, she noted.
In TAILORx, genomic risk for breast cancer recurrence was measured on the Oncotype DX test, with results interpreted as a recurrence score (RS). In that analysis, an RS of 0 to 10 is considered low risk; RS 11 to 25 is considered intermediate; and RS 26 to 100 is considered high risk for breast cancer recurrence.
The main result for the primary endpoint was reported last year; it showed that postmenopausal women with hormone HR+, HER 2+, node-negative breast cancer who had recurrent score (RS) between 11 and 25 did not benefit from the addition of CT to endocrine therapy.
In this analysis, which incorporated clinical risk to genomic risk, high vs low clinical risk was determined from tumor size and histologic grade. Low risk defined a group of patients with tumors 1 cm or smaller and high grade; 2 cm or smaller and intermediate grade; or 3 cm or smaller and low grade. Patients with tumors outside of this categorization were considered to have high-risk tumors.
In the study population (N = 9427), 30% of patients had high-risk tumors and 70% had low-risk tumors.
Clinical risk provides additional prognostic information to RS for distant recurrences, Sparano said. In women with an RS 11 to 25 and receiving endocrine or chemoendocrine therapy, absolute increase in distant recurrence rate was approximately 5% for those with high-risk (vs low-risk) tumors, which translates to a 2.5 to 3 times higher risk for distant recurrence.
For women who had an RS 26 to 100 and who were treated with chemoendocrine therapy, the high-risk tumors were associated with a threefold increased risk and a 10% absolute difference in distant recurrence.
When the team analyzed whether women benefited from CT, they found that women in the group of RS 11 to 25 were not impacted by integrating clinical risk to genomic risk. In that sense, analysis of the primary endpoint remained unchanged, and in this group of women there was no benefit of adding CT to endocrine therapy, implying that it is not clinically useful to determine clinical risk in these women.
However, for women age 50 years or younger who fell in the group with RS 16 to 25, integrating clinical risk distinguished the 50% of patients who derive no CT benefit from the other 50% of patients who derived an absolute benefit of 6% to 9% in distance recurrence.
Sparano also showed that CT benefit is greatest for premenopausal women between the ages of 45 and 50 years who have an RS 16-25. "This effect may be due to a castration effect associated with cytotoxic therapy rather than an effect in eradicating micrometastatic disease," he said.
"Integrated risk provides greater prognostic prediction and may have clinical utility," Sparano concluded.
This approach may be superior to clinical or genomic features used individually, he suggested.
Secondary Endpoints in Clinical Decision Making
In an invited >NEJM editorial, the authors ask how clinicians are to be guided when secondary endpoints from a clinical trial are integrated into clinical decision making.
"The promise of 'precision' medicine has collided with the rather messier world of using all available evidence to try and make educated guesses to improve patient outcomes," write editorialists David J. Hunter, MBBS, from the Nuffield Department of Population Health at the University of Oxford, United Kingdom, and Dan L. Longo, MD, senior physician at Brigham and Women's Hospital in Boston, Massachusetts, and deputy editor of NEJM.
The editorialists support integrating this analysis of secondary endpoints from the TAILORx trial into clinical decision making.
They suggest that these secondary endpoints or even post hoc analysis may incur additional costs and even longer time to set up follow-up trials with a new primary objective generated from these observations. "These follow-up trials are unlikely to be performed," they write.
Those new trials, if they were to be undertaken, would take another 9 years to complete, but clinicians are faced with making treatment decisions now, they point out.
"If we insist on evidence from randomized trials, we will rarely have it, and to the extent we have it, we will have it only for groups of patients who are somewhat similar," Hunter and Longo note.
Additional analyses from important datasets remain a fundamental component of evidence synthesis, they argue. "Distinguishing between results that warrant a change in practice and those that do not will not be a 'precise' process," they write.
However, the editorialists also recognize the market dominance of the Oncotype DX test. Any new and superior panels — based on current approach to breast cancer prognosis — that incorporate a much larger-scale gene expression, genomic copy-number alteration, and epigenetic data will have a much higher hurdle to overcome, they suggest.
"In practice, commercially available prognostic or predictive markers may not be as precise as they could be, and barriers to innovation are high; thus, 'near enough' has become 'good enough,' " they note.
They suggest that even in the era of precision medicine, machine learning, and artificial intelligence, both physicians and patients will face substantial uncertainties, and will have to make "educated guesses" informed by multiple sources of evidence as well as by clinical acumen.
"Medicine in the molecular era will be no more 'precise' than in prior eras — evidence synthesis, clinical judgment, and patient preferences all factor in," Hunter and Longo conclude.
Spa rano reported stock and other ownership interests in Metastat; has consulting or advisory role with Adgero Biopharmaceuticals, AstraZeneca, Cardinal Health, Celgene, Celldex, Genentech/Roche, Juno Therapeutics, Lilly, Merrimack, Novartis, Pfizer, and Prescient Therapeutics; and receives travel/accommodations/expenses from Adgero Biopharmaceuticals, AstraZeneca, Menarini Silicon Biosystems, Myriad Genetics, Pfizer, and Roche/Genentech.
American Society of Clinical Oncology (ASCO) 2019 Annual Meeting: Abstract 503. Presented June 3, 2019.
Source : https://www.medscape.com/viewarticle/914295