Steven E. Vogl, MD
The recently published report of Austrian Breast and Colorectal Cancer Study Group’s Study 18 (ABCSG-18)1 for the secondary endpoint of disease-free survival suggests that denosumab given in a low dose of 60 mg subcutaneously every 6 months during aromatase inhibitor adjuvant therapy is beneficial. And an accompanying editorial by Dr. Marc Lippman of Georgetown University agrees.2 The absolute benefit in disease-free survival at 8 years is 3.1%, with a hazard ratio of 0.82 and a P value of .026.
The conclusions of the paper are technically correct but misleading. When one gives adjuvant systemic therapy to women with resected primary breast cancer, the goal is generally to prevent or delay distant metastases that lead to death. There is no evidence from this trial that denosumab does this. The “disease-free survival” endpoint defined by the STEEP consortium includes a number of other undesirable but nonetheless generally nonlethal events. Further, a larger adjuvant trial (called D-CARE) of higher doses of denosumab given more frequently in a higher risk population failed to show any benefit in disease-free survival.
Disease-Free Survival Is Poor Endpoint for Good-Prognosis Breast Cancer
The disease-free survival events reduced by denosumab were largely new non-breast primary cancers—nothing directly related to the breast cancer at all!
The STEEP collaboration in 2007 tried to standardize what had been confusing and contradictory definitions of disease events in breast cancer in reports of various trials. Many adjuvant trials had included new primary cancers in nonbreast sites as events for disease-free survival analysis, and the STEEP consortium accepted this. One suspects that this was an effort to include information on second primary endometrial cancers caused by adjuvant tamoxifen and to subtract their incidence from the benefits of tamoxifen in preventing breast cancer recurrence and new breast primaries.
Most disease-free survival events in a population like that enrolled onto ABCSG-18 have little to do with the course of the breast cancer for which the patient is being treated.— Steven E. Vogl, MD
Tweet this quote
Disease-free survival is not a bad surrogate for overall survival in patients with very high–risk breast cancer, most of whose events will be distant recurrences that lead to death, generally within a few years after distant metastases are symptomatic. The study population for ABCSG-18 was relatively low risk (primary tumor < 2.1 cm, 72%; node-negative, 71%; grade 1 to 2, 81%; though 6% were HER2-positive). Only 18 of 287 disease-free survival events in the 1,709 women enrolled in the placebo arm of ABCSG-18 were histologically verified distant metastases. Both a new second nonbreast primary cancer and death without antecedent recurrence of breast cancer were considerably more common than distant metastases. This means that most disease-free survival events in a population like that enrolled onto ABCSG-18 have little to do with the course of the breast cancer for which the patient is being treated.
The 3.1% difference in disease-free survival at 8 years is a result of 47 more disease-free survival events in the women assigned to placebo in the trial. This difference is a result of 20 more women in the placebo arm with a histologically proven second primary in another organ, 9 more women who died without an antecedent cancer event, 5 more women with an invasive contralateral breast primary, and 12 more women who had either histologically undocumented distant metastases or histologically undocumented second primaries. It would be of great interest to know how many of this difference of 12 results from second primaries, and the distribution of excess second primaries in the placebo and denosumab groups.
The study chairman points out that an unbiopsied single lesion could be either, and he defends this category (Gnant M, personal communication, May 11, 2019). One suspects that most of these events could be coded based on their presentation. I would code unbiopsied single lung lesions as lung primaries; single bone lesions as benign; single lesions of skin, stomach, and intestines as new primaries or benign; and a single enlarging or new liver lesion as metastasis unless the patient had preexisting liver disease. Even with this uncertain category, though, it is clear that denosumab did not substantially reduce the incidence of distant metastases.
One should consider that, although RANK ligand and NF kappa-B have numerous biologic effects, no one ever thought that giving the RANK ligand inhibitor denosumab to healthy humans would prevent a variety of cancers. ABCSG-18 was not designed to test this hypothesis. If denosumab actually prevents or delays the emergence of cancer in many sites in the human body, this would be of great clinical interest. Since the finding was not confirmed in the larger adjuvant denosumab D-CARE study, this finding in ABCSG-18 most likely is a chance event that will not prove reproducible in another trial designed to look for it. Until we demonstrate that it is reproducible, we should not administer denosumab to patients to achieve an endpoint that we think is a chance event.
D-Care Study of High-Dose Denosumab Finds Neither Improved Disease-Free Survival Nor Decreased New Primary Malignancies
The D-CARE study presented just after ABCSG-18 at the 2018 ASCO Annual Meeting found no effect of denosumab on the incidence of distant metastases or of second primary cancers, which were analyzed as toxic events in D-CARE. This makes it likely that the finding of reduced second primaries in ABCSG-18 is a chance event. The dose of denosumab used in D-CARE was twice that in ABCSG-12, and denosumab was given much more frequently in D-CARE. At the high dose used in D-CARE, denosumab produced an unacceptably high rate of jaw osteonecrosis (5.4%).
Reduction in new cancers should have been greater in D-CARE at higher doses than in ABCSG-18, if the effect on second primary cancers were real. Although denosumab did not reduce second primary cancers or the rate of distant metastases in D-CARE, it did shift the sites of first metastases away from bone to some extent, balanced by an increase in nonbony sites of spread. D-CARE, then, fails to replicate the findings of ABCSG-18 in terms of reducing new nonbreast primary cancers.
Bone-Active Agents in Breast Cancer Adjuvant Treatment
In study after study including all patients with breast cancer, bisphosphonate inhibitors of bone turnover failed to significantly improve disease-free and overall survival.3 These studies included a very large international study called AZURE. It was only after all these studies were done that the investigators noticed a nearly uniform benefit among women who were clearly postmenopausal! A formal meta-analysis confirms this impression.4 The effective agents in the meta-analysis at preventing disease-free survival events were oral clodronate and intravenous zoledronate.
Unfortunately, so far as I know, there is as yet no completed randomized prospective trial confirming the hypothesis that a bone-active agent given to postmenopausal women prevents the emergence of bone or other metastases or prolongs life. The data we now have consist of a collection of subgroup analyses of negative studies hallowed by inclusion in a meta-analysis of subgroup analyses.
The hypothesis of efficacy for older women alone was clearly generated after the data from each study were in place and had been reviewed. None of the very smart investigators for a large number of studies had figured out that postmenopausal women were the correct target population. To conclude from all these analyses that adjuvant bisphosphonates prevent relapse among postmenopausal women is not clearly wrong—however, neither is it clearly right. This was not the hypothesis being tested in any of the trials! I am waiting for the positive report of a trial testing this hypothesis before unequivocally urging my patients who are 55 years of age and older to take adjuvant bisphosphonates to prevent breast cancer events. Guidelines from ASCO have accepted that bisphosphonates are active in this population and recommend them to prevent metastases.5 I have been reluctant to recommend them unless the risk of metastases is very high.
Denosumab Halves the Rate of Fractures on Anastrozole—Offer It for This Purpose
The primary analysis of ABCSG-18 showed that adding 60 mg of denosumab subcutaneously every 6 months during anastrozole therapy reduced the incidence of symptomatic fractures over the 5-year period of treatment from 16% for placebo to 8% for denosumab.6 It makes sense to offer denosumab as a safe, albeit expensive, way to reduce fractures, which are painful and inconvenient but generally not life-threatening.
A conscientious physician would not recommend denosumab as a way to prevent breast cancer metastases or death, or even to prevent local recurrence or new breast primaries. The difference in emphasis matters—preventing the pain and inconvenience of a fracture for 4 weeks is not nearly as important as avoiding lethal metastases. In a value-based payment system, payment would be much less for a drug that benefits 8% of patients a little than for one that prevents lethal metastases and death in 4%. ■
DISCLOSURE: Dr. Vogl has an immediate family member who owns stock in Amgen.
REFERENCES
1. Gnant M, Pfeiler G, Steger GG, et al: Adjuvant denosumab in postmenopausal patients with hormone receptor-positive breast cancer (ABCSG-18): Disease-free survival results from a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 20:339-351, 2019.
2. Lippman M: Adjuvant denosumab in postmenopausal patients with hormone receptor-positive breast cancer. Lancet Oncol 20:312-313, 2019.
4. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG): Adjuvant bisphosphonate treatment in early breast cancer: Meta-analyses of individual patient data from randomised trials. Lancet 386:1353-1361, 2015.
5. Van Poznak C, Somerfield MR, Moy B: Role of bone-modifying agents in metastatic breast cancer: An American Society of Clinical Oncology-Cancer Care Ontario Focused Guideline Update Summary. J Oncol Pract 13:822-824, 2017.
6. Gnant M, Pfeiler G, Dubsky PC, et al: Adjuvant denosumab in breast cancer (ABCSG-18): A multicentre, randomised, double-blind, placebo-controlled trial. Lancet 386:433-443, 2015.
At Microphone 1 is an occasional column written by Steven E. Vogl, MD, of the Bronx, New York. When he is not in his clinic, Dr. Vogl can generally be found at major oncology meetings and often at the microphone, where he stands ready with critical questions for presenters of new data.
The opinions expressed in this column are those of the author. If you would like to share your opinion on this or another topic, please write to editor@ASCOPost.com
