An imperfect surrogate
By definition, a surrogate is “one appointed to act in place of another.” In the case of PREDIX-HER2 and most other NAT studies in HER2-positive breast cancer patients, pCR is a surrogate for the endpoint about which doctors and patients really care – cancer-free survival.
As such, pCR is not a perfect surrogate. Reproducibly, pCR has been highly predictive of disease-free survival and overall survival, especially in the HR-negative subset of HER2-positive patients.
However, despite improvements with dual targeting of HER2 in the TRYPHAENA trial (Eur J Cancer. 2018;89:27-35) and the use of T-DM1 for patients failing to achieve pCR in the immediately practice-changing KATHERINE trial (N Engl J Med. 2019;380:617-28), eventual relapse is seen in 10%-20% of patients in various clinical-pathologic subgroups.
In PREDIX-HER2, the pCR rate for node-positive patients was considerably lower with T-DM1 than with DTP (54.1% vs. 38%), noted Valentina Guarneri, MD, of the University of Padova (Italy), in her discussion of the trial at ESMO Breast Cancer 2020.
Patients with larger initial tumor size and multiple involved axillary nodes at diagnosis remain at increased risk of death because of cancer relapse.
Central nervous system relapse remains a vexing problem. Among patients with triple-positive breast cancer, relapses may occur late, despite pCR.
Patients whose tumors transform from HER2 positive to HER2 negative with NAT, seen in approximately 8% of cases in the KATHERINE trial (ESMO Breast Cancer 2020, Abstract 96O), may be another poor-risk group.
Clinical implications
PREDIX-HER2 is an important study. At the early time point of 2.4 years (especially early since most patients were HR-positive), if pCR is achieved, event-free survival is excellent with T-DM1 or an aggressive multiagent cytotoxic combination plus dual HER2 targeting followed by anthracyclines.
It is ideal to have clinical-pathologic tests to distinguish those patients destined to achieve the surrogate endpoint of pCR from those who will not achieve it.
Despite linkage of TILs to improved outcome for triple-negative and HER2-enriched molecular subtypes (Lancet Oncol. 2018 Jan;19[1]:40-50), analysis of TILs is not standard practice in HER2-positive breast cancer in community settings. Optimal cutoffs are not well established, and TILs have not been linked to the choice of particular treatment options.
Currently, PET-CT scans are not part of National Comprehensive Cancer Network guidelines for pretreatment evaluation, except in patients for whom there is clinical suspicion of distant disease.
For those reasons, the main results of PREDIX-HER2 remain research tools that will focus our attention on the clinical-pathologic correlations Dr. Hatschek highlighted, but the results should have no influence on routine clinical practice at this time.
PREDIX-HER2 was funded by the Swedish Cancer Society, Radiumhemmet of Karolinska Institutet, Region Stockholm, and Roche Sweden. Dr. Hatschek disclosed relationships with Roche Sweden, Pfizer Sweden, and Pierre Fabre Sweden. Dr. Guarneri disclosed relationships with Roche, Novartis, and Eli Lilly.
Dr. Lyss was a community-based medical oncologist and clinical researcher for more than 35 years before his recent retirement. His clinical and research interests were focused on breast and lung cancers as well as expanding clinical trial access to medically underserved populations. He is based in St. Louis. He has no conflicts of interest.
SOURCE: Hatschek T et al. ESMO Breast Cancer 2020, Abstract 97O.