It has been demonstrated that, in early breast cancer, there is no survival benefit to additional treatment once a patient has achieved a pathologic complete response (pCR). Together with the well-established prognostic association between pCR and survival, this presents a strong rationale for escalation/deescalation of neoadjuvant treatment based upon response.

See related article by Spring et al., p. 2838

In this issue of Clinical Cancer Research, Spring and colleagues, in what is the largest individual patient-level meta-analysis to date, examined nearly 28,000 cases across 52 neoadjuvant breast cancer trials. (1) As in multiple other studies, they show pathologic complete response (pCR) to be reliably associated with improved overall survival. The HR of 0.22 (which held true when only randomized clinical trials were included, HR 0.31) is consistent with previously reported work. For triple-negative and HER2+ disease, the probability intervals are very tight, but for hormone-positive disease, the probability intervals were wide. This is likely because hormone-positive disease is heterogeneous, even for larger tumors. In the I-SPY2 trial, a multigene test was used to exclude molecularly low risk disease, where risk of recurrence is often later (after 5 years), and where chemotherapy adds little, if any, benefit (2). In I-SPY2, achieving pCR, regardless of subtype, or type of therapy, including 10 novel targeted agents, was associated with an improved event-free and distant metastasis–free survival with a HR of 0.2 (3), demonstrating the importance of patient selection and biomarkers.

What is most impactful about this new meta-analysis is the comparison of 1,601 patients assigned to additional chemotherapy after achieving pCR with 18,462 who achieved pCR but received no subsequent adjuvant therapy after surgery. In terms of mortality reduction, the authors found no further value of additional therapy in patients who had already achieved pCR. The HR for event-free survival among pCR versus non-pCR patients was 0.36 for both of these groups.

In oncology, we have been wedded to the evidence generated by myriad adjuvant trials. These data have made clinicians reluctant to use anything less than the “standard therapy.” However, the adjuvant approach has limitations. It only allows the study of populations of patients and provides no short-term indicators of response. The neoadjuvant approach provides direct feedback on response, allowing as Spring and colleagues have done, to study women who received the first chemotherapy regimen prior to surgery, and the second regimen after. That further adjuvant therapy after achieving pCR carries no benefit runs counter to the long-held dictum that the full regimen must be delivered. In this case, more is not better, it is simply more toxic.

There has been a fair amount of controversy about whether early endpoints such as complete response should be used to determine the effectiveness of new agents and combinations and whether we can indeed tailor therapy based on response. The debate has even attracted mainstream attention.

In January, an editorial in the New York Times raised concern about the FDA approving new medicines/devices on the basis of surrogate or early clinical endpoints (4). The implication was that surrogate endpoints are potentially hazardous shortcuts.

The purpose of surrogate or early endpoints is to accelerate the pace of learning, or said differently, to make “knowledge turns” shorter. If you are suffering from a life-threatening disease where the treatments are toxic, you hope that your care team and the scientists studying your disease will have a sense of urgency, because your life is on the line. One would hope that there are efforts to move beyond standard ways of testing that require 10–15 years for an effective therapeutic to get to market, to speed development, and get more effective treatments to patients more quickly and efficiently.

In the 1980s, the sense of urgency regarding the AIDS epidemic led to a very successful strategy to speed development, one that was predicated on an early endpoint. T-cell counts became the early surrogate endpoint, rather than death from AIDS. This simple change dramatically accelerated the ability to tell whether a drug was working, and fueled the paradigm shift where AIDS became a treatable chronic disease rather than a death sentence. Today there are more than 200 retroviral drugs approved. Surrogates are still being used to rapidly test promising new therapeutic advances in cardiology to avoid death from fatal heart attack (5). In fact, early endpoints generated from integrity-based trials benefit patients and science.

In breast cancer, the implications of the Spring and colleagues' meta-analysis with the I-SPY2 findings are enormous. The neoadjuvant path is extremely valuable for two major reasons. First, we have proven and continue to demonstrate that response (in clinically and molecularly high-risk disease) is the primary predictor of the patient's long-term outcome. The I-SPY consortium and others have spent 20 years striving to bring effective drug treatment to patients earlier in the course of disease, where excellent response translates into a cure, not just additional months of survival. Complete pathologic response is a reliable early endpoint providing critical information that allows us to learn, early in the course of treatment, whether a drug is working. We are continuing to work to refine and improve the early endpoints using the more granular residual cancer burden, refining and integrating imaging endpoints of response, and emerging biomarkers such as circulating tumor DNA (6–8). Response to treatment is now dictating adjuvant treatments and trials.

Second, we can rapidly evaluate new and promising therapies, and escalate if treatment is suboptimal, avoiding additional toxic therapy if responses are excellent. This is about getting the right drug to the right patient at the right time. If we remove the tumor first, we cannot optimize treatment, and often make surgical therapy unnecessarily aggressive.

Indeed, this can be a watershed moment to pivot and design trials with the goal of optimizing individual outcomes by establishing a pathway to deescalate and escalate therapy in the neoadjuvant setting.

This is an experiment that needs to be run, and the I-SPY team is planning to do just that (Fig. 1). In the near future, we will introduce I-SPY2.2, which is designed to achieve the best outcomes with the least toxicity, and maximize the potential of the targeted therapies in development. Such a strategy allows treatment to stop at the end of the regimen when pCR is predicted and to advance to increasingly targeted therapies when complete response is not achieved.

Figure 1.

I-SPY2.2 is a proposed modification of the current trial design with the goal of optimizing individual outcomes by escalating or deescalating treatment based upon response. While retaining I-SPY's signature adaptively randomized platform design, it incorporates a sequential multiple assignment randomization trial approach that offers the opportunity to rerandomize to targeted therapies for those with suboptimal response, or to forgo further treatment once pCR has been achieved. The first randomization echoes the current I-SPY2 design, serving to evaluate new agents or combinations within predefined subtypes. Based upon response, assessed by MRI and other biological markers, participants predicted to achieve pCR will proceed to surgery to avoid additional toxic therapy. Those with suboptimal response will have the opportunity to be rerandomized to a targeted therapy. Second-line treatments in the second randomization will consist of best-in-class therapies for each subtype, including previous graduates from I-SPY2 and/or the initial randomization. Again, response to second-line treatment determines whether a participant is spared further toxic therapy by going directly to surgery or is rerandomized to a rescue agent. RCB, residual cancer burden; pRCB is predicted RCB based on MRI and biopsy; AC is doxorubicin/cyclophosphamide.

Figure 1.

I-SPY2.2 is a proposed modification of the current trial design with the goal of optimizing individual outcomes by escalating or deescalating treatment based upon response. While retaining I-SPY's signature adaptively randomized platform design, it incorporates a sequential multiple assignment randomization trial approach that offers the opportunity to rerandomize to targeted therapies for those with suboptimal response, or to forgo further treatment once pCR has been achieved. The first randomization echoes the current I-SPY2 design, serving to evaluate new agents or combinations within predefined subtypes. Based upon response, assessed by MRI and other biological markers, participants predicted to achieve pCR will proceed to surgery to avoid additional toxic therapy. Those with suboptimal response will have the opportunity to be rerandomized to a targeted therapy. Second-line treatments in the second randomization will consist of best-in-class therapies for each subtype, including previous graduates from I-SPY2 and/or the initial randomization. Again, response to second-line treatment determines whether a participant is spared further toxic therapy by going directly to surgery or is rerandomized to a rescue agent. RCB, residual cancer burden; pRCB is predicted RCB based on MRI and biopsy; AC is doxorubicin/cyclophosphamide.

Close modal

It is essential that there be a regulatory path forward to make sure that we can get effective and affordable agents to the people most likely to benefit as quickly as possible, and minimize less effective or more toxic treatments. The ability to adapt treatment to response allows us to make clinical trials more patient-centered. There are hundreds of potential new therapies in development. What we need is an efficient system to prioritize the best treatments and eliminate suboptimal options. Fortunately, the FDA has exhibited strong leadership in advancing our regulatory system to meet modern demands.

It is critical that early endpoints be carefully developed, but it is equally important that our scientific advances include the ability to more rapidly measure whether drugs are working, so we increase the chance of success, with less toxicity. The scientific, regulatory, and clinical enterprise, and all of us who work in it, owe this to our patients.

L.J. Esserman is an unpaid board member of the Quantum Leap Healthcare Collaborative that is the sponsor of the I-SPY 2 TRIAL and reports receiving other remuneration from the Medical Advisory Panel for Blue Cross/Blue Shield. No other potential conflicts of interest were disclosed.

This article references the collaborative work of the entire I-SPY2 leadership team and the entire network of investigators. The design improvements to I-SPY2 were supported by a P-01 CA 210961 grant from the Clinical Trials Branch/Cancer Imaging Program of the NCI: “I-SPY2+: Evolving the I-SPY 2 TRIAL to include MRI directed, adaptive sequential treatment to optimize breast cancer outcomes.”

1.
Spring
LM
,
Fell
G
,
Arfe
A
,
Sharma
C
,
Greenup
RA
,
Reynolds
KL
, et al
Pathological complete response after neoadjuvant chemotherapy and impact on breast cancer recurrence and survival: a comprehensive meta-analysis
.
Clin Cancer Res
2020
;
26
:
2838
48
.
2.
Cardoso
F
,
van't Veer
LJ
,
Bogaerts
J
,
Slaets
L
,
Viale
G
,
Delaloge
S
, et al
70-gene signature as an aid to treatment decisions in early-stage breast cancer
.
N Engl J Med
2016
;
375
:
717
29
.
3.
Yee
D
,
DeMichele
A
,
Isaacs
C
,
Symmans
F
,
Yau
C
,
Albain
KS
, et al
Pathological complete response predicts event-free and distant disease-free survival in the I-SPY2 TRIAL [abstract]
.
In
:
Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5–9
;
San Antonio, TX. Philadelphia (PA)
:
AACR
; 
2018
.
Abstract nr GS3-08
.
4.
The F.D.A. is in trouble. Here's how to fix it
. 
2020
.
Available from
: https://www.nytimes.com/2020/01/11/opinion/sunday/fda-commissioner-stephen-hahn.html.
5.
Ray
KK
,
Stoekenbroek
RM
,
Kallend
D
,
Nishikido
T
,
Leiter
LA
,
Landmesser
U
, et al
Effect of 1 or 2 doses of inclisiran on low-density lipoprotein cholesterol levels: one-year follow-up of the ORION-1 randomized clinical trial
.
JAMA Cardiol
2019
;
4
:
1067
75
.
6.
Magbanua
MJM
,
Brown-Swigart
L
,
Hirst
G
,
Yau
C
,
Wolf
D
,
Wu
H-T
, et al
Personalized monitoring of circulating tumor DNA during neoadjuvant therapy in high-risk early stage breast cancer reflects response and risk of metastatic recurrence. [abstract
].
In
:
Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10–14
;
San Antonio, TX. Philadelphia (PA)
:
AACR
; 
2020
.
Abstract nr P5-01-04
.
7.
Yau
C
,
van der Noordaa
M
,
Wei
J
,
Osdoit
M
,
Reyal
F
,
Hamy
A-S
, et al
Residual cancer burden after neoadjuvant therapy and long-term survival outcomes in breast cancer: a multi-center pooled analysis. [abstract]
.
In
:
Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14
;
San Antonio, TX. Philadelphia (PA)
:
AACR
; 
2020
.
Abstract nr GS5-01
.
8.
Li
W
,
Onishi
N
,
Newitt
DC
, et al
Breast cancer subtype specific association of pCR with MRI assessed tumor volume progression during NAC in the I-SPY2 TRIAL
.
The 42nd Annual San Antonio Breast Cancer Symposium
,
San Antonio, USA
,
December 10-14
, 
2019
. https://cancerres.aacrjournals.org/content/80/4_Supplement/PD9-04.