Lajos Pusztai, MD, DPhil, discusses the results from the I-SPY 2 trial, as well as the clinical implications of the findings.
Lajos Pusztai, MD, DPhil, professor of medicine medical oncology and co-leader of Genetics, Genomics, and Epigenetics at Yale Cancer Center
Lajos Pusztai, MD, DPhil
The triplet regimen of durvalumab (Imfinzi), olaparib (Lynparza), and taxane-based chemotherapy demonstrated promising efficacy as neoadjuvant treatment in patients with early-stage, high-risk, HER2-negative stage II/III breast cancer in the phase 2 I-SPY 2 trial, explained Lajos Pusztai, MD, DPhil.
Findings from the trial, which were presented at the 2020 AACR Annual Virtual Meeting I, revealed that the predicted probability of a patient with HER2-negative breast cancer achieving a pCR was 37% (95% CI, 0.27-0.47) with the combination compared with 20% (95% CI, 0.16-0.25) with paclitaxel alone.
Upon subset analysis, patients with hormone receptor (HR)—positive/HER2-negative breast cancer had a respective pCR rate of 28% (95% CI, 0.18-0.38) with the combination versus 14% (95% CI, 0.09-0.19) with chemotherapy alone. Individuals with triple-negative breast cancer (TNBC) had pCR rates of 47% (95% CI, 0.29-0.64) and 27% (95% CI, 0.20-0.34), respectively.
"[These results] add to the growing amount evidence that immunotherapy improves pCR rates in early stage disease," said Pusztai, who is the lead author of the study. "This study represents an important building block of evidence showing that immunotherapy improves pCR rate in early-stage disease in TNBC. Additionally, this is the first time we have seen an improvement in pCR rates in HR-positive patients since the introduction of taxane chemotherapies. We [are aiming to make] chemotherapy more effective for these patients."
In an interview with OncLive, Pusztai, a professor of medicine, director of Breast Cancer Translational Research, and co-leader of the Cancer Center Genomics, Genetics, and Epigenetics Program at Yale Cancer Center, discussed the results from the I-SPY 2 trial, as well as the clinical implications of the findings.
OncLive: Could you speak to the data that have read out with durvalumab and olaparib in breast cancer prior to this trial?
Pusztai: Prior to the study, durvalumab was looked at in 2 other phase 1/2 clinical trials. One of those trials was randomized phase 2 trial investigating whether durvalumab could improve pCR rates in the neoadjuvant setting TNBC.
Most of these studies were intriguing but did not yield conclusive results. The single-arm phase 1/2 study had a pCR rate of 44%, which is a little bit higher than the regimen produced in another similar patient population. In those patients, the pCR rate was around 30%. However, it is in the range that chemotherapy could induce on its own.
The other randomized trial was from Germany and it showed a numerically higher pCR rate, which did not reach statistical significance. This study was small and included about 140 patients.
The I-SPY 2 trial was designed with these 2 studies in the background. The goal was to see whether including a third agent, olaparib, in the combination would improve pCR rates.
I was very pleased to announce these results [at the 2020 AACR Virtual Annual Meeting I] because the pCR rate did improve. In fact, not only did the pCR rate improve, but the entire spectrum of residual disease that we observed at the same time of surgery shifted to a lower value. This means that all patients ended up with less residual cancer than on average with chemotherapy alone.
What makes this study especially exciting is the [improvement] that we saw in the cohort of HR-positive patients. We have a good idea of what subset of patients benefitted from adding these 2 agents to chemotherapy.
What was the rationale for combining immunotherapy with PARP inhibition? Could you shed light on the clinical experience in the MEDIOLA trial?
The preclinical question of combining these 2 classes of agents is quite compelling. There are a number of mechanisms that the PARP inhibitors induce in various cell-line models and preclinical animal models that [would make the tumors responsive to] immunotherapy. This includes increasing the neoantigen load and the overall mutation load [through interferon-gamma], which is 1 of the fundamental mechanisms of action of PARP inhibitors.
The treatment itself increases PD-L1 expression to the cells. It also increases the DNA fragments that are generated by the DNA repair, which can induce the STING pathway. STING is a vital response pathway, which also makes the cells more visible and it activates the local immune system.
These are the 3 major lines of evidence from a series of studies that combining PARP inhibitors with immunotherapy [is effective].
What was the goal of the I-SPY 2 trial? How was the study designed, and what are the requirements for graduation?
The goal of the I-SPY trial was to see whether the pCR rates would improve when durvalumab and olaparib were added to chemotherapy. We used the standard dose of paclitaxel at 80 mg/m2, which was supported by previous trials, which showed that combining PARP inhibitors with chemotherapy is not an easy feat because of the excessive hematologic toxicities that require dose-reduction of olaparib.
Durvalumab was also given at a standard dose. The statistical design and the concept behind I-SPY 2 is rather simple. As the data accumulate, they imply the pCR rates that patients got to in 24 weeks of neoadjuvant therapy. Also, a patient's MRI response predicts whether they will be able to accomplish a pCR, as well as the status of residual disease.
We use the MRI information as it accumulates at weeks 3 and 12, along with the increasing number of patients who have completed surgery, to calculate where the studies are going and the probability that this study would succeed in a phase 3, randomized clinical trial.
If this calculated threshold drops below 10%, then it becomes clear that this is a futile exercise. I believe that 1 or 2 arms [in the I-SPY trial program] have dropped during its 10-year lifespan.
Graduation means that, at that point, accrual stops. What is sometimes confusing, from the presentations that have occurred, is that the final result is going to be different [than what has already been presented] because several patients are still receiving treatment at the time that the graduation occurs. The final results are always slightly different.
What were the results of this analysis?
In the final results, the predicted probability of success of the experimental arm [in a 300-patient randomized trial] was greater than 99% for each of the 3 subsets that were looked at. The predicted probability of the study being successful in a similar outcome beside this patient study dropped to 82% or 83%. It's implied in the statistical design that the final results will be slightly different from the predicted results at the time of graduation.
An interesting finding from this study is that in the early-stage setting, the amount of PD-L1 [mRNA] expression seemed to define the population that benefits [more] from immunotherapy.
The other important finding is that we consistently see in the I-SPY trial, [and in other studies with checkpoint inhibitors], is that HR-positive patients [particularly have improved pCR rates with the addition of checkpoint inhibitors].
What safety signals were found with this combination? Were there any immune-related toxicities of special interest observed?
The most common adverse effects (AEs) are similar to what we would expect in a chemotherapy combination, which are mainly hematologic events. We observed a higher rate of grade 3/4 hematological toxicities. This was not unexpected in the setting of chemotherapy, even with a backbone that is relatively non-myelosuppressive, such as paclitaxel.
The other AEs were gastrointestinal related. Regarding immunological toxicities, we observed significantly more immunological AEs due to the addition of durvalumab. The most common toxicity was kinase function abnormalities. We also saw serious grade 3/4 events of colitis.
Though these events are individually rare, overall hematologic AEs occurred in about 20% of patients.
Could you elaborate on the exploratory analysis, which suggested that potential predictive markers of durvalumab and olaparib benefit over chemotherapy alone?
That is one of the most important biomarker questions. Who are the patients who need immunotherapy on top of another drug and chemotherapy?
There is a lot of hope that [the presence of select] immune markers [benefit from the combination]. Having immune cells in the microenvironment of early-stage disease is a good thing. It predicts for better prognosis regardless of whatever treatment patients received. It also predicts for higher chemotherapy sensitivity and higher response rates with immune checkpoint inhibitors when they are added.
Surprisingly, we found 3 tentative markers in the HR-positive population and the TNBC population by MammaPrint ultra-high group. [Patients with an ultra-high MammaPrint score had the most benefit from the combination]. In TNBC, [low CD3/CD8 gene signature ratio, high macrophage/Tc-Class 2 ratio, and high proliferation signatures were linked with a higher pCR with in the combination].
What are the next steps for this combination?
The next steps are to validate these markers in collaboration with pharmaceutical companies as more data in the randomized neoadjuvant space arrives.
Pusztai L, Han HS, Yau C, et al. Durvalumab in combination with olaparib and paclitaxel in high-risk HER2 negative stage II/III breast cancer: results from the I-SPY 2 trial. Presented at: 2020 AACR Annual Virtual Meeting I; April 27-28, 2020. Abstract CT011.