Higher Residual Cancer Burden Score Linked to Worse Event-Free Survival in TNBC

Updated data of the phase 3 KEYNOTE-522 trial (NCT03036488) demonstrated that an increased residual cancer burden (RCB) score correlated with worse event-free survival (EFS) in patients with triple-negative breast cancer (TNBC). However, patients who were given pembrolizumab (Keytruda) and chemotherapy saw an improvement in EFS vs those in the control arm who showed lower residual cancer burden (RCB) scores.¹

These findings were presented during the 2022 American Society of Clinical Oncology (ASCO) by Lajos Pusztai, MD, DPhil, and showed that patients with early TNBC resulted in significantly longer EFS when given neoadjuvant pembrolizumab plus chemotherapy than when administered neoadjuvant chemotherapy alone.

Patients enrolled into the double-blind study were randomized in a 2:1 ratio where patients with previously untreated stage 2 or 3 TNBC received neoadjuvant therapy with 4 cycles of pembrolizumab at a dose of 200 mg or placebo every 3 weeks in combination with paclitaxel and carboplatin, followed 4 cycles of pembrolizumab or placebo plus doxorubicin/cyclophosphamide or epirubicin/cyclophosphamide. Patients then received adjuvant pembrolizumab (or placebo after definitive surgery every 3 weeks for up to a total of 9 cycles.

According to Pusztai, professor of Medicine (Medical Oncology), and co-leader of Genetics, Genomics, and Epigenetics at Yale Cancer Center, data from prior analyses met the studies primary end point, showing that patients treated with pembrolizumab had pathologic complete responses (pCR) in the first 603 patients out of 1174 compared with chemotherapy alone where the pCR improvement was only approximately 13%.

Further, the second primary end point of EFS was also met during an interim analysis with 85% of patients given pembrolizumab having EFS events compared with 77% of those who received the placebo.

Among the total 1174 patients who were randomized in the trial, 784 were assigned to the pembrolizumab–chemotherapy group while 390 were assigned to the placebo–chemotherapy group. In this fourth interim analysis, the estimated event-free survival at 36 months was 84.5% (95% CI, 81.7-86.9) in the pembrolizumab plus chemotherapy group and 76.8% (95% CI, 72.2-80.7) in the placebo plus chemotherapy group (HR, 0.63; 95% CI, 0.48-0.82; P <0.001).

As for safety, the adverse events that occurred were mostly within the neoadjuvant phase of the study. Safety data were consistent with the established safety profiles of pembrolizumab, and chemotherapy previously reported in studies.

In an interview with Targeted Oncology^TM, Pusztai discussed the updated findings of the phase 3 KEYNOTE-522 clinical trial of pembrolizumab plus chemotherapy in patients with TNBC vs placebo plus chemotherapy.

Targeted Oncology: Can you give a brief overview of the presentation you gave at ASCO 22 on the findings of KEYNOTE-522?

Pusztai: What I presented at the ASCO was the updated results now including pathological complete response rates and the distribution of residual cancer burden in the entire study population. It also showed the event-free survival by residual disease categories.

KEYNOTE-522 is the largest neoadjuvant/adjuvant trial via immune checkpoint inhibitor in triple-negative breast cancer. The trial design included approximately 20 to 24 weeks of neoadjuvant chemotherapy with weekly taxol followed by regular adriamycin/ cyclophosphamide, or epirubicin/cyclophosphamide chemotherapy, either without or with every 3-week pembrolizumab.

After surgery, patients continued with pembrolizumab or a placebo. This was a double-blind, placebo-controlled, phase 3, randomized trial that included 1174 patients. The primary endpoints included both pathologic complete response rate during the neoadjuvant phase and event-free survival in the combined neoadjuvant followed by adjuvant phases of the study. And the preplanned first primary outcome analysis was done after the first 602 patients had pathological complete response.

The study has met its primary endpoint. For this outcome, it showed a pathological complete response rate improvement of approximately 13% compared to chemotherapy alone. And subsequently, in the fourth interim analysis, which was again a preplanned statistical analysis, the second primary end point [of] event-free survival was also met. Overall, 85% or close to 85% of the patients in the pembrolizumab arm were event free compared to 77% in the placebo arm. Again, this represents about 8% improvement in event-free survival.

What were some of the methods used in the study?

We developed a method to quantify the extent of residual cancer, so if you have no residual cancer at all or no viable cancer at the completion of new adjuvant chemotherapy, then we call this a pathologic complete response, or RCB, residual cancer burden 0. But there is a huge range of viable residual cancer in a subset of patients. We developed a method that combines the maximum size of the cancer, the cellularity of the cancer, the number of lymph nodes, which are involved, and the size of the lymph node involvement into a single core, the core score that we call RCB, and this score could be used to categorize patients into RCB0, 1, 2, or 3. Then as the score increases, the extent of the residual disease increases and the expected survival or prognosis of the patient gets worse and worse.

How did the trial come to fruition?

The observation that a lot of immune cells in the microenvironment of breast cancer, particularly triple-negative breast cancer, is associated with good prognosis. It's been known for probably at least 20 years. But it was not clear whether this is some sort of an association between good biology and having lymphocytes around or if it's some sort of cause and effect relationship, that it's the lymphocytes, which control the disease and result in a better outcome by suppressing metastasis and recurrence.

The advent of immunotherapy drugs being highly effective immune checkpoint inhibitors enabled us to ask this question. If the immune response contributes to the good prognosis, then boosting the immune system will improve the prognosis of patients with a moderate amount or even large number of immune cells, you can make that progress better. This was the theoretical background to testing immune checkpoint inhibitors in breast cancer, and lo and behold, it's worked.

There are at least 4 randomized neoadjuvant chemotherapy trials that tested different types of immune checkpoint inhibitors combined with chemotherapy in the preoperative setting. Essentially, all of them showed an improvement in pathologic complete response rate, except 1 study in Italy, the NeoTRIP study [NCT002620280] of ​​atezolizumab [Tecentriq]. We also learned that the pathologic complete response rate improvements are relatively modest and often in single digits. In large studies, these reached significance. In smaller phase 2 studies, they did not have significance, just the numerical improvement.

Even a relatively small phase 2 trial, which tested durvalumab [Imfinzi] under a similar sort of schema with or without development, showed a slight improvement in pathologic complete response rate and a significant improvement in event free survival. I think the data is very consistent that immune checkpoint inhibitors improve the outcome of patients with triple negative disease. KEYNOTE-522 is the largest and most definitive study.

What did the design of the study entail?

KEYNOTE-522 was a randomized, double-blind placebo-controlled trial for stage 2 and stage 3 triple-negative breast cancer. Patients had to have a 2 cm or greater invasive cancer, or they could have a smaller tumor, but the lymph node involvement had to be noted as positive.

What data were you able to conclude from the study?

What we reported at ASCO was that in the final analysis, including all patients, the pathological cognitive response rate difference decreased from 13.6%-7.2%. That's still significant, but we also noticed that there was a consistent shift to lower amounts of residual cancer across the entire spectrum. There were fewer patients in the RCB1 group in the pembrolizumab cohort than in the control cohort because many of them moved into the RCB0. There were fewer RCB2 patients in the pembrolizumab arm because many of them moved to the RCB1 category. There were fewer RCB3’s, which is the poorest prognosis and largest amount of residual cancer, and again, there were fewer of them in the pembrolizumab arm, approximately 5% of patients ended up with still this poor prognosis.

What do you consider to be the most important takeaway from this research?

The most important finding from my presentation is that the benefit from pembrolizumab extended beyond just increasing the pathological complete response rates. Many patients who did not achieve pathologic COVID response also benefited from pembrolizumab. We see this by plotting their event-free survivals by the various residual disease categories for those who had less than a COVID response. The largest clearly significant benefit was seen among those who are in the RCB2 category, which happens to be the largest market subpopulation among those who do not achieve a COVID response. That was approximately 20% of the entire study population on the pembrolizumab arm.

The chemotherapy alone arm combined had this RCB2 extent of residual disease. If they were randomized to the pembrolizumab arm, the event-free survival at 3 years was 75% compared to only 56% in those who received chemotherapy with the placebo. That's almost a 20% improvement in event free survival. This means that we avoided it in 20% of the patients, either local recurrence, distant recurrence, or progression during the new adjuvant phase or death. In fact, the events that pembrolizumab prevented most were distant metastatic recurrences.

REFERENCE:

1. Pusztai L, Denkert C, O'Shaughnessy J, et al. Event-free survival by residual cancer burden after neoadjuvant pembrolizumab + chemotherapy versus placebo + chemotherapy for early TNBC: Exploratory analysis from KEYNOTE-522. J Clin Oncol. 2022; 40 (suppl 16). doi: 10.1200/JCO.2022.40.16_suppl.503