Treatment efficacy score may better predict long-term survival in breast cancer trials

Researchers led by a Yale University team have developed a new measure for neoadjuvant chemotherapy efficacy for use in preoperative breast cancer trials.

The new metric, known as the Treatment Efficacy Score (TES), can gauge the relative effectiveness of an experimental treatment over standard of care by assessing the distribution of the continuous residual cancer burden score — a precise measure of invasive cancer after neoadjuvant chemotherapy calculated by merging tumor size, cancer cellularity and lymph node involvement into one continuous score.

Before TES, the efficacy of two different treatments was measured by the difference in pathologic complete response (pCR) rate between the two arms of a randomized clinical trial. However, this measure does not account for the downstaging effect of treatment on cancers with less than complete response. Furthermore differences in pCR rate between treatments did not always translate into differences in survival.

“There were studies that showed a large improvement in pCR, but they didn’t show any improvement in survival,” Lajos Pusztai, MD, DPhil, scientific co-director of Center for Breast Cancer at Yale Cancer Center who led the study, told Healio. “So, there was often a disconnect — the pCR improved with a given treatment, but the long-term survival and recurrence-free survival didn’t improve significantly.”

Pusztai spoke with Healio about the potential of TES to be a valuable addition to clinical trial measures of neoadjuvant chemotherapy efficacy.

Healio: How has neoadjuvant chemotherapy efficacy traditionally been measured?

Pusztai: The original outcome measure for preoperative chemotherapy was pCR, which is defined as no cancer in the breast and lymph nodes after chemotherapy when a pathologist examines the surgically resected tissues. When someone attains this outcome, they have a very good chance of long-term survival without recurrence. Those who have residual cancer, which is any invasive cancer that survives the initial chemotherapy, generally don’t do so well. The greater the amount of cancer after chemotherapy, the greater the risk for recurrence.

Ultimately, the goal of all systemic treatments in early-stage breast cancer, including preoperative chemotherapy, is not so much to shrink the cancer or make it disappear in the breast, but to eliminate micrometastatic disease that may be hiding in distant organs. We assume micrometastases that may be hiding in the lung, liver or bone have the same chemotherapy sensitivity as the primary tumor, and we use the response to preoperative treatment in the breast as a readout of cancer sensitivity. If a lot of cancer survived the preoperative chemotherapy in the breast or lymph nodes, we unfortunately can assume that micrometastatic cancer also survived outside the breast.

Achieving pCR is very good news for a patient and occurs among 30% to 60% of patients, depending on the type of chemotherapy used and type of breast cancer. The rest have some degree of residual cancer.

More than a decade ago, a breast pathologist colleague, Fraser Symmans, MD, and I designed a very precise measure of this residual cancer that survived the chemotherapy. We called it residual cancer burden. It’s a number that the pathologist can calculate based on the size of the cancer in the breast, the cellularity of the cancer (ie, what percentage of cells in the cancer tissue are actually cancer as opposed to supportive cells), the number of lymph nodes involved with cancer, and the size of the cancer in the lymph node.

Because a patient who achieves pCR has a very good outcome, clinical trials started to compare different therapies in randomized trials to see which produced a higher pCR rate. The pCR rate difference between trial arms has became an established measure of drug efficacy. The higher the pCR rate, the better the drug.

The first drug the FDA approved based on improved pCR rate was pertuzumab (Perjeta, Genentech) for HER2-positive breast cancer in 2012. However, in the past 10 years we have begun to learn that there is a problem with using pCR rate as early surrogate for long-term survival in a trial. Several randomized trials showed treatment regimens that produced impressive double-digit improvements in pCR rate did not always result in similar improvements in recurrence-free survival.

More recently, we saw the opposite relationship in the KEYNOTE-522 trial of preoperative pembrolizumab plus chemotherapy. In that study, final pCR rate improved only a modest 7%, but this translated into a very significant clinically meaningful improvement in recurrence-free survival. This has led to FDA approval of the drug for stage II and stage III triple-negative breast cancer.

Healio: Does this call into question the value of pCR as a clinical trial metric?

Pusztai: Yes. It is clear that it’s not an ideal metric for predicting the impact of a drug on long-term survival. The conundrum is that pCR is a very good prognostic indicator at the patient level, but the pCR rate difference between trial arms is not a precise predictor if the trial arm with the higher pCR rate will end up also showing a lower recurrence rate.

That’s where the TES comes in.

We hypothesized that it’s the distribution of the entire spectrum of residual disease that really matters across a patient population. We already knew from decades of experience that patients with minimal residual cancer burden have outcomes almost as good as patients with pCR, and the smaller the residual cancer burden, the better the prognosis. So, if a drug improves pCR rate by moving patients from the minimal residual disease category to the pCR category but does not reduce the proportion of patients with extensive residual disease, it will not affect survival substantially. In contrast, a treatment that shifts residual cancer burden to smaller values across the entire spectrum may increase pCR rate only modestly but could improve survival quite substantially, because it is better to have 1-cm residual disease than 2-cm, and 2-cm disease is better than 3-cm. So, our idea is to develop a measure that compares the residual cancer burden distribution and captures the shift to lower values. TES is a complex statistical tool to compare two residual cancer burden distributions that we have developed in collaboration with the I-SPY trial investigators using data from more than 1,000 patients who received 10 different types of preoperative chemotherapies.

Healio: What are your plans for this emerging tool?

Pusztai: We made this tool public through a free website so clinical trial investigators could start to independently validate it and improve on it if needed. We hope TES will become an additional way to measure the efficacy of a drug that is tested in a preoperative clinical trial. Our initial validation in I-SPY trial data suggested that TES is a better measure of long-term survival benefit from a new therapy than pCR rate difference. We’re going prospectively to include TES as an additional endpoint measure in the ongoing I-SPY trial. It will be used in conjunction with pCR to complement it. We hope that other investigators will adopt TES as an outcome measure and will independently test its value to identify early a treatment arm that has longer survival with long-term follow up.

HealioIs there anything else you’d like to mention on this topic?

Pusztai: I would like to emphasize that pCR and the residual cancer burden score are patient-level outcome measures, but TES is clinical trial arm-level outcome measure. We calculate TES by summing up all the individual residual cancer burden scores observed in a trial arm.

For more information:

Lajos Pusztai, MD, DPhil, can be reached at Yale Medical Oncology, 300 George St., 1st Floor, New Haven, CT 06520-8028; email: lajos.pusztai@yale.edu.