Model predicts risk for immune-related adverse events in breast cancer subset

SAN ANTONIO — A symptom pattern model predicted risk for immune-related adverse events among women with early-stage, high-risk breast cancer, according to an analysis of the phase 2 I-SPY2 trial presented at San Antonio Breast Cancer Symposium.

The findings could help clinicians be more diligent with monitoring during active treatment and after therapy, researchers concluded.

Background and methods

“Immune checkpoint inhibitors were first introduced in the metastatic breast cancer setting with improved outcome in PD-L1 positive disease. The checkpoint inhibitor pembrolizumab [Keytruda, Merck] is now approved as standard neoadjuvant therapy for high-risk early-stage triple-negative breast cancer, with improvements in both response and event free survival,” Amrita Basu, PhD, assistant professor in the department of urgery at University of California, San Francisco, said during a presentation.

However, these immune checkpoint inhibitors carry risk for immune-related adverse events, some of which are irreversible, Basu added.

“Immune-related adverse events can affect any organ in the body,” she said. “Checkpoint inhibitors given with chemotherapy in the neoadjuvant setting can be associated with multiple adverse events, including rash, thyroid dysfunction and adrenal insufficiency. Combination therapy results in many early symptoms, which may be harbingers of an immune-related adverse event.”

Basu and colleagues developed a standardized method to predict Immune-related adverse events to help patients and providers with treatment decision-making, balancing efficacy with risk for toxicity.

“To do this, we are incorporating side effect burden, duration of side effect and persistence, and quality of life,” she said. “We want patients to achieve excellent efficacy rates and maintain a good quality of life.”

Researchers sought to identify patient demographic characteristics and symptom patterns associated with risk for immune-related adverse events among 482 patients included in the I-SPY2 trial. These patients received at least four doses of immunotherapy in combination with chemotherapy.

Investigators also established a patient-reported outcome cohort (n = 346) in which they used surveys to assess symptoms and quality of life through 24 months.

Researchers compared patients who developed hypothyroidism, adrenal insufficiency, pneumonitis and/or colitis vs. those who did not.

Basu and colleagues used a chi-square test to assess associations between race/ethnicity and immune-related adverse events, and a one-way ANOVA to examine the association between age (>50 years vs. < 50 years) and immune-related adverse events.

Researchers used area under curve to calculate symptom burden score that combined duration of 33 grade 2 or higher symptoms between baseline and week 6 of treatment and grade of adverse event. They used regularized regression using leave-one-out cross-validation to assess early symptoms as predictors and immune-related adverse events as surrogate responses.

Findings

Hypothyroidism represented the most common immune-related adverse event (12%), followed by adrenal insufficiency (8%), pneumonitis (4%) and colitis (1%).

Symptoms associated with development of hypothyroidism included fatigue (15%; mean area under the curve = 11.8 vs. 5.8 among those who did not develop immune-related adverse events), shortness of breath (11%; area under the curve = 4.3 vs. 2.8) and blurry vision (1%; area under curve = 1 vs. 0.12).

Development of adrenal insufficiency appeared linked to early reports of diarrhea (36%; area under the curve = 19 vs. 10.5), shortness of breath (11%; area under the curve = 7.8 vs. 2.6), joint pain (3%; area under the curve = 2.29 vs. 0.58), decreased appetite (3%; area under the curve = 3.55 vs. 0.91) and constipation (1%; area under the curve = 3.6 vs. 0.02).

Researchers observed no significant early symptoms associated with pneumonitis due to a limited number of events.

Common symptoms reported up to week 6 of treatment among those who developed an immune-related adverse event included diarrhea (36%), fatigue (15%), dizziness (12%) and shortness of breath (11%).

“When we removed these symptoms from our model, we found that the model significantly underperformed,” Basu said. “Thus, further predictive understanding of the constellation of these symptoms will help us monitor patients more closely, intervene earlier and may help prevent full-blown disease from occurring.”

Researchers had demographic data for 333 patients, most of whom were white (81%) followed by

Among the 333 patients with available demographic data (81% white, 11% Black, 7% Asian). More white patients (35 of 265) developed hypothyroidism than nonwhite patients (2 of 63; P < .04).

Patients aged older than 50 years appeared more likely than those aged younger than 50 years to develop pneumonitis (P <.01).

Implications

The findings suggest early onset of symptoms may predict subsequent risk for immune-related adverse events, Basu concluded.

“Understanding the risk factors for developing an immune-related adverse event will help to optimize surveillance and potential treatment modification to minimize the impact of toxicity,” she said. “Further confirmation of this model is required for building more interpretable classification algorithms. We are currently analyzing patient reported outcomes and we received a grant to analyze genetic predictors to identify who is at risk for developing severe immune-related adverse events.”