Long-Term Data Add to Prognostic Value of TILs in Treatment De-Escalation in Early HER2+ Breast Cancer

Maria Vittoria Dieci, MD

Although tumor-infiltrating lymphocytes (TILs) should not be the only biomarker utilized when determining if a patient with early HER2-positive breast cancer is a candidate for treatment de-escalation, updated data from the phase 3 ShortHER trial (NCT00629278) add to the evidence supporting TILs as a prognostic marker for this patient population, according to data presented by Maria Vittoria Dieci, MD.

Updated data from ShortHER presented at the 2024 ESMO Congress showed that at a median follow-up of 9.02 years, a 1% increase in TILs reduced the risk of a distant disease-free survival (DDFS) event by 3% (HR, 0.97; 95% CI, 0.96-0.99; P = .001) and reduced the risk of death by 2% (HR, 0.98; 95% CI, 0.96-1.00; P = .016).

“The results reinforce the prognostic value of TILs in HER2-positive early breast cancer,” said Dieci, who is an associate professor at the University of Padua in Italy. “We now have evidence that TILs are not only linked to DFS but also to overall survival [OS], which is the most important end point for patients treated with adjuvant therapy."

In an interview with OncLive^®, Dieci discussed the implications of these long-term findings and emphasized the need to explore combining TILs with other prognostic biomarkers to create comprehensive tools for the identification of patients with early-stage HER2-positive breast cancer who may be candidates for treatment de-escalation.

OncLive: What findings have been previously reported from ShortHER?

Dieci: For patients with early HER2-positive breast cancer, there is an ongoing need to identify prognostic biomarkers that can help guide treatment de-escalation. We had the opportunity to test this option in the context of the ShortHER trial.

The ShortHER trial is a phase 3, randomized trial that tested the non-inferiority of 9 weeks of adjuvant trastuzumab [short arm] compared with the standard 1 year of adjuvant trastuzumab combined with chemotherapy [long arm] for patients with early HER2-positive breast cancer.

This trial enrolled 1253 patients. Although the main results of the trial in terms of disease-free survival did not allow [investigators] to claim the non-inferiority of the short arm vs the long arm, we recently published the updated 10-year OS analysis that showed nearly superimposable outcomes for the 2 treatment arms.

How have TILs been used as a biomarker within the context of the ShortHER trial?

[We were interested to see] whether there was a biomarker that could guide our choice for treatment de-escalation. TILs were assessed in a cohort of 866 patients from the ShortHER trial.

[In 2019], we published in Annals of Oncology a first analysis of the impact of TILs in the context of this trial. We were able to demonstrate that at the median follow-up of 6.1 years that patients with high TILs—defined as a cutoff of [at least] 20%—experienced significantly better distant disease-free survival compared with patients with low TILs.

Moreover, we also found a significant interaction between TILs and treatment arm. Specifically, for patients with low TILs, there was a significant benefit when they were treated in the long arm compared with the short arm. For patients with high TILs, the outcome was good, irrespective of treatment arm. There was even a numerical benefit [for patients with high TILs] when they were treated with the short treatment regimen compared with the long treatment regimen.

What were the key findings from the updated TILs analysis?

At the 2024 ESMO Congress, we had the opportunity to discuss and present the updated analysis at a median follow-up of 9.02 years of the impact of TILs in the ShortHER trial. The follow-up was 3 years longer compared with the previously published data.

This allowed us to consider [an additional] survival end point. DDFS was updated with 107 events. We also studied OS, which is an important clinical end point.

We tested the association between TILs and these survival end points. First, we demonstrated that TIL levels were associated with both DDFS and OS. Specifically, for each 1% increase in TILs, there was a 3% reduction in the risk of a DDFS event [HR, 0.97; 95% CI, 0.96-0.99; P = .001] and a 2% reduction in the risk of death [HR, 0.98; 95% CI, 0.96-1.00; P = .016], and this was highly statistically significant in multivariable models.

Moreover, we stratified patients according to TIL levels, and we explored survival curves for patients with high and low TILs, according to different cutoffs: 20%, 30%, and 50%. The outcome of patients with high TILs [were consistently] better [for DDFS and OS] compared with patients with low TILs for each TIL cutoff. To give you an example, for patients with high TILs with a cutoff of at least 50%, the 10-year OS rate was 98.1% [compared with 88.8% for those with TILs below 50%].

Finally, we also confirmed at this longer follow-up that there is an interaction between TILs and treatment arm for DDFS. For patients with low TILs, there was an evident benefit for the long treatment compared with the short treatment, whereas for patients with high TILs, there was a numerical benefit when they were treated with the de-escalated regimen compared with the long standard therapy, and this was statistically significant.

When we repeated the same analysis, considering OS survival as the main survival end point, the results pointed toward the same direction. There was numerically better OS for patients with low TILs when they received the long treatment compared with the short [treatment], and there was a numerically better OS for patients with high TILs receiving the short treatment compared with the long treatment, and the interaction was borderline [statistically] significant.

What are the implications of these long-term follow-up data?

It is important to highlight that this is the first demonstration that TILs are prognostic in terms of OS for patients with early HER2-positive breast cancer in the context of a clinical trial where all patients received adjuvant chemotherapy and anti-HER2 therapy. [These data] reinforce the prognostic role of TILs. We already knew that TILs were associated with improved DFS, but there was no evidence thus far from a single trial that this biomarker was associated with improved OS, which is the most important survival end point for our patients in the context of patients treated with primary surgery and then adjuvant therapy.

This longer follow-up, these data also reinforce the concept that patients with high TILs who receive the de-escalated treatment—in this specific case, a duration reduction of trastuzumab and a de-escalation in chemotherapy—were not exposed to an increased risk of distant relapse or death.

Our data reinforce the concept that TILs may be an important biomarker to keep in mind when proposing a treatment de-escalation to our patients [with early-stage HER2-positive breast cancer].

With our results, we are not claiming that TILs should be the only biomarker to take into consideration when proposing de-escalated treatment and when identifying patients more suitable for treatment de-escalation. The best way to move the field forward is to develop models and certain tools that integrate all the prognostic and potentially predictive biomarkers in one algorithm. This will be the most helpful way to identify patients for treatment de-escalation. For example, another variable that remains statistically prognostic in our cohort was disease stage. I would say that at [a minimum], TILs and stage should be combined.

There is also a genomic tool that is being tested for its clinical utility in this specific context. It integrates clinical pathological characteristics and gene expression data, including immune-related biomarkers. This is being tested as a tool to help treatment de-escalation [decisions].

Reference

Dieci MV, Bisagni G, Bartolini S, et al. TILs and overall survival (OS) in HER2+ early breast cancer (eBC): 10-year (yr) updated analysis of the ShortHER trial. Ann Oncol. 2024;35(suppl 2):S309-S348. doi:10.1016/annonc/annonc1577