ESTRO 2024 - Abstract Book

S500

Clinical - Breast

ESTRO 2024

Purpose/Objective:

In the randomized DBCG82bc cohort of high-risk breast cancer (BC) patients, high level of tumor-infiltrating lymphocytes (TILs) predicted improved overall survival (OS) after postoperative radiotherapy (RT) 1 . Further studies of the DBCG82bc cohort indicated that the positive association was mediated through superior distant tumor control and dependent on presence of CD8+, cytotoxic T-lymphocytes in the treatment-naïve tumor tissue 2 . In studies of neoadjuvant systemic treatment 3 , the prognostic effect of high TILs in terms of OS differs according to estrogen receptor (ER) status. Interestingly, the association between RT, TILs and OS benefit shown in the DBCG82bc cohort also appeared especially strong in ER-negative (ER-) tumors. However, patients in the DBCG82bc cohort were allocated to endocrine therapy according to menopausal status and not guided by ER-status. The aim of this study was to test, in irradiated BC patients treated according to modern standards (incl. endocrine therapy), if the association between TILs and OS is 1) affected by ER-status and 2) mediated through distant tumor control. We aimed to test these associations in patients with both high and low risk of local and/or distant recurrences.

Material/Methods:

Formalin-fixed, paraffin-embedded from treatment-naïve tumor tissue originating from routine diagnostic procedures was collected from 1329 patients, and Hematoxylin-eosin staining´s were performed.

The patients were part of a total of 2790 patients enrolled in two randomized Danish Breast Cancer Group (DBCG)- cohorts:

From the DBCG-HYPO cohort including relatively low-risk BC patients with T1a-T2, N0 disease, diagnosed 2009 2014, tumor tissue from 349/1329 Danish BC patients were included. They were treated with breast-conserving surgery (BCS) and randomized to normofractionated (50 Gray (Gy)/25 fractions (fr)) versus moderately hypofractionated (40 Gy/15 fr) RT. From the DBCG-IMN2 cohort including high-risk BC patients with node-positive disease, diagnosed 2007-2014, tumor tissue from 980/1461 patients (Central Region of Denmark only) were included. They were treated with BCS/mastectomy, axillary dissection, and loco-regional RT (irradiation of the internal mammary nodes (IMN) only in right-sided BC).

All patients had systemic therapy according to DBCG guidelines.

Stromal TILs were estimated by light microscopy following international guidelines 4 . Endpoints considered were loco-regional recurrence (LRR), distant metastasis (DM) and OS. A case-cohort design was applied where all cases (n=611) and a subset of censored and competing events (n=718) were evaluated for TILs. For the analysis using Kaplan-Meier and Aalen-Johansen estimators and univariate Cox regression analysis, non-cases were upweighted using inverse probability weighting. Due to the case-cohort design, numbers at risk in ER and TILs subgroups were not 100% identical for the three endpoints.

Results: