Characterisation and impact of intratumoural stroma in melanoma and carcinoma brain metastases

Abstract

Research on the tumour microenvironment in brain metastases (BM) has predominantly focused on the immune response, while the presence, morphological patterns, and potential clinical relevance of intratumoural stroma remain less intensively investigated. We retrospectively analysed 604 BM (529 carcinomas, 75 melanomas) from 556 patients. Intratumoural stroma was histomorphologically classified into absent/unclear (Group 0), present without desmoplasia (Group 1) or with desmoplasia (Group 2). Associations with histological features, clinical parameters, and survival were evaluated. Intratumoural stroma was absent in 63.2% of tumours (n = 382), was present without desmoplasia in 14.9% (n = 90), and was desmoplastic in 21.9% (n = 132). Desmoplasia was most frequent in metastases from breast carcinomas and pulmonary squamous cell carcinomas. No significant associations were found between stroma groups and age, sex, brain location, PD-L1 expression, oncogenic mutations in lung carcinoma, or breast-cancer molecular subtype. Independent predictors of poorer survival were increasing age (HR = 1.029, 95% CI: 1.018-1.039, p < 0.001), male sex (HR = 1.492, 95% CI: 1.204-1.849, p < 0.001), and infratentorial location (HR = 1.402, 95% CI: 1.125-1.748, p = 0.003). Stroma groups showed no independent prognostic value in the overall cohort. However, subgroup analysis of non-small cell lung cancer revealed a U-shaped relationship, with Group 1 stroma linked to better survival (p = 0.020). In summary, intratumoural stroma in BM is a carcinoma-specific phenomenon, absent in melanoma. Patient outcomes, however, were primarily determined by demographic and anatomical factors rather than stromal morphology in the overall cohort but may have clinical relevance in particular tumour subgroups.

Keywords: brain metastasis; desmoplasia; histomorphology; intratumoural stroma; survival analysis; tumour microenvironment.

Figure 1

(A–C) Representative examples of three BM (×100 total magnification), assigned to stroma Group 0. These cases show only narrow perivascular connective tissue without the formation of an intratumoural stroma. (A) Clear cell renal cell carcinoma, typically forming small solid nests separated by a delicate capillary‐rich stroma. (B) Breast adenocarcinoma with diffuse infiltration of small tumour cells into the adjacent temporal lobe parenchyma without any intratumoural stroma. (C) Urothelial carcinoma with a predominantly solid growth pattern. (D–F) Representative examples of three BM (×100 magnification) exhibiting a recognisable fibrous stroma (highlighted by arrows) that, while clearly developed, does not appear desmoplastic. This group 1 stroma is typically associated with papillary growth patterns. (D and E) Lung adenocarcinomas. (F) Gastric adenocarcinoma.

Figure 2

Representative examples of six BM (×100 total magnification) exhibiting intratumoural desmoplastic stroma (IDS), corresponding to stroma Group 2. Typically grey‐appearing areas in the broad stroma are highlighted by arrows. (A and B) Breast carcinomas. (C) Colorectal adenocarcinoma with tubular growth pattern. (D) Pulmonary squamous cell carcinomas. (E and F) Pulmonary adenocarcinomas.

Figure 3

Kaplan–Meier survival curves of 501 patients, stratified by (A) sex, (B) age, (C) brain location, (D) stroma groups (overall), (E) primary organ of metastases, and (F) stroma groups within non‐small cell lung cancer (NSCLC). Overall survival in months was defined as the interval between the date of surgical resection and the date of death. Cum survival, cumulative survival.