How does hormone receptor status switch relate to leptomeningeal disease in breast cancer?

Tracking receptor status across the metastatic timeline

The paper asks whether hormone receptor status, and changes in that status, relate to how breast cancer spreads to the central nervous system and specifically to leptomeningeal disease. Leptomeningeal disease means cancer cells infiltrate the leptomeninges covering the brain and spine; the paper describes it as the most aggressive form of central nervous system metastasis, with rapid progression and few effective treatments.

The evidence comes from a retrospective chart review of 141 female patients with leptomeningeal disease from invasive breast cancer treated at Stanford Hospital. Receptor status was recorded at primary tumor diagnosis and, when available, at local recurrence, non-central-nervous-system metastases, brain metastases and leptomeningeal disease. Estrogen and progesterone receptor positivity was defined as at least 1% of tumor nuclei showing immunoreactivity, and HER2 positivity followed the criteria in use at the time of testing.

This design matters because most prior work on receptor status and leptomeningeal disease relied on the primary tumor alone. Here, receptor status at recurrence and distant metastases was entered as a time-dependent covariate, so the analysis could reflect status closer to the moment of central nervous system spread rather than only the original biopsy.

Negative hormone receptors mean earlier leptomeningeal disease

At primary diagnosis, estrogen-receptor-negative patients reached leptomeningeal disease a mean of 32.1 months sooner than estrogen-receptor-positive patients (95% CI 10.6 to 53.7 months; p = 0.004). Progesterone-receptor-negative patients reached it a mean of 31.2 months sooner than progesterone-receptor-positive patients (95% CI 10.6 to 53.7 months; p = 0.003). These estimates come from linear regression, used because every patient in the cohort eventually developed leptomeningeal disease, so the interval to that event was observed for all.

Survival after the leptomeningeal diagnosis moved in the same direction. In Cox proportional hazards models, baseline estrogen-negative and progesterone-negative status each carried an estimated hazard ratio of 1.7 for death after leptomeningeal disease (95% CI 1.1 to 2.5; p = 0.015 for estrogen and p = 0.017 for progesterone), compared with receptor-positive patients.

HER2 status did not show a significant association with time to leptomeningeal disease or with post-diagnosis survival in this cohort. The authors call this partially unexpected and caution against reading it as evidence that HER2 is irrelevant; they suggest heterogeneous exposure to HER2-targeted therapies, which could not be modeled well given cohort size and incomplete treatment data, may have obscured the association. The longitudinal analyses therefore focused on estrogen and progesterone receptors.

Receptor-negative tumors favor the CNS as first metastatic site

A separate comparison asked where the first distant metastasis appeared. Among these patients, estrogen-negative status (p = 0.021), progesterone-negative status (p = 0.043) and HER2-positive status (p = 0.034) at primary diagnosis were each significantly associated with a first metastasis in the central nervous system rather than at a non-central-nervous-system site, by chi-squared analysis.

The paper also normalized central nervous system metastasis cohorts, including its own, against receptor proportions in a primary breast cancer reference population derived from TCGA. Estrogen-negative cancers were overrepresented among central nervous system metastasis cases (odds ratio 3.73, 95% CI 2.66 to 5.22) and so were progesterone-negative cancers (odds ratio 3.60, 95% CI 2.50 to 5.19); HER2-positive cancers were also overrepresented (odds ratio 1.95, 95% CI 1.58 to 2.41). The authors explicitly flag that the TCGA reference is not a matched control cohort and was assembled under different receptor-assessment standards, so these odds ratios should be treated as hypothesis-generating rather than definitive.

This distinction matters for readers: the chi-squared result describes where the first metastasis occurred within a cohort that all eventually developed leptomeningeal disease, while the odds ratios compare receptor frequencies against an external reference. They answer related but not identical questions and should not be substituted for one another.

Receptor status switched in 29% of patients

Receptor status changed after the initial diagnosis in 29% of patients (41 of 141). Hormone receptors switched in 26% (37 of 141) and HER2 switched in 7% (10 of 141). The paper identifies hormone receptor loss as the predominant change after initial diagnosis.

Because status was updated at each newly documented receptor assessment and carried forward until the next assessment or the leptomeningeal diagnosis, the time-dependent models could credit a receptor loss that occurred at a local recurrence or extracranial metastasis before central nervous system involvement. That is the mechanism the authors propose: a switch toward hormone-receptor-negative disease at a non-central-nervous-system metastatic site may accompany or drive accelerated spread to the central nervous system.

One boundary is important here. Receptor status measured directly in leptomeningeal disease tissue was not required for these analyses and was not analyzed as a separate endpoint or covariate. The switches described are therefore those captured at earlier assessed sites, not a direct readout of the leptomeningeal tumor itself.

Time-dependent receptor loss strengthens the CNS signal

When hormone receptor status was modeled as time-dependent rather than fixed at diagnosis, the association with central nervous system spread became stronger. For time from primary diagnosis to leptomeningeal disease, a negative hormone receptor status carried a hazard ratio of 1.7 in the time-dependent estrogen model and 1.6 in the time-dependent progesterone model (both p < 0.001). A first metastasis in the central nervous system carried a hazard ratio of 2.4 in both time-dependent models.

The first-metastasis location also behaved differently depending on how it was modeled. A non-central-nervous-system first metastasis predicted leptomeningeal disease only in the time-dependent model (hazard ratio 1.6, p = 0.005 for estrogen; 1.5, p = 0.008 for progesterone), not when metastasis status was fixed at diagnosis. The authors interpret this as a sign that non-central-nervous-system metastatic events arising during follow-up, together with the receptor discordance attached to them, may contribute to later leptomeningeal progression.

That interpretation is the authors' reading of a statistical pattern, not a demonstrated causal chain. The models adjust for age, race and stage, but the data are retrospective and observational, so the analysis can show association and timing, not that receptor loss causes central nervous system spread.

A single-center cohort with tissue limits

The cohort is 141 patients treated at one institution, identified retrospectively, and every patient had leptomeningeal disease by design. That makes the study informative about progression within this extreme central nervous system-selective group, but it cannot by itself estimate how often receptor loss leads to leptomeningeal disease among all breast cancer patients.

Tissue availability is the central measurement constraint. Receptor status after the primary diagnosis was available in 117 metastases, and brain metastasis receptor status was evaluated in 31 patients, 29 of them by immunohistochemistry or in situ hybridization. Missing outcome values were handled by multiple imputation, with 20 imputed datasets and up to 30 iterations, and estimates were pooled using Rubin's rules.

The HER2 result carries its own caveat, already noted: treatment exposure could not be adequately modeled, so the absence of a significant HER2 association in this cohort should not be read as evidence against HER2's established role in central nervous system metastasis.

Reassessing receptors at recurrence is the practical signal

The paper's most decision-relevant point is a comparison between receptor status fixed at diagnosis and receptor status updated through progression. The time-dependent models, which incorporate hormone receptor loss at recurrence or non-central-nervous-system metastases, showed stronger associations with central nervous system spread than baseline status alone. The authors conclude that this supports reassessing receptor status at recurrence and identifies loss of estrogen or progesterone receptor expression as a potential marker of central-nervous-system-directed progression that warrants further study.

For readers weighing what to do with this, the actionable split is by receptor trajectory rather than by subtype label alone: a patient whose tumor loses estrogen or progesterone receptor expression during progression is the group this paper flags for heightened central nervous system surveillance. That is a hypothesis the authors put forward from observational associations in one cohort, not a validated surveillance threshold.

The paper's own framing is that the consistent relationship between loss of either hormone receptor and central nervous system outcomes suggests dynamic receptor status may reflect biological features linked to central nervous system tropism. Confirming whether that marker holds across other cohorts, and whether acting on it changes outcomes, remains outside what this dataset can show.

Sources used in this answer