Pre-diagnosis green space exposure and survival in women diagnosed with ovarian cancer

Background: Despite treatment advances, the five-year relative survival is 50% for patients with invasive ovarian cancer. Increasing literature demonstrates that distress, depression, and other psychosocial stressors negatively influence survival. Exposure to green spaces - e.g., parks, gardens, forests and street trees - has been associated with lower stress hormone levels and reduced risk of depression. Thus, our objective was to evaluate if greenness exposure, pre-diagnosis, was associated with improved ovarian cancer survival among cases in the Nurses' Health Study (NHS) and NHSII cohorts. Methods: Greenness exposure was assessed seasonally within nearby (270m) and walkable (1230m) distance buffers surrounding NHS and NHSII cohort participants' residential addresses (updated biennially from 1988 to 2018) using the Normalized Difference Vegetation Index (NDVI; 30m 2 resolution)-a satellite-derived indicator of photosynthetic vegetation cover. We used time-varying Cox proportional hazards models to assess cumulative, pre-diagnosis, maximum average NDVI exposure and ovarian cancer specific survival among confirmed cases. Models were adjusted for age and calendar time, US region, population density, and pre-diagnosis factors associated with ovarian cancer survival, including menopause status, family history of breast or ovarian cancer, and parity. Time-varying depression - self-reported antidepressant use or clinician diagnosed depression - was additionally adjusted to evaluate potential mediation of the relationship between greenspace exposure and survival. Results: Analyses included 1,052 women with confirmed ovarian cancer. We observed a suggestive decrease in ovarian cancer mortality risk for a 0.1 unit (10%) increase in cumulative pre-diagnosis 270m NDVI exposure (aHR 0.92; 95% CI 0.79, 1.07) when adjusting for all factors except depression. Adding depression led to a slightly stronger inverse association (aHR 0.90 95% CI 0.76, 1.07), although it did not reach statistical significance. No association was observed for cumulative pre-diagnosis 1230m NDVI exposure. Conclusions: Pre-diagnosis near-residence greenness exposure was suggestively associated with improved survival among ovarian cancer cases. Adjusting for depression status slightly strengthened the inverse association, indicating that the association of greenness with survival is unlikely to be fully explained by depression status; other potential mechanistic pathways (e.g., pollution and temperature reduction, improved sleep quality, physical activity, vitamin D levels and immune function) should be explored. On-going analyses will assess associations by histologic subtype. Future studies are needed to assess the type and quality of nature exposure, including but not limited to green spaces, with risk and survival among patients, with stratification by geographic and ecological regions.