If many drugs are available in RCC, we lack predictive biomarkers of disease recurrence or progression for personalized treatment. Circulating biomarkers (CB) are attractive candidates for disease monitoring. Longitudinal assessment of CB was performed in the NEORAD clinical trial (NCT01715935).
Locally advanced (LA) and metastatic (M) patients with clear cell or papillary RCC received 6 weeks prior-nephrectomy everolimus (Ev). CB assessed were: angiogenesis factors: VEGF-A, VEGFR-1&2, bFGF, SDF1, PlGF; CEC; CTC; HSC; immune cells: CD45, CD14+VEGFR-1/Tie2, Treg, IL-6. Clinical/histological (CH): ECOG-PS, tumor burden (TB), % necrosis, % sarcomatoid. Exploratory analysis (EA): CB and CH (prior to Ev, D22, D42, 4 weeks post-nephrectomy (+ 4 weeks if Ev resumed in M pts) in PD vs non-PD pts used Bayesian Model Averaging (BMA) and regularized Cox regression (LASSO). Three modeling strategies were compared for robustness.
25 patients were included: LA = 14, M = 11. In LA and M cohorts, respectively 2 and 9 pts exhibited progression during the 12m post-surgery follow-up period. Most important predictors upon EA (by order): tumor burden, VEGF-A, LA/M, VEGFR-2, % necrosis, IL-6, VEGFR-1, age, PlGF. Higher TB at baseline and % necrosis were associated with increased risk of 6m post-nephrectomy progression. TB, IL-6, VEGF-A, VEGFR-1&2 exhibited nonlinear relationships suggesting complex underlying pathophysiological mechanisms involved in response to Ev, and are currently explored. Reduced VEGFR-2 at D42 was associated with worse PFS in both cohorts. Upon BMA and LASSO, % necrosis and TB were among the most retrieved predictors prior to nephrectomy, whereas CD45 and CD34 + 45-146+ at 4 weeks post-nephrectomy were the best CB predictors. No significant change in monocyte populations and in Treg was observed.
VEGF related CB, TB and % necrosis were the best candidates to discriminate PD- vs non-PD in LA and M pts. Extensive analysis of data collected using robust non-linear models could contribute to improve our understanding of mechanisms involved and suggest predictive biomarkers to be validated.
Clinical trial identification
Legal entity responsible for the study
All authors have declared no conflicts of interest.