Robust Estimation of Conditional 2D Reference Regions R package TwoDiRef version 0.1.0 Provides tools for constructing conditional two-dimensional reference regions in continuous data, particularly suited for clinical, biological, or epidemiological studies requiring robust multivariate assessment. The implemented methodology combines directional quantiles with median‑based partial correlation models to produce reliable and interpretable reference regions even in the presence of outliers. Key features include robust conditional modeling for two responses conditioned on covariates, directional quantile regions, cross‑validation of coverage, visualization tools, and flexible formula‑based inputs. Software Depends: R (>= 4.0.0) Imports: qgam, mgcv, sp, ggplot2, scales, shiny Suggests: refreg Oscar Lado-Baleato <ladooscar@uniovi.es> Comprehensive R Archive Network (CRAN) Oscar Lado-Baleato [aut, cre], Sara Rodriguez-Pastoriza [aut], Javier Roca-Pardinas [aut] GPL-3 2026-03-30 application/tgz https://CRAN.R-project.org/package=TwoDiRef doi:10.32614/CRAN.package.TwoDiRef