[Bug]: summarize_ancova fails to return contrast rows when ref_group contains regex metacharacters such as `+`

What happened?

Hi,

I found what appears to be a bug in tern::summarize_ancova() when the treatment/reference level contains regex metacharacters, for example:

"Tirzepatide + Placebo"

Expected behavior:
summarize_ancova() should return lsmean_diff, lsmean_diff_ci, and pval for non-reference treatment columns, the same way that direct emmeans::contrast(..., method = "trt.vs.ctrl") does.

Actual behavior:
lsmean is returned, but lsmean_diff, lsmean_diff_ci, and pval are empty in the table output.

What I checked:

• The same model fitted directly with lm() + emmeans() + contrast() returns valid estimates, confidence intervals, and p-values.
• split_cols_by() itself appears to work correctly.
• The failure seems to happen inside summarize_ancova() when contrast labels are matched back to treatment levels.

I believe the issue is here:
https://github.com/insightsengineering/tern/blob/v0.9.8.433/R/summarize_ancova.R#L190

My current understanding is that the reference-group label is used in pattern matching without escaping regex metacharacters, so + is interpreted as regex instead of a literal character.

sessionInfo()

R version 4.4.3 (2025-02-28)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 22.04.5 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so;  LAPACK version 3.10.0

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C              
[10] LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

time zone: Etc/UTC
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices datasets  utils     methods   base     

other attached packages:
[1] BC45538_TLG_1.0.0       tlg_ctassets_0.0.0.9000 tibble_3.2.1            purrr_1.0.2             tidyr_1.3.1             emmeans_1.10.4          dplyr_1.1.4            

loaded via a namespace (and not attached):
 [1] Rdpack_2.6.1          mmrm_0.3.14           tern.mmrm_0.3.3       remotes_2.5.0         testthat_3.2.1.1      rlang_1.1.6           magrittr_2.0.3        compiler_4.4.3        roxygen2_7.3.2        callr_3.7.6           vctrs_0.6.5          
[12] formatters_0.5.11     stringr_1.5.1         pkgconfig_2.0.3       crayon_1.5.3          fastmap_1.2.0         backports_1.5.0       utf8_1.2.4            ps_1.8.0              oceanflow_1.5.0       bit_4.5.0             xfun_0.52            
[23] jsonlite_2.0.0        styler_1.10.3         broom_1.0.8           parallel_4.4.3        R6_2.5.1              stringi_1.8.4         RColorBrewer_1.1-3    parallelly_1.45.0     pkgload_1.4.0         brio_1.1.5            lubridate_1.9.3      
[34] estimability_1.5.1    Rcpp_1.0.13           rtables_0.6.13        assertthat_0.2.1      knitr_1.50            citril.metadata_0.1.8 usethis_3.0.0         R.utils_2.12.3        Matrix_1.7-0          splines_4.4.3         R.cache_0.16.0       
[45] timechange_0.3.0      tidyselect_1.2.1      yaml_2.3.10           processx_3.8.4        rlistings_0.2.12      pkgbuild_1.4.4        lattice_0.22-6        withr_3.0.1           evaluate_1.0.1        citril_0.1.18         desc_1.4.3           
[56] survival_3.8-3        xml2_1.3.6            pillar_1.9.0          checkmate_2.3.2       renv_1.0.10           rex_1.2.1             generics_0.1.3        rprojroot_2.0.4       hms_1.1.3             ggplot2_3.5.1         scales_1.4.0         
[67] xtable_1.8-4          chevron_0.2.12        glue_1.8.0            nestcolor_0.1.3       lazyeval_0.2.2        tools_4.4.3           data.table_1.17.8     forcats_1.0.0         fs_1.6.4              mvtnorm_1.3-1         cowplot_1.1.3        
[78] grid_4.4.3            cyclocomp_1.1.1       rbibutils_2.3         lintr_3.1.2           nlme_3.1-166          cli_3.6.5             gtsummary_2.5.0       dunlin_0.1.12         tern_0.9.9            fansi_1.0.6           gt_1.3.0             
[89] admiraldev_1.4.0.9000 arrow_17.0.0.1        gtable_0.3.6          R.methodsS3_1.8.2     admiral_1.4.1         digest_0.6.37         farver_2.1.2          htmltools_0.5.8.1     R.oo_1.26.0           lifecycle_1.0.4       bit64_4.5.2

Relevant log output

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

Start at R/summarize_ancova.R around line 190 and reproduce the issue with a reference level containing +. Compare the contrast-label matching there with the direct emmeans result. Done means non-reference treatment columns contain lsmean_diff, lsmean_diff_ci, and pval for such labels, with the relevant tests passing.