GATK best pratices for RNA-seq somatic mutation finding

Dear All,

I have followed GATK best practices for finding Somatic mutations from cancer versus normal sample from RNA-seq data using Mutect2 as the final caller and rest all quality control steps as per GATK best practices for RNA-seq. But the resulting filtered sets are false positives as we tried to validate them in the wet lab. So we are now planning to call somatic variants using haplotype caller as per guidelines in GTAK practices, please someone who has experience with this caller comment on this.How is it with Haplotype caller, I would like to know the Sensitivity and specificity rate, also the parameters used in the caller to get true positive variants.
Thank you