The advent of molecular targeted cancer therapy has shed light on the critical role that biomarkers play in identifying responsive patients and fine-tuning personalised cancer care. Bromodomain extra terminal (BET) inhibitors neutralise the bromodomains of BRD2, BRD3, BRD4 and have shown activity in haematological malignancies in Phase I clinical trials. While meaningful and durable responses have been noted, it is currently unclear which patients will respond to BET inhibitor therapy. Growing evidence in preclinical models shows that BET family members are active at discrete super-enhancer elements that direct RNA polymerase II mediated transcription. Unfortunately, super-enhancers are variably formed and require sophisticated methodology such as ChIP-seq to detect. Thus, these elements can be challenging to consistently detect in preserved human tissue and are not amenable as biomarkers in the companion diagnostics setting. Herein, we utilised an epigenetic-focused CRISPR library to identify distinct BET activity in a subset of epigenetically dysregulated colon cancers, characterized by the CpG island methylator phenotype (CIMP). Integrated transcriptomic and genomic analyses defined a discrete super-enhancer in CIMP(+) colon cancers that regulates cMYC transcription. We find that the CCAT1 long non-coding RNA (lncRNA) is transcribed from this super-enhancer and is exquisitely sensitive to BET inhibition. Concordantly, cMYC transcription and cell growth were tightly correlated with the presence of CCAT1 RNA in a variety of tumor types. Similarly, we show that tissue-specific lncRNAs can predict BET activity in other tumor types when expressed from lineage specific super-enhancers at the MYC locus.Taken together, our work provides a strong rationale for exploring the clinical utility of enhancer template lncRNAs in predicting BET inhibitor response.