The prognostic potential of alternative transcript isoforms across human tumors

Area of Science:

• Oncology
• Bioinformatics
• Genomics

Background:

• Cancer progression involves gene expression changes, enabling molecular signatures for tumor staging and prognosis.
• The prognostic value of transcript isoform abundance for clinical stage and survival remains largely unexplored.

Purpose of the Study:

• To systematically analyze the prognostic potential of transcript isoforms in 12 solid tumors.
• To develop novel molecular signatures for cancer stage and prognosis based on transcript isoform abundance.
• To investigate transcript isoform changes in breast and melanoma tumors based on specific phenotypes.

Main Methods:

• Integrated RNA sequencing and clinical data from The Cancer Genome Atlas (TCGA) project.
• Applied information theory and machine learning methods for systematic analysis.
• Developed transcript isoform signatures for stage and survival prediction.

Main Results:

• Transcript isoform signatures accurately distinguished between early and late-stage tumors and metastatic versus non-metastatic tumors.
• Signatures predicted patient survival, particularly in cases with undetermined lymph node invasion or metastatic status.
• Isoform signatures demonstrated comparable or superior accuracy to gene expression signatures and were largely independent of gene expression alterations.

Conclusions:

• New signatures based on transcript isoform abundance characterize tumor phenotypes and progression independently of gene expression.
• Transcript isoform signatures are particularly relevant for assessing lymph node invasion and metastasis.
• These findings may advance precision cancer medicine strategies.