Genomic predictive model for recurrence and metastasis development in head and neck squamous cell carcinoma patients

Abstract

The head and neck squamous cell carcinoma (HNSCC) population consists mainly of high-risk for recurrence and locally advanced stage patients. Increased knowledge of the HNSCC genomic profile can improve early diagnosis and treatment outcomes. The development of models to identify consistent genomic patterns that distinguish HNSCC patients that will recur and/or develop metastasis after treatment is of utmost importance to decrease mortality and improve survival rates. In this study, we used array comparative genomic hybridization data from HNSCC patients to implement a robust model to predict HNSCC recurrence/metastasis. This predictive model showed a good accuracy (>80%) and was validated in an independent population from TCGA data portal. This predictive genomic model comprises chromosomal regions from 5p, 6p, 8p, 9p, 11q, 12q, 15q and 17p, where several upstream and downstream members of signaling pathways that lead to an increase in cell proliferation and invasion are mapped. The introduction of genomic predictive models in clinical practice might contribute to a more individualized clinical management of the HNSCC patients, reducing recurrences and improving patients' quality of life. The power of this genomic model to predict the recurrence and metastases development should be evaluated in other HNSCC populations.

Figure 1

Profile of chromosomal imbalances detected in HNSCC patients using array-CGH technique. Blue represents copy number gains and red copy number losses. The fraction of samples means the fraction of patients that exhibited the imbalance. In X axis, chromosome number, from p arm to q arm (right to left), nucleotide position is represented.

Figure 2

Heatmap with copy number alteration profile in the chromosomal regions used by the different phases of the predictive genomic model, (A) in patients with vs. without recurrence/metastasis - first phase of the predictive model; (B) in patients without recurrence and those unidentifiable - second phase of the predictive model; (C) in patients with recurrence and those unidentifiable - third phase of the predictive model.

Figure 3

Ideogram with chromosomal regions used by predictive genomic model and the highlighted candidate genes in these regions. (A) In patients with vs. without recurrence/metastasis - first phase of the predictive model; (B) in patients without recurrence and those unidentifiable - second phase of the predictive model; (C) in patients with recurrence and those unidentifiable - third phase of the predictive model. Blue represents the proportion of copy number gains and red represents the copy number losses identified in these specific chromosomal regions both in our HNSCC patients and in TCGA database.