20-gene expression profile may improve risk stratification in cSCC
Click Here to Manage Email Alerts
Key takeaways:
- The 20-gene prognostic model had an accuracy of 86% and a sensitivity of 85.7%.
- The model also had a specificity of 86.1% and a positive predictive value of 78.3%.
A 20-gene signature prediction demonstrated high accuracy in diagnosing metastasis risk in patients with cutaneous squamous cell carcinoma, according to a study.
Cutaneous squamous cell carcinoma (cSCC) has a low metastasis rate of 2% to 5%. However, considering it is the most common form of skin cancer, 2% to 5% still constitutes a significant disease burden. Unfortunately, multiple histopathological staging classifications for cSCC, though available, do not accurately predict risk outcomes.
Jun Wang, PhD, reader in genomics and data science at Barts Cancer Institute at Queen Mary University of London, and colleagues have created a gene expression profile (GEP) signature that may improve these outcomes.
"Using an unbiased whole-transcriptomic approach, we harnessed machine learning to create a model that uses the activity of 20 genes to predict the risk of cSCC metastasis with an accuracy of 86%, surpassing current, more subjective, assessment methods," Wang told Healio.
The researchers collected tissue data from 237 immunocompetent patients, 151 with nonmetastasizing cSCC and 86 with metastasizing cSCC.
They developed and validated a 20-gene prognostic model that showed results that were better than pathological staging systems. The 20-gene prognostic model had an accuracy of 86%, a sensitivity of 85.7%, a specificity of 86.1% and a positive predictive value of 78.3%.
The researchers then developed a linear predictor for metastasis that combined the expression values and fold-changes of these 20 genes, showing specific values that correlate to a progressively higher metastatic risk. In other words, the higher the value on the linear predictor, the higher the risk, according to the authors.
Results showed the linear predictor had a highly correlated with metastatic risk. The area under the receiver operating characteristic curve was 0.85 (95% CI, 0.8-0.91) for the training set and 0.88 (95% CI, 0.78-0.99) for the validation set.
“This prognostic signature could significantly improve risk stratification, identifying patients with high-risk cSCC who may benefit from adjuvant treatment and reducing overtreatment for patients with low-risk cSCC," Wang said. "Moreover, some of the genes in our model require better understanding of their roles in cSCC, and they may present promising new targets for therapies. Obviously before we do all these, the immediate next step will involve validating our 20-gene prognostic signature in a larger and more diverse patient cohort."