Can Machine Learning Help to Improve Clinical Outcomes in Hepatocellular Carcinoma?

By: Kayci Reyer
Posted: Thursday, July 20, 2023

According to findings presented in BMC Cancer, a newly identified antigen presenting cells (APC)-T-cell infiltration (TCI)–derived long noncoding RNA signature known as ATLS may be instrumental in improving clinical outcomes in hepatocellular carcinoma. The signature, constructed using multiple machine learning algorithms, may prove to be a useful predictor of overall survival.

“Furthermore, the ATLS also showed the exciting value in assessing the levels of tumor mutation, drug sensitivities, immune infiltration and T[-cell] regulators,” concluded Jianjian Zheng, PhD, of Wenzhou Medical University, China, and colleagues.

Using three public data sets and an external clinical cohort, the study investigators focused on information from 805 patients with hepatocellular carcinoma. A correlation was observed between high ATLS scores and poor prognosis. High levels of tumor mutation, immune activation, T-cell proliferation regulator expression, and anti–PD-L1 response were associated with high ATLS scores. Of note, high ATLS scores were also linked to significant sensitivity to oxaliplatin, fluorouracil, and lenvatinib.

The novel signature was initially built using 15 machine learning integrations powered by 5 machine learning algorithms. The optimal ATLS was then created using the optimal machine learning integration as determined by average C-index in the validation sets. The inclusion of various clinical and molecular traits assisted with the signature’s comparative performance.

“Due to the complex mechanisms and geographic variations of hepatocellular carcinoma, current staging systems have some limitations in the diagnosis and decision-making process for advanced hepatocellular carcinoma,” the authors noted. “Our ATLS model has a superior predictive capacity compared with traditional clinical characteristics and molecular features.”

Disclosure: The study authors reported no conflicts of interest.

BMC Cancer

---