Drug resistance is an important clinical problem affecting the prognosis of colorectal cancer patients with initially unresectable liver metastases. The role of 5-hydroxymethylcytosine (5-hmC) in this process is still unknown.
A single-center study was conducted to enroll colorectal cancer patients with initially unresectable liver metastases at Zhongshan Hospital, Shanghai. The 5-hmC levels of each gene locus in circulating-free DNA (cfDNA) were detected using a highly sensitive sequencing method before the patient received any conversion therapy. Eight weeks after conversion therapy, patients with disease progression according to RECIST 1.1 were defined as drug-resistant group, while patients with complete response, partial response and stable disease were defined as the drug-effective group. By using the gene loci with the most significant difference of 5-hmC level, a predictive model (Logistic regression) was established based on machine-learning method. Then, we validated the model.
From June 2018 to January 2019, 35 patients were enrolled, including 20 patients in drug-effective group and 15 in drug-resistant group. The predictive model was established using the first 100 significantly different gene loci and 10 loci were included in the model. The internal validation results suggested that the sensitivity (predictive accuracy for drug-resistant patients) and specificity (predictive accuracy for drug-effective patients) were 86.7% and 95.0%, respectively, and the area under the ROC curve is 0.990.
Our study screened the gene loci with different 5-hmC level between drug-effective and drug-resistant colorectal cancer patients with initial unresectable liver metastases. Then we established a predictive model for the efficacy of conversion therapy, although this model is still being improved and validated by enrolling more subjects.
Clinical trial identification
Legal entity responsible for the study
Has not received any funding.
All authors have declared no conflicts of interest.