Patient characteristics

93 HCC patients were divided into the non-recurrence group (n = 49, 52.7%) and the recurrence group (n = 44, 47.3%), according to the criteria for disease states. In the recurrence group, the follow-up lasts from 2.5 to 97 months, with a median duration of 30.4 months; and in the non-recurrence group, it varies from 30–113 months, with the median of 78 months, with 32 deaths recorded (34.4%) (Table 1). The Student’s t-test revealed a significant difference between the two groups (P < 0.001), indicating a tight relationship between OS with recurrence and follow-up time? Additionally, Milan criteria, tumor size, TNM stage, number of nodules, microvascular invasion, and antineoplastic prophylaxis after LT were all significantly correlated with recurrence (P < 0.050), but not sex, age, tumor differentiation, hepatitis B, Child–Pugh grade, cirrhosis, and pre-LT serum AFP level (Table 1).

Table 1 Clinical characteristics of patients with and without HCC recurrenceKRAS genotype distribution and the relationship with HCC recurrence

Minor allele frequencies (MAFs) of KRAS rs712 genotypes were more than 0.01 in both recipients and donors, as determined by Hardy–Weinberg equilibrium. When donors and recipients with KRAS rs712 polymorphisms were compared, the latter showed very wide distributions in the non-recurrence and recurrence groups; while, the former did not, which can be seen in Table 2. It is critical to note that patients with the KRAS rs712 TG/GG genotype had a significantly greater incidence of recurrence after LT than those with the TT genotype (79.5% vs20.5%, P < 0.001). However, the donor’s KRAS rs712 genotype had no effect on the recurrence rate of HCC after LT.

Table 2 Donor and Recipient KRAS genotype distribution and the association with HCC recurrenceRisk factors for HCC recurrence after LT

Due to prior research indicating that the histologic grade and AFP level also correlate with the recurrence rate, the study included these two predictive variables in the logistic regression analysis. We conducted the univariate and multivariate logistic regression analysis, shown in Table 3. The Milan criteria, tumor size, TNM stage, microvascular invasion, number of nodules, and recipient KRAS rs712 genotypes are all significant factors of HCC recurrence after hepatic transplantation. To prevent collinearity, the number of nodules and tumor size were omitted, and the remaining variables did not exhibit collinearity (tolerance > 0.1). TNM stage (OR = 12.43, 95% CI 4.58 ~ 39.02; P < 0.001), tumor size (OR = 6.01, 95% CI 2.44 ~ 15.86; P < 0.001), microvascular invasion (OR = 17.12, 95% CI 5.19 ~ 81.75; P < 0.001), number of nodules (OR = 4.43, 95% CI 1.68 ~ 12.94; P = 0.025), Milan criteria (OR = 10.42, 95% CI 3.87 ~ 32.41; P < 0.001), Treatment (OR = 4.86, 95% CI 1.37 ~ 23.98; P = 0.018) and recipient KRAS rs712 genotypes (OR = 0.21, 95% CI 0.04 ~ 0.73; P = 0.018) were related with HCC recurrence (Table 3). Additionally, multivariate analysis revealed that Milan criteria (OR = 0.60, 95% CI 0.36–0.99; P = 0.045) was an independent risk factor of HCC recurrence after liver transplantation.

Table 3 Univariate and Multivariate logistic regression analysis of risk factors associated with HCC recurrence after LTAssociation between recipients rs712 polymorphisms and survival rates

According to univariate Cox regression analysis, recipients’ KRAS rs712 genotype polymorphisms, but not donors’, contributed significantly to the different recurrence in both RFS and OS. The RFS and OS of recipients bearing the donor’s TT, TG, and GG alleles were subsequently examined using Kaplan–Meier survival and the log-rank test to ascertain the impact of the rs712 polymorphism on prognosis in various recipients. Distinctions in the RFS and OS of recipients with the TT, TG, and GG genotypes were detected (P < 0.001 and P < 0.001, respectively; Fig. 1A, B). However, donor rs712 genotypes did not demonstrate an impact on OS and RFS (Supplementary Fig. 1A, B). Simultaneously, the matching rate between donor and recipient genotype polymorphisms has no effect on the patients’ survival rate (Supplementary Fig. 2A, B).

Fig. 1

Kaplan–Meier survival estimates of recurrence-free survival (RFS) (A) and overall survival (OS) (B) among different recipient genotypes (GG, TG and TT). Patients carrying recipient genotype TT had the highest RFS and OS

Risk factors for the survival rate with no recurrence and overall survival rate by Cox regression analysis

We performed multivariate Cox regression analysis on more significant factors, including histologic grade (poorly differentiated), pre-LT serum AFP level (> 400 ng/ml), tumor size (5 cm), TNM stage (stages 3–4), Milan criteria (beyond), number of nodules (3), microvascular invasion, and donor rs712 TG/TT genotype. Tumor size and the number of nodules were excluded to prevent collinearity. There was no evidence of collinearity in the other variables (tolerance > 0.1). The pre-LT serum AFP level (HR = 2.400, 95% CI 1.092 ~ 5.278; P = 0.020), as well as the TNM stage (HR = 4.241, 95% CI 1.252 ~ 14.36; P = 0.010), were deemed to be independent risk factors for RFS (Table 4). Additionally, TNM stage (HR = 3.5249, 95% CI 2.221 ~ 5.594; P < 0.001), tumor size (HR = 2.602, 95% CI 1.645 ~ 4.114; P < 0.001), microvascular invasion (HR = 3.386, 95% CI 2.11 ~ 5.43; P < 0.001), the Milan criteria (HR = 2.654, 95% CI 1.663 ~ 4.236; P < 0.001), and recipients rs712 TG/TT genotype (HR = 0.397, 95% CI 0.218 ~ 0.725; P = 0.002) can be seen as independent risk factors for OS (Table 5).

Table 4 Cox regression analysis of the risk factors for recurrence-free survival time