Prognostic value of a nomogram based on peripheral blood immune parameters in unresectable hepatocellular carcinoma after intensity-modulated radiotherapy

A retrospective study was performed on 309 HCC patients who received IMRT treatment at Guangxi Medical University Cancer Hospital from February 2013 to July 2021. Patients with uHCC were included as per the following criteria: (a) diagnosed pathologically or according to the criteria of the European Association for the Study of the Liver, (b) Child–Pugh grade A or B liver function, and (c) completed peripheral blood cell counts before and after IMRT. Patients who underwent postoperative adjuvant IMRT, failed follow-up, and were unable to have their peripheral blood cell count monitored were excluded. This study was approved by the institutional review board of Guangxi Medical University Cancer Hospital (number LW2022059).

All patients underwent enhanced computerized tomography (CT) scan at 2.5–5 mm slice thickness for the IMRT plan. Gross tumor volume (GTV) was defined as tumor focus. CT-positron emission tomography fusion and magnetic resonance imaging fusion for extrahepatic metastasis and intrahepatic lesions were performed, respectively. The GTV and organs at risk were contoured on the Pinnacle 3 system (Philips, Netherlands) or MIM software (version 6.8; MIM, USA). The planned target volume (PTV) was defined as GTV plus asymmetrical dilation of 1 cm in the craniocaudal direction and 5 mm in the axial direction to set uncertainty and respiratory movement. The IMRT plans were designed using the Pinnacle 3 system or Monaco treatment planning system version 5.1. The final median biologically effective dose, which used α/β ratio = 10 according to the linear-quadratic model, was 67.2 Gy (interquartile range, 60–78 Gy). IMRT was delivered via a 6 MV X-ray linear accelerator (ELEKTA Synergy or ELEKTA Versa-HD, Sweden) using cone-beam CT to correct the positions.

Peripheral blood was collected within 1 week before IMRT to determine lymphocyte, neutrophil, and platelet counts. The ratios and changes in values were calculated as follows: PLR = platelet count/lymphocyte count; NLR = neutrophil count/lymphocyte count; SII = platelet count × neutrophil count/lymphocyte count; post-PLR = post-platelet count/post-lymphocyte count; post-NLR = post-neutrophil count/post-lymphocyte count; post-SII = post-platelet count × post-neutrophil count/post-lymphocyte count; delta-NLR = NLR − post-NLR; delta-PLR = PLR − post-PLR; and delta-SII = SII − post-SII. The optimal cutoff value of immune parameters was selected according to the area under the receiver operating characteristic (ROC) curve.

R language (version 4.0.3) was used for statistical analysis. To validate model performance, 309 uHCC patients were randomly divided into the training and validation cohorts. The differences in baseline characteristics of the patients in these two cohorts were examined by χ² tests and Student’s t-tests for categorical and continuous variables, respectively. The overall survival (OS) was defined as the day of informed consent to the day of death from any cause. The survival curves were drawn using the Kaplan–Meier method, and the difference was compared using the log-rank test. The Cox proportional hazard model was used to determine the prognostic indicators, and the significant variables (P < 0.05) in the univariate Cox model were included in the multivariate model. Backward stepwise selection (P = 0.05 for removal) was used to select characteristics for the multivariable regression model. The nomogram was constructed based on the variables included in the final multivariate Cox proportional hazard model. Moreover, the efficiency of the nomogram was evaluated, and the time-dependent ROC curve and area under the ROC curve (AUROC) were analyzed to compare the discriminatory ability of different models in total survival. Statistical significance was set at P < 0.05.

A total of 338 patients with uHCC were screened, and 309 were included in the study (Fig. 1). The participants included 280 males and 29 females, with a mean age of 55 years, median OS of 16.3 months, and a median follow-up of 38.8 months. Most patients had Child–Pugh grade A liver function (238 cases; 77%), Barcelona clinic liver cancer (BCLC) stage C (250 cases; 80.9%), and radiographic cirrhosis (212 cases; 68.6%). There were 175 cases of splenomegaly. Before IMRT, transcatheter arterial chemoembolization (TACE) (66%) and resection (48.2%) were the main treatment methods. After IMRT, TACE (17.5%), targeted therapy (12%), and immunization (11.7%) were the most common treatment methods. All clinicopathological parameters are shown in Table 1. The patients were randomly divided into the training cohort (n = 216) and validation cohort (n = 93), and no significant difference in baseline characteristics was observed between the two cohorts (P > 0.05).

Peripheral blood was collected 1 week before and 1 month after IMRT. The immune parameters of all patients are shown in Additional file 1: Table S1. Statistical analysis of peripheral blood and inflammatory indicators showed that neutrophil and platelet counts decreased significantly after IMRT (both P < 0.001), while no significant difference was observed in lymphocyte counts. Similarly, a significant increase in inflammatory indicators (PLR, NLR, SII) was observed after IMRT (all P < 0.001).

In the validation cohort, We divided the immune parameters into high and low groups according to the cutoff values and found that all of them affected the prognosis of patients with uHCC after IMRT: PLR ≤ 301.01 and PLR > 301.01 (hazard ratio [HR] 2.48, 95% confidence interval [95% CI] 1.55–3.97, P = 0.000), NLR ≤ 1.3 and NLR > 1.3 (HR 2.79, 95% CI 1.47–5.32, P = 0.002), SII ≤ 1053.96 and SII > 1053.96 (HR 2.04, 95% CI 1.40–2.95, P = 0.000), post-PLR ≤ 602.63 and post-PLR > 602.63 (HR 1.87, 95% CI 1.17–2.97, P = 0.009), post-NLR ≤ 8.26 and post-NLR > 8.26 (HR 1.51, 95% CI 1.08–2.11, P = 0.015), post-SII ≤ 732.52 and post-SII > 732.52 (HR 1.41, 95% CI 1.03–1.93, P = 0.034), delta-PLR ≤ 117.43 and delta-PLR > 117.43 (HR 2.18, 95% CI 1.36–3.51, P = 0.001), delta-NLR ≤ − 0.61 and delta-NLR > -0.61 (HR 1.57, 95% CI 1.14–2.16, P = 0.006), delta-SII ≤ 620.43 and delta-SII > 620.43 (HR 1.88, 95% CI 1.26–2.81, P = 0.002). Kaplan–Meier survival analysis showed a worse HCC prognosis in the group with immune parameters values higher than the cutoff value (Fig. 2a–i). In contrast, neutrophil, lymphocyte, and platelet counts had no significant effect on prognosis. Cox univariate analysis revealed that enlarged spleen, tumor number, tumor size, distant metastasis, BCLC stage (B, C), surgical treatment before IMRT, total bilirubin level, lymphocyte count, NLR, and SII were all prognostic factors, while Sex, age, liver cirrhosis, HBV infection, vascular invasion, TACE before IMRT, RFA treatment and synchronization, and follow-up treatment had no significant effect on prognosis (Table 2). Multivariate analysis further showed that enlarged spleen, BCLC stage (B and C), post-SII, and delta-NLR were independent risk factors for uHCC (Table 2).

Based on the results of multivariate analysis, enlarged spleen, BCLC stage (B, C), post-SII, and delta-NLR were integrated into the nomogram to predict 3- and 5-year survival using the training cohort (Fig. 3). Each category of these variables was assigned a score on the score table, and the overall scores were used to predict survival. Of the variables, the BCLC stage had the most significant effect.

The 3- and 5-year AUROC values of the training cohort were 81.0 (95% CI 74.1–84.9) and 80.4 (95% CI 70.4–90.5), respectively (Fig. 4a, b), which were higher than those of the other models. Meanwhile, the 3- and 5-year AUROC values of the validation cohort were 70.2 (95% CI 53.0–87.4) and 71.1 (95% CI 58.0–84.2), respectively, which were still higher than those of the other models (Fig. 4c, d).

Moreover, in both cohorts, calibration curves showed that the predicted 3-year survival was consistent with the actual observed results (Fig. 4e, d). Due to the small number of cases followed up for more than 5 years, the calibration curve for 5-year survival could not be drawn. In summary, these results suggest that the nomogram based on enlarged spleen, BCLC stage (B and C), post-SII, and delta-NLR can accurately predict the 3- and 5-year survival rates of HCC patients.