Primary liver cancer (PLC) is the 6th most common cancer worldwide in 2020 and the 3rd leading cause of cancer-related death worldwide; hepatocellular carcinoma (HCC) is the most common type of PLC, which accounts for more than 80% of all PLC (1, 2). For HCC, treatment of early patients includes surgical resection, transplantation, radiofrequency ablation, etc. (3–5). Although surgical resection and/or liver transplantation remain the treatment of the first choice for HCC patients, most patients with HCC do not develop symptoms until intermediate and advanced stages, and only 20% of patients have a chance of surgical resection at the time of diagnosis (6).

Like surgery, external beam radiation therapy (EBRT) is one of the most common methods of treating tumors. HCC has moderate to high radiosensitivity to radiation therapy, just inferior to normal organs or tissues that are very sensitive to radiation, such as kidneys, bone marrow, lymphoid tissue, etc. Before the 1990s, due to technical limitations, large-volume liver irradiation often led to hepatotoxicity and even radiation-induced liver disease (RILD), limiting the role of EBRT in HCC patients (7). EBRT technology has undergone a series of advances in recent decades, with the application of three-dimensional conformal radiotherapy (3-DCRT), intensity-modulated radiation therapy (IMRT), and stereotactic body radiotherapy (SBRT), the accuracy of tumor lesion targeting was greatly improved while the radiation dose on the surrounding normal tissues can be substantially reduced, this significantly reduces the incidence of hepatotoxicity, so that EBRT may be one of the promising treatments for HCC patients (8–10).

In recent years, nomogram has been widely used as a prediction method in oncology, which is convenient for clinicians to use for prognosis prediction and has played an important role in promoting personalized medicine (11). As far as we know so far, there are few studies that have developed nomograms to predict prognosis in HCC patients with EBRT, and these were small studies due to the small number of HCC patients treated with EBRT (12–14). For the first time, based on the SEER database and a Chinese cohort, we established and validated such a nomogram to predict OS in HCC patients treated with EBRT. In addition, to provide patients with better medical care, we also established a web-based nomogram model that could help clinicians make better clinical decisions by judging the prognosis of these patients.

Global HCC incidence and mortality continue to rise (15). Due to the insidious onset and rapid progress of HCC, most of them are already advanced at the time of diagnosis; therefore, early local treatment methods such as liver resection, transplantation, and radiofrequency ablation have limited benefits (3–6, 16). With the continuous improvement of EBRT technology, its role has gradually changed from simple palliative treatment to multidisciplinary comprehensive treatment or even radical treatment (8–10, 17–21).

There is currently few complete and valid nomogram model to predict the prognosis of HCC patients treated with EBRT (12–14). Based on the sample data obtained from the SEER database, through univariate and multivariate analysis, a total of seven independent risk factors associated with prognosis were included: T stage, N stage, M stage, AFP status, tumor size, surgery, and chemotherapy; an intuitive prognostic prediction model was constructed, internal and external validation cohorts verified its accuracy. In addition, we designed a web-based nomogram to predict overall survival in those patients, which is expected to provide more evidence for individualized treatment.

It is well known that AJCC (TNM) staging is an important factor affecting the OS of HCC patients and has significant guiding value for its treatment (22). T stage has always been regarded as an important prognostic factor affecting HCC; it has been widely used in traditional HCC staging systems to provide treatment guidance, such as Okuda and BCLC stage (22–24). An Austrian study confirms that tumor cells in lymph nodes can spread into the vascular circulation and metastasize to distant organs (25); meanwhile, the National Comprehensive Cancer Network (NCCN) guidelines and the AJCC (TNM) staging system consider regional lymph node metastases as advanced HCC (26). The 5-year survival rate of patients with early-stage HCC after liver resection exceeds 70%, but once distant metastasis occurs, the median survival time is only 1-1.5 years even after multidisciplinary diagnosis and treatment (27, 28). Similar to previous studies, the T stage, N stage, and M stage in this study were all poor prognostic factors; it can be seen from the nomogram that the occurrence of distant metastasis is the most important factor affecting prognosis.

The serum AFP levels have been the most common laboratory value of HCC for decades. Liu et al. demonstrated that AFP levels are closely related to the degree of differentiation and vascular invasion of HCC (29, 30). In this study, HCC patients with negative AFP levels had lower scores and better prognoses, whereas HCC patients with unknown AFP levels had higher scores and poorer prognoses, probably because this part mainly consisted of patients with positive AFP levels. Tumor size is closely related to the prognosis of HCC, and many guidelines regard tumor size as an important reference for HCC staging and treatment (31). A multicenter study of surgical resection HCC demonstrated that tumors larger than 5cm had a worse prognosis than patients less than or equal to 5cm (32, 33). It is not difficult to see from the nomogram that as the tumor size increases, the higher the score and the worse the prognosis.

Surgery and chemotherapy were independent protective factors in this study. Surgical resection remains the preferred treatment for HCC; however, most HCC patients are asymptomatic in the early stage and are diagnosed in the advanced stage without the opportunity for surgery (6). Our current guidelines do not recommend surgery for advanced HCC (34). However, a retrospective study by Mao et al. showed that even with distant metastases, patients who underwent surgery when disease permitted had better outcomes than those who did not (35). Studies have shown that the combination of surgery and chemotherapy is beneficial for HCC patients (36, 37). Meanwhile, systemic chemotherapy with gemcitabine, doxorubicin or combined regimens also improved HCC patients survival (38). Similar to previous studies, less than 15% of patients in this study were operable. Still, it played a crucial role in the prognosis of HCC patients who received external beam radiation therapy. Likewise, the scores on the nomogram showed that patients who received chemotherapy had a better prognosis than those who did not.

The differences in the prognosis of HCC treated with EBRT among different ethnic groups may be related to the complex socioeconomic factors among ethnic groups and the differences in medical level in residential areas (39). A less differentiated tumor usually indicates a higher degree of malignancy, greater invasiveness, and a worse prognosis (40). There were differences in race and degree of differentiation in univariate analysis in this study, but no difference in multivariate analysis, which may be related to the small number of cases and the unknown degree of differentiation of most HCC patients, which needs further verification.

This study provides sufficient samples and clinical data based on the SEER database; a prognostic prediction nomogram was built and internal validation using eligible patients from the SEER database. Meanwhile, the external validation cohort from China was used for external verification, so the results have high reliability. The overall baseline characteristics of patients in the SEER database and Chinese patients were compared, all the Chinese patients were of the yellow race and the years of diagnosis were after 2014, and there were some differences in the treatment methods. However, The C-index, AUC values, calibration curves of the three cohorts all showed satisfactory results. The DCA curves suggested that the nomogram has good clinical utility. In addition, for the first time, we established a web-based, user-friendly nomogram model that clinicians and patients accessible from any electronic device.

Although the nomogram has good clinical utility, the present study had several limitations. First, data regarding several potential crucial prognosis-related serum markers such as HBsAg, AST, and CEA were unavailable in the SEER database; these will be the main part of our future research. Second, given the international and retrospective nature of the study, we cannot rule out that some clinicopathological characteristics might not have been evaluated uniformly in different institutions. In addition, there might be some selection bias in diagnosis, therapeutic strategies, and follow-up of patients across the institutions; variations in follow-up HCC patient condition changes and treatment plans might result in discrepancies in EBRT outcomes. Finally, compared with the American population-based cohort, the sample size of the external validation cohort from China was small; the validity of our nomogram in Eastern countries needs to be further evaluated.

In conclusion, for the first time, we developed and validated a nomogram to predict overall survival in HCC patients with EBRT; both internal and external validation demonstrated remarkable calibration and discrimination of our nomogram. In addition, our established a web-based nomogram model can help clinicians judge prognosis, make better clinical decisions, and improve individualized survival probability.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the ethics committee at Third Xiangya Hospital, Central South University, Changsha, China. Written informed consent from the patients/participants was not required to participate in this study in accordance with the national legislation and the institutional requirements.

PC designed the research. GZ and HP performed the research and analyzed results. GZ and LZ edited the manuscript. XX provided critical comments and revised the manuscript. YH, XW and ZD collected and organized data. LJ wrote the revised manuscript. All authors contributed to the article and approved the submitted version.