Background:
Second primary malignancies (SPMs) after liver transplantation (LT) are becoming the leading causes of death in LT recipients. The purpose of this study was to explore prognostic factors for SPMs and to establish an overall survival nomogram.
Methods:
A retrospective analysis was conducted of data from the Surveillance, Epidemiology, and End Results (SEER) database on adult patients with primary hepatocellular carcinoma who had undergone LT between 2004 and 2015. Cox regression analysis was used to explore the independent prognostic factors for SPMs. Nomogram was constructed using R software to predict the overall survival at 2, 3, and 5 years. The concordance index, calibration curves, and decision curve analysis were used to evaluate the clinical prediction model.
Results:
Data from a total of 2078 patients were eligible, of whom 221 (10.64%) developed SPMs. A total of 221 patients were split into a training cohort (n=154) or a validation cohort (n=67) with a 7:3 ratio. The 3 most common SPMs were lung cancer, prostate cancer, and non-Hodgkin lymphoma. Age at initial diagnosis, marital status, year of diagnosis, T stage, and latency were the prognostic factors for SPMs. The C-index of the nomogram for overall survival in the training and validation cohorts were 0.713 and 0.729, respectively.
Conclusions:
We analyzed the clinical characteristics of SPMs and developed a precise prediction nomogram, with a good predictive performance. The nomogram we developed may help clinicians provide personalized decisions and clinical treatment for LT recipients.
Key Words: liver transplantation, second primary malignancies, risk factors, nomogram, prognosis prediction, seer
Hepatocellular carcinoma (HCC) has been ranked as the sixth most common malignancy and the third most frequent cause of cancer death worldwide. It is reported that in 2020, the number of new cases and deaths of liver cancer globally were >950,000 and >830,000, respectively.1 Liver transplantation (LT) is the first-choice treatment for patients with end-stage liver disease, acute liver failure, or HCC. Initially, infectious factors were the main cause of death in LT patients, but with the advancement of surgical techniques and the use of anti-infection and antirejection drugs, patients’ survival rates have been significantly improved, and second primary malignancies (SPMs) after transplantation will become one of the main causes of death of LT patients.2–4 An analysis of transplant recipients showed that they had an 11 times higher risk of SPMs than the general population.5 The total incidence rate of SPMs in LT recipients was between 3.1% and 14.4%, and at 5 and 10 years was estimated to be 10% to 14.6% and 20% to 32%, respectively.6,7 While previous studies have focused on the incidence and types of SPMs after LT,8–10 the survival status and related factors affecting the survival of patients with SPMs have not been studied. Thus, clinicians must determine the prognostic factors of SPMs.
The nomogram is an effective and reliable tool, which can combine statistics with clinicopathologic characteristics to explore the prognostic factors of patients and provide the basis for the treatment of patients. Nomogram-based clinical modeling has been currently widely used in clinical research.11–15 The Surveillance, Epidemiology, and End Results (SEER) database is an authoritative source of cancer incidence and survival and covers ∼34.6% of the US population. Therefore, the SEER database can provide sufficient cases for the development of predictive models for SPMs.
In this study, relevant data from the SEER database were analyzed, to determine the factors affecting the survival of patients with SPMs, and to construct a nomogram to predict the overall survival (OS) of those patients. The study outcomes will aid decision-making for clinicians and nurses in the clinical consultation and care management of LT SPM patients.
MATERIALS AND METHODS
Population and Inclusion Criteria
Data were retrieved from the SEER database, which includes patients’ demographic and clinical information from 18 cancer registries, and were released in April 2021 based on the November 2020 submission.
Our study included data from adult patients who were diagnosed with primary HCC and underwent LT between 2004 and 2015. The inclusion criteria were as follows: (1) International Classification of Diseases for Oncology, Third Edition [ICD-O-3] site code C22.0 and histologic type ICD-O-3 codes 8170-8175; (2) patient had undergone LT; (3) diagnosis occurred from between 2004 and 2015, and the age at diagnosis was 18 years and above; and (4) first primary tumor. The exclusion criteria were as follows: (1) the latency of SPMs was <2 months or the survival time was unknown; (2) reporting source was autopsy or death certificate only; (3) unclear follow-up information; and (4) multiple SPMs after transplantation. Data was extracted through the 2020 Submission Database of SEER, and the follow-up date was determined on November 31, 2020.
Data Collection and Definition of Variables
We extracted demographic information (age, sex, year of diagnosis, race, marital status, latency of SPMs), clinicopathological characteristics (grade, American Joint Committee on Cancer [AJCC] sixth TNM staging system, T stage, tumor size), and therapy information (radiotherapy, chemotherapy) from the eligible cases.
For demographic characteristics, we extracted age at initial diagnosis(less than 65, 65 y and above), sex (male, female), year of diagnosis (2004-2008, 2009-2015), race (white, black, Asian, or Pacific Islander and American Indian), marital status (married, unmarried), and latency of SPMs (<60, ≥60 mo). Clinicopathologic characteristics involved initial diagnosis tumor grade (I-II: well-differentiated or moderately differentiated; III-IV: poorly differentiated or undifferentiated), AJCC sixth TNM staging system (I-II, III-IV), T stage (T1-T2, T3-T4), and tumor size (≤20, >20 mm). Treatment information included radiotherapy (yes, no), and chemotherapy (yes, no). Because distant metastasis is a surgical contraindication to LT, we excluded N or M stages positive patients, and all of the SPM patients we selected were N or M stages negative patients. Our study is based on patients with only 1 malignancy after LT, so we exclude patients with multiple malignancies after LT. The latency of SPMs was defined as the interval from LT to the appearance of SPMs. We calculated the interval from LT to SPMs in months. Since the SEER database is a public anonymized database, this study did not require ethical approval. The author gained SEER database permission (User name: 16954-Nov2020).
Statistical Analysis
We categorized all LT patients into only one primary malignancy cohort (n=1857) and SPMs cohort (n=221) according to whether they had SPMs. Categorical variables are presented as frequencies (%) (analyzed by χ2 test or Fisher exact test) and continuous variables are presented as mean±SD or median (interquartile range [IQR]) (analyzed by independent samples t test or Mann-Whitney U test).
In the prognostic analysis of patients with SPMs, 221 patients were randomly assigned at a 7:3 ratio to the training cohort (n=154) or the validation cohort (n=67). In the overall SPMs cohort, Kaplan-Meier analysis was used to generate the survival curve of each factor, and the log-rank test was used to calculate the difference between each factor. In the training cohort, significant variables from univariate Cox regression analysis were incorporated into multivariate Cox regression analysis to identify independent prognostic factors. A nomogram for predicting the 2-, 3- and 5-year OS was constructed based on independent prognostic factors from the training cohort. The performance of the nomogram was measured using the concordance index (C-index). The predictive accuracy of the nomogram was determined by the receiver operating characteristic (ROC) curves, and the area under the curve (AUC). By drawing the calibration curves (1000 bootstrap auto-sampling method), the consistency between the observed and predicted survival was evaluated.16 The net benefit under different threshold probabilities was quantified by decision curve analysis to evaluate the clinical application value of the nomogram17 and the validation cohort conducted external tests on it to prove OS nomogram extrapolation. Further, based on the median risk score, the patient data were divided into high-risk and low-risk groups, and the survival curve of the log-rank test was used to verify the predictive value of the nomogram. This study collected data through SEER-Stat, version 8.3.9 software. SPSS, version 23.0 (IBM Corp.) and R software, version 4.1.2 (https://www.r-project.org/) were used for data cleaning, sorting, and analysis. Two-sided P values<0.05 were considered statistically significant.
RESULTS
Characteristics of the Study Population
According to the above inclusion and exclusion criteria, this retrospective study preliminarily identified 2357 adult patients who were diagnosed with primary HCC and underwent LT between 2004 and 2015 in the SEER database. Of these, 247 were excluded due to unknown race, marital status, AJCC, sixth TNM staging system, T stage, tumor size, and unknown or <2 months survival time. In addition, 32 patients with multiple SPMs after LT were excluded. Finally, data from 2078 patients were included in this study. The data collection and screening process is shown in Figure 1. The demographic and clinicopathological data are shown in Table 1.
FIGURE 1.
Flowchart detailing the selection of the patients in this study. AJCC indicates American Joint Committee on Cancer; HCC, hepatocellular carcinoma; OOPM, only one primary malignancy; SEER, Surveillance, Epidemiology, and End Results; SPM, second primary malignancy.
TABLE 1.
Demographic and Clinicopathologic Characteristics of Patients With Liver Transplantation
| n (%) | ||||
|---|---|---|---|---|
| Variables | Overall (N=2078) | OOPM (n=1857) | SPMs (n=221) | P |
| Sex | ||||
| Female | 438 (21.1) | 406 (21.9) | 32 (14.5) | 0.014 |
| Male | 1640 (78.9) | 1451 (78.1) | 189 (85.5) | |
| Age | ||||
| <65 | 1760 (84.7) | 1586 (85.4) | 174 (78.7) | 0.012 |
| ≥65 | 318 (15.3) | 271 (14.6) | 47 (21.3) | |
| Year of diagnosis | ||||
| 2004-2008 | 847 (40.8) | 732 (39.4) | 115 (52.0) | <0.001 |
| 2009-2015 | 1231 (59.2) | 1125 (60.6) | 106 (48.0) | |
| Race | ||||
| API or AI | 266 (12.8) | 243 (13.1) | 23 (10.4) | 0.243 |
| Black | 213 (10.3) | 195 (10.5) | 18 (8.1) | |
| White | 1599 (76.9) | 1419 (76.4) | 180 (81.4) | |
| Marital status | ||||
| Married | 1434 (69.0) | 1270 (68.4) | 164 (74.2) | 0.091 |
| Unmarried | 644 (31.0) | 587 (31.6) | 57 (25.8) | |
| Grade | ||||
| I-II | 1274 (61.3) | 1144 (61.6) | 130 (58.8) | 0.675 |
| III-IV | 180 (8.7) | 161 (8.7) | 19 (8.6) | |
| Other | 624 (30.0) | 552 (29.7) | 72 (32.6) | |
| AJCC, sixth TNM stage | ||||
| I-II | 1752 (94.3) | 1752 (94.3) | 211 (95.5) | 0.590 |
| III-IV | 115 (5.5) | 105 (5.7) | 10 (4.5) | |
| T stage | ||||
| T1-T2 | 1984 (95.5) | 1769 (95.3) | 215 (97.3) | 0.231 |
| T3-T4 | 94 (4.5) | 88 (4.7) | 6 (2.7) | |
| Size (mm) | ||||
| >20 | 1184 (63.8) | 1184 (63.8) | 155 (70.1) | 0.072 |
| ≤20 | 739 (35.6) | 673 (36.2) | 66 (29.9) | |
| Radiotherapy | ||||
| No | 2010 (96.7) | 1795 (96.7) | 215 (97.3) | 0.770 |
| Yes | 68 (3.3) | 62 (3.3) | 6 (2.7) | |
| Chemotherapy | ||||
| No | 1143 (55.0) | 1019 (54.9) | 124 (56.1) | 0.781 |
| Yes | 935 (45.0) | 838 (45.1) | 97 (43.9) | |
| Latency (mo) | ||||
| >60 | — | — | 80 (36.2) | — |
| ≤60 | — | — | 141 (63.8) | — |
AJCC indicates American Joint Committee on Cancer; API or AI, Asian or Pacific Islander and American Indian; OOPM, only one primary malignancy; SPM, second primary malignancy.
Morbidity and Distributions of SPMs
Among 2078 patients with HCC after LT, 221 (10.64%) developed SPMs after a median follow-up of 88 months (IQR: 54 to 126 mo). The 2-, 3-, and 5-year morbidities were 3.20%, 4.80%, and 8.20%, respectively. The median age at diagnosis of SPMs was 63 years (IQR: 58 to 67 y). Table 2 showed the types of SPMs which occurred in at least 5 cases, overall, the 3 most common SPMs were lung cancer, prostate cancer and non-Hodgkin lymphoma (NHL). The 3 most common SPMs in men were prostate cancer, lung cancer, and NHL; and in women the most common types were NHL, breast cancer, and lung cancer.
TABLE 2.
The Locations of SPMs Which Occurred in At Least 5 Cases
| Location of SPMs | No. cases |
|---|---|
| Lung and bronchus | 35 |
| Prostate | 33 |
| NHL | 28 |
| Melanoma of the skin | 12 |
| Urinary bladder | 11 |
| Liver | 10 |
| Pancreas | 9 |
| Miscellaneous | 9 |
| Tonsil | 6 |
| Kidney and renal pelvis | 5 |
| Other nonepithelial skin | 5 |
| Breast | 5 |
| Intrahepatic bile duct | 5 |
There were 221 sites of SPMs, and the most common site was the lung and bronchus (35), followed by prostate (33), and NHL (28).
NHL indicates non-Hodgkin lymphoma; SPM, second primary malignancy.
Construction and Verification of SPMs Nomogram Prediction Model
Data from 221 eligible SPM patients were included in the prognostic analysis (overall SPMs cohort). The overall SPMs cohort were randomly divided into training cohort (n=154) or validation cohort (n=67) at a ratio of 7:3. The demographic and clinicopathologic characteristics of the SPM patients were shown in Table 3. The overall SPMs cohort included 189 males (85.5%), 47 cases (21.3%) were older than 65 years old, and in 66 cases (29.9%) tumor size was ≤20 mm. The median survival time was 42 months. The Kaplan-Meier analysis was used to generate the survival curve of each factor the overall SPMs cohort (Fig. 2).
TABLE 3.
Demographic and Clinicopathologic Characteristics of SPM Patients in the Training and Validation Cohorts
| n (%) | |||
|---|---|---|---|
| Variables | Training cohort (n=154) | Validation cohort (n=67) | Overall SPMs cohort (N=221) |
| Sex | |||
| Female | 25 (16.2) | 7 (10.4) | 32 (14.5) |
| Male | 129 (83.8) | 60 (89.6) | 189 (85.5) |
| Age | |||
| <65 | 115 (74.7) | 59 (88.1) | 174 (78.7) |
| ≥65 | 39 (25.3) | 8 (11.9) | 47 (21.3) |
| Year of diagnosis | |||
| 2004-2008 | 79 (51.3) | 36 (53.7) | 115 (52.0) |
| 2009-2015 | 75 (48.7) | 31 (46.3) | 106 (48.0) |
| Race | |||
| API or AI | 20 (13.0) | 3 (4.5) | 23 (10.4) |
| Black | 12 (7.8) | 6 (9.0) | 18 (8.1) |
| White | 122 (79.2) | 58 (86.6) | 180 (81.4) |
| Marital status | |||
| Married | 115 (74.7) | 49 (73.1) | 164 (74.2) |
| Unmarried | 39 (25.3) | 18 (26.9) | 57 (25.8) |
| Grade | |||
| I-II | 90 (58.4) | 40 (59.7) | 130 (58.8) |
| III-IV | 15 (9.7) | 4 (6.0) | 19 (8.6) |
| Other | 49 (31.8) | 23 (34.3) | 72 (32.6) |
| AJCC, sixth TNM stage | |||
| I-II | 146 (94.8) | 65 (97.0) | 211 (95.5) |
| III-IV | 8 (5.2) | 2 (3.0) | 10 (4.5) |
| T stage | |||
| T1-T2 | 150 (97.4) | 65 (97.0) | 215 (97.3) |
| T3-T4 | 4 (2.6) | 2 (3.0) | 6 (2.7) |
| Size (mm) | |||
| >20 | 115 (74.7) | 40 (59.7) | 155 (70.1) |
| ≤20 | 39 (25.3) | 27 (40.3) | 66 (29.9) |
| Radiotherapy | |||
| No | 150 (97.4) | 65 (97.0) | 215 (97.3) |
| Yes | 4 (2.6) | 2 (3.0) | 6 (2.7) |
| Chemotherapy | |||
| No | 76 (49.4) | 48 (71.6) | 124 (56.1) |
| Yes | 78 (50.6) | 19 (28.4) | 97 (43.9) |
| Latency (mo) | |||
| >60 | 49 (31.8) | 31 (46.3) | 80 (36.2) |
| ≤60 | 105 (68.2) | 36 (53.7) | 141 (63.8) |
AJCC, American Joint Committee on Cancer; API or AI, Asian or Pacific Islander and American Indian; SPM, second primary malignancy.
FIGURE 2.
Kaplan-Meier survival curves of overall survival based on age (A), marital status (B), year of diagnosis (C), T stage (D), chemotherapy (E), and latency of second primary malignancies (F) in overall second primary malignancy cohort.
The training cohort included 129 males (83.8%), 39 cases (25.3%) were older than 65 years old, and in 39 cases (25.3%) tumor size was ≤20 mm. Significant variables from the univariate Cox regression analysis were included in the multivariate Cox regression analysis to determine the independent prognostic factors (age, marital status, year of diagnosis, and latency) (Table 4). In the overall SPMs cohort, T stage was the significant variables in the Kaplan-Meier analysis, so we included T stage in the construction of the nomogram.
TABLE 4.
Univariate and Multivariate Cox Regression Analyses of SPM Patients Based on Clinicopathologic Characteristics in the Training Cohort
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| Variables | HR | 95% CI | P | HR | 95% CI | P |
| Sex | ||||||
| Female | Reference | |||||
| Male | 0.757 | 0.425-1.348 | 0.345 | |||
| Age | ||||||
| <65 | Reference | Reference | ||||
| ≥65 | 2.008 | 1.270-3.178 | 0.003 | 1.994 | 1.253-3.174 | 0.004 |
| Year of diagnosis | ||||||
| 2004-2008 | Reference | Reference | ||||
| 2009-2015 | 2.019 | 1.249-3.264 | 0.004 | 1.821 | 1.106-2.999 | 0.018 |
| Race | ||||||
| API or AI | Reference | |||||
| Black | 2.188 | 0.792-6.045 | 0.131 | |||
| White | 1.817 | 0.872-3.788 | 0.111 | |||
| Marital status | ||||||
| Married | Reference | Reference | ||||
| Unmarried | 1.846 | 1.158-2.944 | 0.01 | 1.785 | 1.110-2.870 | 0.017 |
| Grade | ||||||
| I-II | Reference | |||||
| III-IV | 1.0378 | 0.491-2.194 | 0.923 | |||
| Other | 0.9926 | 0.615-1.601 | 0.976 | |||
| AJCC, sixth TNM stage | ||||||
| I-II | Reference | |||||
| III-IV | 1.014 | 0.371-2.773 | 0.978 | |||
| T stage | ||||||
| T1-T2 | Reference | |||||
| T3-T4 | 1.401 | 0.344-5.715 | 0.638 | |||
| Size (mm) | ||||||
| >20 | Reference | |||||
| ≤20 | 0.834 | 0.494-1.407 | 0.496 | |||
| Radiotherapy | ||||||
| No | Reference | |||||
| Yes | 1.175 | 0.288-4.794 | 0.822 | |||
| Chemotherapy | ||||||
| No | Reference | |||||
| Yes | 1.357 | 0.878-2.098 | 0.170 | |||
| Latency (mo) | ||||||
| >60 | Reference | Reference | ||||
| ≤60 | 2.415 | 1.463-3.986 | 0.001 | 1.448 | 1.337-3.726 | 0.002 |
AJCC indicates American Joint Committee on Cancer; API or AI, Asian or Pacific Islander and American Indian; HR, hazard ratio; SPM, second primary malignancy.
Based on the prognostic factor determined in the training cohort, the OS nomogram was established (Fig. 3). The C-index of the nomogram was 0.713 (95% CI: 0.667-0.759). The 2-, 3-, and 5-year AUC values were 0.690 (95% CI: 0.576-0.804), 0.767 (95% CI: 0.684-0.849), and 0.867 (95% CI: 0.809-0.926), respectively (Figs. 4A–C). The calibration curve showed a high degree of agreement between the observed and predicted results (Figs. 5A–C). The decision curve analysis demonstrated good practical value (Figs. 6A–C).
FIGURE 3.
Nomogram for predicting 2-, 3-, and 5-year overall survival of second primary malignancy patients.
FIGURE 4.
Receiver operating characteristic curves for 2-, 3-, and 5-year overall survival of second primary malignancy patients in the training and validation cohorts. A–C, Receiver operating characteristic curves of the nomogram for predicting 2-, 3-, and 5-year survival in the training cohort. D–F, Receiver operating characteristic curves of the nomogram for predicting 2-, 3-, and 5-year survival in the validation cohort. AUC indicates area under the curve.
FIGURE 5.
Calibration curves for 2-, 3-, and 5-year overall survival of second primary malignancy patients in the training and validation cohorts. A–C, Calibration curves of the nomogram for predicting 2-, 3-, and 5-year survival in the training cohort. D–F, Calibration curves of the nomogram for predicting 2-, 3-, and 5-year survival in the validation cohort.
FIGURE 6.
Evaluation of the prognostic nomogram. A–C, Decision curve analysis curves for 2-, 3-, and 5-year overall survival of SPM patients in the training cohort. Kaplan-Meier curves of overall survival for risk classification based on the nomogram scores. D, In the training cohort. E, In the validation cohort. F, In overall SPMs cohort. SPM indicates second primary malignancy.
In the validation cohort, we verify the OS prognostic model using the same model data as the established nomogram in the training cohort. The C-index was 0.729 (95% CI: 0.648-0.811). The ROC curves and AUC indicated that the model has a good discrimination ability (AUC were 0.826 for a 2-y OS, 0.825 for a 3-y OS, and 0.850 for a 5-y OS) (Figs. 4D–F). The calibration curve showed that the predicted survival rate is in good agreement with the actual survival rate (Figs. 5D–F).
In addition, according to the median risk score (145.579), all patients with SPMs were divided into high-risk (total score>145.579) and low-risk (total score≤145.579) groups. The Kaplan-Meier analysis showed that patients in the high-risk group had a worse prognosis than the patients in the low-risk group (Figs. 6D–F). These results showed that the validation results of the prediction model in the validation set are excellent, which proved the extrapolation of the model.
DISCUSSION
In recent years, with the progress of diagnosis and treatment technology, the survival time of LT patients has been extended. SPMs are a long-term complication after transplant, and the incidence rate is increasing.18,19 It is crucial to use rigorous statistical prediction methods, to understand patients’ prognosis in advance and achieve early detection, diagnosis, and treatment. In this study, we explored the prognostic factors affecting the survival of SPMs and constructed an OS nomogram. Clinicians could wield the nomogram to evaluate SPM patients’ OS and identify high-risk patients.
In our study using the SEER database, the incidence rate of SPMs in HCC patients after LT was 10.64%, similar to France (10.7%).2 Compared with our study, the prognostic factors of French patients were different. The risk of SPMs was significantly higher in men and in patients transplanted for alcoholic liver disease in France. This may be related to tobacco and alcohol. In our study, the 3 most common SPMS were lung cancer, prostate cancer, and NHL. The 3 most common SPMs in men were prostate cancer, lung cancer, and NHL. These results suggest that lifestyle can affect the occurrence of SPMs. Herrero et al20 showed that smoking increased the incidence of lung, esophageal, and urinary tract cancers after LT. While in women the most common types were NHL, breast cancer, and lung cancer. These results may indicate hormone levels can affect the occurrence of SPMs. Mellemkjær et al21 demonstrated that hormonal changes may stimulate the development of specific cancers.
In the overall SPMs cohort, Kaplan-Meier analysis showed that patients with LT in 2009-2015 had a worse prognosis. Few studies focused on the relationship between the year of transplantation and the prognosis of patients, we only found 2 similar studies. An Italian multicenter study22 from 1985 to 2014 showed that when considering the year of transplantation alone, patients with HCC who received transplantation after 2000 had a higher risk of developing SPMs than patients who received transplantation before 2000, but there was no statistical significance. This study only illustrated the risk of occurrence of SPMs by the year of transplantation but did not explain the prognosis of patients. Benlloch et al23 showed that 772 patients received liver transplants between 1991 and 2001 and that patients who received transplants before 1995 had a longer latency period for SPMs than those transplanted after 1995. Moreover, hematological tumors appeared earlier than solid tumors, with a higher incidence of patients transplanted after 1995 than before 1995 and worse survival. This may be due to the use of more effective first-line immunosuppressive drugs, which has promoted the early development of malignant tumors in recent years. In addition, patients who received chemotherapy had worse survival than those who did not, which might be because patients who received chemotherapy had reached an advanced stage of the disease, so the prognosis was poor. It may also be related to viral reactivation by chemotherapeutic drugs. An in vitro cell model experiment24 has indicated that mitomycin and 5-fluorouracil could directly up-regulate the replication and expression of hepatitis B virus (HBV) in vitro, and the expression of HBV capsid and hepatitis B e antigen in cells increased with increasing concentrations of mitomycin and 5-fluorouracil. Spanish scholars25 have also indicated that cytotoxic chemotherapy or immunosuppressive therapy is associated with the reactivation of HBV infection.
In the prognostic analysis, the prognostic nomogram showed that older age, higher T stage, and shorter latency SPMs were associated with poorer patient outcomes. Older patients are more likely to develop SPMs, which may be related to the degeneration of various body functions, low immune function, and more cardiovascular and cerebrovascular diseases.26 Some studies also indicated that age is always negatively associated with survival. A study on the survival of HCC showed that age over 65 is an independent predictor of poor prognosis.27 Another retrospective cohort study also pointed out that the age of patients is an important factor affecting the prognosis of pancreatic cancer, and the survival period of pancreatic cancer patients is negatively correlated with the age of diagnosis.28 In addition, several studies have shown that longer latency of SPMs are associated with better prognosis.29,30 Similar results were found in our results, where patients with a latency period of >60 months had a better prognosis compared with those with a latency period of ≤60 months. This may be because SPMs are a form of trauma to liver transplant patients. The shorter the interval since surgery, the worse the survival rate. There are many systems related to the staging of HCC,31 and the TNM staging system can be applied in the SEER database. Studies on the prognosis of patients with HCC showed that the T stage was an important factor related to prognosis, and the later T stage was, the worse the prognosis was.32,33 In the overall SPMs cohort, Kaplan-Meier analysis indicated that T stage (P=0.034) was a significant factor, and a higher T stage was also associated with worse prognosis. In addition, when T stage was not included in the nomogram, the C-index of the nomogram was 0.707 (95% CI: 0.683-0.731); when T stage was included into the nomogram, the C-index of the nomogram was 0.713 (95% CI: 0.667-0.759), which was slightly higher than that without T stage, so we included T stage into the nomogram.
Married patients have better prognosis than unmarried patients, and some studies have similar results to ours. Studies on marital status and prognosis of patients with HCC34,35 show that marital status is an independent prognostic factor for survival of patients with HCC, and married patients have better 5-year tumor-specific survival than unmarried patients. In addition, a questionnaire from Germany36 shows that marital status affects the health-related quality of life of liver transplant patients. Unmarried patients may lack the help of family members and friends, and their quality of life may decline, which may affect the prognosis of patients.
The C-index of the nomogram was 0.713 (95% CI: 0.667-0.759) in the training cohort and 0.729 (95% CI: 0.648-0.811) in the validation cohort, indicates good discriminability. The calibration curves and ROC curves indicate that the nomogram has a predictive effect. According to the median risk score of the model, the prognosis of the low-risk group is significantly better than that of the high-risk group, again indicating good discriminability.
In addition, virus infection37,38 and the use of immunosuppressants22,39 will also lead to the occurrence of SPMs. The use of immunosuppressants after transplantation have been found to affect the occurrence of SPMs and the survival of patients. Schnitzbauer et al40 showed that mammalian target of rapamycin inhibitors have better antitumor effects than calcineurin inhibitor. The latest consensus also stated that41 calcineurin inhibitor in liver transplant patients should be reduced to the lowest possible dose to reduce the risk of SPMs; mammalian target of rapamycin does not seem to increase cancer risk and may prevent SPMs after liver transplantation. We were unable to study the effect of immunosuppressive agents on patient survival as no relevant data were recorded in the SEER database. Preoperative status of transplant recipients is also associated with SPMs. A retrospective study42 showed that patients with higher preoperative alpha-fetoprotein (AFP) levels had lower postoperative survival. Berry and Ioannou43 also showed that preoperative AFP level could predict the survival rate of patients after transplantation, and patients with higher AFP level had worse survival rate after transplantation. Straś et al44 found that preoperative microvascular infiltration and an increase in neutrophil-to-lymphocyte ratio may contribute to the recurrence of HCC after liver transplantation. Li et al45 indicated that tumor size is an important independent risk factor affecting the prognosis of patients with liver cancer after liver transplantation. Patients with small tumor size had significantly better OS than those with large tumor size. So, possibly one could prioritize offering transplants to smaller tumors. This may be because the larger the tumor diameter, the easier it is to break through the capsule and the postoperative recurrence rate.
However, our research also has limitations. First, the SEER database does not provide sufficient data to assess liver function, including the Child-Pugh score and prothrombin time/international normalized ratio, Model for End-stage Liver Disease score, which is important for the clinical decision-making process. Second, the SEER database lacks detailed data on treatments, such as chemotherapy regimens, targeted therapy, and use of immunosuppressants. In addition, in our study many of the factors in the nomogram are not modifiable—that is, one can’t correct age or marital status. Last, our study is a retrospective study, which needs further external multicenter prospective validation.
CONCLUSIONS
In this study, we analyzed the prognostic factors of SPM patients. OS nomogram was constructed and validated based on independent prognostic factors to predict survival, and the proposed nomogram resulted in a good performance. The nomogram can help clinicians evaluate the poor prognostic factors of SPMs, and guide the clinic to formulate the best-individualized treatment strategy.
ACKNOWLEDGMENTS
Thanks to all the staff of The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital who contributed to the study. The authors are also very grateful to the SEER program for approving the registration and the SEER database.
Footnotes
The authors declare no conflicts of interest.
Contributor Information
Qingbao Ding, Email: dingqingbao1016@126.com.
Keyu Wang, Email: 894043440@qq.com.
Yupeng Li, Email: liyupeng1008@126.com.
Peng Peng, Email: pengpeng_mph@126.com.
Dongyuan Zhang, Email: sjtuzdy@sina.com.
Donglei Chang, Email: 309700283@qq.com.
Wentao Wang, Email: qfsgandanwaike@163.com.
Lei Ren, Email: renleipy@163.com.
Fang Tang, Email: tangfangsdu@126.com.
Ziqiang Li, Email: 1358@sdhospital.com.cn.
REFERENCES
- 1. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–249. [DOI] [PubMed] [Google Scholar]
- 2. Sérée O, Altieri M, Guillaume E, et al. Longterm risk of solid organ de novo malignancies after liver transplantation: a French National Study on 11,226 patients. Liver Transpl. 2018;24:1425–1436. [DOI] [PubMed] [Google Scholar]
- 3. Buell JF, Gross TG, Woodle ES. Malignancy after transplantation. Transplantation. 2005;80(suppl):S254–S264. [DOI] [PubMed] [Google Scholar]
- 4. Chapman JR, Webster AC, Wong G. Cancer in the transplant recipient. Cold Spring Harb Perspect Med. 2013;3:a015677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Zhou J, Hu Z, Zhang Q, et al. Spectrum of de novo cancers and predictors in liver transplantation: analysis of the Scientific Registry of Transplant Recipients Database. PLoS One. 2016;11:e0155179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Engels EA, Pfeiffer RM, Fraumeni JF, Jr, et al. Spectrum of cancer risk among US solid organ transplant recipients. JAMA. 2011;306:1891–1901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Piselli P, Verdirosi D, Cimaglia C, et al. Epidemiology of de novo malignancies after solid-organ transplantation: immunosuppression, infection and other risk factors. Best Pract Res Clin Obstet Gynaecol. 2014;28:1251–1265. [DOI] [PubMed] [Google Scholar]
- 8. Oweira H, Schmidt J, Helbling D, et al. Impact of liver transplantation on the risk of second malignant tumors among hepatocellular carcinoma patients. Expert Rev Gastroenterol Hepatol. 2017;11:865–869. [DOI] [PubMed] [Google Scholar]
- 9. Park B, Yoon J, Choi D, et al. De novo cancer incidence after kidney and liver transplantation: results from a nationwide population based data. Sci Rep. 2019;9:17202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Wahab MA, Abdel-Khalek EE, Elshoubary M, et al. Predictive factors of de novo malignancies after living-donor liver transplantation: a single-center experience. Transplant Proc. 2021;53:636–644. [DOI] [PubMed] [Google Scholar]
- 11. He T, Chen T, Liu X, et al. A web-based prediction model for cancer-specific survival of elderly patients with early hepatocellular carcinoma: a study based on SEER database. Front Public Health. 2021;9:789026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yang Z, Liu X, Yang X, et al. Second primary malignancies after ovarian cancer: a SEER-based analysis (1975-2016). Taiwan J Obstet Gynecol. 2022;61:80–85. [DOI] [PubMed] [Google Scholar]
- 13. Choi YH, Lee SH. Editorial on “A new nomogram from the SEER database for predicting the prognosis of gallbladder cancer patients after surgery”. Ann Transl Med. 2020;8:522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Li H, He Y, Huang L, et al. The nomogram model predicting overall survival and guiding clinical decision in patients with glioblastoma based on the SEER database. Front Oncol. 2020;10:1051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lu YJ, Yang Y, Yuan YH, et al. A novel nomogram based on SEER database for the prediction of liver metastasis in patients with small-cell lung cancer. Ann Palliat Med. 2020;9:3123–3137. [DOI] [PubMed] [Google Scholar]
- 16. Steyerberg EW, Vickers AJ, Cook NR, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21:128–138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Vickers AJ, Cronin AM, Elkin EB, et al. Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers. BMC Med Inform Decis Mak. 2008;8:53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Kim S, Rovgaliyev B, Lee JM, et al. Clinical significance of de novo malignancy after liver transplant: a single-center study. Transplant Proc. 2021;53:200–206. [DOI] [PubMed] [Google Scholar]
- 19. Kwak BJ, Kim DG, Han JH, et al. Clinical outcome of 1,000 consecutive cases of liver transplantation: a single center experience. Ann Surg Treat Res. 2018;95:267–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Herrero JI, Pardo F, D’Avola D, et al. Risk factors of lung, head and neck, esophageal, and kidney and urinary tract carcinomas after liver transplantation: the effect of smoking withdrawal. Liver Transpl. 2011;17:402–408. [DOI] [PubMed] [Google Scholar]
- 21. Mellemkjær L, Eibye S, Albieri V, et al. Pregnancy-associated cancer and the risk of second primary cancer. Cancer Causes Control. 2022;33:63–71. [DOI] [PubMed] [Google Scholar]
- 22. Shalaby S, Taborelli M, Zanetto A, et al. Hepatocellular carcinoma and the risk of de novo malignancies after liver transplantation—a multicenter cohort study. Transpl Int. 2021;34:743–753. [DOI] [PubMed] [Google Scholar]
- 23. Benlloch S, Berenguer M, Prieto M, et al. De novo internal neoplasms after liver transplantation: increased risk and aggressive behavior in recent years? Am J Transplant. 2004;4:596–604. [DOI] [PubMed] [Google Scholar]
- 24. Ruan J, Sun S, Cheng X, et al. Sun D. Mitomycin, 5-fluorouracil, leflunomide, and mycophenolic acid directly promote hepatitis B virus replication and expression in vitro. Virol J. 2020;17:89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Manzano-Alonso ML, Castellano-Tortajada G. Reactivation of hepatitis B virus infection after cytotoxic chemotherapy or immunosuppressive therapy. World J Gastroenterol. 2011;17:1531–1537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Durand F, Levitsky J, Cauchy F, et al. Age and liver transplantation. J Hepatol. 2019;70:745–758. [DOI] [PubMed] [Google Scholar]
- 27. Ding J, Wen Z. Survival improvement and prognosis for hepatocellular carcinoma: analysis of the SEER database. BMC Cancer. 2021;21:1157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Wang H, Liu J, Xia G, et al. Survival of pancreatic cancer patients is negatively correlated with age at diagnosis: a population-based retrospective study. Sci Rep. 2020;10:7048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Zheng R, Li H, Ye Y, et al. Clinicopathological features and prognostic analysis of 77 patients with multiple primary cancers. J BUON. 2020;25:2110–2116. [PubMed] [Google Scholar]
- 30. Goldfarb M, Rosenberg AS, Li Q, et al. Impact of latency time on survival for adolescents and young adults with a second primary malignancy. Cancer. 2018;124:1260–1268. [DOI] [PubMed] [Google Scholar]
- 31. Selçuk H. Prognostic factors and staging systems in hepatocellular carcinoma. Exp Clin Transplant. 2017;15(suppl 2):45–49. [DOI] [PubMed] [Google Scholar]
- 32. Yan B, Su BB, Bai DS, et al. A practical nomogram and risk stratification system predicting the cancer-specific survival for patients with early hepatocellular carcinoma. Cancer Med. 2021;10:496–506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Liu K, Huang G, Chang P, et al. Construction and validation of a nomogram for predicting cancer-specific survival in hepatocellular carcinoma patients. Sci Rep. 2020;10:21376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Zhang W, Wang X, Huang R, et al. Prognostic value of marital status on stage at diagnosis in hepatocellular carcinoma. Sci Rep. 2017;7:41695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Liang Y, Wu X, Lu C, et al. Impact of marital status on the prognosis of liver cancer patients without surgery and the critical window. Ann Palliat Med. 2021;10:2990–2999. [DOI] [PubMed] [Google Scholar]
- 36. Heits N, Meer G, Bernsmeier A, et al. Mode of allocation and social demographic factors correlate with impaired quality of life after liver transplantation. Health Qual Life Outcomes. 2015;13:162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Razonable RR. Cytomegalovirus infection after liver transplantation: current concepts and challenges. World J Gastroenterol. 2008;14:4849–4860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Polak WG, Fondevila C, Karam V, et al. Impact of COVID-19 on liver transplantation in Europe: alert from an early survey of European Liver and Intestine Transplantation Association and European Liver Transplant Registry. Transpl Int. 2020;33:1244–1252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Wimmer CD, Angele MK, Schwarz B, et al. Impact of cyclosporine versus tacrolimus on the incidence of de novo malignancy following liver transplantation: a single center experience with 609 patients. Transpl Int. 2013;26:999–1006. [DOI] [PubMed] [Google Scholar]
- 40. Schnitzbauer AA, Filmann N, Adam R, et al. mTOR Inhibition is most beneficial after liver transplantation for hepatocellular carcinoma in patients with active tumors. Ann Surg. 2020;272:855–862. [DOI] [PubMed] [Google Scholar]
- 41. Colmenero J, Tabrizian P, Bhangui P, et al. De novo malignancy after liver transplantation: risk assessment, prevention, and management-guidelines from the ILTS-SETH Consensus Conference. Transplantation. 2022;106:e30–e45. [DOI] [PubMed] [Google Scholar]
- 42. Mahmud N, John B, Taddei TH, et al. Pre-transplant alpha-fetoprotein is associated with post-transplant hepatocellular carcinoma recurrence mortality. Clin Transplant. 2019;33:e13634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Berry K, Ioannou GN. Serum alpha-fetoprotein level independently predicts posttransplant survival in patients with hepatocellular carcinoma. Liver Transpl. 2013;19:634–645. [DOI] [PubMed] [Google Scholar]
- 44. Straś WA, Wasiak D, Łągiewska B, et al. Recurrence of hepatocellular carcinoma after liver transplantation: risk factors and predictive models. Ann Transplant. 2022;27:e934924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Li X, Huang L, Leng X. Analysis of prognostic factors of more/equal to10 years of survival for liver cancer patients after liver transplantation. J Cancer Res Clin Oncol. 2018;144:2465–2474. [DOI] [PMC free article] [PubMed] [Google Scholar]