Malignant mesothelioma (MMe) is an uncommon and aggressive cancer (1) mostly associated with asbestos exposure (2, 3). Other causes of mesothelioma include exposure to other types of mineral fibers, radiation, chronic pleural inflammation, or germline and somatic inactivation of the BRCA1-associated protein 1 gene (BAP1) (4, 5). Most cases originate from the pleura, with a lower prevalence at other anatomical sites such as the peritoneum, genital organs, soft tissue, and other organs of the body (6).

Previous studies mostly focused on malignant mesothelioma originating in the pleura or peritoneum; and age, gender, histology, differentiation, lymph node status, chemotherapy, and other characteristics have been demonstrated to be independent prognostic factors (7–11).

It is worth mentioning that the overall concordance between biopsies and surgical resection specimens is 80.6% (12). Clinical N0 stage (cN0) malignant pleural mesothelioma (MPM) patients with lymph node metastasis have a poor prognosis (13). In addition, Leuzzi et al. found that the pathological N stage (pN), rather than the clinical N stage (cN), was the potential predictive factor affecting the overall survival (OS) (14). To avoid this error, all candidates enrolled in our research were diagnosed by positive histology recorded in the SEER database.

The prediction model of MMe has been reported in several population-based studies. Some studies have established a prognosis model of malignant pleural mesothelioma after surgical treatment or systemic chemotherapy (7, 15), and others have proposed a diagnostic model for early recognition of MM through machine learning technology (16). However, these analyses have often been confined to the pleura, peritoneum, or other specific sites and no comparison in prognosis has been drawn across the various anatomic sites. In our study, the survival outcomes of MMe patients from all sites documented in the SEER database were analyzed and the cohort from the SEER database was utilized to develop a novel nomogram to predict the overall survival (OS), thus providing a basis for the individualized management of MMe in various anatomic sites.

MMe is a rare malignancy that mainly originates from the pleura but may also occur in other sites. So far, the majority of available prognostic nomograms are restricted to MMe arising in the pleura (7, 10). Our study has compared the OS of patients with MMe across multiple sites and found that the overall survival of malignant mesothelioma in different primary sites was different and found that patients with MMe occurring in soft tissues, hollow organs, or pleura have a worse prognosis than those originating from the peritoneum, solid organs, or genital organs.

The prognosis for MMe is dismal, and its diagnosis is notoriously difficult (3, 17). For example, the median survival time for malignant pleural mesothelioma is about 10-12 months (18), which is worse than reported in our study ( Table 2 ). It may be because most (80%) patients with malignant pleural mesothelioma are typically diagnosed when the disease has progressed to an advanced stage (19). Individualized surgical treatment is not available for these patients (17, 20). Chemotherapy and other treatments also have a limited effect (21), thus leading to poor prognosis (19). The severe phase of MMe can be avoided by early identification of risk factors for the disease. Talha and Latif reported that early identification of asbestos exposure and its duration, erythrocyte sedimentation rate, and pleura-serum LDH ratio helps prevent patients from entering a high risk of MMe (22, 23). Based on co-expression patterns between ceRNA and immune cells, it was found that has-miR-582-5p, CASP9, dendritic cell rest, ANIX2, T cell CD8, and T cell CD4 memory rest may be associated with mesothelioma bone metastasis (24). Lai J also validated that alternative splicing (AS) events associated with the immune system can help clinicians diagnose MMe earlier and more accurately assess disease activity, and possibly even a potential therapeutic target for MMe (25). It is necessary to establish predictive models for the diagnosis, treatment or prognosis of malignant mesothelioma and its complications. Zhang X used four routinely detected variables, namely carcinoembryonic antigen (CEA) levels, monocyte counts, N-terminal parent-B natriuretic peptide (NT-pro-BNP) levels, and pleural effusion chloride levels at admission as predictors and developed the CONCH prognostic scoring system, which can guide the choice of malignant pleural effusion (MPE) interventions and management (26). Also, Talha ‘s research leverages the latest datasets from the Public Repository (UCI) to present a machine learning-based framework that eliminates the need for expensive testing and painful procedures and can help us make early diagnoses and better treat malignant mesothelioma through important prognostic factors (27). In our research, we developed the first predictive model for the survival outcome of MMe from different sites. This undoubtedly improved our understanding of MMe and provided a basis for the management and care of patients with MMe in different parts.

To identify the independent risk factors included in the variables of this study, the Cox proportional hazards model was employed. Our analysis found that old age, male, tumor originating from pleura, soft tissue or hollow organs, and LNR > 0.45 or no lymph node dissection were associated with shorter OS, whereas young age, female, localized disease, epithelial histological type, high differentiation, no lymph node metastasis and receiving chemotherapy favor a better prognosis ( Figure 3 ).

A recent study described the differences in survival time between young and old malignant mesothelioma patients (28). Older MMe patients have a worse prognosis than younger patients (7, 10, 29). The multivariate analysis in our study yielded similar conclusions, with MMe patients older than 65 years having a relatively poor OS, whereas patients younger than 65 years had longer overall survival. Our findings corroborated a previous study identifying that the female gender was an independent factor associated with good survival outcomes (17, 30–33). This may be because estrogen in female patients in combination with the tumor suppressor estrogen receptor (ER)-β inhibits the progression of malignant mesothelioma (34, 35).

In this study, there was no inconsistency between clinical staging and actual pathological staging, as the pathological staging was obtained from patient information recorded in the Seer database (13, 14). Recent studies have shown that tumor volume in MMe may be closely related to patient staging, assessing treatment response, and predicting survival (36–38). Unfortunately, the present study did not validate this conclusion as tumor volume in MMe patients was not recorded in the SEER database. Some studies have shown that tumor TNM stage, size, and distant metastasis may be related to the survival time of MMe patients (10). However, no statistical significance was found between tumor TNM stage, T stage, and M stage and overall survival time of MMe patients in our study. Several previous studies have reported an association between positive lymph node pathology and reduced survival in MMe patients (39–41). Our study also found that MMe patients with lymph node stage N+ had a shorter survival time than patients without lymph node metastases. In addition, the prognosis of MMe patients was not significantly associated with the number of positive lymph nodes but was significantly associated with the ratio of the number of positive lymph nodes to the total number of detected lymph nodes. A multicenter study by Ilir Hysi et al. also demonstrated that overall survival was not related to the total number of lymph nodes involved and that a lower lymph node ratio (≤13%) implied a considerable improvement in median survival time. However, the lymph node ratio (LNR) was not an independent prognostic factor for patient survival outcomes. This may be attributed to only 99 patients being enrolled in their analysis and incomplete information such as recurrence, cohort heterogeneity, etc. (42).

Previous studies have reported that the histological type is an important factor affecting the survival time, and the prognosis of MMe patients with epithelial histological type is better than the non-epithelial type (43–47). Similarly, our study also found that the survival outcomes of MMe patients were closely related to their tissue type. Specifically, patients with epithelial histology had a better prognosis than those with fibroid and biphasic histology. Patients with epithelial histopathology have a better prognosis and are more responsive to treatment, but other types of mesothelioma patients should not be excluded from first-line trials (48). In addition, malignant mesothelioma (MMe) in situ may be a separate category (48, 49). Our study also found that the extent of diseases had a statistically significant difference, such as limited disease and regional MMe patients had a relatively good prognosis than those with distant metastases. Moreover, our study revealed the relationship between the prognosis of patients with MMe and the grade of tumor differentiation. A higher degree of tumor differentiation was associated with a better prognosis. Conversely, patients with poorly differentiated or undifferentiated tumors generally have shorter survival (7, 10, 44).

Some reports indicate that the incidence of MMe in some areas may have reached a peak. However, the long-term survival time of existing patients is still dismal, so the prognosis and quality of life of MMe patients cannot be ignored (31, 50, 51). And surgery-based multimodal therapy has a promising prognosis in selected patients with malignant mesothelioma (20, 52). All patients included in our study underwent surgery, but there was no significant difference in outcome between complete resection and palliative resection (53). It is worth mentioning that chemotherapy is an irreplaceable part of the current treatment of malignant mesothelioma. For patients who can undergo surgery, chemotherapy can further prolong their survival (32). For unresectable or advanced malignant mesothelioma, the current first-line chemotherapy regimen is pemetrexed plus cisplatin (54–56), and a complete chemotherapy regimen is associated with better survival outcomes (57). In contrast, Barsky et al. proposed a high BED (biologically effective dose) SBRT (stereotactic body radiation therapy) regimen for oligometastatic malignant pleural mesothelioma (58). Radiation therapy has also been shown to be useful after preoperative neoadjuvant therapy (20), extrapleural pneumonectomy (EPP) and less aggressive surgery (59). However, radiotherapy does not improve outcomes for patients with MMe (60). Our study also confirms that the prognosis of patients with malignant mesothelioma is closely related to chemotherapy, but not to radiotherapy. The first-line efficacy evaluation of ipilimumab plus nivolumab combination therapy showed promise for immune checkpoint inhibition and cell-based therapy (61). This was also confirmed by the recent approval of nivolumab plus ipilimumab combination therapy as a first-line treatment marker in patients with malignant pleural mesothelioma (55). A recent study evaluating CHECKMATE-743 found that the impact of this first-line therapy on quality of life and overall survival in patients with MPM remains controversial (62). Furthermore, tumor programmed death ligand 1 (PD-L1) expression is not predictive in clinical trials. This may be because MPM is a low tumor mutational burden (TMB) tumor (63). In addition, approximately 40% of patients treated with nivolumab plus ipilimumab experienced grade 3 or higher adverse events (AEs) (64).

Due to the low incidence of MMe, few studies have focused on the clinical characteristics and prognosis of MMe. However, the following limitations remain: (1) Our study is a retrospective analysis based on the Seer database, and selection bias is unavoidable. (2) The SEER database lacks detailed clinical data such as surgical methods, specific radiotherapy, and chemotherapy regimens, which may imply other unknown variables. (3) With advances in surgical (65–67) and treatment modalities (54, 61), patients with recently diagnosed MMe may have a better prognosis than those included in our study. (4) For the management of MMe patients, no normative standard has yet been established. There is no unified staging standard for MMe of different locations, which may avoid the overfitting of the model due to the inclusion of too many variables and affect the ease of use.

A large prospective trial with strict design needs to be carried out to verify our results. Information not recorded in the Seer database can be included in the trial, such as specific treatment plans, molecular pathology or blood indicators. Besides, external validation is needed to demonstrate its performance in clinical practice.

In conclusion, this study is based on the Seer database, including large sample size, clear pathological diagnosis, and complete follow-up. Our study found that there were differences in prognosis for malignant mesothelioma at different anatomical sites. Further, age, sex, histological type, location, differentiation, the extent of disease, N status, ratio, and chemotherapy were identified as independent factors affecting the overall survival (OS) of MMe patients. Based on these factors, we built a nomogram that can be used to assess the prognosis of individual MMe patients at different sites. Prognostic model established in our study provides a reference for clinical decision-making and helps medical workers to individualize the care of patients with malignant mesothelioma.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

SS, YL and YW conceived and designed the experiments. SS, LS, KQ and TY collected and analyzed data. SS, YW and XJ wrote this manuscript. All authors contributed to the article and approved the submitted version.

Funding from Outstanding Doctoral Project Funding Program of Air Force Military Medical University (2021D08) and Shandong Province Medical and Health Science and Technology Development Plan Project (2018WS383).

The authors are grateful for the invaluable support and useful discussions with other members of the Department of Thoracic Surgery, and thank Home for Researchers (www.home-for-researchers.com) for language enhancement of this paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.