Differentiating malignant pleural mesothelioma and metastatic pleural disease based on a machine learning model with primary CT signs: A multicentre study

Our plain-English summary of Differentiating malignant pleural mesothelioma and metastatic pleural disease based on a machine learning model with primary CT signs: A multicentre study, by Y. Li, B. Cai, B. Wang, et al (Heliyon, 2022).

The study

• Authors: Y. Li, B. Cai, B. Wang, Y. Lv, W. He, X. Xie, D. Hou
• Journal: Heliyon
• Published: 2022;8(11):e11383
• Cited by: 13 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Rationale And Objectives: It is still a challenge to make confirming diagnosis of malignant pleural mesothelioma (MPM), especially differentiating from metastatic pleural disease (MPD). The aim of this study was to develop a model to distinguish MPM with MPD based on primary CT signs.

Materials And Methods: We retrospectively recruited 150 MPM patients and 147 MPD patients from two centers and assigned them to training (115 MPM patients and 113 MPD patients) and testing (35 MPM patients and 34 MPD patients) cohorts. The images were analyzed for pleural thickening, hydrothorax, lymphadenopathy, thoracic volume and calcified pleural plaque (CPP). The selected clinical characteristics and primary CT signs comprised the model by multivariate logistic regression in the training cohort. Then the model was tested on the external testing cohort. ROC curve and F1 score were used to validate the capability of the model in both two cohorts.

Results: There were significant differences between two groups: (1) carcinoembryonic antigen (CEA); (2) nodular and mass pleural thickening; (3) the enhancement of pleura; (4) focal, diffuse and circumferential pleural thickening; (5) the thickest pleura; (6) thickening of diaphragmatic pleura; (7) multiple nodules and effusion of interlobar pleura; (8) hilar LN and ring enhancement of LN; (9) punctate and stipe CPP. The AUC and F1 score of the model were 0.970 and 0.857 in the training cohort, 0.955 and 0.818 in the testing cohort.

Conclusion: The model holds promise for use as a diagnostic tool to distinguish MPM from MPD.

© 2022 The Author(s).

In plain English

This study from two Chinese hospitals asked whether chest CT findings and clinical information could help tell malignant pleural mesothelioma (MPM) apart from metastatic pleural disease (MPD), cancer spread to the pleura from another organ. In this study every MPD patient had lung cancer. The authors built a logistic regression model using 115 mesothelioma and 113 MPD patients at Beijing Chest Hospital (2013 to 2018). They tested it on 35 mesothelioma and 34 MPD patients at Harbin Chest Hospital (2019 to 2022). Compared with MPD patients, mesothelioma patients had lower levels of the blood marker CEA. They more often showed diffuse or circumferential pleural thickening and punctate calcified pleural plaques. Focal thickening and enlarged hilar lymph nodes were more common in MPD. The model's AUC, a measure of how well it separates the groups, was 0.970 in training and 0.955 in testing. The authors concluded that the model has value in distinguishing the two diseases. They noted that the study was retrospective and not large. The paper did not describe how asbestos exposure was assessed. The authors wrote that only several patients had a known history of it.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Retrospective two-centre diagnostic study. A logistic regression (machine learning) model was built on CT signs and clinical characteristics in a training cohort and tested on an external cohort.
• Who was studied: 150 patients with malignant pleural mesothelioma and 147 patients with metastatic pleural disease, all with pathological confirmation. All MPD patients had lung cancer. The training cohort was 115 MPM and 113 MPD patients at Beijing Chest Hospital, May 2013 to December 2018. The testing cohort was 35 MPM and 34 MPD patients at Harbin Chest Hospital, January 2019 to January 2022.
• Exposure: The study assessed CT signs, including calcified pleural plaques, and clinical characteristics. The paper did not describe how asbestos exposure was assessed. The authors noted that only several patients had a history of asbestos exposure.
• Main results:
  • CEA differed significantly between MPM and MPD patients in both cohorts. Median CEA was 1.61 nmol/L for MPM versus 7.64 nmol/L for MPD in the training cohort, and 1.79 versus 7.69 in the testing cohort.
  • Diffuse and circumferential pleural thickening were more frequent in MPM than in MPD in both cohorts (p less than 0.05). Circumferential thickening was seen in 54 of 115 MPM versus 16 of 113 MPD patients in the training cohort. Focal thickening was more common in MPD.
  • Mass pleural thickening was seen in 23 of 115 MPM patients (20.00%) and 2 of 113 MPD patients (1.77%) in the training cohort.
  • Enlarged hilar lymph nodes were more frequent in MPD patients: 45 of 113 versus 13 of 115 MPM patients in the training cohort.
  • Punctate and stripe calcified pleural plaques differed significantly between groups in the training cohort. In the testing cohort, stripe plaques did not differ significantly. Patchy plaques did not differ in either cohort.
  • The density and volume of pleural fluid (hydrothorax) did not differ significantly between the groups in either cohort.
  • In the training cohort, the model had an AUC of 0.970, an F1 score of 0.857 and accuracy of 0.842. In the testing cohort it had an AUC of 0.955, an F1 score of 0.818 and accuracy of 0.826.
• The authors' conclusion: The authors concluded that their model, based on clinical characteristics and primary CT signs, has significant value in differentiating MPM from MPD and could be a noninvasive complement to diagnosis.
• Limitations the authors state:
  • The study was retrospective and did not include a large number of patients.
  • There was not enough pathological proof of lymph node metastasis, because radiologic-pathologic correlation was impossible for individual lymph nodes.
  • The model used only primary CT signs. The authors said future work should use quantitative analysis.

Topics

Medical terms in this study, explained in our Lexicon: Pleural Mesothelioma, Pleural Thickening, Lymphadenopathy, Pleural Plaques, Carcinoembryonic Antigen (CEA), Pleura.

Related studies

• Distant Metastatic Pattern and Its Prognostic Significance in Malignant Pleural Mesothelioma: A Population-Based Study Based on a Machine Learning Model — J. Yu, C. Peng, et al. (2025)
• The use of a next-generation sequencing-derived machine-learning risk-prediction model (OncoCast-MPM) for malignant pleural mesothelioma: a retrospective study — M.G. Zauderer, A. Martin, et al. (2021)
• Cancer Risk Prediction Using Machine Learning for Supporting Early Cancer Diagnosis in Symptomatic Patients: A Systematic Review of Model Types — F. Pennisi, S. Borlini, et al. (2025)
• Cancer-Associated Fibroblasts Influence Survival in Pleural Mesothelioma: Digital Gene Expression Analysis and Supervised Machine Learning Model — S. Borchert, A. Mathilakathu, et al. (2023)
• Immunostaining in lung cancer for the clinician. Commonly used markers for differentiating primary and metastatic pulmonary tumors — J.M. Carney, A.M. Kraynie, et al. (2015)
• Prediction of Radiation Pneumonitis With Machine Learning in Stage III Lung Cancer: A Pilot Study — M. Yakar, D. Etiz, et al. (2021)

Browse more research in Heliyon, or return to medical research.