Predicting Overall Survival for Patients with Malignant Mesothelioma Following Radiotherapy via Interpretable Machine Learning

Our plain-English summary of Predicting Overall Survival for Patients with Malignant Mesothelioma Following Radiotherapy via Interpretable Machine Learning, by Z. Wang, V.R. Li, F.I. Chu, et al (Cancers, 2023).

The study

• Authors: Z. Wang, V.R. Li, F.I. Chu, V. Yu, A. Lee, D. Low, D. Moghanaki, P. Lee, X.S. Qi
• Journal: Cancers
• Published: 2023;15(15)
• Cited by: 5 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Purpose/objectives: Malignant pleural mesothelioma (MPM) is a rare but aggressive cancer arising from the cells of the thoracic pleura with a poor prognosis. We aimed to develop a model, via interpretable machine learning (ML) methods, predicting overall survival for MPM following radiotherapy based on dosimetric metrics as well as patient characteristics.

Materials/methods: Sixty MPM (37 right, 23 left) patients treated on a Tomotherapy unit between 2013 and 2018 were retrospectively analyzed. All patients received 45 Gy (25 fractions). The multivariable Cox regression (Cox PH) model and Survival Support Vector Machine (sSVM) were applied to build predictive models of overall survival (OS) based on clinical, dosimetric, and combined variables.

Results: Significant differences in dosimetric endpoints for critical structures, i.e., the lung, heart, liver, kidney, and stomach, were observed according to target laterality. The OS was found to be insignificantly different (p = 0.18) between MPM patients who tested left- and right-sided, with 1-year OS of 77.3% and 75.0%, respectively. With Cox PH regression, considering dosimetric variables for right-sided patients alone, an increase in PTV_Min, Total_Lung_PTV_Mean, Contra_Lung_Volume, Contra_Lung_V20, Esophagus_Mean, and Heart_Volume had a greater hazard to all-cause death, while an increase in Total_Lung_PTV_V20, Contra_Lung_V5, and Esophagus_Max had a lower hazard to all-cause death. Considering clinical variables alone, males and increases in N stage had greater hazard to all-cause death; considering both clinical and dosimetric variables, increases in N stage, PTV_Mean, PTV_Min, and esophagus_Mean had greater hazard to all-cause death, while increases in T stage and Heart_V30 had lower hazard to all-cause-death. In terms of C-index, the Cox PH model and sSVM performed similarly and fairly well when considering clinical and dosimetric variables independently or jointly.

Conclusions: Clinical and dosimetric variables may predict the overall survival of mesothelioma patients, which could guide personalized treatment planning towards a better treatment response. The identified predictors and their impact on survival offered additional value for translational application in clinical practice.

In plain English

This retrospective study analyzed 60 patients with malignant pleural mesothelioma who had radiotherapy after pleurectomy/decortication surgery at the University of California, Los Angeles, between 2013 and 2018. Thirty-seven had right-sided and 23 had left-sided disease. All received 45 Gy in 25 fractions by helical IMRT on a Tomotherapy unit. The study did not examine asbestos exposure. Left-sided patients received higher doses to the heart and stomach, and right-sided patients higher doses to the liver. Overall survival did not differ significantly by side (p = 0.18): 1-year survival was 77.3% for left-sided and 75.0% for right-sided patients. In the model using clinical variables, being male (hazard ratio 4.23, confidence interval 1.22 to 14.63) and higher N stage were linked to higher risk of death. In the model combining clinical and dose variables, the male hazard ratio was not statistically significant. Survival did not differ between patients with and without radiation pneumonitis (p = 0.86). The authors concluded that dose measures together with clinical variables could predict survival (c-index 0.62 for Cox PH, 0.64 for sSVM). They noted that a larger sample may be needed.

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 cohort study of radiotherapy patients, using Cox proportional hazards regression and survival support vector machine models with 5-fold cross-validation
• Who was studied: 60 malignant pleural mesothelioma patients treated at UCLA, 2013 to 2018, all after pleurectomy/decortication. 37 were right-sided (average age 69.3) and 23 left-sided (average age 66.0). 35 (58.3%) had epithelioid histology and the rest biphasic or mixed. One patient was excluded from model building.
• Exposure: Post-operative radiotherapy: helical IMRT on a Tomotherapy unit, 45 Gy in 25 fractions. The study looked at radiation doses to the target and nearby organs. Asbestos exposure was not studied.
• Main results:
  • Overall survival did not differ significantly between left-sided and right-sided patients (p = 0.18). One-year overall survival was 77.3% for left-sided and 75.0% for right-sided patients.
  • Left-sided patients had higher mean heart and stomach doses (median heart mean dose 25.6 Gy vs 19.5 Gy for right-sided). Right-sided patients had higher liver doses (median liver mean dose 23.4 Gy vs 10.8 Gy for left-sided).
  • In a single-factor (univariate) comparison, overall survival differed between male and female patients (p = 0.02).
  • In the Cox model with clinical variables only, male gender had hazard ratio 4.23 (CI 1.22 to 14.63, p = 0.02) and N stage had hazard ratio 1.60 (CI 1.07 to 2.38, p = 0.02). In the combined clinical and dose model, male gender had hazard ratio 4.54 (CI 0.78 to 26.37, p = 0.09).
  • In the combined clinical and dose model, higher N stage was linked to greater hazard of death (HR 2.46, CI 1.38 to 4.36). Higher T stage was linked to lower hazard (HR 0.31, CI 0.14 to 0.70).
  • No difference in overall survival was found between patients with and without radiation pneumonitis (p = 0.86). 14 of 60 patients (23.3%) were reported to have definitive radiation pneumonitis.
  • With clinical variables, the Cox PH and sSVM models had c-indexes of 0.70 and 0.65 in 5-fold cross-validation.
• The authors' conclusion: The authors concluded that patient-specific dose measures together with clinical variables could predict overall survival (c-index 0.62 for Cox PH and 0.64 for sSVM). They considered this predictive performance fair.
• Limitations the authors state:
  • No stable model could be fitted with dose variables for left-sided patients, likely due to limited data in that subgroup.
  • Confounders not included in the study may hide the true relationship between the predictors and survival.
  • The study was small, and the authors say a larger sample may be needed to validate the findings.

Topics

Medical terms in this study, explained in our Lexicon: Overall Survival, Mesothelioma, Pleural Mesothelioma, Pleura, Prognosis.

Related studies

• Predicting cancer using supervised machine learning: Mesothelioma — A. Choudhury (2021)
• Identification of glycolysis genes signature for predicting prognosis in malignant pleural mesothelioma by bioinformatics and machine learning — Y. Xiao, W. Huang, et al. (2022)
• Cancer-Associated Fibroblasts Influence Survival in Pleural Mesothelioma: Digital Gene Expression Analysis and Supervised Machine Learning Model — S. Borchert, A. Mathilakathu, et al. (2023)
• 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)
• Prediction of Radiation Pneumonitis With Machine Learning in Stage III Lung Cancer: A Pilot Study — M. Yakar, D. Etiz, et al. (2021)
• Differential diagnosis of pleural mesothelioma using Logic Learning Machine — S. Parodi, R. Filiberti, et al. (2015)

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