Predicting survival for patients with mesothelioma: development of the PLACE prognostic model

Introduction The overall survival of patients with mesothelioma is poor and heterogeneous. At present, the prediction model for Chinese patients needs to be improved. We sought to investigate predictors of survival in malignant pleural mesothelioma and develop prognostic prediction models. Methods This Two-center retrospective cohort study recruited patients with pathologically diagnosed mesothelioma at Beijing Chao-Yang Hospital and Beijing Tong-Ren Hospital. We developed a new prognostic prediction model based on COX multivariable analysis using data from patients who were recruited from June 1, 2010 to July 1, 2021 in Beijing Chao-Yang Hospital (n = 95, development cohort) and validated this model using data from patients recruited from July 18, 2014 to May 9, 2022 in Beijing Tong-Ren Hospital (n = 23, validation cohort). Receiver operating characteristic analysis was used to estimate model accuracy. Results The parameters in this new model included PLT > 289.5(10^9/L) (1 point), Lymphocyte > 1.785(10^9/L) (-1point), Age > 73 years old (1 point), Calcium > 2.145(mmol/L) (-1point), Eastern Cooperative Oncology Group performance status (ECOG PS) > 2 (2 points). When the sum of scores < 0, it is recognized as a low-risk group; when the score is 0 ~ 3, it is recognized as a high-risk group. The survival rate of patients in the high-risk group was significantly lower than that in the low-risk group (hazard ratio [HR], 3.878; 95% confidence interval [CI], 2.226–6.755; P < 0.001). The validation group had similar results (HR,3.574; 95%CI,1.064–12.001; P = 0.039). Furthermore, the areas under the curve 6 months after diagnosis in the two cohorts were 0.900 (95% CI: 0.839–0.962) and 0.761 (95% CI: 0.568–0.954) for development and validation cohorts, respectively. Conclusion We developed a simple, clinically relevant prognostic prediction model for PLACE by evaluating five variables routinely tested at the time of diagnosis. The predictive model can differentiate patients of Chinese ethnicity into different risk groups and further guide prognosis.