P25.05 The Brims Decision Tree Model for Prognosis in a UK Malignant Mesothelioma Group

Abstract

Malignant Mesothelioma has traditionally carried a poor prognosis. A number of prognostic tools have been developed to try explain these variations and to guide clinicians and patients to the most appropriate therapy and to try answer difficult questionsregarding survival time.The BTS 2018 guideline recommends the Brim’s decision tree analysis as the most useful guide. Prior partial presentation: British Thoracic Oncology group in January 2020 We conducted a local retrospective database analysis to identify patients diagnosed with malignant pleural mesothelioma (MPM) and evaluate the usefulness of the Brims score. Patients with an MDT diagnosis of MPM were included in the study (including thosewithout histological confirmation) between January 1st 2011 until December 31st 2019. Results were compared against the Brims’ decision tree analysis to assess how survival groups compared in a different geographical cohort. There were 233 patients, 82% (n=192) male, with a median diagnostic age of 76.7 (IQR 71-83) years. Epithelioid was the most common histological diagnosis (34%) followed by sarcomatoid (12%), and biphasic (10.1%). 26.2% of patients had no definite pathology and 18% just a radiological diagnosis.The overall median survival was 11.3 months compared to 9.7 in the Brims validation cohort and 12.7 months in the Brims derivation cohort.Table 1 shows the median survival of each predicted Brims group in comparison to the predicted survival, as well as the proportion of patients whose survival fell with +/- 33% of that prediction.Table 1MPM risk groupNumber of patientsMedian survival, months, IQRPredicted survical% within +/- 33% predicted1513.6 (9.1-18.5)34 (22.9-47)026815.1 (4.6-19.9)17.7 (11.6-25.9)25.734713.9 (5.5-19)12 (6-20.6)36.741135 (1.2-9)7.4 (3.3-11.1)23.1 Open table in a new tab The Brim’s decision tree aid can be a clinically useful tool using readily available clinical information to help clinicians identify those patients at highest risk of dying within 6 months of diagnosis. However, its ability to give accurate prognostic advice in our cohort is limited as due to the wide variation and the low proportion of patients within 33%+/- the predicted survival.