A score incorporating clinical and biological markers could help predict the risk of death in patients with malignant pleural effusion and their likelihood of responding well to pleurodesis, according to a study published online in The Lancet Oncology.
The researchers used five separate and independent datasets from three previous multicenter randomized controlled trials – TIME-1, TIME-2, and TIME-3 – to identify 17 candidates for biomarkers of survival at 3 months and 7 candidates for biomarkers of pleurodesis success at 3 months.
They combined these with clinical, radiological, and biological variables to develop the clinical PROMISE model, which included relative protein expression of tissue inhibitor of metalloproteinases 1, platelet-derived growth factor, vascular endothelial growth factor, cadherin 1 and interleukin 4. The pleurodesis dataset included tumour necrosis factor alpha, TNF-beta, interleukin 6, and fibroblast growth factor 2.
The model was then externally validated using complete case data from 162 individuals with malignant pleural effusion, just over one-third of whom had died before 3 months.
The researchers also developed a biological model based on prognostic factors for survival, including use of previous chemotherapy and radiotherapy, baseline ECOG performance status, cancer type, hemoglobin, and white blood cell count.
The researchers found that the PROMISE scores showed “good discrimination,” and based on that, they came up with four categories – A, B, C, and D, representing less than 25% risk, 25%-49% risk, 50%-74% risk, and 75% or more risk of death by 3 months. However, none of the biomarkers associated with pleurodesis outcomes could be validated.
“All parameters included in the PROMISE score are independently associated with survival, and thus the identified markers permit some speculation as to their biological role in survival in malignant pleural effusion,” wrote Ioannis Psallidas, MD, of the Oxford Centre for Respiratory Medicine at Oxford (England) University Hospital, and his coauthors.
For example, they noted that patients who had previously been treated with chemotherapy and radiotherapy may have a poorer prognosis because of the development of more aggressive cancer after treatment. Similarly, white blood cell counts and C-reactive protein are markers of inflammation and are linked to poor tumor-specific immunity.
“Evidence suggests that clinical judgment alone is imprecise for estimation of patient survival, highlighted by the fact that physicians are ineffective in excluding study participants with poor prognosis from large clinical trials,” the authors wrote. “Although future confirmatory studies are required, the PROMISE score (either clinical or biological) could potentially be used in everyday clinical practice as a method to improve patient management and reduce associated health-care costs, and as an enrichment strategy for future clinical trials.”
The study was supported by the Oxford Respiratory Trials Unit, Medical Research Funding-University of Oxford, Slater & Gordon Research Fund, and Oxfordshire Health Services Research Committee Research Grants. No conflicts of interest were declared.
SOURCE: Psallidas I et al. Lancet Oncol. 2018 Jun 13. doi: 10.1016/S1470-2045(18)30294-8.