RESUMEN
BACKGROUND: In patients with peripheral pulmonary lesions (PPLs), nondiagnostic bronchoscopy results are not uncommon. The conventional approach to estimate the probability of cancer (pCA) after bronchoscopies relies on dichotomous test assumptions, using prevalence, sensitivity, and specificity to determine negative predictive value. However, bronchoscopy is a multidisease test, raising concerns about the accuracy of dichotomous methods. RESEARCH QUESTION: By how much does calculating pCA using a dichotomous approach (pCAdichotomous) underestimate the true pCA when applied to multidisease tests like bronchoscopy for the diagnosis of PPL? METHODS: In this meta-analysis of cohort studies involving radial endobronchial ultrasound for PPL, Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines were followed, constructing 2 × 2 contingency tables for calculating pCAdichotomous. For the multidisease test approach, 3 × 3 contingency tables for calculating probability of malignancy for a test that can have multiple test different categories of results and can diagnose multiple diseases (pCAmultidisease) using the likelihood ratio (LR) method for nondiagnostic results (LR(T0)) was used. Observed malignancy rates in patients with nondiagnostic results were compared with pCAdichotomous and pCAmultidisease. RESULTS: In 46 studies (7,506 patients), malignancy was the underlying diagnosis in 76%, another specific disease in 13%, and nonspecific fibrosis or scar in 10%. The percentage of patients with nondiagnostic results who had malignancy matched pCAmultidisease across all studies. In contrast, pCAdichotomous consistently underestimated cancer risk (median difference, 0.12; interquartile range, 0.06-0.23), particularly in studies with a higher prevalence of nonmalignant disease. The pooled LR(T0) was 0.46 (95% CI, 0.40-0.52; I2 = 76%; P < .001) and correlated with the prevalence of nonmalignant diseases (P = .001). INTERPRETATION: Conventional dichotomous methods for estimating pCA after nondiagnostic bronchoscopies underestimate the likelihood of malignancy. Physicians should opt for the multidisease test approach when interpreting bronchoscopy results.
RESUMEN
BACKGROUND: Evidence-based guidelines recommend management strategies for malignant pleural effusions (MPEs) based on life expectancy. Existent risk-prediction rules do not provide precise individualized survival estimates. RESEARCH QUESTION: Can a newly developed continuous risk-prediction survival model for patients with MPE and known metastatic disease provide precise survival estimates? STUDY DESIGN AND METHODS: Single-center retrospective cohort study of patients with proven malignancy, pleural effusion, and known metastatic disease undergoing thoracentesis from 2014 through 2017. The outcome was time from thoracentesis to death. Risk factors were identified using Cox proportional hazards models. Effect-measure modification (EMM) was tested using the Mantel-Cox test and was addressed by using disease-specific models (DSMs) or interaction terms. Three DSMs and a combined model using interactions were generated. Discrimination was evaluated using Harrell's C-statistic. Calibration was assessed by observed-minus-predicted probability graphs at specific time points. Models were validated using patients treated from 2010 through 2013. Using LENT (pleural fluid lactate dehydrogenase, Eastern Cooperative Oncology Group performance score, neutrophil-to-lymphocyte ratio and tumor type) variables, we generated both discrete (LENT-D) and continuous (LENT-C) models, assessing discrete vs continuous predictors' performances. RESULTS: The development and validation cohort included 562 and 727 patients, respectively. The Mantel-Cox test demonstrated interactions between cancer type and neutrophil to lymphocyte ratio (P < .0001), pleural fluid lactate dehydrogenase (P = .029), and bilateral effusion (P = .002). DSMs for lung, breast, and hematologic malignancies showed C-statistics of 0.72, 0.72, and 0.62, respectively; the combined model's C-statistics was 0.67. LENT-D (C-statistic, 0.60) and LENT-C (C-statistic, 0.65) models underperformed. INTERPRETATION: EMM is present between cancer type and other predictors; thus, DSMs outperformed the models that failed to account for this. Discrete risk-prediction models lacked enough precision to be useful for individual-level predictions.