Abstract: |
Objectives: Immunocytochemistry is often required in the cytologic assessment of malignant serous effusion, particularly for differentiating metastatic carcinoma from mesothelioma. To summarise the diagnostic performance of the monoclonal antibody B72.3, a systematic review and meta-analysis was conducted. Methods: Five databases were searched for relevant studies and reviewed for data extraction and risk of bias assessment. Pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR) and area under curve of summary receiver operating characteristics (AUC-SROC) were calculated for the diagnostic performance of B72.3. Heterogeneity and publication bias were assessed by the I2 index and Deeks' funnel plot. Results: In total, 19 studies (1159 cases) were included. Overall pooled sensitivity and specificity were 0.76 (0.72–0.79) and 0.90 (0.74–1.00), respectively. The NLR, PLR and DOR were 0.27 (0.21–0.34), 7.66 (< 0.001–20.46) and 28.26 (0–75.96), respectively. The AUC-SROC was 0.98, indicating a good overall diagnostic accuracy for B72.3. Subgroup analysis for adenocarcinoma (0.75, 0.71–0.79), mesothelioma (0.92, 0.85–0.98) and benign/reactive mesothelial cells (0.96, 0.93–1.00) showed similar sensitivity and specificity, while the sensitivity for adenocarcinomas of the gastrointestinal/hepatobiliary tract (0.56, 0.41–0.71) and breast (0.55, 0.38–0.71) was significantly lower. High heterogeneity was observed in the majority of our analyses, while no evidence of publication bias was identified. Conclusions: B72.3 has an acceptable performance with low sensitivity. With a good specificity, B72.3 may find use in an immunocytochemical panel for excluding benign mesothelial processes. © 2025 The Author(s). Cytopathology published by John Wiley & Sons Ltd. |