Accuracy for cytological evaluation in the detection of breast cancer among patients with pathologic nipple discharge: a PRISMA-compliant meta-analysis

BACKGROUND: Nipple discharge cytology is a simple non-invasive method that may provide valuable information for detecting underlying malignancy. Several studies have investigated the diagnostic value of cytology in breast cancer patients with pathological nipple discharge, but the results have been highly variable. Herein we presented a systematic review and meta-analysis of published studies pertaining to the diagnostic capacity of nipple discharge cytology in patients with breast cancer. METHODS: A systematic literature search was performed (Medline/PubMed, Embase, Cochrane Library databases, and Google Scholar) to identify studies that investigated the diagnostic capacity of cytology with regard to breast cancer in patients with pathologic nipple discharge. Two independent researchers identified articles that assessed the sensitivity and specificity of cytological evaluation for breast cancer detection in patients with pathologic nipple discharge published between January 2000 and October 2018. Articles were only included in the meta-analysis if they met predetermined criteria. The characteristics of each study and the data they yielded were summarized. Quality assessment of all articles included was performed using the Methodological Index for Non-randomized Studies Criteria (MINORS) and the Quality Assessment of Diagnostic Accuracy Study 2 (QUADAS-2). Heterogeneity was tested via Cochran Q test and the I(2) statistic using Stata 12.0 and Meta-DiSc 1.4 software, and meta-analysis was performed. RESULTS: A total of 286 articles were identified, of which 12 articles including a total of 1476 patients were deemed eligible for inclusion in the meta-analysis. A random-effects model assessing the capacity of nipple discharge cytology to predict breast cancer yielded pooled sensitivity 63% (95% confidence interval [CI]: 53%–72%), specificity 95% (95% CI: 87%–98%), positive likelihood ratio 12.35 (95% CI: 4.87–31.34), and negative likelihood ratio 0.39 (95% CI: 0.30–0.50). The diagnostic odds ratio was 31.88 (95% CI: 11.30–89.98). The area under the summary receiver operating characteristic curve was 0.79 (95% CI: 0.75–0.82). CONCLUSION: The current meta-analysis suggests that nipple discharge cytology is a useful diagnostic modality for detection of breast cancer in patients with pathological nipple discharge, with moderate sensitivity and high specificity.