Predictive values of gynaecological cancer alarm symptoms in a general population

Aim: The aim of this study was to determine predictive values of gynaecological alarm symptoms in the general population. Materials and Methods: A population-based prospective cohort study combining questionnaire data with register data of gynaecological cancer and premalignant conditions in a one-year follow-up period. Results: Among 25,866 women, some 2,957 (11.4%) reported at least one newly onset gynaecological cancer alarm symptom within a four-week time period. A total of 71 women were diagnosed with a gynaecological cancer or a premalignant condition. The positive predictive values (PPVs) ranged between 0.1% (95%-CI: 0.1-0.1) for cancer among women reporting at least one of the 18 unspecific and specific gynaecological cancer alarm symptoms and 0.3% (95%-CI: 0.2-0.4) for dysplasia among women reporting at least one of the specific alarm symptoms that lead to a GP contact. The positive likelihood ratios (LR+s) ranged between 0.9 (95%-CI: 0.7;1.2) for cancer among women reporting at least one unspecific symptom and 1.7 (95%-CI: 1.1-2.5) for dysplasia among women reporting at least one specific alarm symptom. Conclusion: The PPVs and LR+s of gynaecological alarm symptoms in the general population are very low, and only few of the patients diagnosed with gynaecological cancer reported specific alarm symptoms. Thus targeted diagnosis of gynaecological cancer cannot be based on cancer alarm symptoms only.