Introduction/Background*Recurrent bleeding can suggest a presence of cancer and has been incorporated into some risk prediction models for cancer in patients with PMB. However uniformed descriptions of PMB are not consistently used. We aimed to assess if the frequency or duration of bleeding allows identification of women at higher risk of cancer undergoing hysteroscopy for PMB

Methodology A retrospective review of hysteroscopy records from Royal Derby Hospital. Bleeding was defined in terms of number of episodes (1, 2, 2+) and duration and was compared with histopathological records. Cases were identified between 2017 and 2019.

Result(s)*Between 2017 and 2019, 1101 women underwent a hysteroscopy for PMB. Seven were excluded as histology results were not obtained. Of the 1094 women included, 98 cancers were identified (9%).

Regarding bleeding frequency; 184 women were excluded as there was insufficient data recorded. Of the remaining 910 women no significant difference was seen in the rate of cancer with different bleeding episodes; one episode (9.6%), two episodes (5.6%) or more than two episodes (9%) (p>0.05).

Regarding character of bleeding; 556 were excluded as they did not have an explicit duration of bleeding recorded. Of the remaining 538; 409 had a short (≤7 days) duration, with 8 (2%) cancers identified, 108 had an intermediate (1-4 weeks) duration with 12 (11%) cancers identified, and of 21 with a long (>4 weeks) duration of bleeding 4 (19%) had a cancer identified. These results were highly significant (p =<0.00001)

In women who described unspecified cyclical/regular bleeding (n=98) 10 cancers were detected (10%)

Conclusion*Accurately characterising bleeding duration is a more meaningful predictor of malignancy than episodes of bleeding. Standards should be developed to enable clinical history as well as radiological findings to triage care.