Thresholds and timing of pre-operative thrombocytosis and ovarian cancer survival: Analysis of laboratory measures from electronic medical records

Background: Thrombocytosis has been associated with poor ovarian cancer prognosis. However, comparisons of thresholds to define thrombocytosis and evaluation of relevant timing of platelet measurement has not been previously conducted. Methods: We selected Tumor Registry confirmed ovarian, primary peritoneal, and fallopian tube cancer cases diagnosed between 1995-2013 from the Vanderbilt University Medical Center. Laboratory measured platelet values from electronic medical records (EMR) were used to determine thrombocytosis at three thresholds: a platelet count greater than 350, 400, or 450 × 10⁹/liter. Timing was evaluated with 5 intervals: on the date of diagnosis, and up to 1, 2, 4, and 8 weeks prior to the date of diagnosis. Cox regression was used to calculate hazard ratios (HR) and confidence intervals (CI) for association with overall survival; adjustment included age, stage, grade, and histologic subtype of disease. Results: Pre-diagnosis platelet measures were available for 136, 241, 280, 297, and 304 cases in the five intervals. The prevalence of thrombocytosis decreased with increasing thresholds and was generally consistent across the five time intervals, ranging from 44.8-53.2 %, 31.6-39.4 %, and 19.9-26.1 % across the three thresholds. Associations with higher grade and stage of disease gained significance as the threshold increased. With the exception of the lowest threshold on the date of diagnosis (HR₃₅₀: 1.55, 95 % CI: 0.97-2.47), all other survival associations were significant, with the highest reaching twice the risk of death for thrombocytosis on the date of diagnosis (HR₄₀₀: 2.01, 95 % CI: 1.25-3.23). Conclusions: Our EMR approach yielded associations comparable to published findings from medical record abstraction approaches. In addition, our results indicate that lower thrombocytosis thresholds and platelet measures up to 8 weeks before diagnosis may inform ovarian cancer characteristics and prognosis.