The impact of surgical guidelines and periodic quality assessment on the staging of endometrial cancer.

ABSTRACT

OBJECTIVE: To determine the impact of surgical guidelines and transparent periodic assessment of surgical quality on endometrial cancer (EC) staging by gynecologic oncologists in a single institution and to identify process-of-care, patient-specific, and disease-specific risk factors that influence surgical quality. METHODS: In January 2004, a prospective treatment algorithm was implemented for EC at our institution. The number of nodes harvested was a surrogate, and staging quality from 2004 to 2008 (quality assessment [QA] interval) was compared with the previous 5 years (pre-QA interval). Since 2004, low-risk cases based on frozen section examination had not undergone lymphadenectomy and were excluded. Independent patient-specific, disease-specific, and surgery-related risk factors influencing lymphadenectomy quality during both intervals were identified with multivariable logistic regression analysis. RESULTS: Pelvic and para-aortic lymph node dissection (LND) in surgical EC management before QA (n=420) were 77.9% and 48.8% vs 89.3% and 83.4% during the QA (n=561) (P<.001). The median number of pelvic and para-aortic nodes harvested in LND was 29 and 10 before QA vs 34 and 16 during the QA interval (P<.001). With acceptance of stringent criteria for defining systematic LND (mean node count-1 SD) during the QA, systematic pelvic (≥22 nodes) and para-aortic (≥10 nodes) LNDs occurred in 57.4% and 25.7% of cases before QA vs 77.9% and 70.7% during the QA interval (P<.001). In patients with LND, rates of systematic pelvic and para-aortic LND were 73.7% and 53.0% before vs 87.2% and 84.8% after QA (P<.001). Multivariable logistic regression analysis showed independent factors influencing systematic pelvic and para-aortic LND (P<.01) surgeon and stage during the pre-QA interval vs surgical approach; intraoperative ascites; body mass index; surgeon; patient age; and myometrial invasion after QA implementation. CONCLUSION: Inclusion of detailed surgical guidelines and transparent periodic assessment of surgical quality translated to dramatic improvement in quality of surgical EC staging. This implementation was associated with a transition to more patient-specific risk factors influencing systematic LND. Although surgical quality metrics were markedly enhanced during QA, persistent variability observed among surgeons and the change in surgical approach render continuous QA and improvement obligatory.