Correlation of surgeon radiology assessment with laparoscopic disease site scoring in patients with advanced ovarian cancer.

BACKGROUND: Radiographic triage measures in patients with new advanced ovarian cancer have yielded inconsistent results. OBJECTIVE: To determine the correlation between surgeon radiology assessment and laparoscopic scoring by disease sites in patients with newly diagnosed advanced stage ovarian cancer. METHODS: Fourteen gynecologic oncology surgeons from a single institution performed a blinded review of pre-operative contrast-enhanced CT imaging from patients with advanced stage ovarian cancer. Each of the patients had also undergone laparoscopic scoring assessment, between April 2013 and December 2017, to determine primary resectability using the validated Fagotti scoring method, and assigned a predictive index value score. Surgeons were asked to provide expected predictive index value scores based on their blinded review of the antecedent CT imaging. Linear mixed models were conducted to calculate the correlation between radiologic and laparoscopic score for surgeons individually, and as a group. Once the model was fit, the inter-class correlation and 95% CI were calculated. RESULTS: Radiology review was performed on 20 patients with advanced stage ovarian cancer who underwent laparoscopic scoring assessment. Surgeon faculty rank included assistant professor (n=5), associate professor (p=4), and professor (n=5). The kappa inter-rater agreement was -0.017 (95% CI -0.023 to -0.005), indicating low inter-rater agreement between radiology review and actual laparoscopic score. The inter-class correlation in this model was 0.06 (0.02-0.21), indicating that surgeons do not score the same across all the images. When using a clinical cut-off point for the predictive index value of 8, the probability of agreement between radiology and actual laparoscopic score was 0.56 (95% CI 0.49 to 0.73). Examination of disease site sub-scales showed that the probability of agreement was as follows: peritoneum 0.57 (95% CI 0.51 to 0.62), diaphragm 0.54 (95% CI 0.48 to 0.60), mesentery 0.51 (95% CI 0.45 to 0.57), omentum 0.61 (95% CI 0.55 to 0.67), bowel 0.54 (95% CI 0.44 to 0.64), stomach 0.71 (95% CI 0.65 to 0.76), and liver 0.36 (95% CI 0.31 to 0.42). The number of laparoscopic scoring cases, tumor reductive surgery cases, or faculty rank was not significantly associated with overall or sub-scale agreement. CONCLUSIONS: Surgeon radiology review did not correlate highly with actual laparoscopic scoring assessment findings in patients with advanced stage ovarian cancer. Our study highlights the limited accuracy of surgeon radiographic assessment to determine resectability.

Compliance with research standards within gynecologic oncology fellowship: A Gynecologic Oncology Fellowship Research Network (GOFRN) study.

OBJECTIVES: Participation in clinical and basic science research is emphasized in gynecologic oncology training. We sought to identify trends in adherence to expected research practices and reasons for non-adherence among gynecologic oncology fellows. METHODS: An anonymous 31-question online survey assessing academic behaviors, including IRB compliance, authorship assignment, data sharing, and potential barriers to non-adherence was distributed to all SGO gynecologic oncology fellow members in July 2016. Descriptive statistics and univariate analyses were performed. RESULTS: Of 190 members, 35.3% (n=67) responded. 73% (n=49) of respondents reported personal non-compliance and 79.1% (n=53) reported having witnessed others being non-complaint with at least one expected research practice. Areas of compliance failure included changing a research question without appropriate IRB amendment (20%; n=14), conducting research under a nonspecific IRB (13.9%; n=9), and performing research without IRB approval (6.1%; n=4). Longer institutional time for IRB approval was significantly associated with IRB non-adherence (p<0.05). First year fellows were more likely to use a nonspecific IRB (p=0.04) or expand a question without amending the IRB (p=0.04). When asked about storage of protected health information (PHI) for research, 53% reported non-secure storage with 17.1% (n=6) having done so for >1000 patients. Thirty respondents (45.5%) assigned authorship to someone who failed to meet ICMJE criteria and twelve (18.5%) accepted authorship without meeting ICMJE criteria. Most commonly cited reasons for non-adherence were: cumbersome IRB processes (80.3%), pressure from senior authors (78.8%), fear of someone else publishing first, (74.2%) and lack of support navigating appropriate research practices (71.2%). CONCLUSIONS: Fellow non-compliance with expected research practices is high, particularly with regards to secure storage of PHI and appropriate authorship assignment. Time-consuming and cumbersome IRB procedures, perceived pressure from senior authors, and lack of research support contribute to non-adherence. Further support and education of gynecologic oncology fellows is needed in order to help address these barriers.