CHAPTER 3 62 This is the first study to address the prognostic relevance of molecular profiling in low-grade EC. Our study consisted of a large study population, with known LN status in FIGO earlystage to prevent bias by undiagnosed stage III. Furthermore, our results are comparable with the data of the TCGA research network.3 A few limitations must be reported due to the retrospective character of the study. First, differences in the methodology between the included cohorts exists. More than 80% was performed with complete molecular profiling and less than 20% with the immunohistochemistry surrogates of molecular profiling according to the ProMisE criteria. Though, immunohistochemistry surrogate analysis has been established as a reliable alternative for molecular profiling.14 Second, the original diagnosis was used without centralized pathology review, however slides were from large referral hospitals and diagnoses were made by expert gynecological-pathologist. This makes our study applicable to daily practice. Third, race or ethnicity has not been reported in our study. Although we fully agree that these patients’ information might be impact outcome in several diseases, within Europe it is not routinely documented in patient files.23 In order to evaluate whether race and ethnicity might have impacted our results, we performed additional analyses within the Kandoth open access database. Race was not statistically different between low- and high EC patients and between EC-related mortality (data not shown). However, in patients with Black or African American race, TP53-mutant appeared to be more frequently present supporting previous findings of Lu et al. that these women more often were diagnosed with non-endometrioid EC.24 It seems that molecular subgroups overrates the prognostic relevance of race. Fourth, patients were diagnosed between 1994 and 2018, a time spanning over 24 years (median 2006), this could have biased the survival because of different treatment strategies over the time. Including the diagnostic year in the multivariable cox-regression analyses did not change the results of the cox-regression analyses as presented in the result section (data not shown). Finally, although there were significantly more low-grade EC patients in the excluded cases, the DSS for excluded cases showed similar favorable outcome for all molecular subgroups within low-grade EC (data not shown). CONCLUSIONS Routine molecular profiling is not beneficial in low-grade EC patients due to their excellent prognosis independent of molecular subgroup. Our data demonstrate the importance of primary diagnostic tumor grading and do not support routine molecular profiling in lowgrade EC as cost-effective approach.
RkJQdWJsaXNoZXIy MTk4NDMw