• 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br ACCEPTED MANUSCRIPT br regardless of


    272 regardless of time of sampling prior to endometrial cancer diagnosis (Table 2). 273 We then employed LASSO regression analysis on this set of 10 prioritized, case versus
    274 control protein alterations resulting in the selection of 6 proteins: complement factor B,
    275 serotransferrin, catalase, proteasome subunit beta type-6, beta-2-microglobulin, and
    277 The abundance trends for these six L-Nicotine were then assembled into an aggregate score
    278 to predict the relative risk of having endometrial cancer (Fig. 3). The integrated diagnostic score
    282 diagnostic score was evaluated using ROC analysis among case group 1 and all controls (Fig. 4).
    286 We also evaluated the integrated diagnostic score stratified by body mass index (BMI):
    290 there was a slight increase in the predictive accuracy based on integrated score alone compared
    294 Principal Findings
    295 In this case-control study nested within the PLCO Cancer Screening Trial, we analyzed
    296 prediagnostic serum from postmenopausal women to investigate for biomarkers that could aid in
    297 prediction or detection of endometrial cancer. Forty-seven proteins were significantly altered
    298 between cases and controls ≤ 2 years of diagnosis. A predictive protein set of 6 proteins
    299 (complement factor B, serotransferrin, catalase, proteasome subunit beta type-6, beta-2-
    300 microglobulin, and protocadherin-18) distinguished cases from controls with ROC analysis
    301 showing an AUC of 0.80 among patients with blood draw ≤ 2 years from cancer diagnosis.
    303 The results of this study are consistent with previous publications. The most highly
    304 abundant proteins among cases ≤ 2 years of diagnosis corresponded to different subunits of the
    305 cellular proteasome. Chymotrypsin-like activity of proteasomes from multiple types of cancers is
    306 higher among tumor compared to normal tissue with the largest proteasome activity seen within
    307 endometrial cancer (10). Additionally, the abundance of serotransferrin and apolipoprotein A-1
    308 were both lower among cases ≤ 2 years of diagnosis compared to controls, which is consistent
    309 with previous investigation that focused on validating serum biomarkers for early detection of
    311 This study also provides clarification where there is conflicting data in the literature. The
    312 excess production of reactive oxygen species or inadequate antioxidant protection can lead to
    313 oxidative stress, believed to play a role in carcinogenesis (12). Catalase is an important enzyme
    314 that protects the cell from oxidative stress, but there is conflicting data on whether catalase
    315 activity in endometrial cancer differs from benign endometrium (12, 13). Although our study did
    316 not investigate tissue, we show that serum levels of catalase was 19% higher among cases. 317 There is also conflicting data on whether human serum amyloid A (SAA) is higher
    318 among endometrial cancer patients. Cocco et al. using a bead-based immunoassay showed higher
    319 concentrations of serum SAA among endometrioid and serous tumors compared to controls (14,
    320 15). They also demonstrated higher SAA levels among grade 3 compared to grade 1 or 2
    321 endometrioid tumors (15). However, another study showed no difference in serum SAA levels
    322 among endometrioid cancer versus healthy patients (16). Our investigation demonstrated no
    323 difference in SAA concentration between cases or controls ≤ 2 years of diagnosis [logFC (-0.25),
    325 Our findings differ significantly with previous studies that have examined serum from
    326 endometrial cancer patients and controls (17-24). Previous reports have shown serum
    327 concentrations of DJ-1, visfatin, OVX1, sperm- associated antigen 9, and circulating soluble Fas
    329 did not detect these proteins. Preoperative serum levels of YKL-40 have been shown to be higher
    330 among endometrial cancer cases compared to controls (21, 22), but our analysis suggested no
    331 difference in abundance of YKL-40 among cases ≤ 2 years of diagnosis and controls [logFC
    333 Moreover, Trabert et al. also analyzed serum from both endometrial cases and controls
    334 from the PLCO Cancer Screening Trial in which they measured serum levels of 64
    335 inflammation-related biomarkers, with 22 of these markers being associated with endometrial
    336 cancer risk. None of the biomarkers seen in this study are consistent with our findings; however,
    337 this study used immunoassays to evaluate levels of targeted inflammation-related biomarkers and
    338 their analysis did not stratify cases by timing of blood draw prior to cancer diagnosis (18).