Do U.S. Screening Guidelines Contribute to Higher Stage and Worse Survival in our Youngest and Oldest Cervical Cancer Patients? (019)

Abstract

<b>Objectives:</b> To investigate age-related differences in stage at diagnosis and survival in cervical cancer patients to identify possible gaps in current screening guidelines. <b>Methods:</b> Cervical cancer patients diagnosed between 2004-2014 in the population-based SEER-18 program or the hospital-based National Cancer Database (NCDB) were eligible. Analyses in both cohorts were stratified in adolescent/young adults (AYAs: 18-39 years old), middle-aged adults (MAs: 40-64 years old), or the elderly (≥65 years old). The proportion of late (II-IV) versus early (I) stage at diagnosis was compared using the Chi-square test. Survival was compared with log-rank test and Cox modeling. The probability of cancer-related deaths (CRD) and non-cancer deaths (NCD) was estimated in SEER by cumulative incidence function. Early-stage diagnosis, 5-year survival, and adjusted hazard ratio for all-cause death in NCDB cases were estimated for each 5-year increment in age. A dominance analysis from logistic modeling was used to rank the relative importance of demographic and socioeconomic factors on the late stage of diagnosis in NCDB. <b>Results:</b> There were 29,635 cases from SEER (29% AYAs, 55% MAs, and 16% elderly). Late-stage diagnosis increased with age from 33% in AYAs to 54% in MAs and 71% in the elderly (p<i><0.0001)</i>. Five-year survival rates were incrementally worsening with age and were 80%, 65%, and 41% for AYAs, MAs, and the elderly, respectively (p<i><0.0001)</i>. Figures 1A-1C show 5-year CRD (NCD) rates of 19% (1%) in AYAs, 31% (3%) in MAs, and 47% (12%) in the elderly, respectively (p<i><0.0001</i>). These relationships were then further investigated in the 87,834 cases in the NCDB cohort (27% AYAs, 57% MAs and 17% elderly). Late-stage diagnosis again increased with age from 30% in AYAs to 52% in MAs and 68% in the elderly (p<i><0.</i>0001). Figure 1D displays the incremental inferior survival in MAs and the elderly relative to AYAs (p<i><0.</i>0001). Five-year survival for AYAs, MAs, and elderly patients in the NCDB cohort was estimated to be 80%, 67% and 43%, respectively (log-rank p<0.0001). Figures 1E-1F show that the age-dependent reductions in stage I diagnosis and 5-year survival rates mirrored each other. Within AYAs, the youngest patients diagnosed at 18-24 had a similar mortality compared to those diagnosed at 25-29 years old, but had a 25% higher adjusted risk of death over those diagnosed between 30-39 years ld. Within MAs, the adjusted risk of death increased by 15% when diagnosed at ages 45-49 versus 40-44 years, and 31% higher when diagnosed at ages 60-64 versus 40- 4 years old. The adjusted risk of death increased 20%, 47% and 120% in the elderly diagnosed at 70-74, 75-79 and ≥80 relative to 65-69 years old, spectively. A final analysis ranked the relative importance of factors contributing to the late stage at diagnosis (II-IV disease) showing that age and insurance status contributing 62% and 29%, respectively, with minor contributions from socioeconomic status (4.7%), race/ethnicity (2.4%), and comorbidity score (1.2%).Fig. 1 <b>Conclusions:</b> The youngest AYAs and the whole elderly population with cervical cancer in SEER and NCDB had a higher stage at diagnosis and worse survival, suggesting a need to propose age- and insurance-related refinements in screening guidelines to mitigate these disparities.