Although effective screening for cervical cancer exists and can detect pre-malignant lesions and early-stage (stage I) disease (1,2), many women in the United States (US) present with late-stage disease (stages II-IV) (3). Following the introduction and widespread adoption of the Papanicolaou (Pap) smear test in the 1940’s, cervical cancer incidence and mortality have fallen significantly(4,5). However, incidence rates have plateaued since 2012, and rates of invasive cervical cancer have actually increased in recent decades (5). Furthermore, from 2015–2019, nearly half of cervical cancers in the United States were diagnosed late-stage (6). Prior research suggests that women ≥65 had a greater burden of cervical cancer with higher incidence and more late-stage diagnoses (7–10).

The American Cancer Society (ACS), the US Preventive Services Task Force, and the American College of Obstetricians and Gynecologists all recommend that cervical cancer screening end for women >65 years with adequate negative prior screening (3 consecutive normal Pap tests, two consecutive negative HPV tests, or two consecutive negative cotests with Pap and HPV within the prior 10 years, with the most recent screening occurring within the previous five years, and no diagnosis of a precancerous lesion in the past 25 years) (1,11,12). However, 23.2% of women in the U.S. ≥18 are not up-to-date on recommended cervical cancer screening (13). For example, disadvantaged subgroups in the US, including uninsured women and those of lower socioeconomic status were the least likely to report being up to date with cervical cancer screening compared to women of higher educational attainment (13,14). Additionally, screening adherence may decrease as women approach 65, thereby increasing the likelihood that women have not been adequately screened prior to the upper age cutoff (15). As many as 58% of women 64 to 66 years old in a national database failed to meet the criteria to exit screening (16).

We therefore sought to examine cervical cancer stage at diagnosis and relative survival in women ≥65. Previous studies have focused on differences in incidence, mortality, and therapeutic management for younger women (<65) diagnosed with cervical cancer vs. older groups (7,9,15). However, they have not considered adjusted sociodemographic and clinical characteristics associated with late-stage cervical cancer in women ≥65 or associated relative 5-year survival by stage at diagnosis compared to younger age groups within the screening criteria. Therefore, this study utilized large population-based data from the California Cancer Registry (CCR) to examine cervical cancer relative survival by stage at diagnosis and characteristics associated with late-stage disease among women ≥65.

Among 12,442 women diagnosed from 2009–2018 with a first primary cervical cancer, 17.4% (2,171) were aged ≥65 years (Table 1). The proportion of women diagnosed late stage increased with increasing age through age 75 to 79 (Figure 1). Among women ≥65 years, the proportion diagnosed late stage ranged from 60.2% to 70.6% compared to 33.5% to 58.7% for those ages 21 to 64 years. Among women diagnosed with early-stage disease, 5-year relative survival was lowest for women 80 years and older at 51.6%. For those 20–39, 40–59, 60–64, 65–69, 70–74, and 75–79,5-year relative survival was 93.6%, 92.9%, 89.3% 81.5%, 86.2%, and 72.0%respectively. Likewise, among those with late-stage cervical cancer, 5-year relative survival was lowest for women 80 years and older at only 23.2%. For those 20–39, 40–59, 60–64, 65–69, 70–74, and 75–79, 5-year relative survival was 51.5%, 47.5%, 41.5%, 36.8%, 39.1%, and 30.8% respectively (Figure 2).

Of 2,171 women ≥65, most patients were aged 65–69 (34.6%) followed by 70–74 (24.7%), 80 plus (23.1%) and 75–79 (17.5%). More women with early-stage disease at diagnosis were younger (65.7% 65–74 years and 34.3% ages 75 years and older) compared to women with lates-stage disease at diagnosis (56.7% 65–74 years and 43.3% ages75 years and older). The most common histologic subtype for both early-stage and late-stage diagnoses was squamous cell carcinoma (67.0% early-stage; 69.3% late-stage), followed by adenocarcinoma (27.1% early-stage; 19.3% late-stage), and other subtypes (5.9% early-stage; 11.3% late-stage) (Table 1). Compared to patients diagnosed at early-stage, those diagnosed at late-stage had ≥ 2 comorbidities, were unmarried, and had non-adenocarcinoma histologic subtypes.

In multivariable logistic regression models, among patients ≥65, factors associated with late-stage included older age at diagnosis (increase with each additional year of age, OR 1.02; CI, 1.01, 1.04), ≥ 2 comorbidities (vs. no comorbidities: OR, 1.59; CI, 1.21, 2.08), squamous cell carcinoma histology (OR, 1.38; CI 1.10, 1.74), or other subtypes (OR, 2.52; CI 1.68, 3.79) vs. adenocarcinoma (Table 2). Women of Hispanic ethnicity were less likely to be diagnosed with late-stage disease (OR 0.76; CI 0.60, 0.97) compared to non-Hispanic White women. Year of diagnosis, marital status, health insurance status, nSES, and urbanicity were not shown to be significantly associated with late-stage cervical cancer.

In sensitivity analyses including unknown stage at diagnosis with late-stage, unmarried women were more likely to be diagnosed late-stage (OR 1.27; CI 1.04, 1.56) (Table S1). Late-stage disease was associated with older age, comorbidity, non-adenocarcinoma histology and less associated with Hispanic ethnicity, consistent with the main analysis that excluded unknown stage.

In our large population-based study in California, nearly 1 in 5 new cervical cancer cases diagnosed from 2009–2018 were in women ≥65, and these older women had lower 5-year relative survival for both early- and late-stage diagnoses than younger women, with women 80 years and older having the lowest survival of all age groups. We also observed that the proportions of late-stage diagnoses increased up to age 79 years. Among women ≥65, those who were older, had non-adenocarcinoma histology, and had comorbidities were more likely to be diagnosed with late-stage disease.

Our study confirms findings from prior US-based studies that have noted substantial burden of cervical cancer, high rates of late-stage disease, and worse survival in women ≥65 (7–10,15). However, to our knowledge, prior studies have not examined cervical cancer 5-year relative survival by age group and stage at diagnosis. Our study found worsening 5-year relative survival with each increasing age grouping category for both early and late-stage diagnoses.

While we cannot determine the reason for these age disparities with our data, several factors may contribute. More late-stage diagnoses may be due to inadequate screening in women approaching 65 as noted previously (15). Some reasons cited for forgoing screening include discomfort, pain, embarrassment, and the intrusiveness of speculum-based exams (25,26). Another reason can be lack of follow-up after an abnormal screen. As many as 50% of women with abnormal results do not receive follow-up care (27,28). Comorbidities can also contribute to late-stage cancer diagnoses. Comorbidities increase with age and can result in a delayed diagnosis because of distraction from other health issues (29,30). Worse survival with older age can result from more late-stage diagnoses but can also result from less aggressive treatments in women ≥65. Eggemann et al. reported that patients ≥ 61 were less likely to undergo surgery and radiochemotherapy compared to patients <61 (31). Diver et al. likewise found that women ≥65 were less likely to receive surgery (7).

Another issue that could contribute to late-stage diagnosis is the use of supracervical hysterectomy. This procedure leaves the cervix intact and accounted for approximately 7.1% of laparoscopic hysterectomies in 2016 in the US (32). Unfortunately, some women do not realize the need to continue screening; Mattingly et al. found that only 67% of women who had undergone minimally invasive hysterectomy correctly identified whether their cervix had been removed and if they needed screening (33). McHale et al. found that 5.3% of women undergoing a secondary resection of a retained cervix after supracervical hysterectomy had cervical cancer (34).

Despite adequate prior screening, some women ≥65 are still diagnosed with cervical cancer (35,36). Pap testing can be difficult post-menopause when cytology tests may become less sensitive due to retraction of the squamocolumnar junction and vulvovaginal and cervical epithelium atrophy (26,37). Additionally, Pap testing is more effective at detecting squamous cell carcinoma and its precursors than adenocarcinoma which has been increasing in incidence (35,38,39). However, our study found that women with non-adenocarcinoma histologies were more likely to be diagnosed late-stage.

HPV testing provides increased sensitivity for cervical cancer precursors compared to Pap smear testing (11,40,41), but there is evidence that many women approaching age65 have not received HPV testing (42) and acceptance of it has been found to be low (43). Less intrusive options for HPV testing are becoming available including self-testing through vaginal swabs and urine collection and have been successfully used in other countries (44–47). Self-testing has been shown to be accurate and a good alternative to speculum-based exams that can be a reason to delay or altogether avoid testing as noted (25,48).

Year of diagnosis was not associated with stage at diagnosis, indicating that the proportion of late-stage cervical cancer diagnosis has not changed over time among women ≥65 in California. This differs from a previous US study that found that the proportion of regional and distant cervical cancer has in fact increased from 2001–2009 in most states, which might have been driven by the removal of pre-malignant lesions and the resulting increase in the proportion of invasive cervical cancer in unscreened women and those who did not receive adequate follow-up (49). However, our data are more recent and restricted to women ≥65, for many of whom advances in screening technologies and changes in guidelines were not applicable. This highlights the importance of ensuring women ≥65 have met the screening criteria prior to exiting as well as strategies aimed at enhancing follow-up after abnormal screening tests.

Prior studies of younger women have found increased late-stage cervical cancer diagnoses among women of Hispanic ethnicity, African American/Black race, and lower SES (49–51). Our study did not observe these associations and instead found that older Hispanic women were less likely than non-Hispanic White women to be diagnosed late-stage. Consistent with prior studies (8), our sensitivity analyses suggest that unmarried women ≥65 were diagnosed more often with late-stage cervical cancer. Increased late-stage diagnosis in unmarried patients could be related in part to differences in exposure to sexually transmitted infections between unmarried women and married women, as HPV is the causative agent in approximately 91% of cervical cancer cases (52). While the percent of HPV cases attributable to reactivation remains constant with age (18%−36%), the remaining HPV infections are caused by sexual exposure (53). The risk of acquiring HPV increases for women with new, casual, or concurrent sexual partners (53). Additionally, unmarried patients may be diagnosed late-stage more often due to economic disadvantage and less social, emotional, and practical support compared to married patients (54).

Our study had some limitations. We were unable to determine adherence to screening guidelines of our cohort or capture any HPV information, such as new or recurring infection. Additionally, we excluded 10.3% of women ≥65 (n=249) with unknown stage from our main analysis. However, when we included these women in our sensitivity analysis, our logistic regression results were similar. Despite these limitations, we utilized high-quality, large, population-based registry data to evaluate the stage at diagnosis and associated relative survival of women ≥65 diagnosed with cervical cancer.

In California, nearly one fifth of cervical cancers were diagnosed in women ≥65 and the majority were late-stage. Late-stage cervical cancer diagnoses increased with age and were associated with low 5-year relative survival of 36.8% to 23.2%. Our findings highlight the need to better understand how the current screening paradigm might be failing women ≥65. Future work should focus on determining past screening history of older women, determining lapses in follow-up care, and non-invasive testing approaches for women nearing age 65 or those who might need catch up screening.