Uterine cancer is the most common gynecologic malignancy in the United States, with an estimated 67,880 cases and 13,250 deaths projected in 2024. Over the past thirty years, both the incidence and death rate from uterine cancer have risen dramatically. A National Cancer Institute report noted an average annual percentage increase of 0.8% for uterine cancer cases from 2014 to 2018, while mortality increased 1.9% annually from 2015 to 2019, the highest for any malignancy in women.
The rising incidence is multifactorial. The aging of the US population has contributed to the trend since uterine cancer predominantly affects older women. Concurrently, the increasing prevalence of overweight and obesity has played a significant role, as adipocytes produce estrogen that stimulates the endometrium, making obesity one of the strongest risk factors for uterine cancer. However, aging and obesity alone do not fully explain the changing trends in incidence and mortality.
The researchers developed the Columbia University Uterine Cancer Model (UTMO), a state-transition microsimulation model created as part of the National Cancer Institute's CISNET consortium. The model begins at age 18 and uses a one-month cycle length, simulating non-Hispanic White and non-Hispanic Black women separately to account for well-documented racial disparities in uterine cancer incidence, histologic type, and mortality.
The model includes four general states: healthy, precursor lesion (endometrial intraepithelial neoplasia or EIN), cancer, and death. It separately models endometrioid and non-endometrioid tumor types. The model incorporates population-level prevalence of hysterectomy using NHANES data, and obesity is modeled as a categorical variable by age, race, and birth cohort, with higher BMI increasing the probability of transitioning to undetected cancer states.
Model calibration was performed in two phases: a multicohort phase and a cohort-specific phase. The primary calibration target was SEER-18 uterine cancer incidence stratified by age, birth cohort, race, histology, and AJCC stage for patients aged 18 to 84. The researchers used simulated annealing, an optimization algorithm, running 1,000,000 iterations with progressively smaller step sizes to iteratively improve model performance.
Validation was performed by comparing model-estimated cancer incidence and incidence-based mortality to SEER estimates among women aged 40 and older. The model was also stress-tested using the CISNET-recommended maximum clinical incidence reduction (MCLIR) scenario analysis, which tested four hypothetical screening and treatment interventions to assess model robustness.
The model closely fit SEER incidence and mortality data available through 2018. For uterine cancer incidence, the model achieved a normalized sum-squared error (NSSE) of 0.064 in White women and 0.084 in Black women. For incidence-based mortality, the NSSE was 0.124 in White women and 0.333 in Black women, demonstrating excellent model validity.
Predicted 5-year survival in 2013 was 92.6% for White and 86.4% for Black patients with endometrioid tumors, closely matching actual survival rates. For non-endometrioid tumors, predicted 5-year survival was 45.9% for White and 37.5% for Black patients. The model-predicted median age of diagnosis matched actual data closely across all histologic subtypes and racial groups.
From 2020 to 2050, the incidence of uterine cancer is projected to continue increasing in both Black and White women. In White women, incidence will reach 74.2 cases per 100,000 by 2050, up from 57.7 in 2018. In Black women, the rate will rise to 86.9 per 100,000, up from 56.8 in 2018. Among White women, incidence-based mortality will increase from 6.1 to 11.2 per 100,000, while in Black women it will increase from 14.1 to 27.9 per 100,000.
When incorporating projected hysterectomy and obesity trends, the projections are even higher. In White women, incidence could reach 92.9 cases per 100,000 by 2050, while in Black women it could reach 102.0 per 100,000. The corresponding mortality projections also increase substantially under these scenarios, with Black women experiencing a disproportionate burden of the disease.
When stratified by histology, endometrioid tumors are projected to increase considerably in both racial groups. However, a critical disparity emerges with non-endometrioid tumors: while White women will experience only a slight increase, the incidence rate in Black women increases substantially from 22.5 per 100,000 in 2018 to 36.3 per 100,000 in 2050. Non-endometrioid tumors are aggressive subtypes that account for a disproportionate number of deaths from uterine cancer.
By 2050, the incidence-based mortality rate from uterine cancer is expected to be nearly three times greater in Black compared to White women. Black women more often face delays in diagnosis, barriers in accessing care, present with more advanced stage disease, and receive suboptimal care. The disparate rise in non-endometrioid cancers in Black women is a key driver of the increasing mortality gap.
Currently there is no routine screening or prevention strategy for uterine cancer, which is most commonly detected when women present with symptoms such as abnormal bleeding. The model's stress testing using MCLIR methodology showed that interventions applied to women at age 55 produced larger and longer-lasting reductions in cancer incidence compared to those applied at age 45, with effects persisting for 15 to 16 years.
New tests leveraging genomic changes in serum and exfoliated uterine cells have shown promising preliminary results for early detection of uterine cancer. The natural history model developed in this study provides a framework for evaluating the potential role of such tests in screening and early detection strategies. Prior studies have suggested that conventional screening with transvaginal ultrasound and endometrial biopsy are not cost-effective for asymptomatic women.
As with any microsimulation model, the UTMO relies on population-level estimates of risk factors, protective factors, cancer incidence, and mortality. The model does not currently account for all potential risk factors, including reproductive factors such as parity, age at menopause, and age at menarche. Additionally, forecasting future rates of hysterectomy and obesity introduces uncertainty into the long-term projections.
The current iteration of UTMO only includes Black and White women and does not incorporate uterine sarcomas. Future model development will expand to other racial and ethnic groups and include sarcoma subtypes. Despite these limitations, the model estimates closely align with SEER data, supporting the reliability of both endometrioid and non-endometrioid projections.
The incidence and mortality of uterine cancer are projected to increase substantially over the next three decades. Black women will experience a disproportionate increase in the disease, with the racial gap in mortality expected to widen considerably. The number of new cases of uterine cancer in the US increased by more than 50% between 2010 and 2020, and the model predicts these trends will continue through 2050.
These population-level trends support the urgent need to develop and implement novel primary and secondary prevention strategies for uterine cancer. The natural history model provides a powerful tool for evaluating potential interventions and projecting the future burden of disease, which is essential for cancer control planning and resource allocation.