• 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • br We used the exponential tumor growth


    We used the exponential tumor growth model to model tumor volume-doubling time [12]. Our model assumed that patients’ tumor characteristics were known at the time of treatment, and we obtained a woman’s tumor size at diagnosis and predicted the number of nodes at diagnosis using a Poisson linear model, given age and tumor size based on Surveillance, Epidemiology, and End Results (SEER) registry data [13]. We simulated hormonal status independently, allowing 75% to be either ER- or PR-positive [14e17] and 25% to be HER2-positive [18], and determined the stage of disease using the tumor-node-metastasis staging system, given tumor size and the predicted number of nodes [19].
    Model Validation
    We validated our natural history model of breast cancer with the US breast cancer incidence rate between 1973 and 1979, repre-senting a time 174722-31-7 in which the observed breast cancer inci-dence was not altered by widespread use of screening mammography [20]. We extracted the breast cancer incidence rate from SEER data for 1973 to 1979, which was reported at 5-year age intervals, and applied linear extrapolations to derive age-specific breast cancer incidence at 1-year age intervals. We cali-brated model parameters related to age-specific mortality and age-specific breast cancer incidence against the age-specific incidence derived from the SEER data. Figure 1 compares the age-specific breast cancer incidence rates generated from our natural history model with those obtained from SEER data for 1973 to 1979. As shown, incidence rates generated from our calibrated natural history model match well with the observed age-specific incidence in the United States.
    Modeling the Impact of Screening and Diagnostic Procedures
    The clinical parameters in our model are presented in Table 1. The sensitivity of mammography increases with tumor size and age at screening, and the specificity of mammography increases with age at screening [21e26]. Building on the range of sensitivity and specificity of digital mammography reported in a recent system-atic review [27], we modeled age-specific and tumor sizeespecific sensitivity and age-specific specificity for mammography using a logit model [26,28] and accounted for uncertainty using a beta distribution for each parameter of sensitivity and specificity. We
    Fig. 1 – Comparison of Model-Generated Breast Cancer Incidence and Age-Specific Incidence Observed in the SEER data. Note: SEER ¼ Surveillance, Epidemiology, and End Results. Vertical lines are 95% confidence interval of age-specific incidence estimated from the microsimulation model.
    Parameter Values Data source
    Clinical parameters
    Performance characteristics 40 y age < 50 y
    tumor sizeedependent)
    Diagnostic MMG sensitivity
    Diagnostic MMG specificity
    Hazard reduction
    Hormonal therapy, ER-/PR-positive
    Trastuzumab for HER2-positive
    Health utility parameters: utility value and associated duration
    Terminal stage, breast cancer 0.29, last 3 mo of life
    Terminal stage, other diseases 0.375, last 3 mo of life
    Cost parameters
    Screening-related costs
    Mammography (digital; bilateral)
    Diagnostic mammography (digital; unilateral)
    Breast biopsy
    Treatment-related costs
    Tamoxifen for 5 y
    As adjuvant therapy/y
    For metastatic breast cancer
    Annual costs by phase and stagez See Table 2
    Centers for Disease Control and Prevention [9]
    Centers for Medicare & Medicaid Services [42]
    Centers for Medicare & Medicaid Services [42]
    Centers for Medicare & Medicaid Services [50]
    Note. Random variation is added using a beta distribution for all sensitivity and specificity estimates. See description under “Modeling the Impact of Screening and Diagnostic Procedures” subsection in Methods. Several parameters are presented as ranges because the sensitivity and
    specificity of screening mammography depend on the tumor size and/or the patient’s age.
    MMG, mammography; RT, radiation; SEER, Surveillance, Epidemiology, and End Results. * Range based on the internal quartiles of sensitivity and specificity estimated from the authors’ statistical models palynology used a beta distribution to account for uncertainty. y Calculated on the basis of the dosage for an average patient with height 170 cm and weight 70 kg and using the average sales price plus 6%
    markup initial dose of 4 mg/kg over 90-min intravenous infusion, then 2 mg/kg over 30-min intravenous infusion weekly for 52 wk for adjuvant breast cancer and 7.4 mo for metastatic breast cancer [7].
    z Authors’ analysis of SEER-Medicare data.
    used diagnostic mammography as the form of the initial workup for abnormal findings from screening mammography, with sensitivity and specificity estimated from the Breast Cancer Sur-veillance Consortium [29,30]. A biopsy was used to confirm the disease after a positive finding from the initial workup. If a woman has a symptomatically detected tumor, her disease status would also be confirmed with both diagnostic mammography and a breast biopsy. Women in the preclinical state who are never diagnosed with breast cancer are assumed to die of other causes. For women receiving false-positive mammography, a follow-up mammography is provided 6 months after the last mammog-raphy [31].