Disease progression modeling analysis of the change of bone mineral density by postoperative hormone therapies in postmenopausal patients with early breast cancer
SeokKyu Yoon (1), Kyun-Seop Bae (1), Yong-Soon Cho (1), Sunpil Han(1), Hyeongsub Kim(1), and Hyeong-Seok (1)
(1) Department of Clinical Pharmacology and Therapeutics, College of Medicine, University of Ulsan, Asan Medical Center, Pungnap-2-dong, Seoul, Republic of Korea
Introduction: The adverse effects of AI except on the musculoskeletal system are less than tamoxifen in breast cancer patients, but the probability of osteoporosis is much higher in the postmenopausal patients. [1],[2]
In contrast, tamoxifen is known to reduce the risk of bone loss in postmenopausal women. Because Tamoxifen is a partial antagonist/agonist of estrogen, tamoxifen possesses the anti-osteoclastic effects of estrogen, whereas, AI suppresses estrogen and inhibits its anti-osteoclastic effects.
Objectives:
The major purposes of this study were followed as:
To investigate the effects of tamoxifen and third-generation AI (letrozole or anastrozole) on BMD after breast cancer surgery in the postmenopausal breast cancer patients using disease progression modeling.
Methods:
The clinical data of the postmenopausal patients who had received postoperative, hormonal therapy and met the predefined selection criteria were retrospectively collected in an anonymized way. Disease-progression modeling and simulations were performed using NONMEM® version 7.42. In this study, we first implemented empirical models used in previous papers for disease progression model and then modified.[3] A first-order deterioration model with a combination of a symptomatic model (when a drug effect provides a transient bad effect by offsetting the severity of the disease) and a disease-modifying model (when a drug affects the disease progression rate) was used.[3]
The model was validated by graphs as well as statistical criteria. Standard goodness-of-fit plots including dependent variable (DV) versus population predicted values (PRED) or individual predicted values (IPRED) and the individually weighted residuals versus PRED/IPRED were used for the diagnosis. The predictability of the final model was performed through a visual predictive check (VPC) with 1,000 replicates with simulation. A nonparametric bootstrap was performed with 1,000 replicates, and the mean with 95% confidence interval (CI) was calculated. Based on the final model constructed, the changes in BMD over time depending on therapy with postoperative third-generation aromatase inhibitors, tamoxifen, and no hormone was predicted and compared by Monte-Carlo simulation.
Results:
While AI affected BMD, tamoxifen did not significantly affect BMD in the disease-progression modeling. Hence, we did not include tamoxifen in the final model. The change rate (KDET) was -0.00075 T-score per month, AI effect (AIEFF) was -0.21 T-score, vitamin D supplement effect (VITDEFF) was 0.0067 T-score per month, and calcium supplement effect (CALEFF) was -0.088 T-score. The VPC predicted a 95% prediction interval reasonably well between the model and the real data.
An 84-month (7 years) simulation with the group receiving AI and the group without receiving AI was performed. The simulation showed that the BMD was low in the group receiving AI compared to the group receiving tamoxifen group or the group receiving neither of the treatments . Also, simulation showed that the probability of osteoporosis increased by 10% in the AI group compared with other groups.
Conclusions:
In this study, the effects of AI and tamoxifen on BMD was quantitatively analyzed. AI was found to lower BMD by about -0.21 in T-score during its treatment period. Tamoxifen, on the other hand, had no significant effect on BMD and vitamin D supplement had a disease-modifying protective effect in osteoporosis. The simulation demonstrated that AI has a negative bone mineral effect and the probability of osteoporosis increased by 10% in the AI group compared with other groups. The current study suggested that compared to tamoxifen, AI had a worse effect on osteoporosis which is consistent with the previous reports.
References:
[1] Pamela Taxel, Palak Choksi, and Catherine Van Poznak. The Management of Osteoporosis in Breast Cancer Survivors. j.maturitas.2012;73(4):275-279
[2] Bhuvaneswari Ramaswamya. Osteopenia and osteoporosis in women with breast cancer. Seminars in Oncology. 2003;30(6):763-775
[3] Sarah F. CookRobert R. Bies : Disease Progression Modeling: Key Concepts and Recent Developments., Pharmacometrics Volume 2, Issue 5, pp 221–230