The following studies depict two epidemiological hallmarks of cancer:
1. Bi-modal hazard rate
2. A declining hazard rate from the third year and on.

Hazard rates of recurrence following diagnosis of primary breast cancer
Ismail Jatoi1 , Anna Tsimelzon , Heidi Weiss , Gary M. Clark and Susan G. Hilsenbeck
Breast Cancer Research and Treatment 89: 2 2002 pp 173-178

The authors estimated hazard rates for the time of recurrence after breast cancer surgery in 3921 patients. Three curves ( Figure.1 in their publication) are displayed below.

Figure 1. Hazard function for the time of recurrence produced with Nelson–Aalen
method: (a) all patients; (b) patients with different number of nodes
involved; c) different tumor size (cm)

The curves display a bi-modal hazard that declines from the third year and onward. The hazard is proportional to the number of involved lymph nodes, and tumor size. From the third year and on hazard rate continually declines.

--------------------------------

Breast Cancer Mortality Trends in the United States According to Estrogen Receptor Status and Age at Diagnosis
Ismail Jatoi, Bingshu E. Chen, William F. Anderson, Philip S. Rosenberg
Journal of Clinical Oncology, Vol 25, No 13 (May 1), 2007: pp. 1683-1690

According to the hypothesis presented here, peak height is proportional to tumor dependence of the patient. The higher the peak the more dependent she is. As the tumor evolves it becomes less and less differentiated (more anaplastic). ER positive tumors are therefore younger than ER negative. The study shows, that when carrying a young (ER positive) tumor, patient depends less on her tumor than when the tumor is ER negative. Later on when metastasis spreads out the patient becomes less dependent on her primary tumor and the hazard rate declines. Finally when secondary cancer manifestations dominate the patient, the hazard rate rises until she dies.