BACKGROUND: Cisplatin with definitive radiotherapy (RT) is considered the standard of care for cervical cancer; however, older women are frequently undertreated and have worse outcomes compared with younger patients. Because women aged ≥65 years have been disproportionately underrepresented in clinical trials, uncertainties exist regarding how much they benefit from the addition of cisplatin to RT. PATIENTS AND METHODS: Women aged ≥65 years with nonmetastatic cervical cancer treated with definitive external-beam RT and brachytherapy were identified in the SEER-Medicare database. Death attributable to cervical cancer (cancer-specific mortality [CSM]) was evaluated against competing risks of death using Gray's test. Propensity score analysis and the Fine-Gray multivariable regression model were used to adjust for baseline differences, including comorbidity. RESULTS: The total cohort comprised 826 patients, of whom 531 (64%) received cisplatin, 233 (28%) were FIGO stage I, 374 (45%) were stage II, and 219 (27%) were stage III-IVA. Older age and chronic kidney disease significantly predicted omission of cisplatin. Virtually all cisplatin dosing was weekly, with a median of 5 cycles. Death from cervical cancer was significantly lower with cisplatin than without (5-year CSM, 31% vs 39%; P=.02; adjusted hazard ratio, 0.72; P=.02), which persisted in propensity score analysis. Receiving ≥5 cycles was required for benefit, as no difference in CSM was seen in patients receiving 1 to 4 cycles versus no cisplatin. Subgroup analyses revealed that the benefit of cisplatin persisted in women aged ≥75 years and those with early-stage disease. Incidence of cytopenia, nausea/vomiting, and hypovolemia increased in patients treated with cisplatin. CONCLUSIONS: Administration of cisplatin with definitive RT in women aged ≥65 years was associated with a significant benefit in the incidence of death attributable to cervical cancer, despite competing risks for mortality in an older population. Receiving at least 5 cycles of weekly cisplatin was required for benefit.
BACKGROUND:Cisplatin with definitive radiotherapy (RT) is considered the standard of care for cervical cancer; however, older women are frequently undertreated and have worse outcomes compared with younger patients. Because women aged ≥65 years have been disproportionately underrepresented in clinical trials, uncertainties exist regarding how much they benefit from the addition of cisplatin to RT. PATIENTS AND METHODS: Women aged ≥65 years with nonmetastatic cervical cancer treated with definitive external-beam RT and brachytherapy were identified in the SEER-Medicare database. Death attributable to cervical cancer (cancer-specific mortality [CSM]) was evaluated against competing risks of death using Gray's test. Propensity score analysis and the Fine-Gray multivariable regression model were used to adjust for baseline differences, including comorbidity. RESULTS: The total cohort comprised 826 patients, of whom 531 (64%) received cisplatin, 233 (28%) were FIGO stage I, 374 (45%) were stage II, and 219 (27%) were stage III-IVA. Older age and chronic kidney disease significantly predicted omission of cisplatin. Virtually all cisplatin dosing was weekly, with a median of 5 cycles. Death from cervical cancer was significantly lower with cisplatin than without (5-year CSM, 31% vs 39%; P=.02; adjusted hazard ratio, 0.72; P=.02), which persisted in propensity score analysis. Receiving ≥5 cycles was required for benefit, as no difference in CSM was seen in patients receiving 1 to 4 cycles versus no cisplatin. Subgroup analyses revealed that the benefit of cisplatin persisted in women aged ≥75 years and those with early-stage disease. Incidence of cytopenia, nausea/vomiting, and hypovolemia increased in patients treated with cisplatin. CONCLUSIONS: Administration of cisplatin with definitive RT in women aged ≥65 years was associated with a significant benefit in the incidence of death attributable to cervical cancer, despite competing risks for mortality in an older population. Receiving at least 5 cycles of weekly cisplatin was required for benefit.