Impact of differences in treatment strategy on relative survival of elderly patients with operable, hormone receptor positive breast cancer

Abstract

progression). At baseline all patients underwent a geriatric screening using G8 and the Flemish version of the Triage Risk Screening Tool (TRST) followed by a uniform Comprehensive Geriatric Assessment (CGA) including social data (marital status, living situation, educational level), Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), presence of falls, Mobility-Tiredness Test (MOB-T), Mini Mental State Examination (MMSE), Geriatric Depression Scale (GDS-15), Mini Nutritional Assessment (MNA), ECOGPerformance Status (ECOG-PS), Charlson Comorbidity Index (CCI) and polypharmacy assessment. Functionality of patients was followed 2 to 3 months after treatment decision by repeating ADL, IADL and presence of falls. Worsening of ADL and IADL was calculated on total scores. Worsening of ADL was defined as an increase of ≥2 points at follow-up compared to baseline score. Worsening of IADL was defined as a decrease of ≥1 points at follow-up compared to baseline score. In the group of patients receiving chemotherapy, treatment related toxicity grades III–IV were recorded using the NCI CTCAE, version 4.0 for adverse events. Univariate analysis was performed, followed by multivariate logistic regressions to identify predictive markers. Results: A total of 937 patientswere included from10/2009 till 07/2011 with a median age of 76 years and 63.5% women. Breast cancer was the most common malignancy (40.4%) followed by colorectal cancer (20.6%), hematological malignancies (15.9%), prostate cancer (9%), lung cancer (7.8%) and ovarian cancer (6.3%). Data from 811 patients were available at follow-up. Worsening of ADL and IADL was observed in 17.2% and 38.9% respectively. During follow-up, 17.5% of the patients had at least 1 fall. ADL worsening was predicted in multivariate model by baseline IADL, MNA and Flemish version of the TRST (all pb0.05). Worsening of IADL was predicted by baseline ECOG-PS, Flemish version of the TRST, GDS-15, and chemotherapy (all pb0.05). Presence of falls during follow-up could be predicted (pb0.05) by presence of falls during the year before inclusion, baseline ADL and G8, living situation and disease setting (new diagnosis or progression) at baseline. In the chemotherapy group (n=411), hematologic and nonhematologic grade III–IV toxicities developed both in 16% of the patients. Hematological disease (versus solid tumors) was the only baseline variable in univariate analysis predictive for severe hematological toxicity (pb0.05). G8 for hematologic toxicities, and MNA and IADL for non-hematologic toxicities were borderline associated (pb0.1). Conclusion: Parameters from CGA before treatment in older cancer patients can identify at-risk patients for decline in functionality and development of falls. Severe chemotherapy toxicity is not well predicted by baseline CGA parameters, but this effect might be obscured by the utilization of many different chemotherapy regimens with different toxicity profiles.