Real-world data suggest that geriatric assessment fails to stratify risk in treating older adults with lung cancer

In the coming decades the elderly population will make up about 20% of the total population and the percentage of patients over 75 and over 85 will increase disproportionately. The importance of this demographic trend is magnified by the sharp increase in the incidence of most cancers after age 60. Currently, approximately 50 percent of all malignancies and 70 percent of cancer deaths occur in those ≥65 years of age and this fraction tends to increase.1 2 3

Older adult (OA) patients require special attention to the risks of chemotherapy (treatment-related toxicities and quality of life issues). Aging is associated with decreased organ functions, muscle wasting, multiple comorbidities, polypharmacy with likelihood of harmful drug interactions and risk of adherence to therapy.4 5 6 7 8 9 10 11 12 13 14 15 16 For critically ill patients with severe comorbid disease that results in an estimated short life span (particularly one year or less) the risks of chemotherapy often outweigh their benefits. Older patients, then, are at risk for decompensation due to exposure to severe metabolic stresses such as chemotherapy and most of these patients should receive supportive care rather than chemotherapy, with attention to symptom palliation.17

Thus, age and comorbidities may limit the ability of full cancer treatment for advanced non-small cell lung cancer (NSCLC). The median age of newly diagnosed NSCLC patients in developed countries is approximately 68 years and up to 40% are aged ≥70 years at diagnosis.18 In addition, patients with limited or low performance status (PS) [ECOG≥2 or Karnofsky Performance Status (KPS) ≥70] comprise 30 to 40 percent of patients with advanced NSCLC.19 About one third of patients with metastatic disease and most patients who are in the other stages of the disease are not able to undergo a potentially curative surgical resection.20

The most current first-line palliative treatment options for these patients comprise oral target therapy, immunotherapy and chemotherapy combinations, all with robust survival benefits when compared to chemotherapy alone.21 22 23 24 25 26 27 Target therapies, tyrosine kinase inhibitor (TKI) against EGFR, ALK, ROS1, and NTRK, require specific molecular tests to detect the mutations and predict response to these drugs.21 22 23 24 The indication of isolated or combined checkpoint inhibitors also requires specific tests to evaluate expression of PD-L1.25 26 And, once target TKI mutations or high PD-L1 expression are excluded, combinations of chemotherapy with immunotherapy have become the standard treatment.27 However, these treatments require specific tests that are not accessible and, in fact, new less toxic therapies are very restricted for the vast majority of patients in emerging countries, imposing the use of traditional chemotherapies or first-generation target therapies as the only treatment option despite all the recent advances in lung cancer therapies.

Regarding chemotherapy treatment for lung cancer in older patients, some studies show that appropriately selected patients and those with borderline performance status (PS 2), chemotherapy can improve overall survival without adversely affecting quality of life compared to the best supportive care for advanced NSCLC.28 29 30 The American Society of Clinical Oncology Guideline for Geriatric Oncology and a panel of European experts have not definitively resolved the issue of combination versus mono chemotherapy for patients with advanced NSCLC who are older or have low PS.31 32

The aim of this study was to evaluate the clinical data and outcomes of an elderly population with lung adenocarcinoma undergoing palliative treatment with chemotherapy or first-generation target therapy after comprehensive geriatric assessment (GA) to attempt to determine detectable prognostic factors in a scenario of scarce resources, as is the case of most oncology services in emerging countries, which do not always have access to molecular testing, lastgeneration TKI and/or immunotherapy.

A retrospective analysis was performed by collecting clinical data from the electronic medical records of patients with stage IIIB and IV lung adenocarcinoma aged 70 years or older diagnosed and treated with chemotherapy or first- generation TKI in 1^st or 2^nd line between January 2007 to December 2015 in the AC Camargo Cancer Center undergoing GA. The definition of GA for the study was evaluation of functional capacity, comorbidities, polypharmacy, and nutritional status. As a basis for analysis, an evaluation table was used containing epidemiological data (age, gender, ethnicity, and care network), specific pathological data (staging, whether initial or relapsing, and number of metastasis sites), known predictive and prognostic data (hemoglobin, ECOG, and presence of driver mutation), comorbidities (smoking, cardiovascular disease, diabetes mellitus, chronic kidney disease, neurological diseases, respiratory disease, smoking, incontinence and fall, polypharmacy (≥ 5 medications), Charlson score index, and Katz and Lawton functional assessments and nutritional status), cancer treatment (chemotherapies or TKI used in 1^st and 2^nd palliative lines, response rates, and toxicities), and progression or death. Data were tabulated and statistically analyzed to describe the population with frequencies, means and medians. Overall survival (OS) was calculated according to the Kaplan-Meier method, and we performed univariate and multivariate analysis using the Cox method to describe potential prognostic factors for OS, with forced entry of functional capacity.

Between the periods from January 2007 to December 2015 in AC Camargo Cancer Center, 54 patients aged ≥70 years with pulmonary adenocarcinoma who underwent GA were treated palliatively. Demographical data are presented in Table 1.

The median age was 76 years, ranging from 71 to 88 years. The distribution by gender was equal. Regarding performance status, 22% had ECOG≥2 and all were functionally classified according to activities of daily living (ADL) with 76% of patients classified as Katz A and 50% as Lawton <27. In the assessment of nutritional status, 68.5% were at risk of malnutrition or malnourishment (weight loss >5% during prior 6 months or >2% for patients with low BMI or sarcopenia). The average age-adjusted Charlson comorbidity index was 10, with 33% of the population studied above this average; 61% had at least 2 comorbidities and 48% had polypharmacy (≥5 drugs). Regarding cancer disease, 26 patients had at least 2 metastatic sites. The central nervous system (CNS) was affected in 7.5% of cases and the liver in 9.3%. At treatment evaluation, 79.6% of patients underwent chemotherapy and the others received TKI. After a 29-month follow-up, the mean OS was 17.1 months (13.5-27.8 months). In the univariate analysis (Table 2), we observed significantly lower survival results for patients with ECOG≥2 (Hazard Ratio [HR]=5.9; p<0.001), number of metastatic sites ≥2 (HR=2.0; p=0.04) and presence of liver metastasis (HR=12.6; p<0.001). In multivariate analysis (Table 2), male, ECOG≥2, more than 2 metastatic sites and the presence of liver metastases were associated with higher risk of death. We did not observe any difference between patients classified according to worse functional capacity (Katz non-A) in univariate or multivariate analysis (HR=1,6; p=0,18 and HR=0,64; p=0,25, respectively).

Concerned about optimized NSCLC care in OA who does not have access to new diagnostic tests involving PD-L1 expression, target mutations for TKI and new less toxic treatments, such as immunotherapy and last-generation TKI, we made an assessment of survival outcomes, to correlate them with prognostic factors to define how to continue cancer treatment while offering the only accessible therapy, although outdated, respecting physiological and morbid limitations of this more frail population, seeking to improve the general situation with interventions and minimize toxicities and inadequacies of treatment according to GA.

We note that most patients are at risk of malnutrition or are malnourished and half of them have polypharmacy and do not have maximum independence for instrumental activities of daily living (Lawton <27). In spite of that, comorbidities are not very prevalent (only 40% had at least 2 comorbidities) and only a third of our patients received a Charlson score above 10. The end result is that almost 80% of our population has an ECOG 0 or 1. Even though this is not such a frail population, they was at a time immediately preceding access to immunotherapy and the most modern tyrosine kinase inhibitors. Unfortunately, this is still the very reality in most cancer treatment sites in emerging countries. As a result, about 80% of these patients underwent palliative treatment with single-drug or combination chemotherapy. Unlucky, we were unable to obtain quality data regarding the toxicity of treatments that would allow better correlation with survival outcomes.

In our study, we found that the main prognostic factors that negatively impacted the OS on univariate analysis were ECOG≥2, more than 2 sites of metastasis and presence of liver metastasis, which was present in a minority of the patients. This could explain the beyond expected median survival of 17.1 months after a median follow-up of 29 months. In multivariate analysis, these factors plus male gender were associated with higher risk of death. Malnutrition, although very prevalent in our sample, did not impact survival. After all, the tumor biology and the effectiveness of chemotherapy may have a higher prognostic impact when compared to clinical features that compose GA.

We understand the limitations of this work such as the lack of formalization in the quality of life assessment, the use of the less broad concept of GA, which could also include psychological evaluation, geriatric, neuro-cognitive and socio-financial syndromes, so called comprehensive geriatric assessment (CGA). Still, we found no prognostic impact of functional capacity and this is in agreement with a multicenter, open-label, phase III ESOGIA study that evaluated over four hundred patients ≥70 years with a PS of 0 to 2 and stage IV NSCLC and randomly assigned them between chemotherapy allocation based on performance status (standard arm) and age or CGA (carboplatin-based double for fit patients, docetaxel for vulnerable patients, and best supportive care for frail patients). This trial failed to show differences in treatment failure-free survival (primary endpoint) and OS, but patients experienced significantly less toxicity across all grades (85.6% vs. 93.4%; p=0.015) and fewer failures in treatment as a result of toxicity (4.8% vs. 11.8%; p=0.007) compared with patients in the standard arm.33

This results do not minimize the importance of GA and ideal patient care demand participation of a multidisciplinary team attentive to the care of older cancer patients carrying out a broader geriatric assessment, evaluating and taking care of patients' comorbidities and nutrition, always taking into account the risks of polypharmacy, with dynamically personalizing and individualizing treatment that can provide a better quality of life.34

Finally, since GA and CGA proved to be an insufficient tool for selecting systemic treatment with more obsolete therapies against NSCLC, promoting broad access to new prognostic and predictive tests and modern therapies could allow not only less toxicity to this fragile population, but mainly gains in overall survival and quality of life. Thus, we should observe a greater governmental effort in favor of this cause, with investments in medical science, reducing the costs of new drugs, promoting greater access to new diagnostic tests and investing in infrastructure of oncology centers and in the training of its specialists. These measures would certainly advance the care of older patients with metastatic lung cancer in emerging countries.