Impact of comorbidity on short-term mortality and overall survival of head and neck cancer patients

Department of Otorhinolaryngology and Head and Neck Surgery, Erasmus Medical Centre, Rotterdam, The Netherlands.
Head & Neck (Impact Factor: 3.01). 01/2009; 32(6):728-36. DOI: 10.1002/hed.21245
Source: PubMed

ABSTRACT In 2001, we presented a Cox regression model that is able to predict survival of the newly diagnosed patient with head and neck squamous cell carcinoma (HNSCC). This model is based on the TNM classification and other important clinical variables such as age at diagnosis, sex, primary tumor site, and prior malignancies. We aim to improve this model by including comorbidity as an extra prognostic variable. Accurate prediction of the prognosis of the newly diagnosed patient with head and neck cancer can assist the physician in patient counseling, clinical decision-making, and quality maintenance.
All patients with HNSCC of the oral cavity, pharynx, and larynx diagnosed in the Leiden University Medical Centre between 1981 and 1998 were included. From these 1371 patients, data on primary tumor site, age at diagnosis, sex, TNM classification, and prior malignancies were already available. Comorbidity data were collected retrospectively according to the ACE27 manual. The prognostic value of each variable on overall survival was studied univariately by Kaplan-Meier curves and the log-rank test. The Cox regression model was used to investigate the impact of these variables on overall survival simultaneously. Furthermore, univariate analyses were performed to investigate the impact of comorbidity severity on short-term mortality and to investigate the impact of organ-specific-comorbidity on short-term mortality.
Comorbidity was present in 36.4% of our patients. Mild decompensation was seen in 17.4%, moderate decompensation in 13.5%, and severe decompensation in 5.5%. Most frequently observed ailments were cardiovascular, respiratory, and gastrointestinal. In univariate analyses, all prognostic variables, including comorbidity, contributed significantly to overall survival. Their contribution (except sex) remained significant in the multivariate Cox model. Internal validation of this model showed a concordance index of 0.73, indicating a good predictive value. Short-term mortality was seen in 5.7% of our patients. Cardiovascular comorbidity, respiratory comorbidity, gastrointestinal comorbidity, and diabetes showed a significant relationship with short-term mortality.
Comorbidity impacts overall survival of the newly diagnosed patient with HNSCC. There is a clear distinction between the impact of the 4 ACE27 severity grades. The impact of an ACE27 grade 3 is comparable to the impact of a T4 tumor or an N2 neck. Comorbidity impacts short-term mortality as well. Especially cardiovascular comorbidity, respiratory comorbidity, gastrointestinal comorbidity, and diabetes show a strong relationship.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The use of alternative modeling techniques for predicting patient survival is complicated by the fact that some alternative techniques cannot readily deal with censoring, which is essential for analyzing survival data. In the current study, we aimed to demonstrate that pseudo values enable statistically appropriate analyses of survival outcomes when used in seven alternative modeling techniques.
    PLoS ONE 06/2014; 9(6):e100234. DOI:10.1371/journal.pone.0100234 · 3.53 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Modern modelling techniques may potentially provide more accurate predictions of binary outcomes than classical techniques. We aimed to study the predictive performance of different modelling techniques in relation to the effective sample size ("data hungriness"). We performed simulation studies based on three clinical cohorts: 1282 patients with head and neck cancer (with 46.9% 5 year survival), 1731 patients with traumatic brain injury (22.3% 6 month mortality) and 3181 patients with minor head injury (7.6% with CT scan abnormalities). We compared three relatively modern modelling techniques: support vector machines (SVM), neural nets (NN), and random forests (RF) and two classical techniques: logistic regression (LR) and classification and regression trees (CART). We created three large artificial databases with 20 fold, 10 fold and 6 fold replication of subjects, where we generated dichotomous outcomes according to different underlying models. We applied each modelling technique to increasingly larger development parts (100 repetitions). The area under the ROC-curve (AUC) indicated the performance of each model in the development part and in an independent validation part. Data hungriness was defined by plateauing of AUC and small optimism (difference between the mean apparent AUC and the mean validated AUC <0.01). We found that a stable AUC was reached by LR at approximately 20 to 50 events per variable, followed by CART, SVM, NN and RF models. Optimism decreased with increasing sample sizes and the same ranking of techniques. The RF, SVM and NN models showed instability and a high optimism even with >200 events per variable. Modern modelling techniques such as SVM, NN and RF may need over 10 times as many events per variable to achieve a stable AUC and a small optimism than classical modelling techniques such as LR. This implies that such modern techniques should only be used in medical prediction problems if very large data sets are available.
    BMC Medical Research Methodology 12/2014; 14(1):137. DOI:10.1186/1471-2288-14-137 · 2.17 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Survival is poorer in elderly patients with head and neck squamous cell carcinomas [HNSCCs] than in younger patients. Possible explanations include a contribution of co-morbidities to mortality, frequent refusal of standard therapy, and the use of suboptimal treatments due to concern about toxicities. The Comprehensive Geriatric Assessment [CGA] is a multidimensional assessment of general health that can help to customise treatment and follow-up plans. The CGA has been proven effective in several health settings but has not been evaluated in randomised studies of patients with cancer. Our aim here was to assess the impact of the CGA on overall survival, function, and nutritional status of elderly patients with HNSCC.Methods/design: EGeSOR is an open-label, multicentre, randomised, controlled, parallel-group trial in patients aged 70 years or older and receiving standard care for HNSCC. The intervention includes four components: the CGA conducted by a geriatrician before cancer treatment, participation of the same geriatrician in cancer treatment selection, a standardised geriatric therapeutic intervention designed by the same geriatrician; and geriatric follow-up for 24 months. The primary endpoint, assessed after 6 months, is a composite criterion including death, functional impairment [Activities of Daily Living score decrease >=2], and weight loss >=10%. Secondary endpoints include progression-free survival, unscheduled admissions, quality of life, treatment toxicities, costs, and completion of the planned cancer treatment. A centralised online system is used to perform 1:1 randomisation with a minimisation algorithm for centre, age, T and N stages, and tumour site [oral, oropharyngeal, hypopharyngeal, or laryngeal]. The estimated sample size is 704 patients, who are being recruited by 14 centres in 9 French cities.
    BMC Cancer 06/2014; 14(1):427. DOI:10.1186/1471-2407-14-427 · 3.32 Impact Factor


Available from