Discriminant analysis in predicting survival of patients with cancer of the larynx or hypopharynx

Abstract

In order to assess the feasibility of computer-assisted prognostication, step-wise discriminant analysis was applied to data retrospectively obtained from 243 patients who underwent surgery for cancer of the larynx. In all, 62 variables were studied in each patient. Ninety-four out of 243 patients had complete information on the 62 variables studied. With data from these 94 patients a linear discriminant function was obtained, with an overall accuracy of 87.2%. The rate of correct prediction was 84.6% for cancer recurrence, and 88.2% for survival without recurrence. Only 11 variables of the 62 used for analysis were necessary to obtain these results. This paper confirms the usefulness of multivariate analysis in clinical practice.

Prognostic determinants in supraglottic carcinoma: Univariate and Cox regression analysis

“…Performance status and the development of a metachronous tumor were additional parameters correlated with the RR of death for absolute survival. Although limited by the retrospective nature of the study, the results presented here are in keeping with much of the data reported so far 7,9,10,12,17,27,28,32 and could well represent a valid starting point for new prospective confirmatory studies.…”

“…46 The strong prognostic value of factors related to N status has been demonstrated by several multivariate analyses. 6,7,10,14,16,17,19,21,22,26,28,32,33 Although extracapsular spread has been unanimously found to have a significant impact on survival, 10,17,28,32,33 conflicting results have been obtained on the number of metastatic lymph nodes. 6,16,17,32,33 Similarly, the level of lymph node involvement was a significant predictor of outcome in a study by Moe et al, 33 whereas no prognostic relevance was reported for this variable by Kowalski et al 32 All the factors related to N status significantly affected survival in our univariate analysis, whereas only extracapsular spread was associated with a higher RR of death in the final multivariate model.…”

“…Degree of differentiation is the most extensively investigated histopathologic parameter, 6,7,10,12,14,17,19,22,26,27,29,30,32 but only a few studies have been able to document its impact on prognosis. 6,26,29 Even in our study, degree of differentiation was significantly correlated with survival in the univarate (Table 3) but was not related with an increased RR of death in the multivariate analysis.…”

“…However, only a few authors have documented a significant relationship between parameters expressing the intra-and extralaryngeal extent of the primary lesion and survival. 7,13,24,32 In our study, 22 parameters related to the extent of the tumor were tested by univariate analysis ( Table 3). Some of them were significantly associated with survival (marginal versus vestibular tumor site, extralaryngeal invasion, aryepiglottic fold involvement, arytenoid involvement, ventricle involvement, thyroid cartilage invasion, extralaryngeal soft tissue invasion, oropharynx invasion), and others showed a trend (p < .2) toward such an association (false vocal cord involvement, true vocal cord involvement, preepiglottic space involvement, hypopharynx invasion, fixed vocal cord, involvement of the medial wall of the pyriform sinus).…”

Prognostic factors in head and neck cancer: Histologic grading, dna ploidy, and nodal status

“…T h e prognosis for patients with squamous cell carcinoma of the head and neck (HNSCC) is determined by patient factors (ie, age, sex, and status of performance), tumor factors (ie, site of primary tumor, tumor size, and possible metastatic spread), and treatment factors (ie, the response to surgery, radiotherapy, and chemotherapy). [1][2][3][4] Until now, the TNM classification system has been the most important determinant in estimating prognosis and therefore in deciding type of treatment, the rule being that large tumors with evidence of metastatic spread have worse prognosis as compared with small and locally restricted tumors. ',' There is, however, a wellknown difficulty to predict individual prognosis for cases of HNSCC, as patients and tumors in similar condition, comparably staged and treated, have extremely disparate clinical outc o m e~.~ Descriptive histopathologic morphology and biologic behavior has been correlated with prognosis in a histologic malignancy grading system introduced by Jakobsson et al 5 This system is based on 4 different morphologic characteristics for the tumor cell population and 4 characteristics for the tumor-host relationship.…”