Letter to the editor “Tumor volume as a prognostic factor of locally advanced laryngeal cancer”

Abstract

With great interest, I read a recent study about identifying the role of tumor volumes (TVs) in locally advanced laryngeal cancer (LC) [1]. de Andrade NMM and colleagues analyzed the data about TVs of 145 patients with T3 and T4a LC who underwent surgical treatment from 2008 to 2017. They concluded that the TV was considered as a prognostic factor of locally advanced LC. However, this conclusion remains uncertain due to insufficient statistical power and literature support. In the authors’ Table 2 and Table 3, the multivariate analysis obviously showed that the TV did not have significant differences on overall survival (p = 0.094) or diseasefree survival (p = 0.569). In other words, after controlling confounding factors with the multivariate analysis, the TV was proved to be not a prognostic factor. Surprisingly, the authors chose to accept this fetal problem and firmly suggested the TV as a prognostic factor of locally advanced LC. Actually, mimicking randomized controlled trials, propensity score matching (PSM) can elegantly redress imbalanced distributions of the confounding factors and reduce selection bias [2, 3]. Then, before conducting survival analyses about the TV, paired groups (“TV > 14.2 cm3” and “TV > 14.2 cm3” groups) can be created using PSM to ensure that the groups are comparable, which can strengthen the credibility of their research. As the authors stated, “The variables that presented p < 0.1 in the univariate analysis were submitted to multivariate analysis”, which attached my attention. Why chose p < 0.1? Moreover, encouragingly, the authors did a subgroup analysis of patients with T4-stage LC. But the authors only did the univariate analysis and the survival analysis that was related to overall survival. The multivariate analysis was not performed. The local recurrence and disease-free survival were not analyzed. Furthermore, a subgroup analysis of patients with T3-stage LC may also better be performed. A retrospective cohort study of Timmermans AJ et al. proved that the TV was not significantly correlated with local control, locoregional control, or overall survival in the surgically treated group [4]. Moreover, in a study of Eskiizmir G et al. log-rank tests and bivariate Cox regression analyses both indicated that the TV did not have significant differences on 5 year overall survival, 5 year disease-free survival, 5 year locoregional recurrence-free survival, and 5 year distant metastasis-free survival [5]. Thus, these previous studies strongly concluded that the TV may be not a prognostic factor of locally advanced LC. In conclusion, the current evidence of de Andrade NMM’s study may be insufficient. The role of the TV in locally advanced LC needs to be further explored. I am looking forward to their positive response.