Develop and validation a nomogram to predict the recurrent probability in patients with major salivary gland cancer
OBJECTIVES/HYPOTHESIS To find the risk factors of lymph node (LN) metastasis of salivary gland cancer and draw a scheme for LN management. STUDY DESIGN Hospital-based retrospective study. METHODS The records of salivary gland cancer patients treated at the Department of Head and Neck Surgery, Cancer Hospital, Fudan University, were entered in a database, and 219 consecutive patients with carcinomas of major salivary glands primarily operated on between January 1998 and January 2011 were chosen for univariate and multivariate analysis to identify risk factors for LN involvement. RESULTS Fifty-eight (26.5%) patients had LN involvement. Factors associated with cervical LN involvement on univariate analysis included pathologic type, male sex, shorter duration of preoperative course, facial paralysis, advanced T stage, and major nerve, soft tissue, lymphatic/vascular (L/V), neural/perineural, and extracapsular invasion. Multivariate analysis identified major nerve invasion, histologic type, L/V invasion, and extracapsular invasion as significant factors for LN involvement. The proportion of patients with LN involvement with low (105), middle (61), high (34), and super high (19) predictive index scores based on the four risk factors were 3.8%, 27.9%, 55.9%, and 94.7%, respectively. CONCLUSIONS A predictive index using the clinicopathologic factors described in this report can effectively stratify patients into risk groups for nodal metastasis. Comprehensive management based on this risk index should improve treatment outcomes for patients with salivary gland cancer.