362PD - Development and validation of a nomogram to predict the probability of recurrence in patients with major salivary gland cancers

Date 16 December 2016
Event ESMO Asia 2016 Congress
Session Head and neck cancer
Topics Head and Neck Cancers
Presenter Chia-Yen Hung
Citation Annals of Oncology (2016) 27 (suppl_9): ix112-ix122. 10.1093/annonc/mdw587
Authors C. Hung1, C. Lu2, C. Liu3, P. Chang4, K. Yeh4, S. Li3, Y. Lin5, T. Yeh6, Y. Hung5, W. Chou5
  • 1Department Of Hematology-oncology, Mackay Memorial Hospital, 251 - New Taipei City/TW
  • 2Department Of Medical Oncology, Chang Gung Memorial Hospital, Chiayi, 613 - Chiayi/TW
  • 3Department Of Medical Oncology, Chang Gung Memorial Hospital-Kaohsiung, 833 - Kaohsiung/TW
  • 4Department Of Medical Oncology, Chang Gung Memorial Hospital, Keelung, 204 - Keelung/TW
  • 5Department Of Medical Oncology, Chang Gung Memorial Hospital-Linkou, 333 - Taoyuan/TW
  • 6Department Of Surgery, Chang Gung Memorial Hospital-Linkou, 333 - Taoyuan/TW

Abstract

Background

Assessing the post-surgical risk of relapse in patients with major salivary gland carcinomas (MSGCs) is an important but difficult task because of the broad spectrum of histological tumor subtypes and diverse clinical behaviors. We aimed to develop and validate a nomogram, which would allow us to predict the probability of recurrence in patients with MSGCs.

Methods

Two hundred and thirty-one patients who had undergone radical surgery for MSGC treatment between 2002 and 2014 from a specific medical center were used as a training set. Clinicopathological variables with the most significant values in the multivariate Cox regression analysis were used to build into a nomogram capable of estimating the 5-year recurrence probability. An independent validation set, composed of 139 patients treated during the same period in three other hospitals, were selected for external validation and calibration.

Results

The 5-year recurrence rate was 23.2% for the training set and 21.4% for the validation set. The nomogram was developed based on six significant predictive factors of tumor recurrence retained in the multivariate Cox model, including smoking history, tumor grade, perineural invasion, lymphatic invasion, and pathological T- and N-classifications. The nomogram had a highly predictive performance, with a bootstrap-corrected concordance index of 0.82 for the training set and 0.78 for the validation set. The nomogram showed good calibration in predicting 2-year and 5-year recurrence probability in both the training and validation set.

Conclusions

We developed and externally validated an accurate nomogram to predict the tumor relapse probability in patients who had been surgically treated for MSGCs. This nomogram may be used to assist clinicians and patients in specifying the potential risk of relapse and thereby providing more information for the selection of appropriate adjuvant treatments.

Clinical trial indentification

Legal entity responsible for the study

N/A

Funding

N/A

Disclosure

All authors have declared no conflicts of interest.