A novel web-based dynamic nomogram for recurrent laryngeal nerve lymph node metastasis in esophageal squamous cell carcinoma

Ting-Ting Chen; Hao-Ji Yan; Xi He; Si-Yi Fu; Sheng-Xuan Zhang; Wan Yang; Yu-Jie Zuo; Hong-Tao Tang; Jun-Jie Yang; Pei-Zhi Liu; Hong-Ying Wen; Dong Tian

doi:10.3389/fsurg.2022.898705

A novel web-based dynamic nomogram for recurrent laryngeal nerve lymph node metastasis in esophageal squamous cell carcinoma

Front Surg. 2022 Aug 23:9:898705. doi: 10.3389/fsurg.2022.898705. eCollection 2022.

Authors

Ting-Ting Chen^{1

2

3}, Hao-Ji Yan³, Xi He⁴, Si-Yi Fu⁵, Sheng-Xuan Zhang⁵, Wan Yang⁵, Yu-Jie Zuo⁵, Hong-Tao Tang⁵, Jun-Jie Yang⁵, Pei-Zhi Liu⁵, Hong-Ying Wen², Dong Tian^{1

2

6}

Affiliations

¹ Department of Thoracic Surgery, West China Hospital, Sichuan University, Chengdu, China.
² Department of Cardiothoracic Intensive Care Unit, Affiliated Hospital of North Sichuan Medical College, Nanchong, China.
³ College of Medical Imaging, North Sichuan Medical College, Nanchong, China.
⁴ Department of Radiological Sciences, Graduate School of Biomedical Sciences, Nagasaki University, Nagasaki, Japan.
⁵ College of Clinical Medicine, North Sichuan Medical College, Nanchong, China.
⁶ Academician (Expert) Workstation, Affiliated Hospital of North Sichuan Medical College, Nanchong, China.

Abstract

Background: Patients with esophageal squamous cell carcinoma (ESCC) are liable to develop recurrent laryngeal nerve (RLN) lymph node metastasis (LNM). We aimed to assess the predictive value of the long diameter (LD) and short diameter (SD) of RLN lymph node (LN) and construct a web-based dynamic nomogram for RLN LNM prediction.

Methods: We reviewed 186 ESCC patients who underwent RLN LN dissection from January 2016 to December 2018 in the Affiliated Hospital of North Sichuan Medical College. Risk factors for left and right RLN LNM were determined by univariate and multivariate analyses. A web-based dynamic nomogram was constructed by using logistic regression. The performance was assessed by the area under the curve (AUC) and Brier score. Models were internally validated by performing five-fold cross-validation.

Results: Patients who underwent left and right RLN LN dissection were categorized as left cohort (n = 132) and right cohort (n = 159), with RLN LNM rates of 15.9% (21/132) and 21.4% (34/159), respectively. The AUCs of the LD (SD) of RLN LN were 0.663 (0.688) in the left cohort and 0.696 (0.705) in the right cohort. The multivariate analysis showed that age, the SD of RLN LN, and clinical T stage were significant risk factors for left RLN LNM (all P < 0.05), while tumor location, the SD of RLN LN, and clinical T stage were significant risk factors for right RLN LNM (all P < 0.05). The dynamic nomograms showed reliable performance after five-fold cross-validation [(left (right), mean AUC: 0.814, range: 0.614-0.891 (0.775, range: 0.084-0.126); mean Brier score: 0.103, range: 0.084-0.126 (0.145, range: 0.105-0.206)], available at https://mpthtw.shinyapps.io/leftnomo/ and https://mpthtw.shinyapps.io/rightnomo/.

Conclusion: The LD and SD of RLN LN are inadequate to predict RLN LNM accurately, but online dynamic nomograms by combined risk factors show better prediction performance and convenient clinical application.

Keywords: diameter of lymph node; dynamic nomogram; esophageal squamous cell carcinoma; lymph node metastasis; recurrent laryngeal nerve.