Rectal cancer: can T2WI histogram of the primary tumor help predict the existence of lymph node metastasis?

Lanqing Yang; Dan Liu; Xin Fang; Ziqiang Wang; Yue Xing; Ling Ma; Bing Wu

doi:10.1007/s00330-019-06328-z

Rectal cancer: can T2WI histogram of the primary tumor help predict the existence of lymph node metastasis?

Eur Radiol. 2019 Dec;29(12):6469-6476. doi: 10.1007/s00330-019-06328-z. Epub 2019 Jul 5.

Authors

Lanqing Yang¹, Dan Liu¹, Xin Fang¹, Ziqiang Wang², Yue Xing³, Ling Ma⁴, Bing Wu⁵

Affiliations

¹ Department of Radiology, West China Hospital, Sichuan University, Guoxue Xiang No. 37, Chengdu, 610041, Sichuan, People's Republic of China.
² Department of Colorectal Surgery, West China Hospital, Sichuan University, Guoxue Xiang No. 37, Chengdu, 610041, Sichuan, People's Republic of China.
³ Sir Peter Mansfield Imaging Center, School of Medicine, Nottingham, UK.
⁴ GE Healthcare, Shanghai, People's Republic of China.
⁵ Department of Radiology, West China Hospital, Sichuan University, Guoxue Xiang No. 37, Chengdu, 610041, Sichuan, People's Republic of China. bingwu69@163.com.

PMID: 31278581
DOI: 10.1007/s00330-019-06328-z

Abstract

Objectives: To explore if there is a correlation between T2WI histogram features of the primary tumor and the existence of regional lymph node (LN) metastasis in rectal cancer.

Methods: Eighty-eight patients with pathologically proven rectal adenocarcinoma, who received direct surgical resection and underwent preoperative rectal MRIs, were enrolled retrospectively. Based on pathological analysis of surgical specimen, patients were classified into negative LN (LN-) and positive LN (LN+) groups. The degree of differentiation and pathological T stage were recorded. Clinical T stage, tumor location, and maximum diameter of tumor were evaluated of each patient. Whole-tumor texture analysis was independently performed by two radiologists on axial T2WI, including skewness, kurtosis, energy, and entropy.

Results: The interobserver agreement was overall good for texture analysis between two radiologists, with intraclass correlation coefficients (ICCs) ranging from 0.626 to 0.826. The LN- group had a significantly higher skewness (p < 0.001), kurtosis (p < 0.001), and energy (p = 0.004) than the LN+ group, and a lower entropy (p = 0.028). These four parameters showed moderate to good diagnostic power in predicting LN metastasis with respective AUC of 0.750, 0.733, 0.669, and 0.648. In addition, they were both correlated with LN metastasis (r_s = - 0.413, - 0.385, - 0.28, and 0.245, respectively). The multivariate analysis showed that lower skewness was an independent risk factor of LN metastasis (odds ratio, OR = 9.832; 95%CI, 1.171-56.295; p = 0.01).

Conclusions: Signal intensity histogram parameters of primary tumor on T2WI were associated with regional LN status in rectal cancer, which may help improve the prediction of nodal stage.

Key points: • Histogram parameters of tumor on T2WI may help to reduce uncertainty when assessing LN status in rectal cancer. • Histogram parameters of tumor on T2WI showed a significant difference between different regional LN status groups in rectal cancer. • Skewness was an independent risk factor of regional LN metastasis in rectal cancer.

Keywords: Computer-assisted image analysis; Lymphatic metastasis; Magnetic resonance imaging; Rectal neoplasms.

MeSH terms

Adenocarcinoma / diagnostic imaging*
Adenocarcinoma / pathology*
Adult
Aged
Aged, 80 and over
Female
Humans
Lymph Nodes / diagnostic imaging
Lymphatic Metastasis / diagnostic imaging*
Magnetic Resonance Imaging / methods*
Male
Middle Aged
Predictive Value of Tests
Rectal Neoplasms / diagnostic imaging*
Rectal Neoplasms / pathology*
Reproducibility of Results
Retrospective Studies