Skip to main content

Table 1 Features chosen by CR pipeline. Features from intratumoral region are denoted by (I) and features from 4 mm peritumoral region are denoted by (P). There was only one clinical feature selected, and the remainder were radiomic features (5 from intratumoral region, 2 from peritumoral region)

From: Preoperative prediction of lymph node metastasis using deep learning-based features

Category

Features

Clinical (n = 1)

Lymphovascular invasion

Shape (n = 2)

Compacity (I), Extent (I)

GLCM (n = 1)

Wash Out: Entropy (log2) (I)

NGLDM (n = 2)

Wash Out: Contrast (I)

SER: Coarseness (P)

Laws (n = 2)

SER: Energy_8 (I)

Skewness_9 (P)

  1. GLCM grey-level co-occurrence matrix, NGLDM neighborhood grey-level different matrix