Skip to main content
. 2018 Jul 27;5(1):e000217. doi: 10.1136/bmjgast-2018-000217

Figure 3.

Figure 3

Classification systems and reported frequencies of colorectal cancer liver metastases (CRLM) growth patterns. (A) Overall reported frequencies of CRLM growth patterns. Heterogeneous expressions used to describe growth patterns between studies were grouped by similarity with respect to their individual terms and associated descriptions into ‘consensus terms’ (related to figure 2). (B) Number of growth patterns considered in each study. Note that some studies regarded the presence/absence of a fibrotic capsule as a parameter independent of the histological growth patterns—that is, ‘desmoplastic’ metastases could be simultaneously classified as ‘pushing’ or ‘replacement’. ‘Mixed’ patterns are not included, as they do not define a specific category. (C) Percentage frequencies of reported growth patterns by study, according to consensus terms. For studies (in italics) in which a capsule was assessed independently, its frequency is represented in a separate bar. Note: In some studies, single metastases could not be classified by the reporting authors, in which case the bars do not add up to 100%. Although Terayama et al 42 considered a total of four growth patterns the description of ‘sinusoidal’ was considered to closely represent the ‘replacement’ pattern in this specific study and counted as such. GP, growth pattern.