Preoperative systemic chemotherapy alters the histopathological growth patterns of colorectal liver metastases

Abstract

Histopathological growth patterns (HGPs) are a reliable, reproducible, and strong prognostic biomarker that can be assessed on haematoxylin and eosin-stained sections of resected colorectal liver metastases (CRLM). Assessment estimates the relative fraction of the tumour-liver interface for each of the three growth patterns; the desmoplastic HGP reflects good prognosis. Whether preoperative chemotherapy affects the HGP is currently unclear. The present international multicentre study evaluates this in an original cohort of 877 consecutive patients treated in the Netherlands, an external validation cohort of 1,203 consecutive patients treated in the USA, and a post hoc analysis from the phase III randomised controlled European Organization for Research and Treatment of Cancer (EORTC) 40983 trial (n = 70). All patients underwent resection of CRLM with or without preoperative systemic chemotherapy. Trial patients were randomised between perioperative chemotherapy and resection or resection alone. HGPs were determined according to consensus guidelines and compared for preoperative treatment status. Data from three separate tumour regression grading systems were available for the trial cohort. These were correlated with HGP stratified for treatment arm. In the original cohort, the average presence of desmoplastic HGP was 43% for chemo-naïve versus 67% for preoperatively treated patients (p < 0.001). A significant association between chemotherapy and desmoplastic HGP was found on multivariable analysis (β [95% confidence interval, CI]: 24.57 [18.28-30.87], p < 0.001). In the validation cohort, the average presence of desmoplastic HGP was 40% for chemo-naïve versus 63% for preoperatively treated patients (p < 0.001). This association remained on multivariable analysis (β [95% CI]: 24.18 [18.70-29.66], p < 0.001). In the EORTC 40983 trial, the average desmoplastic HGP presence was 33% in the resection arm versus 61% in the chemotherapy arm (p = 0.005). Chemotherapy was independently associated with an increase in desmoplastic HGP (β [95% CI]: 23.29 [1.78-44.79], p = 0.022). All three tumour regression gradings were significantly associated with the desmoplastic HGP in the chemotherapy arm (all p < 0.04). None were associated in the resection arm (all p > 0.11). Preoperative chemotherapy induces histopathological changes that alter the HGP of CRLM.

Keywords: colorectal cancer; colorectal liver metastases; histopathological growth patterns; systemic chemotherapy.

Figure 1

Examples of the distinct HGPs. (A) Example of replacement type HGP in which tumour cells ‘replace’ hepatocytes and infiltrate the liver parenchyma with direct tumour–liver cell contact. (B) Example of pushing type HGP in which the liver parenchyma is ‘pushed’ aside but is not infiltrated. No direct tumour–liver cell contact is present. (C) Example of desmoplastic type HGP, in which the tumour is separated from the liver parenchyma by a desmoplastic capsule. No direct tumour–liver cell contact is present.

Figure 2

Distribution of HGPs in the original cohort of the Erasmus MC Cancer Institute stratified for preoperative treatment status. (A) Distribution of HGPs in the chemo‐naïve cohort. (B) Distribution of HGPs in the preoperatively treated cohort. (C–E) Average observed proportion of replacement type HGP (C), desmoplastic type HGP (D), and pushing type HGP (E) in chemo‐naïve patients compared to preoperatively treated patients.

Figure 3

Distribution of HGPs in the external validation cohort of the MSKCC stratified for preoperative treatment status. (A) Distribution of HGPs in the chemo‐naïve cohort. (B) Distribution of HGPs in the preoperatively treated cohort. (C–E) Average observed proportion of replacement type HGP (C), desmoplastic type HGP (D), and pushing type HGP (E) in chemo‐naïve patients compared to preoperatively treated patients.

Figure 4

Distribution of HGPs in the EORTC 40983 trial stratified for preoperative treatment status. (A) Distribution of HGPs in the resection only arm. (B) Distribution of HGPs in the preoperatively treated arm. (C, D) Average observed proportion of replacement type HGP (C) and desmoplastic type HGP (D) in chemo‐naïve patients compared to preoperatively treated patients.

Figure 5

(A–F) Results of the multivariable linear regression models investigating three separate gradings of tumour regression in patients randomised to either resection only (A–C) or perioperative chemotherapy with resection (D–F) within the EORTC 40983 phase III trial. The dots resemble individual patients; dark grey dots represent a non‐desmoplastic and light grey dots a desmoplastic phenotype, respectively. The regression line represents the association for one of the three tumour regression gradings with the dHGP on multivariable analysis, with the ribbon representing the 95% CI of the estimate. dHGP, desmoplastic HGP; NHR, no histological response; PHR, partial histological response; MjHR, major histological response.

Figure 6

(A–J) Representative examples of resected CRLM exhibiting a desmoplastic HGP in chemo‐naïve (A–E) and pre‐treated (D–J) patients.