HER2 overexpression/amplification status in colorectal cancer: a comparison between immunohistochemistry and fluorescence in situ hybridization using five different immunohistochemical scoring criteria

Abstract

Objective: Although HER2 has gradually become an important therapeutic target for colorectal cancer (CRC), a unified and standard HER2 scoring system was still not established in CRC, and the debatable results of immunohistochemistry and fluorescence in situ hybridization (FISH) in CRC requires further exploration.

Methods: In this study, we use five immunohistochemical (IHC) scoring criteria (i.e., IRS-p, IRS-m, GEA-s, GEA-b and HERACLES) and two FISH criteria to evaluate HER2 status, and further evaluate the correlation between HER2 status and clinicopathological features, survival in a large, unselected Chinese cohort of 664 CRCs.

Results: Finally, we set HER2/CEP17 ratio ≥ 2.0, or an average HER2 copy number ≥ 6.0 as FISH-positive threshold and the amplification rate of HER2 gene was 7.08% (47/664).The HER2 positivity (IHC 3+) was 2.71%, 3.16%, 2.56%, 2.71% and 3.16%, according to the IHC scoring criteria of IRS-p, IRS-m, GEA-s, GEA-b and HERACLES, respectively. Set FISH results as the golden standard; receiver-operating characteristic analysis showed that IRS-p had both high sensitivity and specificity than other IHC scoring systems to evaluate HER2 status. Based on IRS-p criterion, There were significant differences in tumor differentiation (p = 0.038), lymphatic vascular invasion (p = 0.001), pN stage (p value = 0.043), and overall survival (p < 0.001) among IHC score 0-3 + groups. Meanwhile, there were significant differences in pT stage (p = 0.031), pN stage (p = 0.009) and overall survival (p < 0.001) among FISH subgroups.

Conclusion: The IRS-p criterion was more suitable for assessing the HER2 status in CRC patients than other IHC criteria. Whereas for FISH scoring system, only HER2/CEP17 < 2.0, meanwhile HER2cn < 4.0 and HER2cn ≥ 6.0 were subgroups with unique clinicopathological characteristics.

Keywords: Colorectal carcinomas; Fluorescence in situ hybridization; HER2; Immunohistochemistry.

Fig. 1

HER2 evaluation using five immunohistochemistry (IHC) assessment methods in TMA of colorectal cancer and their FISH test results. Representative immunostaining pattern (SP), intensity (SI) of tumor cells, and their percentage of positivity (PPT): A shows a case with a strong IHC cell membrane circumferential staining and its local magnification (B), FISH test (C) and corresponding scoring results; D shows a case with moderate lateral and basolateral staining and its corresponding local magnification (E), FISH test (F) and scoring results; G shows a case with faint/weak lateral or basolateral staining and its corresponding local magnification (H), FISH test (I) and scoring results; and J shows a negative staining case and its corresponding local magnification (HK), FISH test (L) and scoring results

Fig. 2

HER2 evaluation by fluorescence in situ hybridization (FISH) in TMA of colorectal cancer. Representative FISH pattern of tumor cells HER2/CEP17 < 2.0 and HER2cn < 4.0 (A), Her2/CEP17 < 2.0 and 4.0 ≤ HER2cn < 6.0 (B), HER2/CEP17 ≥ 2.0 meanwhile HER2cn < 4.0 (C), HER2/CEP17 ≥ 2.0 meanwhile 4.0 ≤ HER2cn < 6.0 (D), HER2/CEP17 ≥ 2.0 meanwhile HER2cn ≥ 6.0 (E) and Her2/CEP17 < 2.0 meanwhile HER2cn ≥ 6.0 (F). A, B were classified into FISH group 1 and group 2, respectively. C, D were classified into FISH group 3, and E, F were classified into FISH group 4

Fig. 3

Kaplan–Meier curves of overall survival among FISH subgroups (A), HER2/CEP17 < 2.0 and HER2cn < 4.0 vs. HER2cn ≥ 6.0 (B), HER2/CEP17 < 2.0 and HER2cn < 4.0 vs. HER2/CEP17 ≥ 2.0 and HER2cn < 6.0 (C), HER2/CEP17 < 2.0 and HER2cn < 4.0 vs. Her2/CEP17 < 2.0 and 4.0 ≤ HER2cn < 6.0 (D), HER2/CEP17 ≥ 2.0 and HER2cn < 6.0 vs. HER2cn ≥ 6.0 (E), and Her2/CEP17 < 2.0 and 4.0 ≤ HER2cn < 6.0 vs. HER2cn ≥ 6.0 (F). The log-rank test was used to calculate the P value

Fig. 4

Receiver operator characteristic curve plotting test sensitivity in relation to specificity between immunohistochemical scoring criteria of IRS-p, IRS-m, GEA-s, GEA-b and HERACLES

Fig. 5

Kaplan–Meier curves of overall survival among IRS-p score 0–3 + (A), 0 vs. 3 + (B), 0 vs. 2 + (C), 0 vs. 1 + (D), 2 + vs. 3 + (E), and 1 + vs. 3 + (F). The log-rank test was used to calculate the P value