The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients

Abstract

Objectives:

Over the past decade, some publications have reported that Immunoscore was associated with the prognosis of several cancers. To better understand this issue, we conducted this pooled analysis.

Methods:

We systematically searched PubMed, Embase, Web of Science, and the Cochrane Library from their inceptions to 15 May 2019 to identify relevant articles. The pooled hazard ratio (HR) and 95% confidence interval (CI) was estimated for overall survival, disease-free survival, and disease-specific survival.

Results:

A total of 26 cohort studies with 10,328 patients involving eight cancer specialties were evaluated mainly by the consensus Immunoscore. The pooled analysis indicated that a lower Immunoscore was associated with a poor overall survival (HR 2.23, 95% CI 1.58, 2.70), disease-free survival (HR 2.40, 95% CI 1.96, 2.49), and disease-specific survival (HR 2.81, 95% CI 2.10, 3.77) for all cancers. The same convincing results were found in colorectal cancer, gastric cancer, and non-small cell lung cancer (especially the consensus Immunoscore for colon cancer). In five other types of cancer the results were similar, but the sample sizes were limited.

Conclusions:

These findings support that Immunoscore is significantly associated with the prognosis of patients with cancer. It provides a reliable estimate of the risk of recurrence in patients with colon cancer. However, more high-quality studies are necessary to assess the prognostic value of Immunoscore in non-colon cancers.

Keywords

Immunoscore cancer prognosis pooled analysis

Introduction

The American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) tumor node metastasis (TNM) classification is the most common system to classify the extent of the spread of cancer. It has been a cornerstone of cancer care and research for decades and plays a critical role in cancer control,² providing prognostic information and guiding individual treatment decisions.¹ However, TNM does not meet all prognostic needs because of incomplete prognostic information.² It has put the emphasis on tumor biology, while the prognosis is influenced by many factors such as age, sex, co-morbidity, performance status, host immune response, and so on.² Patients with the same TNM stage have had various clinical outcomes; a previous study showed that advanced stage patients can remain stable for years while TNM I/II-stage patients suffered from relapse, deterioration, and even rapid death.³

Tumor cells live in a complex microenvironment, which is closely connected with tumor growth and metastasis.⁴ Recently, an increasing number of studies have found that the adaptive immune system has a critical role in tumor development, which puts an emphasis on the importance of the systemic and local immunological biomarkers especially the tumor-infiltrating lymphocytes (TILs) in the prognosis of patients with cancer.^3,5,6 In 1921, the immune cell infiltration of cancers was suspected to be a positive factor for patients with gastric cancer.⁷ Until 2018, there were more than 250 articles that investigated the association between the immune cell infiltrates of tumors and prognosis in 28 different cancers. Approximately 98% of studies found that the CD8+ T cell infiltration was associated with a good prognosis, and most studies were conducted on colorectal cancer.⁸ In order to investigate the influences of immune parameters on survival, Pagès et al. established the Immunoscore as a clinically prognostic marker, which based on the numeration of two lymphocyte populations (CD3+/CD45RO+, CD3+/CD8+, or CD8+/CD45RO+) both in the center of tumor (CT) and the invasive margin (IM).^9–11 Later, some studies found that the high Immunoscore was related to a better prognosis and it was validated by the international Immunoscore consortium, which involved 14 centers of 13 countries with 3539 TNM stage I–III colon cancer patients, in which the consensus Immunoscore was determined by the density of CD3+ and CD8+ T-cell in CT and IM. The conclusion was that the consensus Immunoscore could be used as a new component of a TNM-immune classification of colon cancer.^8,12 Also, the Immunoscore can accurately predict the recurrence of colon cancer after 5 years of treatment,¹³ and can provide a more accurate prognosis compared with other parameters.¹⁴

Over the past decade, some publications have reported that Immunoscore is associated with the prognosis of several cancers.⁸ However, few papers have evaluated the impact of Immunoscore on the prognosis of cancer patients. Although the paper by Sun et al.¹⁵ evaluated the prognostic value of Immunoscore in patients with colorectal cancer, the number of studies was limited and several of the latest studies were not included. Therefore, we systematically searched the existing publications and conducted a systematic review and meta-analysis to investigate the prognostic value of Immunoscore in all type cancers.

Methods

Literature search

We systematically searched PubMed, Embase, Web of Science, and the Cochrane Library from the inception until 15 May 2019 to identify the articles that examined the prognostic value of Immunoscore in cancer patients. We used the following search strategy: (“cancer” OR “neoplasm” OR “carcinoma” OR “tumor” OR “adenocarcinoma” AND “Immunoscore” OR “immune score”). We also reviewed the references of related studies. We performed the meta-analysis under the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRSIMA).¹⁶

Study selection

The inclusion criteria for the articles were those that: (a) illustrated the evaluation of Immunoscore; (b) showed overall survival (OS), disease-free survival (DFS) or disease-specific survival (DSS) in the endpoints; and (c) reported the hazard ratio (HR) estimates and related 95% confidence interval (CI) or survival curves. We excluded the reviews, commentaries, studies published only in abstract form, and studies that did not have sufficient primary data to calculate the HR.

Date extraction

The name of first author, year of publication, country, enrollment time, indicator, and geographic region of sampling, the number of participants, type and stage of cancer, duration of follow-up, outcomes, and their HR with 95% CI were extracted from each included study. If the study did not report the HR while the survival curves were available, we would calculate the HR by digitizing the curves using the open-source Engauge Digitizer software reported by Tierney et al.¹⁷

Assessment of study quality

The Quality and Prognosis Studies (QUIPS) was used to assess the quality of included study, which was a valid and helpful tool for a systematic reviewer to critically assess the quality of study.¹⁸ The risk of bias was graded into three levels (including low, moderate and high) based on study participation, study attrition, prognostic factor measurement, outcome measurement, study confounding, and statistical analysis and reporting.¹⁸

Data synthesis and analysis

In this meta-analysis, we calculated the pooled HR and 95% CI to evaluate the prognostic value of the Immunoscore for cancer patients. The I² and Cochran’s Q tests were used to evaluate the heterogeneity of each outcome. We also conducted the subgroup analyses by stratifying on the cancer site, clinical stage, follow-up duration, number of participants, indicators, and geographic region. Sensitivity analyses were conducted and publication bias was accessed. Review Manager 5.3 software (The Nordic Cochrane Centre, Copenhagen, Denmark) was used to conduct the systematic review and meta-analysis.

Results

Study selection and characteristics

In the literature search, we looked at 270 articles from the databases. Finally, 26 articles,^8,19–43 published between 2011 and 2019, were included for quantitative synthesis and meta-analysis. The processes of selection are shown in Figure 1.

Figure 1.

Flowchart of the selection process.

Of these 26 cohort studies, a total of 10,328 (range 32–3539) cases were included from Europe, Asia, America, and Oceanic countries, which involved colorectal cancer (n = 16),^{8,19,22,24,25,27–31,35–38,40,41} gastric cancer (n = 2),^23,39 non-small cell lung cancer (NSCLC) (n = 2),^32,33 pancreatic cancer (n = 2),^26,34 and one each in bladder cancer,⁴² head and neck squamous cell carcinoma (HNSCC),⁴³ hepatocellular carcinoma (HCC),²¹ and ovarian cancer.²⁰ The ranges of follow-up time were from 0 to 267 months. In the Mlecnik et al.,²⁸ Nearchou et al.,³⁰ and Jiang et al. studies,³⁹ where more than one cohort was reported in each study, we recorded them one by one when analysis. The main characteristics of the 26 studies are shown in Table 1.

Table 1.

The characteristics of included studies.

First author	Year	Country	Cancer	Indicator	Region	Survival analysis	No. of patients	Year of recruitment	Stage	Age	Follow-up period (month)
Pagès et al.⁸	2018	13 countries	CC¹	CD3+/CD8+	CT/IM	OS/DFS	3539	2013–2015	I–III	69 (60–77)	111 (105–116)
Wirta et al.³⁸	2017	Finland	CC¹	CD3+/CD8+	CT/IM	OS/DFS/DSS	510	2000–2010	I–IV	73 (64–79)	72 (2.04–108)
Mlecnik et al.²⁹	2018	France	CRC	CD3+/CD8+	CT/IM	OS/DFS	441	2004–2010	IV	NR	NR
Park et al.²²	2017	UK	CRC	CD3+/CD8+	CT/IM	OS/DSS	331	1997–2008	I–III	70 (55–85)	134 (108–170)
Ko & Pyo⁴¹	2019	Korea	CRC	CD3+/CD8+	CT/IM	OS	265	2001–2010	I–IV	NR	0 to 60
Wang et al.³⁶	2018	China	CRC	CD3+/CD8+	CT/IM	OS/DFS	249	2002–2015	IV	NR	NR
Park et al.³¹	2016	UK	CRC	CD3+/CD8+	CT/IM	DSS	246	1997–2008	I–III	NR	150 (87–206)
Van den Eynde et al.³⁵	2018	Belgium	CRC	CD3+/CD8+	CT/IM	DFS	222	NR	NR	NR	NR
Kwak et al.²⁴	2016	Korea	CRC	CD3+/CD8+	CT/IM	OS	196	2003–2009	NR	NR	37.3 (0.8–104.6)
Yomoda et al.⁴⁰	2019	Japan	CRC	CD3+/CD8+	CT/IM	OS/DFS	132	2009–2010	II–III	NR	NR
Anitei et al.¹⁹	2014	France	RC	CD3+/CD8+	CT/IM	OS /DFS	111	1987–2004	I–IV	NA	74 (0–244)
Nearchou et al.³⁰	2019	Japan	CRC	CD3+/CD8+	CT/IM	DSS	170	2002–2004	II	NR	138
Nearchou et al.³⁰	2019	Japan	CRC	CD3+/CD8+	CT/IM	DSS	62	2006–2011	II	NR	103.2
Nearchou et al.³⁰	2019	Japan	CRC	CD3+/CD8+	CT/IM	DSS	56	2002–2003	II	NR	24
Liu et al.²⁵	2018	China	CRC	CD3+/CD8+	CT/IM	OS	60	2013–2016	IV	59.64 ±10.68	0 to 39
Ward-Hartstonge et al.³⁷	2017	New Zealand	CRC	CD3+/CD8+	CT/IM	DFS	32	1995–2006	II	72	NR
Mlecnik et al.²⁸	2011	France	CRC	CD8+/CD45RO+	CT/IM	OS/DFS/DSS	599	1990–2004	I–IV	NR	NR
Mlecnik et al.²⁷	2016	France	CRC	CD8+/CD45RO+	CT/IM	OS/DFS/DSS	270	NR	I–III	NR	NR
Jiang et al.³⁹	2019	China	GC	CD3+/CD8+/CD45RO+/CD66b	CT/IM	OS/DFS	879	2005–2009	I–IV	NR	NR
Kim et al.²³	2017	Korea	GC	CD3+/CD8+	CT/IM/EC/SC	OS	153	2004–2009	NR	NR	NR
Paulsen et al.³²	2015	Norway	NSCLC	CD8+/CD45RO+	CT/IM	DSS	536	1990–2010	I–IIIA	67 (28–85)	86 (34–267)
Paulsen et al.³³	2017	Norway	NSCLC	PD-1/PD-L1	EC/SC	DSS	633	1990–2010	I–IIIA	67 (28–85)	86 (34–267)
Yu et al.⁴²	2018	Canada	BC	CD3+/CD8+	CT/IM	OS/DFS	67	2011–2013	pT1–T4	68.9 (17.5)*	21.9 (15–32.5)
Yao et al.²¹	2017	China	HCC	CD3+/CD8+/CD45RO+	CT/IM	OS/DFS	92	2006–2010	I–IV	46.7 (22–77)	NR
Zhang et al.⁴³	2018	China	HNSCC	CD3+/CD8+	CT/IM	OS/DFS	88	2009–2015	I–III	NR	36
Bosmuller et al.²⁰	2016	Germany	OC	CD3+/CD103+	IC/SC	OS	138	2000–2008	II–IV	35–85	1 to 120
Tahkola et al.³⁴	2018	Finland	PC	CD3+/CD8+	CT/IM	OS/DSS	108	2000–2016	IA–IIB	66.9 (8.2)^#	44 (15.8–57.3)
Miksch et al.¹⁹	2019	Germany	PC	CD3+/CD8+	CC²/SC	OS/DFS	57	NR	NR	70.4	19

BC: bladder cancer; CC¹: colon cancer; CC²: cellular components; CRC: colorectal cancer; CT: core/center of tumor; DFS: disease-free survival; DSS: disease-specific survival; EC: epithelial compartments; GC: gastric cancer; HCC: hepatocellular carcinoma; HNSCC: head and neck squamous cell carcinoma; IC: intraepithelial compartment; IM: invasive margin; NR: not reported; NSCLC: non-small cell lung cancer; OC: ovarian cancer; OS: overall survival; PC: pancreatic cancer; PD-1: programmed death-1 receptor; PD-L1: programmed cell death-ligand 1; RC: rectal cancer; SC: stromal components.

*interquartile range

standard deviation

Survival outcomes in cancer patients

There were 20 studies with 8285 patients that reported the impact of Immunoscore on OS. The pooled HR was calculated with a random effects model because of the present heterogeneity among included studies. The statistical results showed that the lower Immunoscore was associated with a worse OS (HR 2.23, 95% CI 1.58, 2.70; I² = 90%, P_{heterogeneity} < 0.01). Fifteen studies with 7288 patients provided the HRs and 95% CIs for DFS. The pooled results showed that a lower Immunoscore was associated with a worse DFS (HR 2.40, 95% CI 1.96, 2.49; I² = 78%, P_{heterogeneity} < 0.01). Nine studies with 3521 patients reported the outcomes of DSS. A lower Immunoscore was related to a poor DSS (HR 2.81, 95% CI 2.10, 3.77; I² = 74%, P_{heterogeneity} < 0.01) (Figure 2 and Table 2).

Figure 2.

Forest plots of prognostic value of Immunoscore on overall survival (a), disease-free survival (b), and disease-specific survival (c).

Table 2.

Results based on cancer type

Cancer	No.of study	OS				DFS				DSS
Cancer	No.of study	No. of patients	HR (95%CI)	I²,%	P	No. of patients	HR (95%CI)	I²,%	P	No. of patients	HR (95%CI)	I²,%	P
All	26	8285	2.23 (1.58, 2.70)	90	<0.001	7288	2.40 (1.96, 2.93)	78	<0.001	3521	2.81 (2.10, 3.77)	74	<0.001
CRC	16	6703	1.91 (1.47, 2.49)	95	<0.001	6105	2.25 (1.76, 2.87)	81	<0.001	2244	3.07 (2.13, 4.43)	74	<0.001
GC	2	1032	2.78 (2.31, 3.35)	0	0.82	879	2.81 (2.33, 3.38)	0	0.63	–	–	–	–
NSCLC	2	–	–	–	–	–	–	–	–	1169	1.81 (1.12, 2.92)	63	0.1
PC	2	165	3.38 (2.02, 5.66)	11	0.29	57	1.76 (0.87, 3.57)	None	None	108	4.43 (2.21, 2.91)	None	None
HCC	1	92	5.26 (1.16, 25)	None	None	92	0.93 (0.20, 4.35)	None	None	–	–	–	–
HNSCC	1	88	5.52 (2.25, 13.524)	None	None	88	3.92 (1.85, 8.33)	None	None	–	–	–	–
BC	1	67	1.96 (0.34, 11.11)	None	None	67	7.69 (1.32, 50.00)	None	None	–	–	–	–
OC	1	138	2.44 (1.20, 5.00)	None	None	–	–	–	–	–	–	–	–

BC: bladder cancer; CRC: colorectal cancer; OS: overall survival; DFS: disease-free survival; DSS: disease-specific survival; GC: gastric cancer; HCC: hepatocellular carcinoma; HNSCC: head and neck squamous cell carcinoma; HR: hazard ratio; NSCLC: non-small cell lung cancer; OC: ovarian cancer; PC: pancreatic cancer.

We also conducted the subgroup analysis based on the consensus Immunoscore in which the scoring system is based on the density of CD3+ and CD8+ T-cell effectors in the CT and its IM. The results showed that lower Immunoscore was associated with poor OS (HR 2.12, 95% CI 1.60, 2.81; I² = 93%, P_{heterogeneity} < 0.01), DFS (HR 2.22, 95% CI 1.70, 2. 09; I² = 71%, P_{heterogeneity} < 0.01) and DSS (HR 3.14, 95% CI 2.46, 4.02; I² = 23%, P_{heterogeneity} = 0.25) for all included cancer (Table 3).

Table 3.

Results based on the consensus Immunoscore (CD3+ and CD8+ T-cell)

Cancer type	OS				DFS				DSS
Cancer type	No. of patients	HR (95% CI)	I²,%	P	No. of patients	HR (95% CI)	I²,%	P	No. of patients	HR (95% CI)	I²,%	P
All	5926	2.12 (1.60, 2.81)	93	<0.001	5391	2.22 (1.70, 2.90)	71	<0.001	1483	3.14 (2.46, 4.02)	23	0.25
CRC	5663	1.90 (1.40, 2.58)	94	<0.001	5236	2.04 (1.57, 2.63)	70	<0.001	1375	2.76 (2.11, 4.07)	0	0.47
PC	108	4.22 (2.1, 8.17)	None	None	–	–	–	–	108	4.43 (2.21, 2.91)	None	None
HNSCC	88	5.52 (2.25, 13.524)	None	None	88	3.92 (1.85, 8.33)	None	None	–	–	–	–
BC	67	1.96 (0.34, 11.11)	None	None	67	7.69 (1.32, 50.00)	None	None	–	–	–	–

BC: bladder cancer; CRC: colorectal cancer; DFS: disease-free survival; DSS: disease-specific survival; HR: hazard ratio; HNSCC: head and neck squamous cell carcinoma; OS: overall survival; PC: pancreatic cancer.

Survival outcomes in patients with colorectal cancer

Of 26 studies, 16 investigated the prognostic value of Immunoscore on patients with colorectal cancer (6703 patients). The statistical results showed that the lower Immunoscore were significantly associated with poor OS (HR 1.91, 95% CI 1.47, 2.49; I² = 95%, P_{heterogeneity} < 0.01), DFS (HR 2.25, 95% CI 1.76, 2.87; I² = 81%, P_{heterogeneity} < 0.01), and DSS (HR 3.07, 95% CI 2.13, 4.43; I² = 74%, P_{heterogeneity} < 0.01) for patients with colorectal cancer.

More than 60% of the patients had colon cancer. Two studies with 4049 patients provided the HR and 95% CI in OS and DFS.^7,37 A lower consensus Immunoscore was significantly associated with poor OS (HR 1.85, 95% CI 1.47, 2.32; I² = 0%, P_{heterogeneity} = 0.58) and DFS (HR 3.01, 95% CI 1.12, 8.55; I² = 82%, P_{heterogeneity} < 0.01). The same result was found for DSS (HR 4.48, 95% CI 2.49, 8.05) in another study. In 2016, a study included 510 patients with TNM stage I–IV colon cancer found that a higher Immunoscore was related to improved OS, DFS, and DSS. According to the multivariable analysis with the Cox proportional hazard model, the HR of the lowest Immunoscore was 2.47(95% CI 1.66, 3.67) for OS, 5.68 (95% CI 2.42, 23.31) for DFS, and 4.48 (95% CI 2.49, 8.05) for DSS compared with the highest Immunoscore.³⁸ Recently, similar results were found in 3539 patients with stage I–III colon cancer from a 14-center (13 countries) study published in The Lancet, which showed that patients with a high Immunoscore had a lower risk of recurrence at 5 years (rate of recurrence 8%), while higher risk of recurrence in patients with a low Immunoscore (rate of recurrence 32%).

We also conducted a subgroup analysis by stratifying the cancer site, the clinical stage of tumor, the follow-up duration, the number of participants, the indicators, and the geographic region. Similarly, a lower Immunoscore was associated with a bad prognosis despite the cancer site (colorectal, colon, rectal), the number of participants (>300 or < 300), the stage of tumor (I–III or IV) or the maximum follow-up time (>10 years or < 10 years). However, there was no significant difference in DFS for Asian patients (Table S1).

Survival outcomes in patients with gastric cancer

Two studies with 1032 patients identified the influence of Immunoscore for gastric cancer prognosis based on the OS and DFS. The polled HRs were calculated with a fixed-effects model because there was no heterogeneity. Lower Immunoscore was associated with worse OS (HR 2.78, 95% CI 2.31, 3.35; I² = 0%, P_{heterogeneity} = 0.82) and DFS (HR 2.81, 95% CI 2.33, 3.38; I² = 0%, P_{heterogeneity} = 0.63) (Table 2).

Two studies investigated the connection between Immunoscore and prognosis of gastric cancer patients. The Jiang et al.³⁹ study with 879 patients represented that a higher Immunoscore had a significant survival advantage in OS and the Immunoscore was calculated by a specific formula, which depended on the density of CD3, CD8, CD45RO, and CD66b from CT and IM. In another study,²³ the Immunoscore system was calculated based on the density of the CD3+ and CD8+ TILs from both epithelial and stromal compartments of the CT and the IM. However, the results were similar with Jiang et al. study.

Survival outcomes in patients with NSCLC

Two studies with 1169 patients provided the HRs and 95% CI for DSS. The polled results showed that lower Immunoscore was related to poor DSS (HR 1.81, 95% CI 1.12, 2.92; I² = 63%, P_{heterogeneity} = 0.1) (Table 2). Paulsen et al. conducted two studies that included 536 patients with TNM stage I–IIIA to access the prognostic value of Immunoscore for NSCLC. The Immunoscore was based on CD8 +/CD45RO+ TILs in one study³² and PD-1/PD-L1 in the other study.³³

Survival outcomes in other cancer patients

Regarding HCC, HNSCC, pancreatic, bladder and ovarian cancer, the lower Immunoscore was associated with a worse prognosis compared with the higher Immunoscore. The results were similar even though the sample sizes of all studies included were limited. The statistical results are shown in Table 2.

Sensitivity analyses

In order to eliminate the over-dependence of the results on an individual study, we conducted the sensitivity analyses for OS, DFS, and DSS based on the weights in all studies. There were no significant differences influenced by removing any one study that was most weighted.

Quality assessment

According to the QUIPS criteria, about 46.2% articles had a low risk of bias. More than half of the 26 articles had a moderate-to-high risk of bias. Detailed information regarding quality assessment is shown in Table S2.

Publication bias

For the studies included in the OS, DFS, and DSS analyses, the funnel plots suggested evidence of publication bias (Figure S1).

Discussion

The immune system plays a critical role in tumor development.⁴⁴ The protective value of immunity was initially proposed by Paul Ehrlich in 1909; however, it was impossible to access the validity of the prediction because of the limited awareness of the composition and function of the immune system.⁴⁵ In the 2000s, mouse studies⁴⁵ confirmed that the immune system was an effective tumor-suppressor system and affected the development and progression of cancer in which immunodeficient mice developed tumors earlier and with greater frequency compared with the wild-type mice in the same condition.⁴⁶ Later, similar results were confirmed in human colorectal cancer.¹²

In 2005, Pagès et al.⁴⁴ investigated the role of tumor-infiltrating immune cells in the early metastatic invasion of colorectal cancer and demonstrated that patients with a high density of infiltrating memory and effector memory T cells (CD45RO+) were less likely to disseminate to lymphovascular and perineural structures and to regional lymph nodes. Subsequently, they characterized the tumor-infiltrating immune cells in large cohorts of human colorectal cancers by gene expression profiling and in situ immunohistochemical staining. They found that the immunological data (the type, density, and location of immune cells within the tumor samples) seemed to be a better predictor of patient survival than the histopathological methods.¹⁰ In the study, the patients without recurrence had higher densities of CD3+, CD8+, and CD45RO+ immune cells than those patients with recurrent tumors and had a longer survival.¹⁰ For the tumor location, they found that combining the two regions of CT and IM could improve the prediction of patient survival.¹⁰ CD3+_CT/CD3+_IM density especially was the only independent parameter associated with OS and DFS according to multivariate analysis.¹⁰ The theory was that immune cells would release the cytotoxic mediators when stimulated by an antigen to destroy and kill the tumor cells, which was called “adaptive immune response.”^10,44 For example, CD3+CD8+ T cells with cytotoxic granules that contain perforin and granzymes, which were released on interaction with target cells expressing cognate antigen, led to the death of target cells by apoptosis.¹²

In the following years, Galon and colleagues⁴⁷ created the concept of “Immunoscore,” which was used to reflect the prevalence of immune infiltrates in tumor microenvironment and was determined by the density, location, and type of different immune infiltrate-cells such as total T cells (CD3+), cytotoxic T cells (CD8+), memory T cells (CD45RO+),⁴⁷ and the scoring system ranged from Immunoscore 0(I0)—which had low densities of both cell types in both regions—to Immunoscore 4(I4), having high densities of both cell populations in both locations. In order to standardize the measurement of immune infiltrates and promote the utilization of Immunoscore in clinical practice internationally, a working group composed of the Society for Immunotherapy of Cancer (SITC), the Europe Academy of Tumor Immunology (EATI) and “La fondazione Melanoma Onlus” was formed to validate the consensus Immunoscore in clinical practice for patients with stage I–III colon cancer, which was defined as the density of CD3+ and CD8+ T-cell effectors both in CT and IM because the CD3 and CD8 were the two easiest membrane stains.^8,11,48 The results were recently reported by Pagès et al.,⁸ that Immunoscore provided a reliable estimate of the risk of recurrence in patients with colon cancer and the implementation of the consensus Immunoscore as a new component of a TNM-Immune classification of cancer. In this meta-analysis, approximately 87.5% studies were in line with the consensus Immunoscore in terms of the measurement of immune infiltrates.

Many studies showed that the prognosis value of Immunoscore was superior to microsatellite-instability (MSI) staging.^27,28,49 Pagès et al.⁸ analyzed the connection of the Immunoscore and MSI status in 1579 patients with colon cancer and found that patients with high Immunoscore had prolonged OS, DFS, and time to recurrence no matter what their MSI status was. In a multivariate Cox model, Immunoscore was a significant predictor for OS, DFS, and time to recurrence, while MSI remained a significant factor for time-to-recurrence but not for OS and DFS. Similar results were found in other studies.^27,28

All the evidence has proved that Immunoscore could be used as a prognostic biomarker that can be standardized across pathology laboratories especially for colon cancer. Also, it can be used as an actionable predictive biomarker of responsiveness to both chemotherapy and immunotherapy, which has been recently under investigation.^14,50 Finally, the Immunoscore assay could be used widely in clinical settings once the clinical utility and cost effectiveness are demonstrated, which would promote the further development of precision medicine, and would contribute to global cancer prevention and control.⁴⁹

In gastric cancer, the immune infiltration cells have been investigated since the early 1900s.⁷ Increasing evidence has indicated the association between immune infiltration and clinical outcomes.⁵¹ However, there was no consensus in the measurement of Immunoscore. Jiang et al.³⁹ used the least absolute shrinkage and selection operator (LASSO) Cox regression model to build a classifier, which included 5 features out of 27 and calculated Immunoscore depending on four different types of immune cells (CD3+, CD8+, CD45RO+, and CD66b) by a specific formula. Nevertheless, in other bioinformatics studies, high Immunoscore patients had a worse survival inversely because of the different measurement of Immunoscore.^51,52 In order to investigate which measurement can predict the clinical course of gastric cancer, Lee et al.⁵³ combined the different cells of TILs to search the prognostic role of TILs, but the Immunoscore was not mentioned. Moreover, other immune cell (e.g. the plasma cell) macrophages had an effect on the prognosis of gastric cancer as well.⁵¹ A study combined the CD20+/Tbet+ cells to access the association between immune infiltrates and patient survival, and observed that patients with a high density of CD20+ and Tbet+ had the lowest risk of relapse, whereas patients with a low density of both CD20+ and Tbet+ were at the highest risk of relapse.⁵⁴ Considering the predictive value of Immunoscore in post-surgical survival and sensitivity to adjuvant chemotherapy for gastric cancer,⁵⁵ a standard and comprehensive Immunoscore system was needed to make more accurate prognoses for gastric cancer patients.

The immune microenvironment of lung tumors was composed of T cells, B cells, natural killer cells, mature and immature dendritic cells, tumor-associated macrophages (TAMs), neutrophils, and mast cells.⁵⁶ The prognosis value of macrophages on lung cancer was indicated in 1986.⁵⁷ In 2006, it was justified that CD4+/CD8+ was a favorable prognostic factor in NSCLC and high CD4+/CD8+ was associated with a longer survival.⁵⁸ The CD8+ T cells could recognize particular tumor-associated antigens presented on MHC class I molecules at the cancer cell surface and possessed the ability to destroy cancer cells directly.⁵⁸ Then Al-Shibli et al. found the location of TILs would affect the association between immune cells and prognosis; and a high number of stromal CD8+ and CD4+ cells were independent positive prognostic factors for DSS while not of epithelial cells.⁵⁹ Recently, the density of CD8+, CD45RO+, CD103+, PD-L1, PD-1, mature dendritic cell and other cells have been used to describe the extent of TILs for NSCLC.^32,33,60 However, there was no criterion to identify the Immunoscore. The existing evidence shows that the CD8+ T cell, especially the stromal CD8+ T cell density was the most conclusive marker; however, further studies are needed on promising T-cell markers such as CD3, CD4, and CD45RO.^61–63 In our meta-analysis, the Immunoscores of NSCLC were based on epithelial CD45RO+/stromal CD8+ and PD-1/PD-L1 respectively. The uniform measurement of Immunoscore is also needed for NSCLC.

Recently, the value of Immunoscore has been investigated in several other types of cancer, including ovarian cancer, hepatocellular carcinoma, pancreatic cancer, bladder cancer, head and neck squamous cell carcinoma. However, the number of studies in the literature is limited (see Table 1 and Table 2). Bösmüller et al.²⁰ found that the combined assessment of CD103 and CD3 counts improves the prognostic value of TIL counts in high-grade serous ovarian cancer. In this paper, patients with CD3^high/CD103^high tumors showed a 5-year survival rate at 90%, CD3^low/CD103^high at 63%, and CD3^low/CD103^low at 0% (P < 0.001). Yao QW investigated the value of IS staging system in hepatitis B virus-related hepatocellular carcinoma (HBV-HCC), and found IS was correlated significantly with OS. They suggested the IS staging was closely related to the outcome of patients, and it can compensate the TNM tumor classification system in predicting the prognosis of HBV-HCC patients.²¹ Similar results were found for patients with bladder cancer⁴² or head and neck squamous cell carcinoma.⁴³ Given the limited number of studies, further research is needed.

In this meta-analysis and systematic review, we searched existing publications to summarize the evidence of the prognostic value of Immunoscore in all type cancers, and found that Immunoscore is significantly associated with the prognosis of patients with cancer. The pooled results indicated that a lower Immunoscore was associated with a poor OS, DFS, and DSS for all cancers. Also, we conducted a subgroup analysis stratified by cancer sites, clinical stage, period of follow-up, number of participants, indicator, and geographic region for patients with colorectal cancer. The results were similar except for the geographic region, which showed that low Immunoscore was significantly correlated with poor prognosis for Europeans, but not significantly correlated with Asians. The results indicated that there may be racial differences in the impact of Immunoscore on the prognosis of colorectal cancer. Considering the number of studies based on Asian populations is limited, future studies are needed to investigate the difference in survival between different races. In addition, in five other types of cancers, the results were similar, but the sample sizes were limited, and there were different indicators for individual cancers.^26,34 Therefore, more studies are warranted to identify the more valuable and reasonable indicators for those cancers.

We systematically evaluated the prognostic role of Immunoscore in eight types of cancer with a large sample size. However, there were several limitations. First, heterogeneity was found between the included studies. The reasons were the different immune cells, various measurements of Immunoscore, and the different cancer types; heterogeneity may also be caused by different follow-up times in each study. We conducted the subgroup analysis and the sensitivity analysis to reduce the high heterogeneity. Second, there were not enough studies in some cancer specialties—especially ovarian cancer, bladder cancer, and HNSCC. Third, the impact of Immunoscore was different between different cancer sites and clinical stages, but we did not perform a subgroup analysis in other cancers (except the colorectal cancer) because of the limited sample size, and the primary information about the each patient was unavailable. Fourth, the scoring cutoffs were different among studies. Finally, publication bias would have had an impact on the results.

Conclusion

Immunoscore had a prognostic value in predicting the progression for all kinds of cancers included in the study, which can be used as a supplementation of TNM classification especially for colorectal cancer. The consensus Immunoscore could be used as a new component in classification for colon cancer. Also, it is possible to use Immunoscore to predict recurrence and to guide adjuvant treatment allocation for patients with colon cancer. However, more high-quality studies are needed to investigate the association between Immunoscore and prognosis for other types of cancer.

Supplemental Material

sj-pdf-1-jbm-10.1177_1724600820927409 - Supplemental material for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients

Supplemental material, sj-pdf-1-jbm-10.1177_1724600820927409 for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients by Xingxia Zhang, Jie Yang, Liang Du, Yong Zhou and Ka Li in The International Journal of Biological Markers

Supplemental Material

sj-jpg-2-jbm-10.1177_1724600820927409 - Supplemental material for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients

Supplemental material, sj-jpg-2-jbm-10.1177_1724600820927409 for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients by Xingxia Zhang, Jie Yang, Liang Du, Yong Zhou and Ka Li in The International Journal of Biological Markers

Supplemental Material

sj-pdf-3-jbm-10.1177_1724600820927409 - Supplemental material for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients

Supplemental material, sj-pdf-3-jbm-10.1177_1724600820927409 for The prognostic value of Immunoscore in patients with cancer: A pooled analysis of 10,328 patients by Xingxia Zhang, Jie Yang, Liang Du, Yong Zhou and Ka Li in The International Journal of Biological Markers

Footnotes

Declaration of conflicting interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Chinese Medical Board Grant on Evidence-Based Medicine, New York, USA (No. 98-680), National Natural Science Foundation of China (No.30901427) and Sichuan Provincial Science and Technology Support Project (2016SZ0047) and National Natural Science Foundation of China (No.71974135).

ORCID iD

Xingxia Zhang

Supplementary material:

Supplemental material for this article is available online.

References

Galon

Mlecnik

Bindea

, et al. Towards the introduction of the ‘Immunoscore’ in the classification of malignant tumours. J Pathol 2014; 232(2):199–209.

O’Sullivan

Brierley

Byrd

, et al. The TNM classification of malignant tumours-towards common understanding and reasonable expectations. Lancet Oncol 2017; 18(7):849–851.

Mlecnik

Bindea

Pagès

, et al. Tumor immunosurveillance in human cancers. Cancer Metastasis Rev 2011; 30(1):5–12.

Galon

Fridman

Pagès

The adaptive immunologic microenvironment in colorectal cancer: a novel perspective.

Cancer Res 2007; 67(5):1883–1888.

Fernando

Kariman

Bloy

, et al. Immune effectors responsible for the elimination of hyperploid cancer cells.[J] .Oncoimmunology 2018; 7: e1463947.

McMasters

KM.

Predicting survival in melanoma. Ann Surg Oncol 2002, 9: 113–114.

MacCarty

Mahle

Relation of differentiation and lymphocytic infiltration to postoperative longevity in gastric carcinoma. J Lab Clin Med 1921; 6: 473.

Pagès

Mlecnik

Marliot

, et al. International validation of the consensus Immunoscore for the classification of colon cancer: a prognostic and accuracy study. Lancet 2018; 391(10135): 2128–2139.

Pagès

Kirilovsky

Mlecnik

, et al. In situ cytotoxic and memory T cells predict outcome in patients with early-stage colorectal cancer. J Clin Oncol 2009; 27(35): 5944–5951.

10.

Galon

Costes

Sanchez-Cabo

, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science (New York, N.Y.) 2006; 313(5795): 1960–1964.

11.

Galon

Pagès

Marincola

, et al. Cancer classification using the Immunoscore: a worldwide task force. J Transl Med 2012; 10: 205.

12.

Fridman

Pagès

Sautès-Fridman

, et al. The immune contexture in human tumours: impact on clinical outcome. Nature Reviews Cancer 2012; 12(4): 298–306.

13.

Stower

Predicting colon cancer recurrence. Nat Med 2018; 24(7): 898.

14.

Sidaway

Immunoscore provides a more accurate prognosis. Nat Rev Clin Oncol 2018; 15(8): 1.

15.

Sun

Dong

Tang

, et al. The prognostic value of immunoscore in patients with colorectal cancer: A systematic review and meta-analysis. Cancer Med 2019; 8(1): 182–189.

16.

Knobloch

Yoon

Vogt

PM.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement and publication bias. J Craniomaxillofac Surg 2011; 39(2): 91–92.

17.

Tierney

Stewart

Ghersi

, et al. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007; 8: 16.

18.

Hayden

van der Windt

Cartwright

, et al. Assessing bias in studies of prognostic factors. Ann Intern Med 2013; 158(4): 280–286.

19.

Anitei

Zeitoun

Mlecnik

, et al. Prognostic and predictive values of the immunoscore in patients with rectal cancer. Clin Cancer Res 2014; 20(7): 1891–1899.

20.

Bösmüller

Wagner

Peper

, et al. Combined Immunoscore of CD103 and CD3 identifies long-term survivors in high-grade serous ovarian cancer. Int J Gynecol Cancer 2016; 26(4): 671–679.

21.

Yao Q, Bao X, Xue R, et al. Prognostic value of immunoscore to identify mortality outcomes in adults with HBV-related primary hepatocellular carcinoma. Medicine (Baltimore) 2017; 96: e6735.

22.

Park

vanWyk

Roxburgh

CSD

, et al. Tumour invasiveness, the local and systemic environment and the basis of staging systems in colorectal cancer. Br J Cancer 2017; 116(11): 1444–1450.

23.

Kim

Yang

Kim

, et al. Combined prognostic effect of PD-L1 expression and immunoscore in microsatellite-unstable advanced gastric cancers. Oncotarget 2017; 8(35): 58887–58902.

24.

Kwak

Koh

Kim

, et al. Immunoscore encompassing CD3+and CD8+T cell densities in distant metastasis is a robust prognostic marker for advanced colorectal cancer. Oncotarget 2016; 7(49): 81778–81790.

25.

Liu

Peng

, et al. Prognostic value of Immunoscore and PD-L1 expression in metastatic colorectal cancer patients with different RAS status after palliative operation. Biomed Res Int 2018; 2018(8): 1–8.

26.

Miksch

Schoenberg

Weniger

, et al. Prognostic impact of tumor-infiltrating lymphocytes and neutrophils on survival of patients with upfront resection of pancreatic cancer. Cancers (Basel). 2019; 11(1): 10.3390/cancers11010039.

27.

Mlecnik

Bindea

Angell

, et al. Integrative analyses of colorectal cancer show immunoscore is a stronger predictor of patient survival than microsatellite instability. Immunity 2016; 44(3): 698–711.

28.

Mlecnik

Tosolini

Kirilovsky

, et al. Histopathologic-based prognostic factors of colorectal cancers are associated with the state of the local immune reaction. J Clin Oncol 2011; 29(6): 610–618.

29.

Mlecnik

Vanden Eynde

Bindea

, et al. Comprehensive intrametastatic immune quantification and major impact of Immunoscore on survival. J Natl Cancer Inst 2018; 110(1): 10.1093/jnci/djx123.

30.

Nearchou

Lillard

Gavriel

, et al. Automated analysis of lymphocytic infiltration, tumor budding, and their spatial relationship improves prognostic accuracy in colorectal cancer. Cancer Immunol Res 2019; 7(4): 609–620.

31.

Park

McMillan

Edwards

, et al. Comparison of the prognostic value of measures of the tumor inflammatory cell infiltrate and tumor-associated stroma in patients with primary operable colorectal cancer. Oncoimmunology 2016; 5(3): e1098801.

32.

Paulsen

Kilvaer

Khanehkenari

, et al. CD45RO+/CD8+ memory T lymphocytes–a candidate marker for TNM-Immunoscore in squamous non-small cell lung cancer. Neoplasia 2015; 17(11): 839–848.

33.

Paulsen

Kilvaer

Khanehkenari

, et al. Assessing PDL-1 and PD-1 in non-small cell lung cancer: a novel immunoscore approach. Clin Lung Cancer 2017; 18(2): 220–233.e8.

34.

Tahkola

Mecklin

Wirta

, et al. High immune cell score predicts improved survival in pancreatic cancer. Biomed Res Int 2018; 472(4): 653–665.

35.

Van den Eynde

Mlecnik

Bindea

, et al. The link between the multiverse of immune microenvironments in metastases and the survival of colorectal cancer patients. Cancer Cell 2018; 34(6): 1012–1026.e3.

36.

Wang

Lin

Huang

, et al. The Immunoscore system predicts prognosis after liver metastasectomy in colorectal cancer liver metastases. Cancer Immunol Immunother 2018; 67(3): 435–444.

37.

Ward-Hartstonge

McCall

McCulloch

, et al. Inclusion of BLIMP-1(+) effector regulatory T cells improves the Immunoscore in a cohort of New Zealand colorectal cancer patients: a pilot study. Cancer Immunol Immunother 2017; 66(4): 515–522.

38.

Wirta

Seppälä

Friman

, et al. Immunoscore in mismatch repair-proficient and -deficient colon cancer. J Pathol Clin Res 2017; 3(3): 203–213.

39.

Jiang

Zhang

, et al. ImmunoScore signature a prognostic and predictive tool in gastric cancer. Ann Surg 2019; 267(3): 504–513.

40.

Yomoda

Sudo

Kawahara

, et al. The Immunoscore is a superior prognostic tool in stages ii and iii colorectal cancer and is significantly correlated with programmed death-ligand 1 (PD-L1) expression on tumor-infiltrating mononuclear cells. Ann Surg Oncol 2019; 26(1): 415–424.

41.

Pyo

. Clinicopathological significance and prognostic role of tumor-infiltrating lymphocytes in colorectal cancer. Int J Biol Markers 2019; 34(2): 132–138.

42.

Mansure

Solanki

, et al. Presence of lymphocytic infiltrate cytotoxic T lymphocyte CD3+, CD8+, and immunoscore as prognostic marker in patients after radical cystectomy. PLoS ONE 2018; 13 (10). e0205746.

43.

Zhang

Song

Shen

, et al. Prognostic and predictive values of immune infiltrate in patients with head and neck squamous cell carcinoma. Cancer Immunol Immunother 2018; 82: 104–112.

44.

Pagès

Berger

Camus

, et al. Effector memory T cells, early metastasis, and survival in colorectal cancer. N Engl J Med 2005; 353(25): 2654–2666.

45.

Schreiber

Old

Smyth

MJ.

Cancer immunoediting: integrating immunity’s roles in cancer suppression and promotion. Science (New York, N.Y.) 2011; 331(6024): 1565–1570.

46.

Shankaran

Ikeda

Bruce

, et al. IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature 2001; 410(6832): 1107–1111.

47.

Angell

Galon

From the immune contexture to the Immunoscore: the role of prognostic and predictive immune markers in cancer. Curr Opin Immunol 2013; 25(2): 261–267.

48.

Galon

Pagès

Marincola

, et al. The immune score as a new possible approach for the classification of cancer. J Transl Med 2012; 10: 1.

49.

Ogino

Giannakis

Immunoscore for (colorectal) cancer precision medicine. Lancet 2018; 391(10135): 2084–2086.

50.

Jorissen

Sakthianandeswaren

Sieber

OM.

Immunoscore-has it scored for colon cancer precision medicine?

Ann Transl Med 2018; 6(Suppl 1): S23.

51.

Zeng

Zhou

, et al. Gene expression profiles for a prognostic immunoscore in gastric cancer. Br J Surg 2018; 105(10): 1338–1348.

52.

Wen

Wang

, et al. A four-factor Immunoscore system that predicts clinical outcome for stage II/III gastric cancer. Cancer Immunol Res 2017; 5(7): 524–534.

53.

Lee

Chae

Lee

, et al. Prognostic implications of type and density of tumour-infiltrating lymphocytes in gastric cancer. Br J Cancer 2008; 99(10): 1704–1711.

54.

Hennequin

Derangère

Boidot

, et al. Tumor infiltration by Tbet+ effector T cells and CD20+ B cells is associated with survival in gastric cancer patients. Oncoimmunology 2016; 5(2): e1054598.

55.

Wang

ImmunoScore predicts gastric cancer postsurgical outcome. Lancet Oncol 2017; 18(2): e68.

56.

Remark

Becker

Gomez

, et al. The non-small cell lung cancer immune contexture. A major determinant of tumor characteristics and patient outcome. Am J Respir Crit Care Med 2015; 191(4): 377–390.

57.

Takeo

Yasumoto

Nagashima

, et al. Role of tumor-associated macrophages in lung cancer. Cancer Res 1986; 46(6): 3179–3182.

58.

Hiraoka

Miyamoto

Cho

, et al. Concurrent infiltration by CD8+ T cells and CD4+ T cells is a favourable prognostic factor in non-small-cell lung carcinoma. Br J Cancer 2006; 94(2): 275–280.

59.

Al-Shibli

Donnem

Al-Saad

, et al. Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer. Clin Cancer Res 2008; 14(16): 5220–5227.

60.

Mandarano

Bellezza

Belladonna

, et al. Assessment of TILs, IDO-1, and PD-L1 in resected non-small cell lung cancer: an immunohistochemical study with clinicopathological and prognostic implications. Virchows Arch 2019; 474(2): 159–168.

61.

Schalper

Brown

Carvajal-Hausdorf

, et al. Objective measurement and clinical significance of TILs in non-small cell lung cancer. J Natl Cancer Inst 2015; 107(3). DOI:10.1093/jnci/dju435.

62.

Donnem

Kilvaer

Andersen

, et al. Strategies for clinical implementation of TNM-Immunoscore in resected nonsmall-cell lung cancer. Ann Oncol 2016; 27(2): 225–232.

63.

Donnem

Hald

Paulsen

, et al. Stromal CD8+ T-cell density—a promising supplement to TNM staging in non-small cell lung cancer. Clin Cancer Res 2015; 21(11): 2635–2643.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.23 MB

0.11 MB

0.05 MB