Helicobacter pylori serology is associated with worse overall survival in patients with melanoma treated with immune checkpoint inhibitors.

The microbiome is now regarded as one of the hallmarks of cancer and several strategies to modify the gut microbiota to improve immune checkpoint inhibitor (ICI) activity are being evaluated in clinical trials. Preliminary data regarding the upper gastro-intestinal microbiota indicated that Helicobacter pylori seropositivity was associated with a negative prognosis in patients amenable to ICI. In 97 patients with advanced melanoma treated with ICI, we assessed the impact of H. pylori on outcomes and microbiome composition. We performed H. pylori serology and profiled the fecal microbiome with metagenomics sequencing. Among the 97 patients, 22% were H. pylori positive (Pos). H. pylori Pos patients had a significantly shorter overall survival (p = .02) compared to H. pylori negative (Neg) patients. In addition, objective response rate and progression-free survival were decreased in H. pylori Pos patients. Metagenomics sequencing did not reveal any difference in diversity indexes between the H. pylori groups. At the taxa level, Eubacterium ventriosum, Mediterraneibacter (Ruminococcus) torques, and Dorea formicigenerans were increased in the H. pylori Pos group, while Alistipes finegoldii, Hungatella hathewayi and Blautia producta were over-represented in the H. pylori Neg group. In a second independent cohort of patients with NSCLC, diversity indexes were similar in both groups and Bacteroides xylanisolvens was increased in H. pylori Neg patients. Our results demonstrated that the negative impact of H. pylori on outcomes seem to be independent from the fecal microbiome composition. These findings warrant further validation and development of therapeutic strategies to eradicate H. pylori in immuno-oncology arena.

Introduction

Immune checkpoint inhibitors (ICIs) are now at the forefront of the management of patients with melanoma. Anti-PD-1 monotherapy or combination therapy with anti-CTLA-4 has led to unparalleled improvements in overall survival (OS) in patients with metastatic melanoma and represents the standard of care.[1] Despite improved outcomes with these new therapies, there is a need to develop robust biomarkers to adequately predict primary resistance, which occurs in 24–40% of patients.[2,3] Several intrinsic and extrinsic tumor factors such as PD-L1 expression, tumor – infiltrating lymphocytes, and neoantigens have been evaluated as biomarkers of ICI efficacy with moderate performances.[4] The gut microbiome has recently emerged as a biomarker of primary resistance and is now regarded as a hallmark of cancer.[5] Fecal microbiome profiling coupled with preclinical mouse experiments has revealed that beneficial bacteria are associated with ICI-response and immune-related side effect profile.[6-11] Moreover, several strategies to manipulate the microbiota including fecal microbiota transplantation have shown encouraging preliminary clinical results.[12-14]

While the focus has been on fecal commensal bacteria, the gastro-intestinal (GI) tract harbors various ecosystem subjected to different environmental conditions at different anatomical regions along the GI tract. Helicobacter pylori (H. pylori) colonizes the stomach mucosa and is present in over 50% of the world population. This bacterium was first discovered to cause gastric ulcer and MALT lymphoma.[15,16] Beyond the local effect in the stomach, recent evidence suggests that H. pylori has the potential to shift the macrophage polarization and could potentially alter systemic immune response.[17] Recently, Oster et al. demonstrated that neonatal infections with H. pylori in mice inhibited the ICI response through a shift in dendritic cell (DC) presentation to CD8+ T cells independently of the fecal microbiome composition response.[18] The impact of H. pylori on the systemic immune response was further revealed in independent cohorts of patients with non-small cell lung cancer (NSCLC) and gastric cancer treated with ICIs where H. pylori seropositivity was associated with a shorter OS.[18,19]

Considering this observation in NSCLC, we sought to determine the impact of H. pylori on clinical outcomes of patients with advanced melanoma treated with ICI, and if the seropositivity of H. pylori influences the fecal microbiome composition.

Materials and methods

Patients

We conducted a multicenter retrospective cohort study of 97 patients with advanced melanoma. The study was conducted across four academic centers in Canada: Center Hospitalier de l’Université de Montréal (CHUM), the McGill University Health Center in Montreal (MUHC), the Jewish General Hospital in Montreal (JGH) and the Sunnybrook Health Sciences Center in Toronto. This study was approved by the institutional ethics committee (IRB: MP-02-2022-10115).

Inclusion criteria for this study were patients with a histologically proven diagnosis of advanced or unresectable melanoma treated with ICI therapy between 2013 and 2021. Patients with a diagnosis of uveal melanoma were excluded.

Tumor response was evaluated on CT scans using the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.[20] Immune related-adverse events (irAEs) were recorded according to CTCAE v5.[21] All patients were followed until death or until data lock on October 25, 2021.

H. pylori serology test

H. pylori serologic status (H. pylori IgG) was determined with validated commercial ELISA tests (H. pylori IgG ELISA Kit, Genesis Diagnostics). H. pylori IgG antibodies were quantified in the plasma of 97 patients, following the protocol for ELISA tests. The optical density (OD) at 450 nm was read in a microplate reader (Tecan, Infinite M200 Pro). The patients were stratified into H. pylori positive (Pos) or negative (Neg) groups according to their serology. Plasmas were collected as part of institutional biobank (Biobank Montreal Cancer Consortium 16.161 and TIME Study Sunnybrook).

Shotgun sequencing analysis

Fecal samples were obtained at the initiation of the ICI treatment for 44 of the 97 patients. We also examined fecal samples from patients with NSCLC (n = 28) (CRCHUM IRB 18.085) from a cohort published in Oster et al.[18] Feces were collected according to International Human Microbiome Standards (IHMS) guidelines (SOP 03 V1). Isolated DNA was analyzed using shotgun sequencing to investigate the microbial composition in fecal samples.[22] DNA was extracted following Suau et al.’s protocol.[23] The genetic material was subsequently sequenced using pyrosequencing. Resulting reads were then filtered using AlienTrimmer to both remove low quality reads as well as sequencing adapters.[24] The resulting cleaned data was further processed to remove human and other potential DNA contaminants. This was performed by removing any sequences matching to the human, Bos taurus and Arabidopsis thaliana genome with an identity score threshold of 97% using Bowtie 2.[25]

Microbial taxonomic profiling and quantification have been performed on the resulting reads using MetaPhlAn 3.0.[25] The samples were subsequently filtered to remove any samples containing a countable species level below 3 standard deviations from the mean.

Additionally, we compared the role of H. pylori seropositivity in terms of relative abundance of key immunotherapy-related bacteria:[9-11]Bifidobacterium longum, Faecalibacterium prausnitzii, Ruminoccocus genus, Streptococcus parasanguinis and Streptococcus salivarius. Akkermansia muciniphila (Akk) trichotomic analysis was conducted in the NSCLC cohort using the same thresholds as published by Derosa et al. Nature Medicine 2022.[9] All these analyses were conducted using a Mann–Whitney U test. H. pylori Pos and H. pylori Neg groups were subsequently analyzed in terms of these groups using a Chi-square test.

H. Pylori IgG virulent factors testing

We performed the recomLine Helicobacter IgG 2.0 line immunoassay on the serum of 20 patients with melanoma to identify specific antibodies against ten selected antigens of Helicobacter pylori (such as cytotoxin-associated gene A; CagA, Vacuolating cytotoxin; VacA and GroEL).[26] The evaluation of the presence of horizontal gene transfer in the stools was performed by retrieving known virulence factors listed by Chang et al. DNA sequences on NCBI’s Gene Webservice (VacA, CagA, T4SS, BabA, and DupA).[27,28] These sequences were subsequently compared to the patient microbiome data using BLAST and Daisy.[29,30]

Statistically Analysis

Proportions and means were compared by the chi-square test, or t-test, for categorical and continuous variables, respectively. Objective response rate (ORR) was defined as the sum of complete (CR) and partial (PR) responses according to RECIST 1.1. OS was defined as time between ICI initiation and death from any cause. Progression-free survival (PFS) was defined as the time between ICI initiation and tumor progression (defined using RECIST 1.1), or death, whichever occurred first. Survival functions for OS and PFS were estimated with the Kaplan–Meier estimator. Log-rank test was used to compare survival distributions between groups. The association between clinical and biological variables and survival outcomes was assessed with univariate Cox models.

Metagenomic analyses were performed with the phyloseq R package.[31] Shannon and Inverse Simpson indexes was used as alpha-diversity measurements at the species level. The Wilcoxon signed-rank test was used to determine significant differences of these indexes among the different groups. The Bray-Curtis dissimilarity was used as a beta-diversity measurement and compared between groups using the PERMANOVA test. Linear discriminant analysis (LDA) Effect Size (LEfSe) was used to perform differential microbial abundance analysis. All statistics and figures were performed using R v4.1.0 and GraphPad Prism v.8.3.1. A p-value ≤0.05 was considered as statistically significant.

Results

Association between the H. pylori seropositivity status and clinical outcomes in the melanoma cohort

In this cohort of 97 patients with advanced melanoma, the median follow-up was 29.9 months and the majority of patients received ICI in the first-line setting (80%), with 81% of patients treated with anti-PD-1 monotherapy, while 12% received the combination of anti-PD-1 plus anti-CTLA-4. The mean age of the studied population was 61 years, with most patients having an ECOG performance status of 0 (81%) (Table 1).

Table 1.

Baseline characteristics of patients with advanced melanoma (n = 97).

Characteristics	Total	H. pylori Neg	H. pylori Pos	p-Value
	n = 97	n = 76	n = 21
Age at ICI initiation, yearsa	61 (13.5)	61 (13.7)	62 (13.8)	0.74
Sexb
Female	36 (37)	31 (41)	5 (24)	0.15
Male	61 (63)	45 (59)	16 (76)
ECOG performance-statusscoreb
0	74 (76)	59 (78)	15 (71)	0.55
1	23 (24)	17 (22)	6 (29)
Stageb
IIIb	12 (12)	10 (13)	2 (10)	0.75
IIIc	15 (15)	11 (14)	4 (19)
IV	70 (73)	55 (73)	15 (71)
BRAF statusb
Wild-type	56 (64)	43 (63)	13 (65)	0.89
Mutated	32 (36)	25 (37)	7 (35)
NA	9	8	1
LDH at ICI (N < 215)b
< ULN	72 (76)	55 (74)	17 (81)	0.77
> ULN	23 (24)	19 (26)	4 (19)
NA	2	2	1
Brain metastasisb
No	81 (84)	61 (80)	20 (95)	0.18
Yes	16 (16)	15 (20)	1 (5)
First line ICIb
No	19 (20)	15 (20)	4 (19)	1.00
Yes	78 (80)	61 (80)	17 (81)
ICIb
Anti-PD-1	79 (81)	59 (78)	20 (95)	0.42
Anti-CTLA-4	6 (7)	6 (8)	0 (0)
ICI combinaison	12 (12)	11 (14)	1 (5)
Antibioticsb
No	91 (94)	71 (93)	20 (95)	1.00
Yes	6 (6)	5 (7)	1 (5)
Proton Pomp Inhibitorsb
No	83 (85)	63 (83)	20 (95)	0.29
Yes	14 (15)	13 (17)	1 (5)

Notes: aExpressed in mean (SD). Student’s t test p-value. bExpressed in n (%). Chi-square test p-value. ICI: Immune Checkpoint Inhibitor.

H. pylori antigen-directed IgG antibodies were detected in the plasma of 21 (22%) patients. Baseline patient characteristics between H. pylori Pos vs. H. pylori Neg were well balanced between both groups (Table 1). Importantly, there was no difference of antibiotic (ATB) exposure 30 days prior to ICI initiation in both groups (p = 1.00). In addition, 14 patients were on proton pomp inhibitors (PPI) prior to ICI initiation, and it was well balanced between both groups (p = .29) (Table 1).

Next, we assessed the impact of H. pylori serologic status on the clinical outcomes. First, ORR was numerically decreased in the H. pylori Pos compared to H. pylori Neg group, 57% and 75%, respectively (p = .11) (Figure 1a). Second, in H. pylori Pos, OS was significantly shorter compared to H. pylori Neg with a median OS of 44.1 months and not reached (NR), respectively (HR 3.21 95% CI 1.15–8.97, p = .02) (Figure 1b). There was a trend toward shorter PFS in the H. pylori Pos group compared to the H. pylori Neg group 22.2 months vs NR (HR 1.92 95% CI 0.95–3.92, p = .07) (Figure 1c).

Figure 1.

Association between the H. pylori seropositivity status and clinical outcomes in 97 patients with advanced melanoma. A. Stacked barplot between H. pylori seropositivity status in terms of ORR in 97 patients with advanced melanoma. CR; complete response. PR; partial response, SD; stable disease, PD; progressive disease analyzed using Chi-square test. B. Kaplan–Meier curve of overall survival in 97 patients with advanced melanoma. C. Kaplan–Meier curve of progression-free survival in 97 patients with advanced melanoma. D. Stacked barplot between H. pylori seropositivity status in terms of autoimmune toxicities in 97 patients with advanced melanoma.

Then, we addressed the potential association between irAEs and H. pylori seropositivity. Seventeen patients (18%) had grade 3–4 irAEs. There was no association between H. pylori status and the occurrence of irAEs (p = .35) (Figure 1d).

Next, we performed a subgroup analysis taking into consideration only the patients treated in the first-line setting (n = 78). In this sub-group of 78 patients, 17 (22%) patients were H. pylori Pos and baseline characteristics were well balanced between H. pylori groups (Supp. Table 1). ORR was numerically decreased in the H. pylori Pos group compared to H. pylori Neg group (64% vs 82%, p = .38) (Figure 2a). OS was significantly shorter in the H. pylori Pos (HR 5.45, 95% CI 1.47–20.2, p = .005) (Figure 2b). PFS was numerically shorter in the H. pylori Pos patients (22.2 months vs NR, HR 2.13 95% CI 0.87–5.21, p = .09) (Figure 2c). Lastly, there was no association between irAEs and H. pylori status (p = .75) (Figure 2d).

Figure 2.

Association between the H. pylori seropositivity status and clinical outcomes in 78 patients with advanced melanoma treated with first-line ICI. A. Stacked barplot between H. pylori seropositivity status in terms of ORR in 78 patients with advanced melanoma treated with first-line ICI. CR; complete response. PR; partial response, SD; stable disease, PD; progressive disease analyzed using Chi-square test. B. Kaplan–Meier curve of overall survival in 78 patients with advanced melanoma treated with first-line ICI. C. Kaplan–Meier curve of progression-free survival in 78 patients with advanced melanoma treated with first-line ICI. D. Stacked barplot between H. pylori seropositivity status in terms of autoimmune toxicities in 78 patients with advanced melanoma treated with first-line.

H. pylori seropositivity does not impact fecal microbiome diversity

We subsequently sought to understand the potential association between fecal microbiome composition and H. pylori seropositivity. Comparison of the fecal microbiome content was performed on patients with melanoma (n = 43; n = 13 patients H. pylori Pos compared to n = 30 patients H. pylori Neg). There was no difference in alpha diversity metrics between H. pylori Pos and H. pylori Neg groups (Shannon: p = .99 and Inverse Simpson: p = .74, Figure 3a). Beta diversity analysis between both groups showed no significant cluster difference following statistical analysis (p = .20, Figure 3b). Differential abundance analyses showed an increase of Dorea formicigenerans, Eubacterium ventriosum, and Ruminococcus (Mediterraneibacter) torques in the H. pylori Pos group while a significant enrichment of Alistipes finegoldii, Clostridium sp. CAG 58, Blautia producta and Hungatella hathewayi in the H. pylori Neg group was observed (Figure 3c). In addition, we assessed the relative abundance of known bacteria associated with ICI response such as Bifidobacterium longum, Faecaliacterium prausnitzii and Ruminococcus genus between H. pylori Pos and H. pylori Neg patients and showed there was no difference (p = .46, 0.35 and 0.38, respectively) (Figure 3d, 3E Supp. Fig. 1A). Then, we analyzed the content of Streptococcus parasanguinis and Streptococcus salivarius which correlated with ICI primary resistance in melanoma.[11] No difference was observed between H. pylori Pos and Neg groups (p = .62 and p = .78, respectively) (Supp. Fig. 1B and 1C).

Figure 3.

Gut microbiome composition for H. pylori Pos and H. pylori Neg patients in melanoma cohort. A. Alpha diversity of H. pylori Pos and H. pylori Neg patients (n = 43). The bold line represents the median. The bottom and top hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The top whisker extends from the hinge to the largest value no further than 1.5 * interquartile range from the hinge. B. Beta diversity for 43 patients with melanoma stratified by H. pylori seropositivity. C. Differential abundance analysis using LEfSe stratified by H. pylori status. Only species were displayed. Note that the presence of each bacterium on the LEfSe denotes statistical significance (p < .05). D. Relative abundance of Bifidobacterium longum between H. pylori Pos and H. pylori Neg patients. E. Relative abundance of Faecalibacterium prausnitzii between H. pylori Pos and H. pylori Neg patients.

To strengthen these observations, we performed the same analytical pipeline in an independent cohort using data obtained from 29 patients with NSCLC published in Oster et al.[18] In this cohort, eight patients were H. pylori Pos, and the PFS was decreased in the H. pylori Pos patients compared to H. pylori Neg group (2.1 vs 9.2 months HR = 2.39, 95% CI 1.01–5.65, p = .05). When we analyzed the fecal microbiome profiling of these patients according to H. pylori seropositivity, we did not observe any significant change in terms of alpha diversity (Shannon: p = .94 and Inverse Simpson: p = .82, Figure 4a) and absence of unique cluster for beta diversity analysis (p = .58, Figure 4b). The LEfSe analysis on the cohort demonstrated a significant increase in Phascolarctobacterium sp. CAG 266, Hungatella hathewayi, Bacteroides galacturonicus and Clostridium sp. CAG 299 in the H. pylori Pos group. Conversely, in the H. pylori Neg group, we observed an increase in Bacteroides xylanisolvens (Figure 4c). Then, we investigated the relative abundance of Akkermansia muciniphila, where no difference was observed (p = .57) (Figure 4d). Next, we represented Akkermansia muciniphila in a trichotomic stratification as previously published[9] and showed no correlation with H. Pylori status (p = .61) (Figure 4e).

Figure 4.

Gut microbiome composition for H. pylori Pos and H. pylori Neg patients in lung cohort. Alpha diversity of H. pylori Pos and H. pylori Neg NSCLC patients (n = 28). The bold line represents the median. The bottom and top hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The top whisker extends from the hinge to the largest value no further than 1.5 * interquartile range from the hinge. B. Beta diversity for 28 NSCLC patients stratified by H. pylori seropositivity. C. Differential abundance analysis using LEfSe stratified by H. pylori status. Only species were displayed. Note that the presence of each bacterium on the LEfSe denotes statistical significance (p < .05). D. Relative abundance of Akkermansia muciniphila between H. pylori Pos and H. pylori Neg patients. E. Trichotomic analysis of Akkermansia muciniphila between H. pylori Pos and H. pyloriNeg patients.

Altogether, in the melanoma and lung cohorts, our results suggest that beneficial bacteria with known immune-potentiating effects such as Eubacterium, Dorea, Hungatella and Alistipes were increased independently of H. pylori status (Figure 3c, Figure 3d).[7,9-11,32,33]

Finally, we tried to detect H. pylori virulence factor such as VacA, CagA, T4SS on the serum and metagenomic raw data.

In 20 H. pylori seropositive patients, immunoassay revealed that 18 (90%) patients were tested positive for at least one H. pylori virulent antigen. Using the manufacturer brochure, 16 patients tested positive with 9 (56%) patients presented a type I (express virulent CagA or VaCA protein) and 7 (44%) patients presented a type II (does not express virulent CagA or VaCA protein) infection. Eight (40%) patients were positive for CagA, 4 (20%) patients were positive for VacA and 17 (85%) were positive for GroEL. Therefore, there is a possibility that four (20%) patients may have been misclassified as a false-positive with the ELISA test.

To determine if this new analysis could impact our results, we reanalyzed the data and included these four patients in the H. pylori Neg group. Nevertheless, OS was significantly shorter compared to H. pylori Neg with a median OS of 44.1 months and not reached, respectively (HR 3.38 95% CI 0.74–15.5, p = .01) (Supp. Fig. 2A). Of note, we did not find any difference in OS based on the type I and type II infection (p = .59) (Supp. Fig. 2B). However, we were unable to demonstrate that any of the H. pylori sequences were horizontal gene transfer in the fecal samples sequenced.

Discussion

First, the results of our study suggest that H. pylori seropositivity status was associated with shorter OS in patients with melanoma. Second, the worse clinical outcomes observed in H. pylori Pos group were independent of prior ATB exposure, an important finding given that ATB have been associated with worse response to ICI in a meta-analysis of more than 11,000 patients.[34] Third, neither alpha- nor beta-diversities were different in H. pylori groups providing no evidence of a unique microbiome cluster present in each group. Finally, in order to validate that this association was independent on the gut microbiome, we found similar results in one independent NSCLC cohort.

Our finding of the association of H. pylori Pos status with negative outcome supports prior work from two cohorts of 29 and 60 patients with NSCLC amenable to ICI whereby H. pylori Pos patients had decrease ICI efficacy.[18] Oster et al. further demonstrated in preclinical model that H. pylori infection reduced the activation of CD8+T cells by altering the cross-presentation activities of DC and affected the innate and adaptive immune response of the infected mice.[18] Moreover, a recent paper showed in a cohort of 77 patients with gastric cancer treated with anti-PD-1, that H. pylori seropositive was associated with a shorter OS and PFS.[19] Furthermore, from recent large microbiome studies in melanoma and NSCLC patients, we observed bacteria associated with favorable response such as Eubacterium ventriosum, Ruminococcus (Mediterraneibacter) torques, Dorea formicigenerans significantly overrepresented in H. pylori Pos patients.[11,35] Moreover, Hungatella hathewayi was overrepresented in H. pylori Neg patients. This bacterium is known to be increased post-ATB exposure in NSCLC and RCC and was associated with ICI resistance.[32,33]

In the additional metagenomic we performed in the independent NSCLC cohort, we did not observe any evidence of unique bacteria cluster based on H. pylori status. Furthermore, Hungatella hathewayi was, in this cohort, increased in the H. pylori Pos group, however Bacteroides xylanisolvens was increased in H. pylori Neg group. Bacteroides xylanisolvens has recently been associated with resistance in a large cohort of 94 patients with melanoma.[10] The absence of unique microbiome signature according to H. pylori status was previously demonstrated in murine experiments.[18] Murine infections of H. pylori were less responsive to either anti-CTLA4/PD-L1 combination therapy or Ova specific CD8+ T cells (OT-1 cells) vaccine. Two strategies to show an absence of the H. pylori-mediated immunosuppression relied on the microbiome were used. Co-housing of infected and not infected mice did not alter the immunosuppressive phenotype of H. Pylori (co-housing did not induce H. pylori transmission). Next, 16S rRNA did not show any influence in microbiome composition in H. pylori infected animal.[18] Taken together, these findings are in-line with our findings that H. pylori associated with OS in our cohort, and that this association was independent on the gut microbiome.

We observed a discordance between H. pylori seropositivity on ELISA test and on immunoassay. Quantitative IgG detection represents the standard of care in routine clinical setting while immunoassay specific for H. pylori remains a research tool. Only, two patients had no specific IgG response to any virulent factors and the discrepancy in the test is most likely secondary to sensitivity threshold used in the routine ELISA testing or a lost of IgG response to specific virulent factor in an elderly patient population. Nevertheless, the correlation with OS and H. pylori seropositive patients remained significant using both kits.

Questions also remain regarding the impact of different H. pylori strains and geodistribution of the population. Indeed, the H. pylori immune potentiating effect has been associated with refractory immune thrombocytopenia purpura (ITP).[36] Even if society guidelines recommend eradication of H. pylori in this setting, several observational studies from different countries demonstrated that H. pylori eradication did not have the same beneficial effect on ITP leading to the hypothesis that different H. pylori strains may have differing immune effects.[37]

Despite determining the impact of H. pylori on a separate cohort of patients treated with ICI, this study contains several limitations. First, we only tested serum H. pylori and did not distinguish between eradicated vs chronic infection, and our conclusions are only applicable to IgG seropositivity. Second, we only had access to fecal microbiome, and therefore our analysis did not take the upper GI tract microbiome into account. Last, we did not take into account factors which may influence seropositivity, such as socioeconomic factors, diet or other risk factors for H. pylori.

Altogether, our study is the first to report the negative impact of H. pylori on OS in patients with melanoma amenable to ICI and to confirm similar findings in a separate cancer histology. Furthermore, we demonstrated that H. pylori did not seem to have any potent impact on the fecal microbiome and more mechanistic experiments are needed. The data presented in our study only provide hypothesis-generating data and further need to perform additional trials to confirm this observation. Our findings warrant validation in large prospective observational cohort in patients with ICI, comparing the outcomes for melanoma patients between H. pylori Pos and H. pylori Neg patients and reinforce the need to develop novel strategies to eradicate H. pylori in cancer patients.[38]

Click here for additional data file.