Pulmonary lymphangitis carcinomatosis: A peculiar presentation clustering in MET‐amplified gastric cancer

Abstract Background The clinicopathological features of MET‐amplified gastric cancer (GC) and real‐world data on the efficacy of MET‐targeted therapies remain unknown. Pulmonary lymphangitis carcinomatosis (PLC) is a peculiar manifestation of GC, whose management has not been thoroughly described. Methods This study analyzed patients diagnosed with MET‐amplified GC or GC with PLC at any time point of the disease course from 2011 to 2021 in two centers. Clinicopathological features and survival outcomes of MET‐amplified GC were analyzed. The clinical and molecular implications of GC with PLC were discussed. Results Fifty‐eight patients with MET‐amplified GC and 20 patients with GC accompanied by PLC were finally enrolled for analysis (including 13 overlapped patients). GC with PLC was more common in female patients (p = 0.010), diagnosed at a younger age (p = 0.002), presented with a higher baseline ECOG PS (p = 0.016), and was more likely to develop lung metastasis (p < 0.001), and serous effusion (p = 0.026) than GC without PLC. Patients with primary MET‐amplified GC had a worse prognosis than those with secondary MET‐amplified GC (p = 0.005). The application of anti‐MET therapy was associated with numerically prolonged survival, but the association was not statistically significant (p = 0.07). MET amplification was concentrated in patients with PLC, in which anti‐MET therapies elicited a high response rate. Conclusions MET‐targeted therapies are efficacious in real‐world populations with MET‐amplified GC. Patients with PLC have distinct clinical and molecular features and might benefit from MET‐targeted therapies.


| INTRODUCTION
In 2018, gastric cancer (GC) was the fifth most common malignancy (1,034,000 new cases, 5.7% of all cancers) and the third leading cause of cancer-related death worldwide (783,000 deaths, 8.2% of all cancer-related deaths). 1 The prognosis of patients with late-stage GC remains poor.In this context, the development of novel therapies for metastatic or recurrent GC is warranted.
][5][6][7][8] C-Met, a receptor tyrosine kinase (RTK) encoded by the MET gene, and its ligand hepatocyte growth factor (HGF) are targets receiving extensive attention. 94][15] In addition, MET amplification contains two forms: (a) primary amplification is defined as activated MET signaling in treatment-naïve tumors; (b) secondary amplification refers to newly detected MET overactivation in treated tumors (primary and/or metastatic sites). 16It has been suggested in previous research that anti-MET monoclonal antibodies fail to improve survival in patients with MET-overexpressing GC. 17,18 A phase II, multicenter, randomized controlled trial in France found that adding rilotumumab to standard chemotherapy did not bring any benefits to GC patients. 19Despite this, MET tyrosine kinase inhibitors (TKIs), such as crizotinib, may result in a positive response in a specific group of screened patients with MET-amplified GC. [20][21][22][23] These findings suggest that de novo MET amplification could be a pivotal driver of oncogenesis and a druggable target.However, the baseline characteristics of patients could be tremendously different in clinical trials versus those in the real-world.This triggers an interest in investigating the authentic efficacy of anti-MET therapies in unselected populations, which has never been reported.
Pulmonary lymphangitis carcinomatosis (PLC) is a peculiar pattern of lung metastasis featuring lymphatic permeation and lymph capillary infiltration by tumor cells, which portends a miserable prognosis.Previous studies have indicated that breast, gastric, and lung cancers are the most common malignancies that give rise to PLC, while the pathogenic mechanisms of PLC are not understood. 24In addition, patients with PLC generally succumb to poor physical condition and thus are often excluded from clinical trials.Judicious management of PLC remains an unmet need.
We sought to investigate the clinicopathological and prognostic features of MET-amplified GC.Our study also revealed that MET amplification is a molecular event enriched in GC with PLC, which represents a distinct subtype showing responses to molecular targeted therapies.

| Study population
In our two-center retrospective study, eligible patients were diagnosed with recurrent or metastatic GC that was either MET-amplified or manifested with PLC at any time point of the disease course from 2011 to 2021 in two medical centers.Tumor samples were generated from all patients by endoscopy, surgery, or needle biopsy at primary and/or secondary tumor sites.Tissues were fixed with formalin and paraffin-embedded in blocks, which were prepared at the time of biopsy.Patients with available baseline characteristics and clinical data were included in the analysis.All patients were followed until October 2021.

| Confirmation of MET status and PLC
MET gene copy numbers in all tumor samples were detected by next-generation sequencing (NGS).NGS analysis was conducted by Genecast, Inc., a College of American Pathologists (CAP)-accredited laboratory.A customized 769-gene panel (Genecast) was employed to detect point mutations, indels, fusions, rearrangements, and CNVs.Targeted sequencing was performed on the NovaSeq 6000 platform (Illumina).The cutoff for MET amplification was 2.0 times (averaged MET copy number in tumor tissue normalized by normal comparator).The PLC was determined by two independent radiologists based on chest computed tomography (CT) scans.Typical radiological findings include smooth or nodular thickening of interlobular septa, peribronchovascular interstitium, polygonal arcades with thickened limbs, and ground glass opacities.Clinical manifestations (e.g., progressive dyspnea, cough, and tachypnea) also assisted in the diagnosis.

| Immunohistochemistry (IHC)
HER2, programmed death ligand 1 (PD-L1), and mismatch repair (MMR) status were evaluated through IHC.HER2 expression was analyzed using the anti-HER-2 (4B5) rabbit monoclonal primary antibody (Ventana, U.S.).HER2 positivity was defined as IHC 3+ or both IHC 2+ and FISH positive.IHC of MLH1, MSH2, MSH6, and PMS2 was performed with a BenchMark ULTRA slide processing system (Ventana, U.S.) according to the manufacturer's instructions.Tumors were defined as deficient mismatch repair (dMMR) only if at least one of the proteins was absent of staining in the tumor cells; otherwise, proficient mismatch repair (pMMR) was deemed.PD-L1 staining was performed using a 22C3 pharmDx kit (Dako, Denmark).A combined positive score (CPS, defined as the total number of tumor cells and immune cells stained positive divided by the number of viable tumor cells, then multiplied by 100) of 1 or higher was considered PD-L1 positive.

| In situ hybridization (ISH)
HER2 amplification was detected by fluorescence in situ hybridization (FISH) using a PathVysion HER-2 DNA probe kit (Abbott Laboratories, U.S.).Cases with a HER2:centromere 17 ratio >2.0 were considered FISH positive.The presence of EBV infection was determined by ISH using probes against Epstein-Barr virus-encoded RNAs (EBER); (Leica Biosystems).Specimens with dark blue staining in tumor cell nuclei were deemed EBV-positive.

| Statistical analysis
Statistical analyses were performed using SPSS software (v.24.0;IBM Corporation, U.S.) and Prism (v.9; GraphPad Software, U.S.).Categorical variables were evaluated using the chi-square test or Fisher's exact test, as appropriate.Survival curves were plotted using the Kaplan-Meier method, and survival rates were compared between groups using the log-rank test.Cox multivariate regression analyses were used to determine factors that affect survival outcomes.Overall survival (OS) was calculated from the time of diagnosis until death or the last observation for surviving patients.Progression-free survival (PFS) was defined as the interval from the start of therapy to treatment discontinuation for disease progression or death of any cause.All tests were two-sided, and differences were considered significant when p < 0.05.

MET-amplified GC according to PLC manifestation
A total of 937 patients diagnosed with stage IV GC were included for eligibility screening.Of these, 65 patients were eligible and had available clinicopathological information.Fifty-eight patients with MET gene-amplified GC and 20 patients with GC accompanied by PLC were finally enrolled for analysis (including 13 overlapped patients).A flow diagram is shown in Figure S1.Patient demographics and tumor characteristics are summarized in Table S1.Among patients with MET-amplified GC (n = 58, accounted for 6.2% of the total patients), the median age at diagnosis was 53 years (range, 23-84 years), with a male preponderance.Twenty-eight out of 58 (48.3%) patients had serous effusion.Liver, bone, and lung metastasis occurred in 21 (36.2%), 15 (25.9%), and 21 (36.2%)patients, respectively.GC with PLC was more common in females (42.9% vs. 10.8%,p = 0.010), diagnosed at a younger age (median age, 44 vs. 56 years, p = 0.002), and presented with a higher baseline ECOG PS (≥2 points, 46.2% vs. 17.8%,p = 0.016) than GC without PLC.Lung metastasis and serous effusion were more frequent in GC with PLC than in GC without PLC (84.6% vs. 22.2%, p < 0.001; 61.5% vs. 44.4%,p = 0.026).GC with PLC accounted for a higher percentage of bone metastasis and poorly differentiated histology (38.5% vs. 22.2%, p = 0.091; 92.3% vs. 71.1%,p = 0.084) than GC without PLC, though the differences were not significant.For molecular features, the average MET gene copy number was 7.4 (range, 2.1-29.0).No difference was found in MET gene copy number according to PLC (7.2 vs. 8.0, p = 0.76).One out of 13 GC with PLC and 6 out of 45 GC without PLC were HER2 positive.All patients with detectable pathological specimens had pMMR and EBV-negative tumors.Thirty-nine percent of patients (15/38) had PD-L1-positive GC, with CPS ranging from 1 to 98.

MET-amplified GC patients
Survival curves were plotted and compared according to tumor metastatic sites, treatment regimens, clinicopathological parameters, and MET amplification forms.In our cohort, 45 patients had tumors with primary MET amplification, while 13 patients had tumors with secondary MET amplification.Patients with primary MET-amplified GC presented shorter OS than those with secondary METamplified GC (11.2 vs. 25.6 months, p = 0.005); (Figure 1).The presence or absence of liver, bone, lung, or peritoneum metastasis were not correlated with OS (Figure S2).Survival outcome was not different between patients with or without PLC regarding HER2 status or PD-L1 expression.The onset of thorax or peritoneum effusion was not correlated with worse survival (Figure S3).Upon further analysis of the correlation between the therapeutic regimen and OS, we found that patients who received anti-MET therapy had prolonged survival numerically (16.7 vs. 11.2months), while a significant difference was not seen (p = 0.07); (Figure 2).The usage of taxane-based chemotherapy, immunotherapy, or antiangiogenic therapies was not associated with OS.
In Cox regression analysis, age, gender, and prognostic factors with p < 0.15 in univariate analysis (including ECOG PS, presence of PLC, presence of lung metastasis, primary MET amplification, and anti-MET therapy) were added into the multivariate model.Only secondary MET amplification was a predictive factor of a prolonged OS with statistical significance (p = 0.007; hazard ratio, 0.017; confidence intervals, 0.001-0.322).

| Characteristics and treatments of patients with PLC
A total of 20 patients were diagnosed with PLC and included for analysis.The median age at diagnosis was 42.5 years (range 23-77 years).Eleven out of 20 patients were female.Ten patients had an ECOG PS of 0-1, and the other 10 patients had an ECOG PS of 2-3.Seventeen patients had tumors with assessable MET gene status, of which 13 (76.5%)were confirmed MET-amplified and otherwise nonamplified.The genomic variation information of 17 patients with PLC is shown in Figure S4.
Four and seven out of 20 patients were HER2 positive and PD-L1 positive, respectively.All included patients had EBV-negative and pMMR tumors.Tumor mutation burden (TMB) ranged from 1.93 to 14.01 mutations per megabase, denoting that all patients with PLC in our cohort had TMB-low tumors.Thirteen out of 20 patients (65.0%) manifested serous effusion, suggesting its high incidence rates in GC with PLC.The lung, nonregional lymph nodes, liver, adrenal glands, ovary, and bone were common metastatic sites.Other clinicopathological characteristics are presented in Table 2.
Eleven patients received ≥3 lines of treatment for PLC.Ten patients received anti-MET therapies.Regarding drug types, 10 patients received anti-MET therapy with crizotinib, one of whom also received volitinib.Owing to the absence of measurable lesions in a few cases, therapeutic responses were determined from two perspectives (both symptomatically and radiologically).Fourteen patients rapidly alleviated their symptoms (e.g., cough, dyspnea, and fatigue) or obtained radiographical remission after systemic therapies.3).The median OS of patients with PLC was 11.2 months, and the median interval between the first onset of PLC and PLC progression was 2.5 months.

| Analysis of the molecular heterogeneity in MET-amplified GC
Temporal and spatial heterogeneity were found in several patients who progressed on anti-MET therapies (Figure 3).MET copy numbers were sharply reduced in 4 patients.Tumor resampling suggested HER2 and FGFR amplification in two of them, respectively.In addition, one patient with a HER2-negative primary tumor developed HER2-positive brain metastasis.Another patient with a HER2-negative primary tumor developed HER2-positive ovary metastasis.Both patients received anti-HER2 therapies and achieved local control of the new-onset lesions.

| DISCUSSION
MET-amplified GC is a rare tumor subtype that has shown sensitivity to anti-MET TKIs in small-scale studies.No.However, the reported response rates are diverse.In a phase 2 single-arm trial, second-line AMG-337 (a smallmolecule MET inhibitor) resulted in an overall response rate (ORR) of 18% in patients with pretreated METamplified GC, with a median PFS of 3.4 months. 21In the VICTORY umbrella trial, second-line treatment with savolitinib (a highly selective small-molecule MET inhibitor) resulted in an ORR of 50% in 20 metastatic MET-amplified GC patients. 22However, in unselected populations, anti-MET therapies only show suboptimal effects.In molecular unselected GC patients treated with foretinib (an oral, multitarget MET inhibitor), 23% achieved SD with a median duration of 3.2 months. 25Among unselected metastatic GC patients treated with tivantinib (a smallmolecule selective MET inhibitor), 36.7% achieved disease control with a median PFS of only 43 days. 26These inconsistent results indicate the necessity of precise screening of beneficiaries.It should be emphasized that in real-world conditions, MET-amplified GC is typically aggressive or advanced.Patients are commonly assessed with an ECOG PS of 3 or 4 and are generally excluded from trials that recruit participants with fair physical status.Therefore, the authentic efficacy of anti-MET therapies in GC in the realworld setting remains unknown.We report for the first time the clinicopathological and molecular features of MET-amplified GC.The absence of an internationally recognized set of standards has led to the investigation of various cutoff values to define MET amplification.We took a cutoff of 2.0 (copies in tumor tissue normalized by control) in our study.Our previous research used this cutoff, which mirrored the HER2 amplification cutoff. 11In another clinical study that had distinct detection methods and platforms, 6-copy was used as a cutoff. 23To determine the gold standard for MET amplification in GC patients, further studies are definitely required.In our cohort, MET-amplified GC is highly invasive and aggressive, manifests as multiple organ metastases involving the lung, liver, and ovary, and is often accompanied by serous effusions, which is in line with previous reports. 11,13,27Patients with primary MET-amplified GC had a significantly worse prognosis than those with secondary MET-amplified GC.This is comprehensible since acquired MET aberrations often occur in the later course of the disease secondary to therapeutic pressure.The biological behavior is similar to that of MET nonamplified tumors before MET amplification develops, but since then, the disease could be aggressive and lethal.In terms of prognostic markers, we found that distant organ metastasis was not associated with prognosis.Our study also suggests that MET amplification could coexist with HER2 positivity but is rarely accompanied by EBV infection or dMMR status.Twenty-eight patients in our cohort received anti-MET TKIs, of which 20 (71.4%) achieved disease control.Considering that most patients were heavily pretreated or progressed after multiline therapy, these efficacy data are impressive.In addition, patients who have undergone anti-MET treatment have superior OS, which also verifies the effectiveness of the treatment.The small sample size might explain the insignificance of the statistical results.Nevertheless, despite the desirable efficacy, the duration of response is often transient, suggesting there is a rapid development of resistance over time in a few cases.

Sex
Importantly, in this study, we identified that PLC is a unique manifestation of GC.According to past literature, PLC denotes an end-stage condition of malignancy characterized by the infiltration of lymphatic vessels by tumor cells.GC with PLC is a rare condition, with only a few cases reported, and the disease characteristics or management have not been extensively investigated. 28,29n our cohort, we comprehensively depicted the clinicopathological and molecular patterns of GC with PLC.Compared to GC without PLC, GC with PLC has earlier ages at onset, higher tumor burden, more tumor spreading or dissemination, and worse ECOG PS.Patients are inclined to develop compression symptoms due to pleural or peritoneal effusions.No molecular marker (including TMB) alone was found to be significantly associated with the prognosis of patients with PLC, which warrants future investigations in a large-scale cohort.It is surprising that patients with PLC did not have a worse prognosis.The possible explanations are as follows: (1) PLC was a predictor of poor prognosis in the univariate analysis, but not in the multivariate analysis, indicating a potential indirect association between PLC and the prognosis of patients with MET-amplified GC; (2) In our study, we performed Cox regression analysis on 58 patients with MET-amplified GC who themselves had a poor prognosis, as opposed to unselected patients.It is assumed that MET amplification, not PLC, is the determinant of a worse prognosis.
It is noteworthy that GC with PLC accounts for a considerable proportion of MET-amplified GC.In our cohort, 16/58 (27.6%) patients developed PLC at any point during the disease course.Among 17 patients with GC and PLC who had assessable MET status, 13 (76.5%)were confirmed to be MET-amplified, indicating that PLC is enriched in MET-amplified tumors.It is reasonable to infer that aberrant MET signaling potentially gives rise to PLC, which confers profound clinical implications.In this regard, a patient with PLC should be suspected of having a MET-amplified tumor and be referred for genetic testing.Conversely, clinicians should be alert to PLC during the subsequent course of MET-amplified GC for earlier interventions (Figure 4).Given the rarity of MET amplification and the economic cost of molecular testing, it is key to choose an appropriate screening time.According to published reports, acquired MET amplification confers anti-HER2 resistance in patients with EGFR-or HER2-amplified GC. 30,31 MET overactivation also mediates resistance to anti-FGFR therapies. 32Our present study suggests that patients who are resistant to multiple treatments or with extensive organ metastasis (especially PLC) are populations suitable for targeted screening.
In our study, we also revealed that systemic therapies often lead to focal control of lung lesions.For patients with MET-amplified GC, anti-MET TKIs are remarkably efficient, indicated by target lesion shrinkage and/or rapid symptom relief.A vast number of patients with PLC do not have antitumour opportunities due to a poor physical state (e.g., requiring respiratory support).Nevertheless, once PLC is controlled, patients might regain the opportunity for systemic therapies to diminish tumor burden, therefore improving long-term survival.In that case, timely molecular diagnosis is crucial.Considering the poor physical condition of patients with PLC, reexamination of tumor tissue is impractical.Liquid biopsy (circulating tumor DNA or pleural/peritoneal fluid DNA sequencing) might be an ideal alternative due to its low invasiveness.
We also shed light on the heterogeneity of MET status in GC.MET aberrations could coexist with other driver molecular changes simultaneously, such as FGFR or HER2 amplification.Previous literature has demonstrated inconsistencies in molecular characteristics between primary and metastatic lesions. 33,34Targeted therapies could impose selection pressure on evolving tumors, resulting in disease progression via the outgrowth of MET-nonaddicted constituents.Thus, molecular heterogeneity might account for treatment resistance.To counteract this, circulating tumor DNA sequencing could be used to longitudinally monitor molecular alterations.In our cohort, MET copy numbers were reduced in 5 crizotinib-resistant patients, and a novel RTK target was detected in one patient, which suggests the necessity of tumor rebiopsy to assist in clinical decision making.
Our study has several limitations.First, the small-scale and retrospective nature of this study could be an inherent defect.We acknowledge that it is difficult to interpret the survival data due to the limited number of enrolled patients and the variations in treatments.Further validation is warranted to consolidate the relationship between PLC and MET amplification, although the overall rarity and aggressive nature of GC with PLC make prospective studies challenging.Second, each detection method of MET amplification has its flaws.NGS is less sensitive than FISH, which might affect the detection rate of MET amplification.Third, in our cohort, PLC was diagnosed via CT and clinical manifestations, which might lead to underdiagnosis of very mild cases (subclinical or radiographically invisible cases).

| CONCLUSIONS
Our study indicates that MET amplification could be enriched in gastric cancer accompanied by pulmonary lymphangitis carcinomatosis, and the latter shows a high response rate to anti-MET target therapies.
Figure S5 Displays a representative CT image of PLC from a MET-amplified GC patient.In particular, seven out of 10 patients with MET-amplified GC achieved a PR after anti-MET therapies, among which the median PFS was 6.0 months.Nearly half of the patients (9/20) had stabilized PLCs but developed new lesions (Table

F I G U R E 2
Survival curves of MET-amplified GC patients according to different therapies.T A B L E 2 Clinicopathological and molecular properties of patients with PLC.

T A B L E 3
Treatment and responses of patients with pulmonary lymphangitis carcinomatosis.

F I G U R E 3
Molecular heterogeneity in MET-amplified GC.

F I G U R E 4
Association between PLC and MET gene amplification.

Total cases (n = 58) MET-amplified GC with PLC (n = 13) MET-amplified GC without PLC (n = 45) p-Value
F I G U R E 1 Survival curves of MET-amplified GC patients according to different amplification types.