Background

ICBs are today the backbone of non-oncogene addicted NSCLC, both alone and in combination with chemotherapy (CT-ICB). To date, PD-L1 is the only predictive factor for ICBs benefit. Tumor burden is a new emerging additional biomarker to select pts who may derive a benefit from a CT combination compared to single agent ICBs. The biological correlates of tumor burden are largely unknown.

Methods

In this multicentric study, 18F FDG PET scans were performed ≤ 42 days from 1st line initiation (CT, ICBs or CT-ICB). Total MTV was calculated with a threshold of 42% of SUV max. Progression Free Survival (PFS) and Overall Survival (OS) were analyzed with Kaplan Meyer method and log rank test. On a subset of pts with PET performed within 30 days from biopsy, transcriptomic data on fresh frozen tissue were correlated with MTV of the biopsied lesion.

Results

488 pts were enrolled, 162 treated with ICBs, 232 with CT-ICB and 94 with CT. The 3 groups were comparable in terms of MTV, sex, stage and PS, pts in ICBs are older (69y vs 64 for CT-ICB and 62 for CT). MTV was negatively correlated with the outcome of ICI, with a median PFS of 31.8 months for 1^st quartile (1Q), 13.8 for 2Q, 4.7 for the 3Q and 1.8 for the 4Q (p < 0.001). A weaker correlation was found for CT-ICB, with a median PFS of 14.0m 1Q, 7.1 for 2Q, 7.0 for 3Q and 5.5 for 4Q (any PD-L1 expression), p 0.037, No correlation was found for CT ( p 0.696). When comparing ICI to CT-ICB in patients with PD-L1 ≥ 50%, an advantage in PFS was seen for CT-ICB (10.7 vs 3.0m, p 0.039 ) for pts with MTV > median (100cc) as well as in 3 (73% vs 92%, p 0.043) and 6 months OS (66 vs 89%, p 0.033), but not in those with MTV < median. Transcriptome analysis was performed on 48 pts and showed a negative correlation of MTV with GSEA immune signatures, as well as with RNAseq immune infiltrate quantification (Cybersort, p 0.016).

Conclusions

MTV identifies NSCLC pts with a worse prognosis and an increased rate of early progression and death on ICB. In the subgroups of PD-L1≥50%, pts with high MTV may do better on CT-ICB compared to ICB. Transcriptomic analysis suggests that tumor microenvironment become increasingly immune suppressive with the increase in volume.

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

G. Zalcman: Financial Interests, Personal, Invited Speaker: Bristol Myers Squibb, AstraZeneca, Pfizer, Boehringer Ingelheim; Non-Financial Interests, Personal, Invited Speaker: Roche-France, Bristol Myers Squibb, Takeda. J. Remon Masip: Financial Interests, Personal, Invited Speaker: Roche, Pfizer, MSD, Boehringer Ingelheim; Financial Interests, Personal, Advisory Board: AstraZeneca, BMS, Jansse, Takeda, Sanofi; Financial Interests, Personal, Expert Testimony: OSE Immunotherapeutics; Non-Financial Interests, Personal, Principal Investigator, PI of PECATI trial in Thymic malignancies endorsed by a grant by MSD: MSD; Non-Financial Interests, Personal, Other, Co-PI of APPLE trial (EORTC-1525): AstraZeneca. L. Hendriks: Financial Interests, Institutional, Advisory Board: Amgen, Boehringer Ingelheim, Lilly, Novartis, Pfizer, Takeda, BMS, Merck, Janssen, MSD; Financial Interests, Personal, Other, mentorship with key opinion leaders funded by AstraZeneca: AstraZeneca; Financial Interests, Institutional, Invited Speaker, for educational webinar: AstraZeneca, Lilly; Financial Interests, Institutional, Invited Speaker, educational webinar/interview: Bayer; Financial Interests, Institutional, Invited Speaker, educationals: MSD; Financial Interests, Personal, Invited Speaker, for webinars: Medtalks; Financial Interests, Institutional, Advisory Board, one time also personal: Roche; Financial Interests, Institutional, Other, performing interviews at conference: Roche; Financial Interests, Personal, Other, travel support: Roche; Financial Interests, Institutional, Other, podcast on brain metastases: Takeda; Financial Interests, Personal, Invited Speaker, payment for post ASCO round table discussion: VJOncology; Financial Interests, Personal, Invited Speaker, payment for post ASCO/ESMO/WCLC presentations, educational committee member: Benecke; Financial Interests, Institutional, Invited Speaker, payment for post ESMO/ASCO discussion: high5oncology; Financial Interests, Institutional, Other, educational webinar: Janssen; Financial Interests, Personal, Other, member of the committee that revised these guidelines: Dutch guidelines NSCLC, brain metastases and leptomeningeal metastases; Financial Interests, Institutional, Research Grant, for IIS: Roche, Boehringer Ingelheim, AstraZeneca, Takeda; Financial Interests, Institutional, Invited Speaker: AstraZeneca, GSK, Novartis, Merck Serono, Roche, Takeda, Blueprint Medicines, Mirati, AbbVie, MSD, Gilead; Financial Interests, Institutional, Other, drug support for IIS (contract in negotiation): Beigene; Non-Financial Interests, Personal, Other, Chair metastatic NSCLC for lung cancer group: EORTC; Non-Financial Interests, Personal, Other, secretary NVALT studies foundation: NVALT. C. Helissey: Financial Interests, Personal, Invited Speaker: Janssens, Roche, Astellas, AstraZeneca, Sanofi; Non-Financial Interests, Personal, Principal Investigator: Janssen, Sanofi, Roche, Astellas. I. Monnet: Other, Personal, Other, invitation to virtual ASCO 2021 and 2022: Pfizer; Other, Personal, Other, invitation to ESMO congress 2021 and 2022: Takeda. D. Planchard: Financial Interests, Personal, Advisory Board: AstraZeneca, BMS, Merck, Novartis, Pfizer, Roche, Samsung, Celgene, AbbVie, Daiichi Sankyo, Janssen; Financial Interests, Personal, Invited Speaker: AstraZeneca, Novartis, Pfizer, priME Oncology, Peer CME, Samsung, AbbVie, Janssen; Non-Financial Interests, Personal, Principal Investigator, Institutional financial interests: AstraZeneca, BMS, Merck, Novartis, Pfizer, Roche, Daiichi Sankyo, Sanofi-Aventis, Pierre Fabre; Non-Financial Interests, Personal, Principal Investigator: AbbVie, Sanofi, Janssen. F. Barlesi: Financial Interests, Personal, Advisory Board: AstraZeneca, Bayer, Bristol Myers Squibb, Boehringer Ingelheim, Eli Lilly Oncology, F. Hoffmann–La Roche Ltd, Novartis, Merck, Mirati, MSD, Pierre Fabre, Pfizer, Sanofi Aventis, Seattle Genetics, Takeda; Non-Financial Interests, Personal, Principal Investigator: AstraZeneca, BMS, Merck, Pierre Fabre, F. Hoffmann-La Roche Ltd. B. Besse: Financial Interests, Institutional, Funding: 4D Pharma, AbbVie, Amgen, Aptitude Health, AstraZeneca, BeiGene, Blueprint Medicines, Boehringer Ingelheim, Celgene, Cergentis, Cristal Therapeutics, Daiichi Sankyo, Eli Lilly, GSK, Janssen, Onxeo, OSE Immunotherapeutics, Pfizer, Roche-Genentech, Sanofi, Takeda, Tolero Pharmaceuticals; Financial Interests, Institutional, Research Grant: Genzyme Corporation, Chugai pharmaceutical, Eisai, Inivata, Ipsen, Turning Point Therapeutics. All other authors have declared no conflicts of interest.

Background

Adoptive cell therapy (ACT) using tumor-infiltrating lymphocytes (TILs) has proven efficacious in metastatic melanoma (Mel). Nevertheless, long-term clinical efficacy remains unsatisfactory for the majority of patients. Therefore, understanding TILs and the orchestrated tumor-microenvironment (TME) states associated with clinical response is of paramount importance.

Methods

We report the translational analysis of a single-center phase I study to assess feasibility and efficacy of TIL-ACT in Mel patients (NCT03475134). We performed comprehensive translational studies on patients` tumor longitudinal samples including multispectral immuno-fluorescence (mIF) imaging, bulk RNA-sequencing and single-cell RNA-sequencing before and after TILs infusion (30 days).

Results

Thirteen patients successfully completed TIL-ACT therapy and were included in the analysis. Best overall response was 46.1% (6/13) at 3 months, including two patients with ongoing complete response 3 years after infusion. Responders (Rs) exhibited immunogenic tumor cell states with higher inferred CNVs and overexpressed DNA sensing/IFN and class I antigen presentation-related genes. Multiplex IF imaging revealed that Rs had increased densities of polyfunctional intra-epithelial (ie)CD8⁺ T cells marked by higher PD-1 expression at baseline, and these cells persisted 30 days after TILs infusion. Single-cell transcriptomic analyses confirmed higher cytotoxic and exhaustion CD8 T cell states at baseline. PDCD1 (PD-1), CXCL13, TNFRSF9 (CD137), GZMB, HAVCR2 (TIM3) and PRF1 genes were overexpressed in CD8⁺ TILs and associated to clinical responses. Cell-cell interaction predictions corroborated by spatial neighborhood analyses revealed that tumors of Rs were highly enriched for ie-and stromal T-cell networks involving activated myeloid and costimulatory B cells. Successful TIL-ACT therapy reprogrammed the myeloid compartment and further increased CD8 TIL-myeloid cell networks.

Conclusions

Our systematic target discovery study reveals CD8 T-cell network-based biomarkers that could improve patient selection and guide the design of ACT clinical trials.

Clinical trial identification

NCT03475134.

Legal entity responsible for the study

Centre Hospitalier Universitaire Vaudois (CHUV), Ludwig Institute for Cancer Research.

Funding

Centre Hospitalier Universitaire Vaudois, Lausanne Branch of the Ludwig Insitute for Cancer Research.

Disclosure

O.A. Michielin: Financial Interests, Institutional, Invited Speaker: BMS, Amgen, Pierre Fabre, Roche, BMS; Financial Interests, Institutional, Advisory Board: BMS, Amgen, Roche, Novartis, Pierre Fabre, MSD; Financial Interests, Personal, Other, Advisory Role: BMS, MSD, Roche, Novartis, Pierre Fabre, Amgen, GSL; Financial Interests, Institutional, Research Grant: BMS, MSD, Pierre Fabre, Amgen, Merck; Financial Interests, Institutional, Funding: BMS, MSD, Pierre Fabre, Amgen, MSD; Non-Financial Interests, Personal, Principal Investigator: BMS, MSD, Amgen, Roche, Pierre Fabre, Novartis. G. Coukos: Financial Interests, Personal, Invited Speaker: CHUV. All other authors have declared no conflicts of interest.

Background

The PRADO trial (n=99) tested different surgical and adjuvant therapy strategies based on the pathologic response after neoadjuvant IPI 1mg/kg and NIVO 3mg/kg in stage III melanoma patients (pts). The pathologic response rate (pRR: ≤50% viable tumor) was 72%, including 61% major pathologic responses (MPR: ≤10% viable tumor). After a median follow-up of 28.1 months, the 2-year (2y) event-free survival (EFS) rate was 80%. Here, we report the response and EFS data of PRADO according to the IFN-γ signature and TMB.

Methods

TMB and the IFN-γ gene expression signature (GES) were examined in baseline tumor biopsies by whole exome sequencing (n=76) and mRNA sequencing (n=80). Association with pRR, MPR or EFS was examined by logistic/Cox regression analyses. Cutoffs between high (H) and low (L) were calculated using ROC curves.

Results

The table shows the association between IFN-γ GES and TMB with pRR, MPR, and EFS. IFN-γ GES and logTMB were not correlated (R = 0.065; p = 0.579). Pts with a high IFN-γ GES had a higher pRR and MPR rate than pts with a low IFN-γ GES (87% vs 50%, p=0.001 and 77% vs 35%, p<0.001, respectively), and also a higher 2y EFS rate (87% vs 68%, log-rank p=0.015). Pts with a high TMB were more likely to achieve a pathologic response (pRR 83% vs 58%, p=0.024) and MPR (80% vs 38%, p<0.001), but did not have a lower risk of relapse (2y EFS 77% vs 76%, log-rank p=0.531). When combined, the pRR for pts with IFN-γ H/TMB H was 90%, IFN-γ H/TMB L 79%, IFN-γ L/TMB H 67% and IFN-γ L/TMB L 42%. For MPR rates this was 85%, 63%, 67% and 19%, respectively. Table: 6P

Conclusions

Similar to findings in our previous trials, baseline IFN-γ GES and TMB were independent biomarkers for pathologic response and MPR, and the IFN-γ GES was associated with EFS. However, TMB was not associated with risk of relapse, possibly owing to the different response-driven surgical and adjuvant therapy strategies in PRADO.

Clinical trial identification

NCT02977052.

Legal entity responsible for the study

Netherlands Cancer Institute.

Funding

BMS.

Disclosure

R. Saw: Financial Interests, Personal, Invited Speaker: Bristol Myers Squib, Novartis; Financial Interests, Personal, Advisory Board: Novartis, MSD, Qbiotics; Financial Interests, Personal, Other, On Faculty, support of University of Sydney salary: Melanoma Institute Australia. E. Kapiteijn: Financial Interests, Institutional, Advisory Board: BMS, Novartis, Pierre Fabre, Merck, Delcath, Bayer; Financial Interests, Institutional, Invited Speaker: BMS. A.A.M. Van der Veldt: Financial Interests, Institutional, Advisory Board: BMS, MSD, Merck, Roche, Eisai, Pfizer, Sanofi, Novartis, Pierre Fabre, Ipsen. K. Suijkerbuijk: Financial Interests, Institutional, Advisory Board: Novartis, BMS, AbbVie, Pierre Fabre, MSD; Financial Interests, Institutional, Invited Speaker: Roche; Financial Interests, Institutional, Research Grant: Novartis, TigaTx. G. Hospers: Financial Interests, Institutional, Advisory Board: Amgen, Roche, MSD, BMS, Novartis, Pierre Fabre, Pfizer; Financial Interests, Institutional, Research Grant: BMS, Seerave. W. Van Houdt: Financial Interests, Institutional, Invited Speaker: Amgen, BMS; Financial Interests, Institutional, Advisory Board: Belpharma; Financial Interests, Institutional, Expert Testimony: Sanofi, MSD; Financial Interests, Personal, Other, travel grant: Novartis. R. Scolyer: Financial Interests, Institutional, Advisory Board: F. Hoffmann-La Roche, Evaxion, Provectus Biopharmaceuticals Australia, QBiotics, Novartis, MSD, NeraCare, Amgen, BMS, Myriad Genetics, GlaxoSmithKline. A.M. Menzies: Financial Interests, Personal, Advisory Board, advisory board: BMS, MSD, Novartis, Roche, Pierre-Fabre, QBiotics. A.C.J. van Akkooi: Financial Interests, Institutional, Advisory Board: Amgen, Bristol Myers Squibb, Novartis, MSD - Merck, Merck-Pfizer, Pierre Fabre, Sanofi, Sirius Medical, 4SC, Provectus; Financial Interests, Institutional, Research Grant, NIVEC study: Amgen; Financial Interests, Institutional, Research Grant: Merck-Pfizer. G.V. Long: Financial Interests, Personal, Other, Consultant Advisor: Agenus Inc, Amgen Inc, Array Biopharma Inc., Boehringer Ingelheim International GmbH, Bristol Myers Squibb, Evaxion Biotech A/S, Hexal AG (Sandoz Company), Merck Sharpe & Dohme (Australia) Pty Limited, Novartis Pharma AG, OncoSec Medical Australia, Pierre Fabre, Provectus Australia, Qbiotics Group Limited, Regeneron Pharmaceuticals Inc; Financial Interests, Personal, Advisory Board, Consultant Advisor: Highlight Therapeutics S.L. C.U. Blank: Financial Interests, Institutional, Advisory Board: BMS, MSD, Roche, Novartis, GSK, AZ, Pfizer, Lilly, GenMab, Pierre Fabre; Financial Interests, Personal, Expert Testimony: Third Rock Ventures; Financial Interests, Personal, Stocks/Shares: Immagene; Financial Interests, Personal, Stocks/Shares, intention to develop IFN signature algorithm: NewCo, no name yet; Financial Interests, Institutional, Invited Speaker: BMS, Novartis, NanoString, 4SC; Other, Personal, Other, pending patent: WO 2021/177822 A1. All other authors have declared no conflicts of interest.

Background

Tumour-specific antigens are one of the principal targets in immune checkpoint inhibitor (CPI) treatment and fundamental to the development of personalised immunotherapy. The search for immunogenic neoantigens has primarily focused on mutations in protein-coding regions of genes. In this study, we investigate how novel open-reading frames (neoORFs) in the 5’ and 3’ untranslated region (UTR) of genes (generated by premature start-gain and stop-loss mutations respectively) contribute to the immune landscape of cancer.

Methods

We analysed two pan-cancer CPI-treated patient cohorts from the Hartwig Medical Foundation (HMF, n = 384), and Genomics England (GEL, n = 364). The frequency and length of UTR neoORFs for each patient tumour were predicted from whole genome sequencing data using PrimeCUTR, an R-based bioinformatics tool we developed. CPI treatment response was determined based on RECIST 1.1 criteria.

Results

Start-gain neoORFs occurred in 65.4–75.5% (GEL–HMF) of patients, while stop-loss neoORFs were found in 18.6–30.5%. 53.2–65.6% had start-gain neoORFs spanning 20 amino acids or more, with an overall median length of 28 amino acids. Among prevalent cancer groups, lung and head and neck cancer had the longest start-gain neoORFs. CPI response in HMF was associated with longer total length of start-gain (but not frameshift or stop-loss) neoORFs (P = 0.005). This association was also seen subsetting for low tumour mutational burden (TMB) (P = 0.014). In GEL, survival benefit was specifically seen in melanoma (n = 155), where patients with start-gain neoORFs longer than 28 amino acids (n = 60) had greater OS compared to those without (HR 0.59, 95% CI 0.36–0.95, P = 0.03).

Conclusions

UTR neoORFs occur frequently in cancer, yielding a promising source of neoantigens. We designed a computational pipeline for identifying these neoORFs, which will be made available upon peer-review. We show that 5’ UTR start-gain neoORFs were associated with clinical response to CPI, even in the low TMB setting which is typically linked to reduced CPI response. These findings warrant further study to validate UTR neoORFs as a biomarker and target for personalised immunotherapy.

Legal entity responsible for the study

The authors.

Funding

National Institute for Health and Care Research.

Disclosure

K.R. Litchfield: Financial Interests, Personal, Invited Speaker: Roche Tissue Diagnostics; Financial Interests, Personal, Other, Consulting work: Monopteros Therapeutics; Financial Interests, Institutional, Research Grant: Ono/LifeArc; Financial Interests, Institutional, Research Grant, Research funding: Genesis Therapeutics; Non-Financial Interests, Institutional, Proprietary Information, Collaboration on data analysis: BMS All other authors have declared no conflicts of interest.

Background

ACT with TIL is personalized immunotherapy with promising results in MM. Recently, the positive outcome of a multicenter (National Center for Cancer Immune Therapy [Herlev, DK] and the Netherlands Cancer Institute [Amsterdam, NL]) phase 3 trial evaluating TIL in MM patients (NCT02278887) was recently announced. In this trial, the objective response (OR) rate for TIL-treated patients was 49% with 20% obtaining a complete response according to RECIST 1.1. However, there are currently no validated biomarkers that correlate with response to TIL. Identifying cellular subsets associated with response to treatment could in the future help guide clinicians in treating MM patients with TIL.

Methods

In this study, we analyzed the cryopreserved samples of the TIL infusion product from all patients with unresectable stage IIIC-IV MM treated with TIL in the phase 3 trial using multiparametric flow cytometry. The phenotypic characteristics of the TIL infusion product and OR to therapy according to RECIST 1.1 were investigated. The Mann-Whitney U test was used to test unpaired observations for statistical significance.

Results

In total, 80 patients received TIL, with 39 responders (complete + partial response) and 39 non-responders (stable + progressive disease). For 2 patients, the response was unevaluable. A higher number of TILs infused (Table) was significantly associated with response to therapy (p=0.013). Several lymphocyte subsets, including a higher number of CD8+ (p=0.004), CD8+CD28+ (p<0.001), CD8+BTLA+ (p=0.004), and CD8+CD39+ cells (p=0.007) favored response to therapy. Table: 8P

Absolute number of cells infused (x10⁹)

Conclusions

In this study, we present an explorative immune characterization of the melanoma TIL infusion product and identify several cellular subsets that correlate with response to TIL therapy.

Legal entity responsible for the study

National Center for Cancer Immune Therapy (Herlev, DK) and the Netherlands Cancer Institute (Amsterdam, NL).

Funding

The trial was supported by the Dutch Cancer Society, the Netherlands Organization for Health Research and Development, the Dutch Ministry of Health, Stichting Avento, Copenhagen University Hospital, Herlev, the Danish Cancer Society and Capital Region of Denmark Research Foundation.

Disclosure

T. Holz Borch: Financial Interests, Personal, Other, Honorary lectoures: Bristol Meyer Squibb. M. Donia: Financial Interests, Personal, Invited Speaker, Teaching: Novartis, Roche; Financial Interests, Personal, Other, Advisor: Achilles Therapeutics; Non-Financial Interests, Personal, Other, Sub-investigator of clinical trial with connected translational research: Bristol Myers Squibb; Financial Interests, Personal, Proprietary Information, Proprietary Data Access: Bristol Myers Squibb; Other, Institutional, Other, Chariman of the Melanoma and non-melanoma Skin Cancer Scientific Committee: Danish Medicines Council (Medicinrådet). J.B.A.G. Haanen: Financial Interests, Institutional, Advisory Board: Bristol Myers Squibb, Achilles Therapeutics, Ipsen, Merck Sharpe & Dohme, Merck Serono, Pfizer, Molecular Partners, Novartis, Roche, Sanofi, Third Rock Venture, Iovance Biotherapeutics; Financial Interests, Institutional, Advisory Board, SAB member: BioNTech, Immunocore, Gadeta, Instil Bio, PokeAcel, T-Knife; Financial Interests, Personal, Advisory Board, SAB member: Neogene Therapeutics, Scenic; Financial Interests, Personal, Stocks/Shares: Neogene Therapeutics; Financial Interests, Institutional, Research Grant: Bristol Myers Squibb, BioNTech US, Merck Sharpe & Dohme, Amgen, Novartis, Asher Bio; Non-Financial Interests, Personal, Member: ASCO, AACR, SITC; Other, Personal, Other, Editorial Board ESMO Open: ESMO; Other, Personal, Other, Editor-in-Chief IOTECH: ESMO; Other, Personal, Other, Editorial Board: Kidney Cancer. I. Svane: Financial Interests, Personal, Advisory Board: BMS, Pierre Fabre, Novartis; Financial Interests, Personal, Invited Speaker: MSD, Pierre Fabre, Novartis, Roche, BMS, MSD; Financial Interests, Personal, Stocks/Shares, Cofounder and Founder warrents: IO Biotech; Financial Interests, Institutional, Research Grant: Adaptimmune, Enara Bio, Lytix Biopharma, TILT Biotherapeutics; Financial Interests, Institutional, Funding: Evaxion; Non-Financial Interests, Personal, Principal Investigator: BMS, Roche, TILT Biotherapeutics, Lytix Biopharma, Novartis. All other authors have declared no conflicts of interest.

Background

ctDNA kinetics in the first 6 weeks and immune-related gene expression signatures can predict outcomes in patients (pts) treated with anti-PD-1 therapy. The earliest predictive timepoint in ctDNA kinetics for benefit to this therapy is unknown.

Methods

Pts with recurrent/metastatic head and neck squamous cell carcinoma treated with pembrolizumab or nivolumab monotherapy were prospectively enrolled. Plasma was collected at baseline (B), 2, 3, 8, 15, 22 and 29 days after first dose. Whole exome sequencing (WES) and RNA-seq were performed in archival tumor. ctDNA was analyzed using a bespoke 16 gene panel based on matched WES (Signatera^TM). Differential Gene Expression Analysis was performed with DESeq2. Gene Set Enrichment Analysis included different transcription factors (TF), molecular/immune pathways and PREDICT-IO signature (PMID 36055464).

Results

A total of 104 plasma samples from 15 pts were collected. ctDNA was detected in 90/104 (87%). Clinical Benefit (CB) rate, defined as complete (CR), partial response (PR) or stable disease (SD) > 6 months by RECIST 1.1, was 47% (1 CR, 3 PR, 3 SD). There were no differences in ctDNA quantity at B according to local/distant disease, primary site or PD-L1 status. A decrease in ctDNA at day 8 from B was associated with CB (100 vs 14%, p<0.01) and with longer progression-free survival (PFS) adjusted by Holm-correction for multiple comparisons (HR: 22.9 95%CI 2.5-211.3, p=0.02) compared to a stability/increase. This association was also seen at day 15, 22 and 29 but not at day 2 or 3. A steady decline in ctDNA beyond day 8 (↓ΔctDNA) was observed in 4 pts (3 PR, 1 SD). Immune-related pathways and TF were significantly enriched in CR/PR and ↓ΔctDNA. Oncogenic pathways and TF were enriched in pts with increase/stability beyond day 8 (↑ΔctDNA). PREDICT-IO High was associated with longer PFS (HR: 7.7 95%CI 1.5-38, p=0.005). All PREDICT-IO Low pts were ↑ΔctDNA (N=7). Within PREDICT-IO High pts, ↓ΔctDNA showed a trend to longer PFS than ↑ΔctDNA (HR: 6.9 95%CI 0.95-63, p=0.051).

Conclusions

A decrease in ctDNA by day 8 of anti-PD1 therapy is associated with CB and longer PFS. ctDNA kinetics can improve predictive value of PREDICT-IO High signature. Validations in larger pan-cancer cohorts are needed.

Clinical trial identification

NCT04606940.

Legal entity responsible for the study

Princess Margaret Cancer Centre.

Funding

Natera.

Disclosure

G. Laliotis: Financial Interests, Personal, Full or part-time Employment: Natera. A. Spreafico: Financial Interests, Personal, Advisory Board: Merck, Bristol Myers Squibb, Oncorus, Jansen, Medison and Immunocore; Financial Interests, Institutional, Principal Investigator, Clinical trial: Novartis; Financial Interests, Institutional, Principal Investigator: Bristol Myers Squibb, Symphogen, AstraZeneca/MedImmune, Merck, Bayer, Surface Oncology, Northern Biologics, Janssen Oncology/Johnson and Johnson, Roche, Regeneron, Alkermes, Array Biopharma, GSK, Oncorus, Treadwell, Amgen. A.R. Hansen: Financial Interests, Personal, Advisory Board: GSK, Merck, Eisai; Financial Interests, Institutional, Principal Investigator: GSK, Merck, Pfizer, MedImmune/Genetech, Roche, Janssen, BMS, AstraZeneca, Astellas, Boehringer, Bayer. S. Willingham: Financial Interests, Personal, Full or part-time Employment: Natera. M. Liu: Financial Interests, Personal, Full or part-time Employment: Natera. T. Pugh: Financial Interests, Personal, Advisory Board: Illumina, Merck, Chrysalis Biomedical Advisors, Canadian Pension Plan Investment Board; Financial Interests, Institutional, Research Grant: Roche/Genentech. S. Bratman: Financial Interests, Personal, Leadership Role: Adela; Financial Interests, Personal, Stocks/Shares: Adela; Financial Interests, Personal, Royalties: Roche. L.L. Siu: Financial Interests, Personal, Advisory Board: Merck, AstraZeneca, Roche, Oncorus, Seattle Genetics, Voronoi, Arvinas, Tessa, Navire, Relay Therapeutics, Janpix, Amgen, Marengo, InterRNA, Medicenna, Hoopika, Coherus; Financial Interests, Personal, Other, Spouse is co-founder: Treadwell Therapeutics; Financial Interests, Personal, Stocks/Shares, Spouse has stock ownership: Agios; Financial Interests, Institutional, Invited Speaker: Novartis, Bristol Myers Squibb, Pfizer, Boerhinger-Ingelheim, GlaxoSmithKline, Roche/Genentech, AstraZeneca, Merck, Astellas, Bayer, Amgen, Symphogen, Intensity Therapeutics, Shattucks, EMD Serono; Non-Financial Interests, Personal, Advisory Role: ICR. All other authors have declared no conflicts of interest.

Background

Tertiary lymphoid structures (TLS) are ectopic lymphoid formations which are composed predominantly of B cells, T cells and dendritic cells. The presence in the tumor compartment of mature TLS - characterized by mature follicles containing germinal centers - has been strongly associated with improved survival upon cancer immunotherapies for patients with solid tumors. For this reason, TLS could be used as a predictive factor biomarker to identify the patients more likely to benefit from immune checkpoint inhibitors. However, the pathological assessment of the TLS status remains time-consuming and usually requires additional analysis including CD3/CD20 staining. This study aims to show that artificial intelligence techniques applied to digital pathology could offer a fast and cheap mass patient screening solution to assess TLS from Hematoxylin/Eosin (HE) digital pathology slides used in clinical workflow.

Methods

We analyzed a pan-cancer cohort of 289 HE-stained Whole Slide Images (WSI) from Institut Bergonié (43% NSCLC, 12% sarcoma, 45% other solid tumors) on which TLS were manually contoured by expert pathologists. A Deep Learning (DL) model was trained on WSI to predict TLS status at the patient level (presence or absence of TLS). Models were evaluated using five-fold cross validation. The best performing architecture included two main components: a predicted score of TLS presence on small areas (112x112 micrometers) of the WSI followed by an aggregation at the patient level. To assess the transferability of the model, a validation cohort (PEMBROSARC) of 236 sarcoma WSI was used.

Results

The best performing model achieved a ROC AUC score of 0.917 (std of 0.036) in five-fold cross validation on the pan-cancer cohort (specificity of 68% for a sensitivity of 90%). This model transferred very well on the PEMBROSARC study, the first clinical trial implementing TLS status as an inclusion criteria (Italiano et al Nature Medicine 2022) with a ROC AUC score of 0.89 (specificity of 64% for a sensitivity of 90%).

Conclusions

Our study demonstrates the predictive power of DL applied to predict the patient's TLS status from HE images. This model could be implemented in pathology labs as an efficient pre-screening tool.

Legal entity responsible for the study

Institut Bergonie.

Funding

Owkin, Inc.

Disclosure

L. Gillet, J. Le Douget, B. Schmauch, C. Maussion: Financial Interests, Institutional, Full or part-time Employment: Owkin, Inc. All other authors have declared no conflicts of interest.

Background

V-domain Ig suppressor of T cell activation, VISTA, is a type I transmembrane protein that functions as an immune checkpoint. VISTA can be expressed by a wide range of cell types. Malignant pleural mesothelioma (MPM) is a tumor type with the highest levels of VISTA expression. However, it remains unclear if VISTA could be a predictive or prognostic marker in MPM following PD-1 blockade.

Methods

The phase III PROMISE MESO study randomized 144 patients to either chemotherapy (vinorelbine or gemcitabine) or pembrolizumab. We analyzed 62 patients from the pembrolizumab arm and 57 patients from the chemotherapy arm with available tumour tissue. We performed multiplex IHC for the following markers/cell types: calretinin of WT1, CD8, CD14, CD66b, CD11c, VISTA, and DAPI.

Results

Most of the VISTA expression originated from the mesothelioma cells. Bioinformatic analysis of MPM RNAseq data showed enrichment for response to INFa and INFg with VISTA overexpression. Responses to pembro or chemo were not influenced by global VISTA expression. However, VISTA expression on CD66b-positive neutrophils negatively correlated with responses. The VISTA+/CD66b+ median value was 0.56 (IQR: 0.0-3.5), and the rounded median value of 1.0/mm² was used as threshold. Patients with VISTA+/CD66b+ high had an ORR of 5.6% whereas VISTA+/CD66b+ low tumors had 36.8%. Neutrophil infiltration alone did not correlate with responses. The difference in ORR also translated to prolonged PFS in patients with low VISTA+/CD66b+ cell counts (interaction p=0.0051), but no difference in overall survival was detected (interaction p=0.35). Neutrophils were not predictive of PFS or OS. In public single-cell expression data of lung cancer patients, VISTA expression indeed was higher in a subset of neutrophils, specifically in tN1 and tN3 subtypes.

Conclusions

Overall, our data show that VISTA expressing neutrophils correlate with outcomes. Our data underscore the complex biology of VISTA and suggest that specific targeting of VISTA on neutrophils might be a viable therapeutic option in a subset of MPM patients.

Clinical trial identification

NCT02991482. ETOP 9-15 PROMISE-meso trial.

Legal entity responsible for the study

ETOP IBCSG Partners Foundation.

Funding

ETOP IBCSG Partners Foundation, and by Merck Sharpe & Dohme.

Disclosure

All authors have declared no conflicts of interest.

Background

DNA hypomethylating agents (DHA) combined with immune checkpoint inhibitors (ICI) are a promising strategy to improve the immunomodulatory and antitumor activity of ICI-based treatment. Providing initial support to this notion, guadecitabine plus ipilimumab in the phase Ib NIBIT-M4 trial, proved to be feasible, safe, and tolerable, with clinical and biological activity in metastatic melanoma patients (pts) (Di Giacomo AM, CCR 2019).

Methods

Pts enrolled in the NIBIT-M4 trial were assessed for long-term clinical outcomes. Molecular profiles of serial tumor biopsies performed at week 0, 4, and 12, including RNA, Whole Exome and Genome-wide methylation sequencing, were analyzed in an integrative manner and correlated with clinical data; pts were classified as Responder (R) or Non-responder (NR) based on disease control.

Results

Among the 19 pts enrolled (minimum follow-up 45 months), the 5-y OS rate was 29%; median duration of response was 20.6 months (95% CI,12.4-28.8). Analysis of longitudinal biopsies from 14 pts showed that primary/acquired resistance is underpinned by multi-level molecular mechanisms. Indeed, NR pts had higher mutation frequency in NRAS and harbored somatic alterations in genes involved in the Epithelial to Mesenchymal Transition (EMT) and chromatin organization pathways compared to R pts. Moreover, NR pts had a heterogeneous expression phenotype, characterized by activated patterns of proliferation/differentiation/EMT and by deactivation of INF-ɣ/HLA-II/immune-related genes. Conversely, R pts displayed dynamic patterns of activated immune signatures and favorable tumor-infiltrating microenvironment with T/NK cell clonal activation progressively increasing with therapy. We also developed a DNA-RNA immune stratification, simultaneously taking into account both the composition of the tumor microenvironment and the presence of reactive T-cells measured as the ratio between observed and predicted tumor neoantigens.

Conclusions

Our comprehensive longitudinal multi-omics analyses of tumors from the NIBIT-M4 study identify a landscape of biomarkers predictive of response that may guide patient stratification and selection in ICI plus DHA therapies.

Clinical trial identification

EudraCT 2015-001329-17, NCT02608437.

Legal entity responsible for the study

NIBIT Foundation Onlus.

Funding

Fondazione AIRC under 5 per Mille 2018 – ID 21073 program (principal investigator M. Maio).

Disclosure

A.M. Di Giacomo: Financial Interests, Personal, Advisory Board: Bristol Myers Squibb, MSD, Pierre Fabre, Novartis; Financial Interests, Personal, Invited Speaker: Sanofi. M. Maio: Financial Interests, Personal, Advisory Board: MSD, BMS, AstraZeneca, Incyte, Amgen, Pierre Fabre; Financial Interests, Personal, Invited Speaker: Roche, Eli Lilly; Financial Interests, Personal, Ownership Interest: Epigen Therapeutics, Theravance. All other authors have declared no conflicts of interest.

Background

Reactivation of tumor-infiltrating T lymphocytes (TILs) with immune checkpoint inhibitors or co-stimulators has proven to be an effective anti-cancer strategy for a broad range of malignancies. However, epithelial ovarian cancer (EOC) remains largely refractory to current T cell-targeting immunotherapeutics. Therefore, identification of novel immune checkpoint targets and biomarkers with prognostic value for EOC is warranted. We here identified a TIM-3/CXCL13-positive tissue-resident memory (CD8/CD103-positive) T cell (Trm) population in EOC. Analysis of a cohort of ∼175 patients with high-grade serous EOC revealed TIM-3-positive Trm were significantly associated with improved patient survival.

Methods

- Single cell mRNA sequencing data analysis. - Immunohistochemical staining, image acquisition and analysis. Tumor Infiltrated Lymphocyte flow cytometric analysis.

Results

A cohort of EOC core samples was evaluated for the presence of tumor infiltrating CD8/CD103/TIM-3triple-positive T cells and subsequently correlated with patient survival. Interestingly, increased tumor infiltration of CD8/CD103/TIM-3triple-positive cells associated with improved patient survival in EOC, suggesting that CD8/CD103/TIM-3triple-positive TILs can serve as a prognostic marker for EOC. In line with this finding, a single-cell tumor immune transcriptomic dataset revealed upregulation of TIM-3, CXCL13 and CD103 within the terminally exhausted CD8-positive T cell fraction. Confirmatory flowcytometric evaluation of CXCL13 expression on isolated EOC TILs revealed that CXCL13 was predominantly found within the CD8/CD103/TIM-3triple-positive fraction compared to it’s single- and double-positive counterparts.

Conclusions

TIM-3 itself or as surrogate marker for prognostically favorable CXCL13-positive CD8-positive TILs may have prognostic value. As CXCL13-positive CD8-positive T cells have been strongly linked to patient response to anti-PD1 immune checkpoint blockade, combinatorial TIM-3 and PD-1 blockade therapy may be of interest for the (re)activation of anti-cancer immunity in EOC.

Legal entity responsible for the study

The author.

Funding

European Union H2020 Programme.

Disclosure

The author has declared no conflicts of interest.

Background

Immunotherapy has become the standard of care for metastatic non-small cell lung cancer (mNSCLC) without a targetable driver alteration, yet we still lack insight into which patients (pts) will benefit from such treatments. To that end, we investigated characteristics of the immune infiltrate in the tumor microenvironment in relation to immunotherapy response. We report the results of an automated deep learning approach applied to digital H&E whole slide images (WSIs) of pre-treatment biopsies from the PEMBRO-RT clinical trial.

Methods

61 quality-checked H&E WSIs were processed with 3 deep learning algorithms. We extracted a tissue mask using an existing method (Bándi et al., 2019), and detected tumor and immune cells using HoVerNet (Graham et al., 2019). Tumor clusters were identified by combining the output of HoVerNet and tumor segmentation from an nnUnet (Isensee et al., 2021) model that we trained on external NSCLC images. From the output of this pipeline, we extracted immune infiltrate-based density metrics, calculated over all tissue (_allINF), stroma within 500um from the tumor border (_sINF), tumor region (_tINF), and the combination of stroma and tumor (_t+sINF). All metrics were used in ROC analysis after dichotomizing pts as responders and non-responders (response was defined as complete or partial response at any time point or stable disease for ≥12 weeks according to RECIST 1.1 measurement). Differences in metric distributions between the two groups were tested with a two-sided Welch t-test. Kaplan-Meier (KM) analysis was performed on progression-free survival (5-year follow-up).

Results

Our automated analysis reported denser immune infiltrates in responders, although not statistically significant (0.05<p≤0.2). All immune infiltrate metrics showed some predictive value with AUCs > 0.63, where _tINF reported an AUC of 0.70. KM analysis showed p=0.07 if pts were stratified based on the median _tINF, and p=0.02 if stratified based on the optimal operating point of its ROC curve.

Conclusions

Deep learning models that analyze the immune infiltrate density on H&E WSIs can identify mNSCLC responders to pembrolizumab.

Clinical trial identification

NCT02492568.

Legal entity responsible for the study

Netherlands Cancer Institute.

Funding

Dutch Research Council (NWO).

Disclosure

M. van den Heuvel: Research funding paid to their institution by AstraZeneca, Bristol Myers Squibb, Janssen, Merck, Novartis, Stichting Treatmeds, Merck Sharp & Dohme, Pamgene, Pfizer, Roche and Roche Diagnostics; Advisory Board fees paid to their institution by Bristol Myers Squibb, Pfizer, Merck, Merck Sharp & Dohme, AstraZeneca, Novartis and Roche-Genentech, all in the 36 months prior to manuscript submission; Member of the Dutch Association of Physicians in Chest Medicine and Tuberculosis (NVALT) and the Dutch Society for Medical Oncology (NVMO). W. Theelen: Financial Interests, Institutional, Research Grant: AstraZeneca, MSD, Sanofi. F. Ciompi: Financial Interests, Personal, Stocks/Shares: Aiosyn BV; Financial Interests, Personal, Advisory Board: TRIBVN Healthcare. All other authors have declared no conflicts of interest.

Background

There is a lack of effective predictor for pCR of neoadjuvant immunotherapy in esophageal squamous cell carcinoma (ESCC) patients. High-throughput single-cell RNA sequencing (scRNAseq) can define the tumor microenvironment (TME), tumor heterogeneity and immune microenvironment. In this study, we sought to explore the immune pathways related to treatment response of immunotherapy by scRNAseq.

Methods

This was a single-arm phase II clinical trial, planned to enroll 20 locally advanced ESCC patients (cT2-4N0-1M0). Patients received neoadjuvant chemotherapy (paclitaxel-albumin + carboplatin) combined with tislelizumab and followed by minimaly invasive esophagectomy (MIE). We conducted a scRNAseq using fresh tumor tissues taken by endoscopy before neoadjuvant therapy. Histological outcome was divided into pCR (no residual tumor cells in tumor and lymph nodes), MPR (<10% residual tumor – excludes pCR) and non-MPR (>10% residual tumor).

Results

17 patients were enrolled since March,2022. 7 of 17 patients finished neoadjuvant therapy and MIE surgery. Among these 7 patients, 3 achieved pCR, 1 got MPR, and 3 were non-MPR, with a total of 42808 single cells sequenced. Compared to non-MPR, there were significantly increase in T lymphocytes (CD8+, T helper cell, Regulatory T cell) and mature dendritic cells, and significantly decrease in B cells and neutrophils in pCR patients. Morever, pCR patients showed greatly increase on the expression of PD-L1(CD274) in mature dendritic cells.

Conclusions

This study identified novel immunological predictor in baseline tissue for ESCC neoadjuvant immunotherapy. The enrichment of T lymphocytes (CD8+, T helper cell, Regulatory T cell) and mature dendritic cells together with the increase expression of PD-L1(CD274) in mature dendritic cells suggest a promising role of predicting completed pathological response (pCR) induced by neoadjuvant chemotherapy and immunotherapy in locally advanced ESCC patients.

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.