WO2019012147A1 - Outil d'imagerie basée sur la radiomique pour surveiller l'infiltration et de lymphocytes tumoraux et le résultat chez des patients cancéreux traités par agents anti-pd-1/pd-l1 - Google Patents

Outil d'imagerie basée sur la radiomique pour surveiller l'infiltration et de lymphocytes tumoraux et le résultat chez des patients cancéreux traités par agents anti-pd-1/pd-l1 Download PDF

Info

Publication number
WO2019012147A1
WO2019012147A1 PCT/EP2018/069169 EP2018069169W WO2019012147A1 WO 2019012147 A1 WO2019012147 A1 WO 2019012147A1 EP 2018069169 W EP2018069169 W EP 2018069169W WO 2019012147 A1 WO2019012147 A1 WO 2019012147A1
Authority
WO
WIPO (PCT)
Prior art keywords
cancer
tumor
radiomic
glrlm
radiomics
Prior art date
Application number
PCT/EP2018/069169
Other languages
English (en)
Inventor
Charles FERTE
Elaine Johanna LIMKIN
Roger SUN
Eric Deutsch
Original Assignee
Institut Gustave-Roussy
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institut Gustave-Roussy filed Critical Institut Gustave-Roussy
Priority to US16/630,031 priority Critical patent/US20210003555A1/en
Priority to EP18740223.5A priority patent/EP3652534B1/fr
Priority to ES18740223T priority patent/ES2975288T3/es
Priority to CN201880058427.3A priority patent/CN111094977B/zh
Priority to JP2020501259A priority patent/JP2020526844A/ja
Priority to AU2018298890A priority patent/AU2018298890A1/en
Priority to IL271996A priority patent/IL271996B1/en
Priority to CA3069612A priority patent/CA3069612A1/fr
Priority to KR1020207003496A priority patent/KR20200040754A/ko
Publication of WO2019012147A1 publication Critical patent/WO2019012147A1/fr

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
    • G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
    • G01N33/5005—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving human or animal cells
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2800/00—Detection or diagnosis of diseases
    • G01N2800/52—Predicting or monitoring the response to treatment, e.g. for selection of therapy based on assay results in personalised medicine; Prognosis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2800/00—Detection or diagnosis of diseases
    • G01N2800/70—Mechanisms involved in disease identification
    • G01N2800/7023—(Hyper)proliferation
    • G01N2800/7028—Cancer
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30096—Tumor; Lesion

Definitions

  • the present invention proposes a radiomics-based biomarker for detecting the presence and the density of tumor infiltrating CD8 T-cells in a solid tumor without having to use any biopsy of said tumor.
  • the invention also proposes to use this information to assess the immune phenotype of said solid tumor.
  • the invention proposes to predict the survival and / or the treatment efficiency of cancer patients treated with immunotherapy such as anti-PD-1/PD- L1 monotherapy.
  • Radiomics are based on the visual interpretation of simple features such as tumor size and tumor global shape
  • CT computed tomography scans
  • MRI magnetic resonance imaging
  • PET PET
  • This new approach analyzes and translates images computationally into quantitative complex data.
  • These high dimensional data allow more in-depth characterization of the tumor phenotype [1-3], with the underlying assumption that imaging reflects not only tissue architecture but also cellular and molecular composition.
  • the end goal of radiomics is to generate imaging biomarkers to serve as clinical decision support tools and to permit better understanding of cancer biology [4-6].
  • Radiomics has several advantages: (i) non-invasive, (ii) evaluates the tumor and its microenvironment in their entirety, thus characterizing spatial heterogeneity, and (iii) can be repeated over time, allowing the assessment of the changes throughout the evolution of the disease.
  • Immune-inflamed tumors present a dense functional CD8 T-cell infiltration (TIL) reflected by increased IFNv signaling, an expression of checkpoint markers (e.g. PD-L1), and high mutational burden. These tumors tend to respond to immunotherapy [12, 13, 15].
  • TIL CD8 T-cell infiltration
  • TGF-beta signaling TGF-beta signaling, myeloid-derived suppressor cells, angiogenesis, etc.
  • the immune-desert phenotype exhibits limited infiltration of CD8 T-cells, with highly proliferating tumor cells, and the immunotherapeutic treatments are often vain. While immunotherapy is being increasingly used in oncology, of note, the last two randomized trials showed that PD-L1 was not associated with response to immunotherapy [55, 56].
  • Radiomics is a rapidly-emerging discipline with the goal of extracting quantitative data from medical images to be used as clinical decision support tools.
  • information obtained from standard imaging modalities [computed tomography scan (CT), magnetic resonance imaging (MRI), and Positron emission tomography scan (PET)]
  • CT computed tomography scan
  • MRI magnetic resonance imaging
  • PET Positron emission tomography scan
  • imaging information is much richer, and the goal of radiomics is to extract high throughput quantitative features, covering the fields of texture, advanced shape modeling, and heterogeneity, to name a few.
  • imaging biomarkers may be used for and contribute to cancer detection, diagnosis, choice of therapeutic strategy, prognosis inference, prediction of response, and surveillance.
  • cluster D had the lowest PD-L1 + tumor cell and the highest infiltrating CD3 T-cell counts.
  • radiomics could allow a reliable evaluation of tumor immune infiltration and thus lead to the identification of novel predictors of efficacy of immunotherapy.
  • radiomics to predict specifically a patient response to immunotherapy such as anti-PD-1/PD-L1 monotherapy.
  • the present study is the first introducing a radiomic-based biomarker for detecting tumor infiltrating CD8 T-cells which shows a reproducible and significant correlation with pathologic quantification of tumor infiltrating lymphocytes (CD8+ tumor infiltrating T cells), tumor immune phenotype and clinical responses to anti-PD-1/PD-L1 in three independent validation cohorts.
  • the present study is unique in that it reveals the link between standard medical images and gene expression signature of CD8 T-cells, pathologic quantification of TIL, tumor immune phenotype, and patient outcomes, especially when treated with immunotherapy. It confirms that radiomics could be an efficient, non-invasive, cost-effective and repeatable way to evaluate patients for precision medicine.
  • the biomarker of the invention can be obtained easily given the widespread availability and routine utilization of tomography scans (CTs). Detailed description of the invention
  • Radiomics consists of the analysis of quantitative data extracted from standard medical imaging to generate imaging biomarkers.
  • the process of radiomics consists of discrete steps: image acquisition and segmentation, feature extraction, and statistical learning. A considerable number of features can be used to assess the characteristics of a target zone. Features may be classified into several categories. There are quantitatively extracted descriptors of size, shape, and other radiologic terminologies which characterize the tumor surface.
  • First- order statistics are used to study the distribution of voxel values without considering spatial relationships;
  • second- order statistics characterize spatial relationships between voxels, such as the co-occurrence matrix (GLCM), gray-level run length matrix (GLRLM), gray-level size zone matrix (GLZLM), and the neighborhood gray-level different matrix (NGLDM).
  • GLCM co-occurrence matrix
  • GLRLM gray-level run length matrix
  • GLZLM gray-level size zone matrix
  • NLDM neighborhood gray-level different matrix
  • Filter grids such as Gabor and Fourier may be used both in the pre-processing step and for extracting spatial or spatio-temporal features.
  • a limitation is that some extracted values are dependent on the ROIs contoured.
  • the extracted features can be global (one value for the whole ROI), or local (a value per image patch) when inhomogeneous patterns are present in the image, where dimensionality significantly increases if simple concatenation of local descriptors is carried out.
  • more advanced frameworks explore compact statistical representations based on coding structures/dictionaries.
  • texture analysis methods focusing on microscopy images of cells or tissues can be found in [46].
  • Radiomic features in CT imagery cannot be accessed or extracted by pencil and paper or acquired by the human mind. Radiomic features present in CT imagery are sub- visual features that are not visible to the human eye.
  • Std Standard deviation; SRE: short-run emphasis; LRE: long-run emphasis; LGRE: low gray- level run emphasis; HGRE: high gray-level run emphasis; SRLGE: short-run low gray-level emphasis; SRHGE: short-run high gray-level emphasis; LRLGE: long-run low gray-level emphasis; LRHGE: long-run high gray-level emphasis; GLNUr: gray-level non-uniformity for run; RLNU: run-length non-uniformity; RP: run percentage; SZE: short-zone emphasis; LZE: long- zone emphasis; LGZE: low gray-level zone emphasis; HGZE: high gray-level zone emphasis; SZLGE: short-zone low gray-level emphasis; SZHGE: short-zone high gray-level emphasis; LZLGE: long-zone low gray-level emphasis; LZHGE: long-zone high gray-level emphasis; GLNUz: gray-level nonuniformity for zone; ZLNU: zone length non-uniformity;
  • the Table 1 presents a selection of some radiomic features that can be extracted from images.
  • First-order features correspond to conventional indices and features extracted from the intensity histogram.
  • Second-order or textural features can be also extracted from four textural matrices calculated using the LIFEx software (http://www.lifexsoft.org): the Gray-Level Co-occurrence Matrix (GLCM), the Gray-Level Run Length Matrix (GLRLM), the Neighborhood Gray-Level Difference Matrix (NGLDM) and the Gray-Level Zone Length Matrix (GLZLM).
  • GLCM and GLRLM can be computed in 13 directions to account for all independent directions between one voxel and its 26 neighbors. Based on some of these features, the present inventors developed a radiomics-based predictor of the presence of TILs and investigated whether such signature could predict the outcome of patients treated by anti-PD1/PDL1 .
  • the radiomics-based predictor could contain other technical variables which are related to the CT scan such as acquisition marker e.g. kV (kilovoltage peak) or localization markers e.g. VOI (location of the volume of interest ) in order to improve the reproducibility of the signature.
  • acquisition marker e.g. kV (kilovoltage peak)
  • localization markers e.g. VOI (location of the volume of interest ) in order to improve the reproducibility of the signature.
  • a first radiomics-based CD8+ signature was developed using the six radiomics features which had highest performance on random forest. Radiomics features were extracted from CT scans after manual segmentation of tumors on contrast-enhanced CTs of 57 HNSCC patients from the TCGA (The Cancer Genome Atlas) / TCIA (The Cancer Imaging Archive) databases.
  • CD8+ T cells were estimated by the Microenvironment Cell Populations-counter signature.
  • this classifier was applied to an independent cohort of 100 patients for which the pathologic tumor immune infiltrate was postulated as either favorable (lymphoma, melanoma, lung, bladder, renal, MSI+ cancers, and adenopathy; 70 patients) or unfavorable (adenoid cystic carcinoma, low-grade neuroendocrine tumors, uterine leiomyosarcoma; 30 patients).
  • the predictor was applied on a second external cohort of 139 patients prospectively enrolled in anti PD-1/PD-L1 phase 1 trials to infer its relation with patient outcome (overall survival (OS)).
  • the median of the radiomics- based CD8+ score was used to separate patients into two groups.
  • this signature was associated with the postulated tumor immune infiltrate (Wilcoxon test, P ⁇ 0.001).
  • a second radiomics-signature of tumor immune infiltration was developed from CT-scans. Access to both RNA-seq data and images of the biopsied lesion in the MOSCATO trial and clinical data of patients from anti-PD-1 /PD-L1 phase 1 trials were used to assess links between imaging features, transcriptomic data, tumor immune phenotype and clinical responses to immunotherapy.
  • Immune infiltration was still modeled using the Microenvironment Cell Populations-counter signature, and especially the CD8B gene according to the "CD8 T-cells" signature by Becht et al ([37]) which associated specifically this gene with the infiltration of CD8 T-cells and was distinct from both the "Cytotoxic T-cells” signature (which encompasses NK cells), and the "T-cells” signature (which encompasses CD4 and naive T-cells) present in the same R- package ([37]).
  • the two identified by the inventors have been validated with several external and independent cohorts and can be reliably used on other cohorts. They appear as a promising tool to estimate tumor immune infiltrates in solid tumors. They can also be used to infer the outcome of patients suffering from cancer, especially those that are treated with immunotherapeutic treatments such as anti-PD1/PD-L1 .
  • the present invention relates to the use of a radiomics-based signature for evaluating the quantity and /or spatial density and / or heterogeneity and/or changes over time of tumor immune infiltrates in a solid tumor.
  • tumor immune infiltrate is herein synonymous of "tumor-infiltrating immune cells”. These terms encompass white blood cells that have left the bloodstream and that have migrated into the tumor. These white blood cells include T cells and B cells, and also natural killer cells, macrophages, neutrophils, dendritic cells, mast cells, eosinophils, basophils, etc., that can be present in the tumors in variable proportions.
  • the "tumor immune infiltrate” detected by the signature of the invention is composed of tumor infiltrating lymphocytes (TILs), and, more preferably, of CD8+ tumor infiltrating lymphocytes.
  • TILs tumor infiltrating lymphocytes
  • the radiomics-based signature of the invention is used for evaluating the spatial distribution of CD8+ T cells infiltrating the tumors.
  • solid tumor it is herein meant any kind of solid tumor, in particular of epithelial, neuroectodermal or mesenchymal origin. It can be a metastatic cancer or not.
  • the tumor can for example be selected from, without being limited to, the group consisting of squamous cell carcinoma, small-cell lung cancer, non-small cell lung cancer, glioma, gastrointestinal cancer, renal cancer, ovarian cancer, liver cancer, colorectal cancer, endometrial cancer, kidney cancer, prostate cancer, thyroid cancer, neuroblastoma, brain cancer, central nervous system cancer, pancreatic cancer, glioblastoma multiforme, cervical cancer, stomach cancer, bladder cancer, malignant hepatoma, breast cancer, colon carcinoma, head and neck cancer, gastric cancer, germ cell tumor, pediatric sarcoma, rhabdomyosarcoma, Ewing's sarcoma, osteosarcoma, soft tissue sarcoma, sinonasal NK/T-cell lymphoma, myeloma, melanoma, multiple myeloma. Benign solid tumors such as uterine leiomyosarcoma are also encompassed.
  • the radiomics-based signature of the invention is used for evaluating the spatial heterogeneity and/or changes over time of tumor immune infiltrates in carcinoma, and in particular in a head and neck squamous cell carcinoma.
  • a CT scan also known as computed tomography scan, makes use of computer-processed combinations of many X-ray measurements taken from different angles to produce cross- sectional (tomographic) images (virtual "slices") of specific areas of a scanned object, allowing the user to see inside the object without cutting.
  • Other terms include computed axial tomography (CAT scan) and computer aided tomography.
  • Digital geometry processing is used to further generate a three-dimensional volume of the inside of the object from a large series of two-dimensional radiographic images taken around a single axis of rotation.
  • CT computed tomography
  • PET positron emission tomography
  • SPECT single-photon emission computed tomography
  • X-ray tomography a predecessor of CT, is one form of radiography, along with many other forms of tomographic and non-tomographic radiography.
  • the radiomics-based signature of the invention is generated by non-invasive imagining technologies such as scanners, magnetic resonance imaging (MRI) or PET (Positron Emission Tomography).
  • non-invasive imagining technologies such as scanners, magnetic resonance imaging (MRI) or PET (Positron Emission Tomography).
  • the signature of the invention is generated from images taken by a CT scan.
  • the radiomics signature of the invention contains at least six (6) radiomics features or variables, preferably at least seven (7), or eight (8) radiomic features, that can be chosen in any of the matrix categories highlighted in Table 1 .
  • the signature of the invention combines radiomics features of at least two, and more preferably three categories.
  • the signature of the invention contains at least one, preferably two, more preferably three, even more preferably four feature(s) of the Gray-level Run Length Matrix (GLRLM) variables. These features can be combined with conventional indices such as min Values.
  • GLRLM Gray-level Run Length Matrix
  • the signature of the invention contains at least one acquisition marker, for example kVp (kilovoltage peak).
  • acquisition marker for example kVp (kilovoltage peak).
  • This particular marker takes into account the image acquisition variability when image acquisition protocols are heterogenous, and therefore enables to use the signature on CT images from different machine and/or acquired using different image acquisition protocols, as shown in example 2 below.
  • the signature of the invention contains at least one localization marker, for example a VOI feature (location of the volume of interest) as proposed in example 2 below.
  • a VOI feature location of the volume of interest
  • Particularly preferred are the VOI markers of adenopathy and head-and-neck. It may happen that the values of these VOI markers are null when the target tumors are not from adenopathy or head-and-neck.
  • the signature of the invention contains at least one, preferably two, more preferably three Gray-Level Co-occurrence Matrix features, for example EnergyH, correlation, and/or Contrast. These parameters are preferably detected for the tumor ROI, as disclosed in example 1 (tum).
  • the signature of the invention contains features characterizing the peripheral zone of the tumor (ring) and the inside tumor (tum). Conventional indices such as maxValue can be used to characterize these zones, as disclosed in example 1 .
  • SZHGE Short-Zone high gray level emphasis
  • the radiomics-based signature of the invention can be used to predict the outcome of cancer patients. In a preferred embodiment, it is used to aid the skilled cancerologist in the selection of appropriate treatments for maximizing the survival of the patients. Appropriate treatments are for example chemotherapeutic treatments, immunotherapeutic treatments, radiotherapeutic treatments and/or surgery. Preferably, the signature of the invention is generated before initiating a treatment.
  • said patients have been treated or will be treated with anti-cancer drugs.
  • Said anti-cancer agent can be selected from chemotherapy, immunotherapy (or immune checkpoint blocker), anti-cancer vaccine, radiotherapy, and surgery.
  • Said "chemotherapeutic agent” is typically an agent selected for example from an antitumor/cytotoxic antibiotic, an alkylating agent, an antimetabolite, a topoisomerase inhibitor, a mitotic inhibitor, a platin based component, a specific kinase inhibitor, an hormone, a cytokine, an antiangiogenic agent, an antibody, a DNA methyltransferase inhibitor and a vascular disrupting agent.
  • Said antitumor agent or cytotoxic antibiotic can for example be selected from an anthracycline (e.g. doxorubicin, daunorubicin, adriamycine, idarubicin, epirubicin, mitoxantrone, valrubicin), actinomycin, bleomycin, mitomycin C, plicamycin and hydroxyurea.
  • anthracycline e.g. doxorubicin, daunorubicin, adriamycine, idarubicin, epirubicin, mitoxantrone, valrubicin
  • actinomycin bleomycin, mitomycin C, plicamycin and hydroxyurea.
  • Said alkylating agent can for example be selected from mechlorethamine, cyclophosphamide, melphalan, chlorambucil, ifosfamide, temozolomide busulfan, N-Nitroso-N-methylurea (MNU), carmustine (BCNU), lomustine (CCNU), semustine (MeCCNU), fotemustine, streptozotocin, dacarbazine, mitozolomide, thiotepa, mytomycin, diaziquone (AZQ), procarbazine, hexamethylmelamine and uramustine.
  • Said antimetabolite can for example be selected from a pyrimidine analogue (e.g.
  • a fluoropyrimidine analog a fluoropyrimidine analog, 5-fluorouracil (5-FU), floxuridine (FUDR), cytosine arabinoside (Cytarabine), Gemcitabine (Gemzar®), capecitabine); a purine analogue (e.g. azathioprine, mercaptopurine, thioguanine, fludarabine, pentostatin, cladribine, clofarabine); a folic acid analogue (e.g. methotrexate, folic acid, pemetrexed, aminopterin, raltitrexed, trimethoprim, pyrimethamine).
  • a purine analogue e.g. azathioprine, mercaptopurine, thioguanine, fludarabine, pentostatin, cladribine, clofarabine
  • a folic acid analogue e.g. methotrex
  • Said topoisomerase inhibitor can for example be selected from camptothecin, irinotecan, topotecan, amsacrine, etoposide, etoposide phosphate and teniposide.
  • Said mitotic inhibitor can for example be selected from a taxane [paclitaxel (PG-paclitaxel and DHA-paclitaxel) (Taxol ®), docetaxel (Taxotere ®), larotaxel, cabazitaxel, ortataxel, tesetaxel, or taxoprexin]; a spindle poison or a vinca alkaloid (e.g. vincristine, vinblastine, vinorelbine, vindesine or vinflunine); mebendazole; and colchicine.
  • Said platin based component can for example be selected from platinum, cisplatin, carboplatin, nedaplatin, oxaliplatin, satraplatin and triplatin tetranitrate.
  • Said specific kinase inhibitor can for example be selected from a BRAF kinase inhibitor such as vemurafenib; a MAPK inhibitor (such as dabrafenib); a MEK inhibitor (such as trametinib); and a tyrosine kinase inhibitor such as imatinib, gefitinib, erlotinib, sunitinib or carbozantinib. Tamoxifen, an anti-aromatase, or an anti-estrogen drug can also typically be used in the context of hormonotherapy.
  • a BRAF kinase inhibitor such as vemurafenib
  • MAPK inhibitor such as dabrafenib
  • MEK inhibitor such as trametinib
  • a tyrosine kinase inhibitor such as imatinib, gefitinib, erlotinib, sunitinib or carbozantinib.
  • a cytokine usable in the context of an immunotherapy can be selected for example from IL-2 (lnterleukine-2), IL-1 1 (lnterleukine-1 1), IFN (Interferon) alpha (IFNa), and Granulocyte- macrophage colony-stimulating factor (GM-CSF).
  • Said anti-angiogenic agent can be selected for example from bevacizumab, sorafenib, sunitinib, pazopanib and everolimus.
  • the monoclonal antibody can be selected from a anti-CD20 antibody (anti-pan B-Cell antigen), anti-Her2/Neu (Human Epidermal Growth Factor Receptor- 2/NEU) antibody; an antibody targeting cancer cell surface (such as rituximab and alemtuzumab); a antibody targeting growth factor (such as bevacizumab, cetuximab, panitumumab and trastuzumab); a agonistic antibody (such as anti-ICOS mAb, anti-OX40 mAb, anti-41 BB mAb); and an immunoconjugate (such as 90Y-ibritumomab tiuxetan, 131 l-tositumomab, or ado- trastuzumab emtansine).
  • a anti-CD20 antibody anti-pan B-Cell antigen
  • anti-Her2/Neu Human Epidermal Growth Factor Receptor- 2/NEU
  • Said DNA methyltransferase inhibitor can for example be selected from 2'-deoxy-5-azacytidine (DAC), 5-azacytidine, 5-aza-2'- deoxycytidine, 1 -[beta]-D-arabinofuranosyl-5-azacytosine and dihydro-5-azacytidine.
  • Said vascular disrupting agent can for example be selected from a flavone acetic acid derivative, 5,6-dimethylxanthenone-4- acetic acid (DMXAA) and flavone acetic acid (FAA).
  • chemotherapeutic drugs include a proteasome inhibitor (such as bortezomib), a DNA strand break compound (such as tirapazamine), an inhibitor of both thioredoxin reductase and ribonucleotide reductase (such as xcytrin), and an enhancer of the Thl immune response (such as thymalfasin).
  • a proteasome inhibitor such as bortezomib
  • a DNA strand break compound such as tirapazamine
  • an inhibitor of both thioredoxin reductase and ribonucleotide reductase such as xcytrin
  • an enhancer of the Thl immune response such as thymalfasin
  • Said immune checkpoint blocker is typically an antibody targeting an immune checkpoint.
  • an immune checkpoint blocker can be advantageously selected from anti-CTLA4 (ipilimumab and Tremelimumab), anti-PD-1 (Nivolumab and Pembrolizumab), anti-PD-L1 (Atezolizumab, Durvalumab, and Avelumab), anti-PD-L2 and anti-Tim3.
  • Said cancer vaccine can for example be selected from a vaccine composition comprising (antigenic) peptides; a Human papillomavirus (HPV) vaccine (such as Gardasil®, Gardasil9®, and Cervarix®); a vaccine stimulating an immune response to prostatic acid phosphatase (PAP) sipuleucel-T (Provenge®); an oncolytic virus and talimogene laherparepvec (T-VEC or Imlygic®).
  • HPV Human papillomavirus
  • PAP prostatic acid phosphatase
  • T-VEC oncolytic virus and talimogene laherparepvec
  • the radiotherapy typically involves rays selected from X-rays ("XR”), gamma rays and/or UVC rays.
  • the treatment which can include several anticancer agents is selected by the cancerologist depending on the specific cancer to be prevented or treated.
  • said patients have been treated or will be treated with immunotherapy drugs such as anti-PD-1 and/or anti-PD-L1 drugs.
  • the present invention also encompasses a method for evaluating the spatial heterogeneity and/or changes over time of said tumor immune infiltrates in a solid tumor, said method comprising the steps of : a) obtaining a radiological image of a region of a tumoral tissue, said image including a plurality of voxels; b) specifying a region of interest (ROI) in the image; c) extracting a set of at least 6, preferably at least 7, more preferable 8 radiomics features from said ROI; d) calculating a score from said at least 6, preferably at least 7, more preferable 8 radiomics features; e) comparing said score to a reference value; f) concluding from said comparison that the tumor immune infiltrate is present or determining its density or its evolution.
  • ROI region of interest
  • the radiological image includes a set of morphological features acquired using an imaging system, such as a CT system. Accessing the radiological image may include retrieving electronic data from a computer memory, receiving a computer file over a computer network, or other computer or electronic based action.
  • the invention is also drawn to an in vitro method for evaluating the density, spatial heterogeneity and/or changes over time of a tumor immune infiltrate in a solid tumor, said method comprising the steps of : a) specifying a region of interest (ROI) in a radiological image of a region of a tumoral tissue, said image including a plurality of voxels; b) extracting a set of at least 6, preferably at least 7, more preferable 8 radiomics features from said ROI; c) calculating a score from said at least 6, preferably at least 7, more preferable 8 radiomics features; d) comparing said score to a reference value; e) concluding from said comparison that the tumor immune infiltrate is present or determining its density or its evolution.
  • the nature of the radiomics feature have been described above for the signature of the invention. All of them, especially the preferred ones, can be used in the methods of the invention.
  • the region of tumoral tissue can be either located inside the tumor, or at the periphery of said tumor (for the determination of the "ring" variable).
  • the ROI may be annotated by an expert radiologist using a 3D slicer approach or may be annotated using an automated segmentation approach. Other annotation or segmentation approaches or techniques may be employed.
  • the calculated score is a continuous and linear value reflecting the density of the tumor immune infiltrate in the ROI of the analyzed solid tumor. It is calculated by computing the levels of the at least 6 radiomics features with predefined coefficients, and by summing all the computed values (see examples below). The higher the score is, the denser the tumor immune infiltrate is likely to be. Consequently, the higher the score is, the more efficient an anti-tumoral treatment (such as immunotherapy) will be.
  • the term "reference value” is a predetermined value that has been selected on a population having a defined diagnostic and prognostic of cancer. For example, it has been obtained from a population of patients responding efficiently to a define treatment (in this case, the reference value will be high). By contrast, it can be obtained from a population of patients poorly responding to a define treatment (in this case, the reference value will be low).
  • the reference value is preferably determined by training the machine learning classifier with images issued from particular tumors.
  • First testing images include images of a region of a tumor that responded to immunotherapy.
  • Second testing images include images of a region of a tumor that did not respond to immunotherapy.
  • Accessing the testing images may include retrieving electronic data from a computer memory, receiving a computer file over a computer network, or other computer or electronic based action.
  • An "efficient response to a treatment” is usually concluded when the overall survival (OS) of the patient treated with said treatment is of at least one year, preferably two years, more preferably five years. Patients having such OS is also called a "responder”.
  • OS overall survival
  • a Complete Response to a treatment is more preferably defined according to RECIST 1 .1 criteria [62].
  • a Complete Response (CR) is defined as a disappearance of all target lesions. Any pathological lymph nodes (whether target or non-target) must have reduction in short axis to ⁇ 10 mm.
  • a Partial Response (PR) is defined as at least a 30% decrease in the sum of diameters of target lesions, taking as reference the baseline sum diameters.
  • a Progressive Disease (PD) is defined as at least a 20% increase in the sum of diameters of target lesions, taking as reference the smallest sum on study (this includes the baseline sum if that is the smallest on study). In addition to the relative increase of 20%, the sum must also demonstrate an absolute increase of at least 5 mm. (Note: the appearance of one or more new lesions is also considered progression).
  • a Stable Disease (SD) is defined as neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD, taking as reference the smallest sum diameters while on study.
  • a "non-responder" is considered as a patient with a progression disease or a stable disease as defined according to RECIST 1 .1 criteria.
  • the reference value used in the method of the invention is preferably a cut-off obtained from a specific cohort of patients (see in the examples below). It can be determined from the area under the receiver operating characteristic curve (AUC) that describes the relationship between the sensitivity and the complement of the specificity for each possible value taken by the signature as a discrimination threshold.
  • Optimal cut-off corresponds to the most effective signature discrimination threshold and it could be measured by several methods. One method to find the optimal cut-off could be by using the Youden index. It corresponds to the value that maximizes the Youden index as defined as the sum of the sensitivity and specificity for each possible value of the signature. 95% confidence intervals were determined according to the Delong method in this example, but can be also assessed by a bootstrapping method.
  • the method of the invention, and the radiomics-based signature of the invention are able to discriminate between responders and non-responder patients suffering in particular from carcinomas, toward immunotherapeutic treatments such as anti-PD1 and anti-PD-L1 immune checkpoint blockers.
  • the method of the invention is advantageously used to predict the outcome of cancer patients. It requires the use of non-invasive imagining technologies such as scanners, magnetic resonance imaging (MRI) or PET (Positron Emission Tomography), so as to generate the "radiological image" of the ROI.
  • non-invasive imagining technologies such as scanners, magnetic resonance imaging (MRI) or PET (Positron Emission Tomography), so as to generate the "radiological image" of the ROI.
  • the method of the invention can contain other steps such as using genomic data, feature extraction of tumor and rim, and machine-learning. Preferably, it does not contain any invasive step. In particular, it does not require any biopsy nor blood collection step.
  • Said method uses computational medical imaging and can be also called “radiomics-based CD8+ score” (in the context of the invention, the term “radiomics” stands for “medical image processing and analysis”). It is a novel marker of the efficacy of immunotherapy in patients in need for said treatment.
  • the present inventors identified at least two different signatures of at least 6 radiomics features able to detect the tumor immune infiltrates and having prognostic value as exposed above. These two signatures are detailed in examples 1 and 2 below.
  • the first signature (example 1) has been determined and validated on a cohort of patients suffering from a specific cancer, a head and neck carcinoma. It is therefore more useful for this particular category of patients.
  • the second signature (example 2) has been determined and validated on cohorts containing patients suffering from different types of cancers (Tables 9 to 14). It is therefore more generic and can be applied to any cancer patients suffering from a solid tumor.
  • the signatures can be reproduced as explained in the examples. Briefly, CTs were selected if the slice-thickness was ⁇ 5 mm and images were reconstructed using soft or standard reconstruction algorithms. Volumes of interest (VOI) for the radiomic analysis consisted of the tumor volume and a peripheral ring, which was created around the tumor margins using 3D expansion and shrinkage of 2 mm on both sides (i.e. inner and outer), resulting in a 4 mm thickness 3D ring around the tumor. An image processing step has been done to normalize the images. Voxels were resampled to 1x1x1 mm3, and Hounsfield values of the images were regrouped into one discrete value for every 10UH (absolute discretization).
  • Variables computationally extracted from the images consisted of radiomics features from tumor and ring, VOI location and variables related to the imaging acquisition.
  • a machine learning algorithm was used to train a radiomic signature to predict the estimation of CD8 T-cells and validation datasets were used to validate this signature on a genomics, phenotype and clinical level.
  • the signature of the invention contains 5 radiomics features from the tumor, and one radiomic feature from the ring of the tumor.
  • a peritumoral region may be defined as the region surrounding the tumoral region at a specified distance from the tumor borders.
  • the peritumoral region may be defined as the region extending 2mm from the external and internal tumoral borders creating a form of 4 mm in diameter enclosing the tumor borders.
  • the peritumoral region may be the region extending 5mm from the tumoral boundary, or 10mm from the tumoral boundary.
  • the peritumoral region may be defined by a distance measured in mm, or in other units, including pixels or voxels.
  • tumor specific features or “variables” can consist in energyH, correlation, maxValue, Contrast.1 , and SZHGE and the peripheral feature can consist in maxValue, as defined in the LIFEX [43].
  • Table 2 Said radiomic score is calculated by means of a random forest type calculation, then compared with a reference value so as to determine, for each analyzed patient:
  • an anti-cancer treatment e.g., an immunotherapeutic treatment as defined above
  • the signature of the invention can contain 2 radiomics features from the tumor, 3 radiomic features from the ring of the tumor, 2 radiomics features from the localization variables (VOI) and one acquisition-dependent variable.
  • VOI localization variables
  • tumor specific features or “variables” can be minValue, and GLRLM_SRHGE.
  • the ring specific features or “variables” can be GLRLM_SRLGE, GLRLM_LGRE and GLRLM_LRLGE.
  • the localisation variables can be VOI_Adenopathy and VOI_head_and_neck.
  • the acquisition variable can be kVp (kiloVoltage peak).
  • the radiomic-based signature of the invention may specifically contain:
  • Gray-level Run Length Matrix (GLRLM) variable - at least one acquisition marker such as kVp,
  • - at least one localization marker such as VOI at least one localization marker such as VOI.
  • the radiomic-based signature of the invention may more specifically contain:
  • Gray-level Run Length Matrix (GLRLM) variables - one acquisition marker such as kVp,
  • the radiomic-based signature of the invention may even more specifically contain: - the minValue for the tumor as conventional variable,
  • Gray-level Run Length Matrix (GLRLM) variables consisting of : GLRLM_SRHGE, GLRLM_SRLGE, GLRLM_LGRE and GLRLM_LRLGE,
  • kVp and VOI locations were included by design to account for radiomic information likely related only to the organ analyzed since it is recognized that tissue origin exerts a major impact on the tumor immune contexture [58].
  • the machine learning model retained three groups: adenopathy, head and neck, and the rest. A sensitivity analysis was performed without the non-radiomic variables and showed poorer results, underscoring the importance of adjusting to these parameters when developing radiomics predictive tools (Figure 8). Said radiomic score is then compared with a reference value so as to determine, for each analyzed patient:
  • an anti-cancer treatment e.g., an immunotherapeutic treatment as defined above
  • the method of the invention extract radiomic features (e.g. quantitative image descriptors) from radiological images to generate predictive and prognostic information, and thus provide non- invasive biomarkers for treatment response, monitoring patients, disease prognosis, and personalized treatment planning.
  • the extracted radiomic features are then preferably provided to a machine learning predictor which predicts a quantitative score of the CD8 T-cells infiltration in the region of interest.
  • the classification can then be achieved by the user in view of this score and according to a cut-off which is to be determined depending on the cohort of patient analyzed.
  • the methods of the invention enable to generate a personalized treatment plan.
  • the personalized treatment plan is based on the high or low score of CD8+ T cells .
  • the personalized treatment plan may include an immunotherapy recommendation, an immunotherapy schedule, an immunotherapy dosage value, a follow up treatment schedule, or other action.
  • the personalized treatment plan may include information that facilitates developing a precision treatment plan for a patient associated with the radiological image. For example, upon determining that a tumor is classified as a responder, the method of the invention may control the personalized cancer treatment system to generate a first personalized treatment plan that indicates a first type of therapy. Upon determining that the tumor is classified as a non-responder to this first therapy, the method may generate a second, different personalized treatment plan that proposes a second, different type of therapy.
  • the methods of the invention produce the concrete, real-world technical effect of reducing the amount of unnecessary biopsies or other invasive procedures for patients who are unlikely to benefit from immunotherapy treatment. Additionally, these methods reduce the expenditure of time, money and therapeutic resources on patients who are unlikely to benefit from the treatment. They thus improve on conventional approaches to predicting response to immunotherapy in a measurable, clinically significant way.
  • the present invention relates to a computer-readable storage device that may store computer executable instructions that if executed by a machine (e.g., computer, processor) cause the machine to perform the methods of the invention.
  • a computer-readable storage device may include, but are not limited to, a floppy disk, a flexible disk, a hard disk, a magnetic tape, other magnetic medium, an application specific integrated circuit (ASIC), a compact disk (CD), other optical medium, a random access memory (RAM), a read only memory (ROM), a memory chip or card, a solid state device (SSD), a memory stick, a data storage device, and other media from which a computer, a processor, on the cloud in a SAS mode or other electronic device can read.
  • ASIC application specific integrated circuit
  • CD compact disk
  • RAM random access memory
  • ROM read only memory
  • SSD solid state device
  • a data storage device and other media from which a computer, a processor, on the cloud in a SAS mode or other electronic device can read.
  • the apparatus of the invention provides a timely and intuitive way for a human pathologist, a personalized cancer treatment system, or a diagnostic system to more accurately predict response to immunotherapy, thus improving on conventional approaches to predicting response to treatment.
  • Figure 3. Importance of the different features in the random forest model of example 1 .
  • Figure 4. Internal validation of the signature of example 1 .
  • Figure 5. Study design of the assay of example 2.
  • FIG. 1 Performance of the radiomic signature of example 2 in training and validation datasets.
  • AUC Area under the curve of the receiver operator characteristic of radiomic-scores in MOSCATO training set, TCGA validation set, immune phenotype-based dataset.
  • B Objective response to anti-PD-1/anti-PD-L1 monotherapy at 6 months according to the CD8 T-cells radiomic score, and overall survival of patients according to the radiomic score (high/low defined by the median value).
  • PD stable disease
  • DC disease control
  • SD stable disease
  • PR partial response
  • CR complete response.
  • TILs tumor infiltrating lymphocytes
  • BLCA liver hepatocellular carcinoma
  • LAD lung adenocarcinoma
  • LUSC lung squamous cell carcinoma
  • HNSCC CD8 T-cells radiomic score according to the tumor type in head and neck tumors.
  • ACC adenoid cystic carcinoma
  • UCNT undifferentiated carcinoma of nasopharyngeal type
  • Example 1 Contrast-enhanced computer tomography (CT) imaging of head and neck carcinomas Training set
  • Each primary tumor was manually delineated on CT-scan images by two radiation oncologists using ISOgray® segmentation solution (Dosisoft, Cachan, France).
  • ISOgray® segmentation solution Dosisoft, Cachan, France.
  • a peripheral ring was created around the tumor margins using 3D expansion and shrinkage of 2 mm on both sides (inner and outer) to take into account the peri-tumoral stroma and the invasion margins. Large vessels and neighboring organs were excluded from the ring if they were not involved by the tumor (not shown).
  • Radiomics feature extraction was performed using the LIFEx software (Local Image Feature Extraction, www.lifexsoft.org) [38]. Thirty-eight first and second order features were extracted for each of the two volumes of interest (VOI, tumor and ring) leading to a total of 76 features. Hounsfield-units (HU) values in the VOIs were then resampled into 400 discrete values using absolute discretization. The minimum and maximum bounds of the resampling intervals were set to -1000 and 3000 HU, leading to a bin size of 10 HU.
  • GLCM Gray-Level Co-occurrence Matrix
  • GLRLM Gray-Level Run Length Matrix
  • NLDM Neighborhood Gray-Level Different Matrix
  • GLZLM Gray-Level Zone Length Matrix
  • Feature selection was done by first eliminating correlated and redundant variables (Pearson's rho > 0.75) using the Caret R package (version 6.0-76) [40]. Then, a random forest model was used on the remaining radiomic feature variables to predict the CD8 T cells abundance estimated by MCP-counter, using the RandomForest R package (version 4.6-12), and 500 trees [41 ]. The optimal number of variables randomly sampled as candidates at each split (mtry) was defined using the "tuneRF" function in this package. Variables were ranked according to their importance estimated by the IncNodePurity approach, which corresponds to the total decrease in node impurities, measured by the Gini index from splitting on the variable, averaged over all trees. The six radiomic features with the largest IncNodePurity were selected to train the final random forest model to predict a CD8 T cells quantitative score.
  • the first consists of 100 randomly selected patients from our institute database, for which the pathologic abundance of tumor immune infiltrate was postulated as either favorable based on documented good response to immunotherapy in the literature or if the VOI was a lymph node (lymphoma, melanoma, lung, bladder, renal and MSI+ cancers; 70 patients), or unfavorable if otherwise (adenoid cystic carcinoma, low-grade neuroendocrine tumors, uterine leiomyosarcoma; 30 patients).
  • This cohort was previously described and published elsewhere [44].
  • a total of 139 patients who had a baseline contrast-enhanced CT-scan images with a slice thickness ⁇ 5 mm were included.
  • One target lesion defined by the radiologist according to RECIST 1 .1 was delineated for each patient.
  • follow-up and survival times were calculated from the baseline date.
  • the median of the radiomics-based CD8+ score was used to separate patients from this cohort into two groups.
  • the signature was applied on two clinical cases from the cohort of patients treated with immunotherapy to assess its relationship with spatial heterogeneity through correlations with sequential pathological analyses, and with temporal heterogeneity by repeated measure or the radiomics-based score through the course of the disease.
  • CD8+ T cell abundance by MCP-counter 38.39 [15.68, 101.54]
  • the final CD8+ radiomics-based signature obtained after random forest consisted of five features from the tumor (energy, correlation, max value, NGLDM contrast, short-zone high grey-level emphasis [SZHGE]) and one from the ring (max value) (FIGURE 3).
  • N 100 30 70
  • this signature was applied for two patients with uncommon responses to immunotherapy from the cohort of patients treated with immunotherapy.
  • Patient 1 is a 42 year-old woman with colon cancer and liver metastases treated with immunotherapy. She underwent stereotactic radiation therapy of one liver metastasis (segment 6) with the aim of inducing abscopal response. Three months later, the liver metastases progressed, but the morphology of the irradiated lesion significantly changed, showing core necrosis and a rim of hypodensity. These changes were likewise seen in the non-irradiated lesion (segment 4) (not shown).
  • Patients 2 had metastatic head and neck cancers and was treated with anti-PD-1 . He exhibited dissociated responses, with progression of cervical lesions, but marked objective responses of non-cervical secondary lesions (pulmonary lesions). Radiomics analysis of cervical nodes and pulmonary lesions at baseline and during the follow-up showed persistently high CD8+ scores for pulmonary lesions and low score for cervical node (not shown).
  • Radiomic features were extracted from contrast-enhanced CTs of 135 patients with advanced solid malignant tumors from the prospective trial MOSCATO. For each patient, RNA-seq data were used to quantify CD8 T-cells. From 84 variables (78 radiomic features, 5 location variable, 1 technical variable), a radiomic-based predictor of CD8 T-cell expression was built using elastic- net. The primary objective was to confirm the relationship of this predictor with gene expression in an independent cohort of 1 19 patients from The Cancer Genome Atlas (TCGA).
  • TCGA Cancer Genome Atlas
  • This radiomic-signature of CD8 T-cells was shown to be relevant in three independent cohorts. It provided promising means to assess tumor immune phenotype and to infer outcomes for cancer patients treated with anti-PD-1/PD-L1 .
  • Three cohorts refer to patients treated at Gustave Roussy while the last came from The Cancer Genome Atlas (TCGA) (https://cancergenome.nih.gov/) and The Cancer Imaging Archive (TCIA) databases (http://www.cancerimagingarchive.net/).
  • TCGA Cancer Genome Atlas
  • TCIA The Cancer Imaging Archive
  • the MOSCATO dataset contains data from patients included in a precision medicine trial [42] and was used to train the radiomic signature. Three cohorts were used for validation.
  • the TCGA dataset was constituted of RNA-seq data from The Cancer Genome Atlas, and the corresponding imaging data and pathology slides from The Cancer Imaging Archive and The Cancer Digital Slide Archive.
  • the immune-phenotype based cohort contained tumors labeled as either immune-desert or inflamed.
  • the IO (Immuno-Oncology) treated dataset was constituted of patients included in anti-PD-1/PD-L1 monotherapy phase 1 trials.
  • the training dataset used to build the radiomic-signature of CD8 T-cells consisted of patients of the prospective MOSCATO trial (NCT01566019) where genomic information was obtained through computed tomography (CT)-guided biopsy ([42]).
  • CT computed tomography
  • RNA-seq data were available, allowing estimation of CD8 T-cells using RNA-seq data and radiomic analysis of the corresponding biopsied tumor images.
  • the TCGA dataset included patients for whom baseline preoperative imaging data with required quality standards and corresponding transcriptomic data were available ([36]). Five collections were used (Table 9): head and neck squamous cell carcinoma (TCGA-HNSC), lung squamous cell carcinoma (TCGA-LUSC), lung adenocarcinoma (TCGA-LUAD), liver hepatocellular carcinoma (TCGA-LIHC), and bladder endothelial carcinoma (TCGA-BLCA). This dataset was used to validate the radiomic-signature.
  • TCGA-HNSC head and neck squamous cell carcinoma
  • TCGA-LUSC lung squamous cell carcinoma
  • TCGA-LUAD lung adenocarcinoma
  • TCGA-LIHC liver hepatocellular carcinoma
  • TCGA-BLCA bladder endothelial carcinoma
  • Genomic-based CD8 T -cells score (median [IQR]) 1.00 [0.52, 1.56]
  • the "immune-phenotype dataset” consisted of randomly selected patients from our institute's database, representing the two extreme tumor immune phenotypes: inflamed or immune-desert, irrespective of treatment delivered. Inflamed tumors either had recognized sensitivity to immunotherapy, or had lymph nodes as volumes-of-interest (VOI) (lymphoma, melanoma, lung, bladder, renal and microsatellite instability-high (MSI+) cancers). Immune-desert tumors were those typically known to have poor lymphocyte infiltration (adenoid cystic carcinoma, low-grade neuroendocrine tumors, uterine leiomyosarcoma) (Table 10). Labeling of tumors was made by CF and EJL. This dataset was used to evaluate concordance of the radiomic-signature with tumor immune phenotype.
  • the IO-treated cohort consisted of consecutive patients enrolled in five immuno-oncology phase 1 trials (anti-PD-1/PD-L1 monotherapy) at Gustave Roussy (NCT01375842, NCT01358721 , NCT01295827, NCT02054806, NCT01693562), details of which are available in previous publications ([44]).
  • This cohort was used to infer the relationship of the radiomic-signature with patient response according to Response Evaluation Criteria In Solid Tumors (RECIST) version 1 .1 , progression-free survival (PFS) and overall survival (OS).
  • a peripheral ring was created using 2 mm 3D dilation and erosion from the tumor boundaries, resulting in a 4 mm 3D ring. Large vessels, neighboring organs and air cavities were excluded if not invaded. Textural pattern can differ depending on tissue macrostructure.
  • VOI locations were introduced as a parameter and labelled as "adenopathy” for node metastasis, "head and neck” for primary or secondary lesions of the pharynx, larynx, oral cavity or salivary glands, "lung” and “liver” for primary or secondary lesions of the lung or the liver respectively, and “other” for subcutaneous or abdominal lesions. Tumor volume was included in analyses as a potential confounding factor.
  • Radiomic feature extraction was performed using LIFEx software version 3.44 (www.lifexsoft.org) ([38]). Images were resampled to 1x1x1 mm 3 voxels using 3D Lagrangian polygon interpolation. Hounsfield-units (HU) values through all the images were then resampled into 400 discrete values (called "bins") using absolute discretization from -1000 to 3000 HU, leading to a fixed bin size of 10 HU.
  • Four gray-level matrices were calculated in 3D resulting to 39 radiomic features (first- and second-order features and volume) for each of the two VOIs (tumor and ring) (Table 1 ). Values of extracted radiomic features were normalized linearly in the range 0 to 1 .
  • the Table 1 presents the radiomic features that have been extracted from the images.
  • First-order features correspond to conventional indices and features extracted from the intensity histogram.
  • Second-order or textural features have been also extracted from four textural matrices calculated using the LIFEx software (http://www.lifexsoft.org): the Gray-Level Co-occurrence Matrix (GLCM), the Gray-Level Run Length Matrix (GLRLM), the Neighborhood Gray-Level Difference Matrix (NGLDM) and the Gray-Level Zone Length Matrix (GLZLM).
  • GLCM and GLRLM have been computed in 13 directions to account for all independent directions between one voxel and its 26 neighbors. Each textural feature extracted from these two matrices corresponds to the average value over the 13 directions.
  • RNA-seq data were quantified using TPM (Transcript Per Million) method by the Salmon tool ⁇ [43].
  • TPM Transcript Per Million
  • TCGA Pan-cancer project V4.6 of the TCGA Pan-cancer project [49] was used (https://www.synapse.org/#!Synapse:syn1701959).
  • Data consisted of 20530 genes obtained with lllumina HiSeq RNASeqV2 (lllumina, San Diego, CA, USA) and quantified using RPKM (Reads Per Kilobase Million) method. All RNA-seq data were rescaled to have the same mean and variance as the training set, stratified by VOI location.
  • Variable selection, machine learning Input variables for the machine learning method consisted of 84 variables: 78 radiomic features, five locations (labeled as binary variables), and one global imaging variable, the peak kilovoltage (kVp), given its established impact on radiomic output [51 ].
  • a linear elastic-net model was used as regression method using the GLMNET R package version 2.0-10, for feature selection and model building [52].
  • the regularization parameter ⁇ was defined using cross-validation and the a penalty was set to 0.5 after a grid search.
  • AUC Area under the curve
  • the MOSCATO training dataset used to build the radiomic-signature consisted of 135 patients included between May 1 , 2012 to March 31 , 2016 (TABLE 12). Category Overall
  • the TCGA validation dataset included 1 19 patients among the 435 patients available for screening at the time of inclusion (June 30, 2017) (patient characteristics and flowchart in Table 13).
  • the "immune-phenotype dataset” consisted of 100 patients randomly selected from August 24, 2005 to November 19, 2015, with 70 (70%) tumors recognized to belong to the immune-inflamed group, and 30 (30%) tumors recognized to belong to the immune-desert group (Table 14).
  • Table 14 VOI: volume of interest, IQR: interquartile range, MSI: microsatellite instability
  • Kidney 1 1 (8.0) 5 (7.2) 6 (8.8)
  • Lymphoma 7 (5.1 ) 1 (1.4) 6 (8.8)
  • the IO-treated cohort consisted of 137 consecutive patients enrolled in anti-PD-1/PD-L1 phase 1 trials between December 1 , 201 1 and January 31 , 2014 (TABLE 15).
  • VOI volume of interest
  • IQR interquartile range
  • RMH score Royal Marsden Hospital prognostic score
  • Radiomic feature from the tumor Ring: radiomic feature from the peripheral ring around the tumor;
  • VOI location of the volume of interest;
  • GLRLM Gray-Level Run Length Matrix;
  • SRHGE short-run high gray-level emphasis;
  • LGRE low gray-level run emphasis;
  • SRLGE short-run low gray-level emphasis;
  • LRLGE long-run low gray-level emphasis;
  • kVp kilovoltage peak.
  • PD stable disease
  • DC disease control
  • SD stable disease
  • PR partial response
  • CR complete response
  • Radiomic-score of CD8 T-cells High 0.52 0.35 - 0.79 0.0022
  • Cytotoxic lymphocytes 0,64 ⁇ 1 e-05 0,32 0.00037
  • T cells 0,65 ⁇ 1 e-05 0,34 0.00025
  • NK cells 0,48 ⁇ 1 e-05 0, 15 0.087
  • Endothelial cells 0,29 9.00E-04 0,24 0.0088
  • Cytotoxic lymphocytes 0,8 ⁇ 1 e-05 0,23 0.03
  • NK cells 0,66 ⁇ 1 e-05 0, 1 0.4
  • Myeloid dendritic cells 0,57 ⁇ 1 e-05 0,092 0.4
  • Endothelial cells 0,32 0.00044 0, 16 0.18
  • HNSC head and neck tumors with different histologies
  • ACC from the phenotype-based dataset
  • UCNT undifferentiated nasopharyngeal cancers
  • the radiomic-predictor could significantly discriminate the different histologies in accordance with pathologically-assessed tumor immune infiltration such as ACC (lower infiltration than HNSCC), HNSCC and UCNT (higher infiltration) (FIGURE 7B).
  • This second signature comprises textural features from the GLRLM matrix ([47]).
  • This matrix reflects homogeneity or heterogeneity of an image.
  • An intuitive interpretation of the signature is that relatively homogeneous and hypodense tumors and peripheral rings were associated with a high CD8 T-cells score (see in [47] & [48]). These patterns could be representative of inflammatory infiltrate, while heterogeneity and high gray levels might be more representative of heterogeneous and intertwined processes like chaotic vascularization and necrosis.
  • the signature included kVp and VOI locations. Retaining kVp highlights the need to account for image acquisition variability when cohorts are heterogenous, given that textural features are highly dependent on it [51 , 57].
  • VOI locations were included by design to account for radiomic information likely related only to the organ analyzed since it is recognized that tissue origin exerts a major impact on the tumor immune contexture [58].
  • the machine learning model retained three groups: adenopathy, head and neck, and the rest. A sensitivity analysis was performed without the non-radiomic variables and showed poorer results, underscoring the importance of adjusting to these parameters when developing radiomics predictive tools (Figure 8).
  • lymphocytes in the vertical growth phase of primary cutaneous melanoma Cancer 1996; 77(7):1303-1310.
  • Galloway MM Texture analysis using gray level run lengths.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Immunology (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • Molecular Biology (AREA)
  • Primary Health Care (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Pathology (AREA)
  • Chemical & Material Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Urology & Nephrology (AREA)
  • Hematology (AREA)
  • Biotechnology (AREA)
  • Microbiology (AREA)
  • Cell Biology (AREA)
  • Tropical Medicine & Parasitology (AREA)
  • Food Science & Technology (AREA)
  • Medicinal Chemistry (AREA)
  • Databases & Information Systems (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Data Mining & Analysis (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Medicines That Contain Protein Lipid Enzymes And Other Medicines (AREA)

Abstract

La présente invention propose un biomarqueur basé sur la radiomique pour détecter la présence et la densité de lymphocytes T CD8 infiltrés dans des tumeurs solides sans avoir à utiliser de biopsie de ladite tumeur. L'invention propose également d'utiliser ces informations pour évaluer le phénotype immunitaire de ladite tumeur solide. Dans un mode de réalisation particulier, l'invention propose de pronostiquer la survie et/ou l'efficacité de traitement de patients cancéreux traités par immunothérapie telle qu'une monothérapie par agent anti-PD-1/PD-L1.
PCT/EP2018/069169 2017-07-13 2018-07-13 Outil d'imagerie basée sur la radiomique pour surveiller l'infiltration et de lymphocytes tumoraux et le résultat chez des patients cancéreux traités par agents anti-pd-1/pd-l1 WO2019012147A1 (fr)

Priority Applications (9)

Application Number Priority Date Filing Date Title
US16/630,031 US20210003555A1 (en) 2017-07-13 2018-07-13 A Radiomics-Based Imaging Tool to Monitor Tumor-Lymphocyte Infiltration and Outcome in Cancer Patients Treated by Anti-PD-1/PD-L1
EP18740223.5A EP3652534B1 (fr) 2017-07-13 2018-07-13 Outil d'imagerie par analyse radiomique pour surveiller l'infiltration des lymphocytes tumoraux et la survie de patients cancéreux traités par anti-pd-1/pd-l1
ES18740223T ES2975288T3 (es) 2017-07-13 2018-07-13 Una herramienta de imagen basada en radiómica para monitorizar la infiltración de linfocitos tumores y el resultado en pacientes con cáncer tratados con anti-PD-1/PD-L1
CN201880058427.3A CN111094977B (zh) 2017-07-13 2018-07-13 监测抗pd-1/pd-l1治疗的肿瘤患者中肿瘤淋巴细胞浸润和预后的基于影像组学的成像工具
JP2020501259A JP2020526844A (ja) 2017-07-13 2018-07-13 抗ｐｄ−１／ｐｄ−ｌ１によって処置されたがん患者における腫瘍リンパ球浸潤及び転帰を監視するためのラジオミクスに基づくイメージングツール
AU2018298890A AU2018298890A1 (en) 2017-07-13 2018-07-13 A radiomics-based imaging tool to monitor tumor-lymphocyte infiltration and outcome in cancer patients treated by anti-PD-1/PD- L1
IL271996A IL271996B1 (en) 2017-07-13 2018-07-13 A radiomics-based imaging tool for monitoring lymphocyte infiltration into the tumor and its outcome in cancer patients treated with anti-PD-1/PD-L1 therapy
CA3069612A CA3069612A1 (fr) 2017-07-13 2018-07-13 Outil d'imagerie basee sur la radiomique pour surveiller l'infiltration et de lymphocytes tumoraux et le resultat chez des patients cancereux traites par agents anti-pd-1/pd-l1
KR1020207003496A KR20200040754A (ko) 2017-07-13 2018-07-13 항-pd-1/pd-l1에 의해 치료되는 암 환자의 종양-림프구 침윤 및 결과를 모니터링하기 위한 라디오믹스 기반 영상 툴

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US201762532139P 2017-07-13 2017-07-13
US62/532,139 2017-07-13
EP18305680.3 2018-06-01
EP18305680 2018-06-01

Publications (1)

Publication Number Publication Date
WO2019012147A1 true WO2019012147A1 (fr) 2019-01-17

Family

ID=62683137

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2018/069169 WO2019012147A1 (fr) 2017-07-13 2018-07-13 Outil d'imagerie basée sur la radiomique pour surveiller l'infiltration et de lymphocytes tumoraux et le résultat chez des patients cancéreux traités par agents anti-pd-1/pd-l1

Country Status (1)

Country Link
WO (1) WO2019012147A1 (fr)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112037167A (zh) * 2020-07-21 2020-12-04 苏州动影信息科技有限公司 一种基于影像组学和遗传算法的目标区域确定系统
EP3770908A1 (fr) * 2019-07-22 2021-01-27 Koninklijke Philips N.V. Évaluation des risques à base de radiologie pour les résultats obtenus par biopsie
WO2021016502A1 (fr) * 2019-07-24 2021-01-28 Genentech, Inc. Détection de tumeurs à programmation neuronale à l'aide de données d'expression
WO2021028763A1 (fr) * 2019-08-09 2021-02-18 International Business Machines Corporation Classification de bactéries
WO2021030784A1 (fr) * 2019-08-15 2021-02-18 H. Lee Moffitt Cancer Center And Research Institute Inc. Signature radiomique pour prédire la réponse d'immunothérapie du cancer du poumon
WO2021228888A1 (fr) 2020-05-12 2021-11-18 Asylia Diagnostics Biomarqueurs pour maladie hyperprogressive et réponse thérapeutique après immunothérapie
US11254744B2 (en) 2015-08-07 2022-02-22 Imaginab, Inc. Antigen binding constructs to target molecules
CN114190958A (zh) * 2021-11-08 2022-03-18 南方医科大学南方医院 一种基于影像组学的pd-1疗效预测模型及其构建方法
WO2023001874A1 (fr) 2021-07-22 2023-01-26 Asylia Diagnostics Bv Biomarqueurs pour réponse thérapeutique après immunothérapie
US11568992B2 (en) 2020-07-24 2023-01-31 Onc.Ai, Inc. Predicting response to immunotherapy treatment using deep learning analysis of imaging and clinical data
US11959797B2 (en) 2019-08-20 2024-04-16 Beijing Taifang Technology Co., Ltd. Online detection device and method for piezoelectric device
WO2024104931A1 (fr) * 2022-11-14 2024-05-23 F. Hoffmann-La Roche Ag Nouveau biomarqueur pour prédire l'efficacité d'une immunothérapie anticancéreuse

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016083791A1 (fr) * 2014-11-24 2016-06-02 The Institute Of Cancer Research: Royal Cancer Hospital Évaluation de l'infiltration de la tumeur par des lymphocytes
US20170193175A1 (en) * 2015-12-30 2017-07-06 Case Western Reserve University Prediction of recurrence of non-small cell lung cancer

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016083791A1 (fr) * 2014-11-24 2016-06-02 The Institute Of Cancer Research: Royal Cancer Hospital Évaluation de l'infiltration de la tumeur par des lymphocytes
US20170193175A1 (en) * 2015-12-30 2017-07-06 Case Western Reserve University Prediction of recurrence of non-small cell lung cancer

Non-Patent Citations (68)

* Cited by examiner, † Cited by third party
Title
AERTS HJWL; VELAZQUEZ ER; LEIJENAAR RTH ET AL.: "Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach", NAT. COMMUN., vol. 5, 2014, pages 4006
ANSELL SM.: "Hodgkin Lymphoma: Diagnosis and Treatment", MAYO CLIN. PROC., vol. 90, no. 11, 2015, pages 1574 - 1583
ANTONIA SJ; VILLEGAS A; DANIEL D ET AL.: "Durvalumab after Chemoradiotherapy in Stage III Non-Small-Cell Lung Cancer", N ENGL J MED, vol. 377, 2017, pages 1919 - 29
ARKENAU H-T; BARRIUSO J; OLMOS D ET AL.: "Prospective validation of a prognostic score to improve patient selection for oncology phase I trials", J CLIN ONCOL, vol. 27, 2009, pages 2692 - 6
BALERMPAS P; MICHEL Y; WAGENBLAST J ET AL.: "Tumour-infiltrating lymphocytes predict response to definitive chemoradiotherapy in head and neck cancer", BR. J. CANCER, vol. 110, no. 2, 2014, pages 501 - 509
BECHT E; GIRALDO NA; LACROIX L ET AL.: "Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression", GENOME BIOL., vol. 17, no. 1, 2016, pages 218
BRAMAN NM; ETESAMI M; PRASANNA P ET AL.: "Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI", BREAST CANCER RES, vol. 19, 2017, pages 57
CATALDO SD; FICARRA E: "Mining textural knowledge in biological images: applications, methods and trends", COMPUT STRUCT BIOTECHNOL J, 2016
CHAMPIAT S; DERCLE L; AMMARI S ET AL.: "Hyperprogressive Disease Is a New Pattern of Progression in Cancer Patients Treated by Anti-PD-1/PD-L1", CLIN. CANCER RES., vol. 23, no. 8, 2017, pages 1920 - 1928
CHEN DS; MELLMAN I.: "Elements of cancer immunity and the cancer-immune set point", NATURE, vol. 541, no. 7637, 2017, pages 321 - 330
CHEN R-Y; LIN Y-C; SHEN W-C ET AL.: "Associations of Tumor PD-1 Ligands, Immunohistochemical Studies, and Textural Features in18F-FDG PET in Squamous Cell Carcinoma of the Head and Neck", SCI REP, vol. 8, 2018, pages 105
CLARK K; VENDT B; SMITH K ET AL.: "The Cancer Imaging Archive (TCIA): maintaining and operating a public information repository", J. DIGIT. IMAGING, vol. 26, no. 6, 2013, pages 1045 - 1057, XP035345958, DOI: doi:10.1007/s10278-013-9622-7
CLEMENTE CG; MIHM MC JR; BUFALINO R ET AL.: "Prognostic value of tumor infiltrating lymphocytes in the vertical growth phase of primary cutaneous melanoma", CANCER, vol. 77, no. 7, 1996, pages 1303 - 1310, XP000647023, DOI: doi:10.1002/(SICI)1097-0142(19960401)77:7<1303::AID-CNCR12>3.0.CO;2-5
EISENHAUERA EA; THERASSEB P; BOGAERTSC J ET AL.: "New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1", EUR J CANCER, vol. 45, 2009, pages 228 - 247, XP025841550, DOI: doi:10.1016/j.ejca.2008.10.026
FLETCHER JG; LENG S; YU L; MCCOLLOUGH CH: "Uncertainty in CT Images", RADIOLOGY, vol. 279, 2016, pages 5 - 10
GAJEWSKI TF; WOO S-R; ZHA Y ET AL.: "Cancer immunotherapy strategies based on overcoming barriers within the tumor microenvironment", CURR. OPIN. IMMUNOL., vol. 25, no. 2, 2013, pages 268 - 276, XP055454050, DOI: doi:10.1016/j.coi.2013.02.009
GALLOWAY MM: "Texture analysis using gray level run lengths", COMPUTER GRAPHICS AND IMAGE PROCESSING, vol. 4, 1975, pages 172 - 9, XP055086925, DOI: doi:10.1016/S0146-664X(75)80008-6
GANDHI L; RODRIGUEZ-ABREU D; GADGEEL S ET AL.: "Pembrolizumab plus Chemotherapy in Metastatic Non-Small-Cell Lung Cancer", N ENGL J MED, 2018
GILLIES RJ; KINAHAN PE; HRICAK H.: "Radiomics: Images Are More than Pictures", THEY ARE DATA. RADIOLOGY, vol. 278, no. 2, 2016, pages 563 - 577
GROSSMANN P; STRINGFIELD O; EL-HACHEM N ET AL.: "Defining the biological basis of radiomic phenotypes in lung cancer", ELIFE, 2017, pages 6
HEGDE PS; KARANIKAS V; EVERS S.: "The Where, the When, and the How of Immune Monitoring for Cancer Immunotherapies in the Era of Checkpoint Inhibition", CLIN. CANCER RES., vol. 22, no. 8, 2016, pages 1865 - 1874
HELLMANN MD; RIZVI NA; GOLDMAN JW ET AL.: "Nivolumab plus ipilimumab as first-line treatment for advanced non-small-cell lung cancer (CheckMate 012): results of an open-label, phase 1, multicohort study", LANCET ONCOL., vol. 18, no. 1, 2017, pages 31 - 41, XP029868855, DOI: doi:10.1016/S1470-2045(16)30624-6
HERBST RS; SORIA J-C; KOWANETZ M ET AL.: "Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients", NATURE, vol. 515, no. 7528, 2014, pages 563 - 567, XP055262130, DOI: doi:10.1038/nature14011
HUGO W; ZARETSKY JM; SUN L ET AL.: "Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma", CELL, vol. 165, no. 1, 2016, pages 35 - 44, XP029473850, DOI: doi:10.1016/j.cell.2016.02.065
JOYCE JA; FEARON DT.: "T cell exclusion, immune privilege, and the tumor microenvironment", SCIENCE, vol. 348, no. 6230, 2015, pages 74 - 80
KIM JM; CHEN DS.: "Immune escape to PD-L1/PD-1 blockade: seven steps to success (or failure", ANN. ONCOL., vol. 27, no. 8, 2016, pages 1492 - 1504, XP055390469, DOI: doi:10.1093/annonc/mdw217
KUHN M; WING J; WESTON S ET AL., CLASSIFICATION AND REGRESSION TRAINING, 2017
LEIJENAAR RTH; NALBANTOV G; CARVALHO S ET AL.: "The effect of SUV discretization in quantitative FDG-PET Radiomics: the need for standardized methodology in tumor texture analysis", SCI. REP., vol. 5, 2015, pages 11075
LIAW A; WIENER M.: "Classification and Regression by randomForest", R NEWS, vol. 2, no. 3, 2002, pages 18 - 22, XP055305332
LIMKIN EJ; SUN R; DERCLE L ET AL.: "Promises and challenges for the implementation of computational medical imaging (radiomics) in oncology", ANN. ONCOL., vol. 28, no. 6, 2017, pages 1191 - 1206
MASSARD C; MICHIELS S; FERTE C ET AL.: "High-Throughput Genomics and Clinical Outcome in Hard-to-Treat Advanced Cancers: Results of the MOSCATO 01 Trial", CANCER DISCOV., vol. 7, no. 6, 2017, pages 586 - 595
MELERO I; ROUZAUT A; MOTZ GT: "Coukos G. T-cell and NK-cell infiltration into solid tumors: a key limiting factor for efficacious cancer immunotherapy", CANCER DISCOV., vol. 4, no. 5, 2014, pages 522 - 526
MILES KA; GANESHAN B; GRIFFITHS MR; YOUNG RCD; CHATWIN CR: "Colorectal cancer: texture analysis of portal phase hepatic CT images as a potential marker of survival", RADIOLOGY, vol. 250, 2009, pages 444 - 52
MOTZER RJ; ESCUDIER B; MCDERMOTT DF ET AL.: "Nivolumab versus Everolimus in Advanced Renal-Cell Carcinoma", N. ENGL. J. MED., vol. 373, no. 19, 2015, pages 1803 - 1813
NATHANIEL M. BRAMAN ET AL: "Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI", BREAST CANCER RESEARCH, vol. 19, no. 1, 18 May 2017 (2017-05-18), XP055506030, DOI: 10.1186/s13058-017-0846-1 *
NIOCHE C; ORLHAC F; BOUGHDAD S ET AL.: "A freeware for tumor heterogeneity characterization in PET, SPECT, CT, MRI and US to accelerate advances in radiomics", J NUCL MED, vol. 58, no. 1, 2017, pages 1316 - 1316
NIOCHE C; ORLHAC F; BOUGHDAD S; REUZE S; GOYA-OUTI J; ROBERT C; PELLOT-BARAKAT; C, SOUSSAN M; FROUIN F; BUVAT I: "LIFEx: a freeware for radiomic feature calculation in multimodality imaging to accelerate advances in the characterization of tumor heterogeneity", CANCER RES., 29 June 2018 (2018-06-29)
O'CONNOR JPB; ROSE CJ; WATERTON JC ET AL.: "Imaging intratumor heterogeneity: role in therapy response, resistance, and clinical outcome", CLIN. CANCER RES., vol. 21, no. 2, 2015, pages 249 - 257
OGINO T; SHIGYO H; ISHII H ET AL.: "HLA class I antigen down-regulation in primary laryngeal squamous cell carcinoma lesions as a poor prognostic marker", CANCER RES., vol. 66, no. 18, 2006, pages 9281 - 9289
ORLHAC F; NIOCHE C; SOUSSAN M: "Buvat I. Understanding Changes in Tumor Texture Indices in PET: A Comparison Between Visual Assessment and Index Values in Simulated and Patient Data", J NUCL MED, vol. 58, 2017, pages 387 - 92
OROOJI M; RAKSHIT S; BEIG N ET AL.: "Computerized textural analysis of lung CT to enable quantification of tumor infiltrating lymphocytes in NSCLC", J. CLIN. ONCOL., vol. 34, no. 15, 2016, pages 11584 - 11584
PAGES F; BERGER A; CAMUS M ET AL.: "Effector memory T cells, early metastasis, and survival in colorectal cancer", N. ENGL. J. MED., vol. 353, no. 25, 2005, pages 2654 - 2666, XP002390433, DOI: doi:10.1056/NEJMoa051424
PAO W; OOI C-H; BIRZELE F ET AL.: "Tissue-Specific Immunoregulation: A Call for Better Understanding of the 'Immunostat' in the Context of Cancer", CANCER DISCOV, vol. 8, 2018, pages 395 - 402
PATRICK GROSSMANN ET AL: "Defining the biological basis of radiomic phenotypes in lung cancer", LIFE 2017;6:E23421, 21 July 2017 (2017-07-21), pages 1 - 48, XP055506071, Retrieved from the Internet <URL:https://doi.org/10.7554/eLife.23421.001> [retrieved on 20180911], DOI: 10.7554/eLife.23421 *
POSTOW MA; CHESNEY J; PAVLICK AC ET AL.: "Nivolumab and ipilimumab versus ipilimumab in untreated melanoma", N. ENGL. J. MED., vol. 372, no. 21, 2015, pages 2006 - 2017, XP002761464, DOI: doi:10.1056/NEJMoa1414428
PRAWIRA A; DUFORT P; HALANKAR J ET AL.: "Development of a predictive radiomics signature for response to immune checkpoint inhibitors (ICIs) in patients with recurrent or metastatic squamous cell carcinoma of the head and neck (RM-SCCHN", ANN. ONCOL., 2016
Q WEN ET AL: "Radiomic CT Features for Evaluation of PD-L1, CD8+TILs and Foxp3+TILs Expression Status in Patients with Stage I NSCLC", INTERNATIONAL JOURNAL OF RADIATION ONCOLOGY, VOLUME 99, ISSUE 2, SUPPLEMENT,, 1 October 2017 (2017-10-01), pages E738 - E738, XP055506121, Retrieved from the Internet <URL:https://www.redjournal.org/article/S0360-3016(17)33431-4/fulltext> [retrieved on 20180911] *
R CORE TEAM: "R: A Language and Environment for Statistical Computing", R FOUNDATION FOR STATISTICAL COMPUTING, 2017
R SUN ET AL: "A novel radiomic based imaging tool to monitor tumor lymphocyte infiltration and outcome of patients treated by anti-PD-1/PD-L1", ANNALS OF ONCOLOGY, VOLUME 28, ISSUE SUPPL.5, 1 September 2017 (2017-09-01), pages 1 - 1, XP055506096, Retrieved from the Internet <URL:https://academic.oup.com/annonc/article/28/suppl_5/mdx390.004/4109674> [retrieved on 20180911] *
RECK M; RODRIGUEZ-ABREU D; ROBINSON AG ET AL.: "Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer", N. ENGL. J. MED., vol. 375, no. 19, 2016, pages 1823 - 1833
REUZE S; ORLHAC F; CHARGARI C ET AL.: "Prediction of cervical cancer recurrence using textural features extracted from 18F-FDG PET images acquired with different scanners", ONCOTARGET, vol. 8, no. 26, 2017, pages 43169 - 43179
RIBAS A; SHIN DS; ZARETSKY J ET AL.: "PD-1 Blockade Expands Intratumoral Memory T Cells", CANCER IMMUNOL RES, vol. 4, no. 3, 2016, pages 194 - 203, XP055380539, DOI: doi:10.1158/2326-6066.CIR-15-0210
ROBERT C; LONG GV; BRADY B ET AL.: "Nivolumab in previously untreated melanoma without BRAF mutation", N. ENGL. J. MED., vol. 372, no. 4, 2015, pages 320 - 330
ROGER SUN ET AL: "A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: an imaging biomarker, retrospective multicohort study", THE LANCET ONCOLOGY, VOLUME 19, ISSUE 9, SEPTEMBER 2018, 1 September 2018 (2018-09-01), pages 1180 - 1191, XP055505301, Retrieved from the Internet <URL:https://reader.elsevier.com/reader/sd/0903C829240DE915841D582E9F4CF731F114E5EAEB4865030EDAE1AB88C85AC2C671936084BAD06E6D1DDAE8476099DD> [retrieved on 20180906] *
ROSENBERG JE; HOFFMAN-CENSITS J; POWLES T ET AL.: "Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial", LANCET, vol. 387, no. 10031, 2016, pages 1909 - 1920, XP029530539, DOI: doi:10.1016/S0140-6736(16)00561-4
SALMON H; FRANCISZKIEWICZ K; DAMOTTE D ET AL.: "Matrix architecture defines the preferential localization and migration of T cells into the stroma of human lung tumors", J. CLIN. INVEST., vol. 122, no. 3, 2012, pages 899 - 910
SATO E; OLSON SH; AHN J ET AL.: "Intraepithelial CD8+ tumor-infiltrating lymphocytes and a high CD8+/regulatory T cell ratio are associated with favorable prognosis in ovarian cancer", PROC. NATL. ACAD. SCI. U. S. A., vol. 102, no. 51, 2005, pages 18538 - 18543, XP002687398, DOI: doi:10.1073/PNAS.0509182102
SUN R; CHAMPIAT S; DERCLE L ET AL.: "Baseline lymphopenia should not be used as exclusion criteria in early clinical trials investigating immune checkpoint blockers (PD-1/PD-L1 inhibitors", EUR J CANCER, vol. 84, 2017, pages 202 - 11
SUN R; ORLHAC F; ROBERT C ET AL.: "In Regard to Mattonen et al. Int. J. Radiat", ONCOL. BIOL. PHYS., vol. 95, no. 5, 2016, pages 1544 - 1545
TANG C; AMER A; HOBBS B ET AL.: "Pathology-Based Non-Small Cell Lung Cancer Radiomics Signature Describing the Local Tumor Immune Environment: Discovery and Validation", INT. J. RADIAT. ONCOL. BIOL. PHYS., vol. 96, no. 2, 2016, pages S42 - S43
TANG C; HOBBS B; AMER A ET AL.: "Development of an Immune-Pathology Informed Radiomics Model for Non-Small Cell Lung Cancer", SCI REP, vol. 8, 2018, pages 1922
TOKITO TAKAAKI ET AL: "Predictive relevance of PD-L1 expression combined with CD8+ TIL density in stage III non-small cell lung cancer patients receiving concurrent chemoradiotherapy", EUROPEAN JOURNAL OF CANCER, ELSEVIER, AMSTERDAM, NL, vol. 55, 6 January 2016 (2016-01-06), pages 7 - 14, XP029428431, ISSN: 0959-8049, DOI: 10.1016/J.EJCA.2015.11.020 *
TUMEH PC; HARVIEW CL; YEARLEY JH ET AL.: "PD-1 blockade induces responses by inhibiting adaptive immune resistance", NATURE, vol. 515, no. 7528, 2014, pages 568 - 571, XP055247294, DOI: doi:10.1038/nature13954
WANSOM D; LIGHT E; THOMAS D ET AL.: "Infiltrating lymphocytes and human papillomavirus-16--associated oropharyngeal cancer", LARYNGOSCOPE, vol. 122, no. 1, 2012, pages 121 - 127
WEBER JS; KUDCHADKAR RR; YU B ET AL.: "Safety, efficacy, and biomarkers of nivolumab with vaccine in ipilimumab-refractory or -naive melanoma", J. CLIN. ONCOL., vol. 31, no. 34, 2013, pages 4311 - 4318, XP055259778, DOI: doi:10.1200/JCO.2013.51.4802
WEINSTEIN JN; COLLISSON EA ET AL.: "Nat Genet", vol. 45, 2013, CANCER GENOME ATLAS RESEARCH NETWORK, article "The Cancer Genome Atlas Pan-Cancer analysis project", pages: 1113 - 20
ZOU H; HASTIE T.: "Regularization and variable selection via the elastic net", J R STAT SOC SERIES B STAT METHODOL, vol. 67, 2005, pages 301 - 20
ZULEY, M. L.; JAROSZ, R.; KIRK, S.; LEE, Y.; COLEN, R.; GARCIA, K.; AREDES, N. D.: "Radiology Data from The Cancer Genome Atlas Head-Neck Squamous Cell Carcinoma [TCGA-HNSC] collection", THE CANCER IMAGING ARCHIVE, 2016, Retrieved from the Internet <URL:http://doi.org/1 0.7937/K9/TCIA.2016.LXKQ47MS>

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11254744B2 (en) 2015-08-07 2022-02-22 Imaginab, Inc. Antigen binding constructs to target molecules
EP3770908A1 (fr) * 2019-07-22 2021-01-27 Koninklijke Philips N.V. Évaluation des risques à base de radiologie pour les résultats obtenus par biopsie
WO2021013858A1 (fr) * 2019-07-22 2021-01-28 Koninklijke Philips N.V. Évaluation de risque basée sur la radiologie pour des scores basés sur une biopsie
WO2021016502A1 (fr) * 2019-07-24 2021-01-28 Genentech, Inc. Détection de tumeurs à programmation neuronale à l'aide de données d'expression
GB2600891A (en) * 2019-08-09 2022-05-11 Ibm Bacteria classification
WO2021028763A1 (fr) * 2019-08-09 2021-02-18 International Business Machines Corporation Classification de bactéries
GB2600891B (en) * 2019-08-09 2023-02-08 Ibm Bacteria classification
WO2021030784A1 (fr) * 2019-08-15 2021-02-18 H. Lee Moffitt Cancer Center And Research Institute Inc. Signature radiomique pour prédire la réponse d'immunothérapie du cancer du poumon
US20220170909A1 (en) * 2019-08-15 2022-06-02 H. Lee Moffitt Cancer Center And Research Institute, Inc. Radiomic signature for predicting lung cancer immunotherapy response
US11959797B2 (en) 2019-08-20 2024-04-16 Beijing Taifang Technology Co., Ltd. Online detection device and method for piezoelectric device
WO2021228888A1 (fr) 2020-05-12 2021-11-18 Asylia Diagnostics Biomarqueurs pour maladie hyperprogressive et réponse thérapeutique après immunothérapie
CN112037167B (zh) * 2020-07-21 2023-11-24 苏州动影信息科技有限公司 一种基于影像组学和遗传算法的目标区域确定系统
CN112037167A (zh) * 2020-07-21 2020-12-04 苏州动影信息科技有限公司 一种基于影像组学和遗传算法的目标区域确定系统
US11568992B2 (en) 2020-07-24 2023-01-31 Onc.Ai, Inc. Predicting response to immunotherapy treatment using deep learning analysis of imaging and clinical data
WO2023001874A1 (fr) 2021-07-22 2023-01-26 Asylia Diagnostics Bv Biomarqueurs pour réponse thérapeutique après immunothérapie
CN114190958B (zh) * 2021-11-08 2022-11-22 南方医科大学南方医院 一种基于影像组学的pd-1疗效预测模型及其构建方法
CN114190958A (zh) * 2021-11-08 2022-03-18 南方医科大学南方医院 一种基于影像组学的pd-1疗效预测模型及其构建方法
WO2024104931A1 (fr) * 2022-11-14 2024-05-23 F. Hoffmann-La Roche Ag Nouveau biomarqueur pour prédire l'efficacité d'une immunothérapie anticancéreuse

Similar Documents

Publication Publication Date Title
EP3652534B1 (fr) Outil d&#39;imagerie par analyse radiomique pour surveiller l&#39;infiltration des lymphocytes tumoraux et la survie de patients cancéreux traités par anti-pd-1/pd-l1
WO2019012147A1 (fr) Outil d&#39;imagerie basée sur la radiomique pour surveiller l&#39;infiltration et de lymphocytes tumoraux et le résultat chez des patients cancéreux traités par agents anti-pd-1/pd-l1
Sun et al. A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: an imaging biomarker, retrospective multicohort study
Khorrami et al. Changes in CT radiomic features associated with lymphocyte distribution predict overall survival and response to immunotherapy in non–small cell lung cancer
Valentinuzzi et al. [F] FDG PET immunotherapy radiomics signature (iRADIOMICS) predicts response of non-small-cell lung cancer patients treated with pembrolizumab
Chen et al. Radiomic features analysis in computed tomography images of lung nodule classification
Banna et al. The promise of digital biopsy for the prediction of tumor molecular features and clinical outcomes associated with immunotherapy
US9984199B2 (en) Method and system for classification and quantitative analysis of cell types in microscopy images
Yip et al. Impact of experimental design on PET radiomics in predicting somatic mutation status
Yip et al. Relationship between the temporal changes in positron-emission-tomography-imaging-based textural features and pathologic response and survival in esophageal cancer patients
ES2730384T3 (es) Puntuación de la infiltración linfocitaria de los tumores
Lu et al. Predicting therapeutic antibody delivery into human head and neck cancers
Yachoui et al. Pulmonary MALT lymphoma in patients with Sjögren’s syndrome
Sadot et al. Cholangiocarcinoma: correlation between molecular profiling and imaging phenotypes
US20210210169A1 (en) Morphometric genotyping of cells in liquid biopsy using optical tomography
JP2023537743A (ja) 連続バイオマーカー予測のための電子画像を処理するためのシステム及び方法
Li et al. Prognostic and predictive values of metabolic parameters of 18F-FDG PET/CT in patients with non-small cell lung cancer treated with chemotherapy
JP2024509576A (ja) 明細胞腎細胞がんを有する患者における治療に対する応答の予測
Castello et al. Tumor heterogeneity, hypoxia, and immune markers in surgically resected non-small-cell lung cancer
Park et al. Prognostic utility of FDG PET/CT and bone scintigraphy in breast cancer patients with bone-only metastasis
Hörst et al. Histology-based prediction of therapy response to neoadjuvant chemotherapy for esophageal and esophagogastric junction adenocarcinomas using deep learning
Zhang et al. Development of machine learning models integrating PET/CT radiomic and immunohistochemical pathomic features for treatment strategy choice of cervical cancer with negative pelvic lymph node by mediating COX-2 expression
Wang et al. Exploring tumor heterogeneity in colorectal liver metastases by imaging: Unsupervised machine learning of preoperative CT radiomics features for prognostic stratification
Xu et al. Using histopathology images to predict chromosomal instability in breast cancer: a deep learning approach
Liu et al. Prediction of tumor mutation load in colorectal cancer histopathological images based on deep learning

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18740223

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 3069612

Country of ref document: CA

Ref document number: 2020501259

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018298890

Country of ref document: AU

Date of ref document: 20180713

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2018740223

Country of ref document: EP

Effective date: 20200213