ABSTRACT
Anti-PD(L)-1 inhibition combined with platinum doublet chemotherapy (Chemo-IO) has become the most frequently used standard of care regimen in patients with non-small cell lung cancer (NSCLC). The negative impact of antibiotics on clinical outcomes prior to anti-PD(L)-1 inhibition monotherapy (IO) has been demonstrated in multiple studies, but the impact of antibiotic exposure prior to initiation of Chemo-IO is controversial. We assessed antibiotic exposures at two time windows: within 60 days prior to therapy (-60 d window) and within 60 days prior to therapy and 42 days after therapy (-60 + 42d window) in 2028 patients with advanced NSCLC treated with Chemo-IO and IO monotherapy focusing on objective response rate (ORR: rate of partial response and complete response), progression-free survival (PFS), and overall survival (OS). We also assessed impact of antibiotic exposure in an independent cohort of 53 patients. Univariable and multivariable analyses were conducted along with a meta-analysis from similar studies. For the -60 d window, in the Chemo-IO group (N = 769), 183 (24%) patients received antibiotics. Antibiotic exposure was associated with worse ORR (27% vs 40%, p = 0.001), shorter PFS (3.9 months vs. 5.9 months, hazard ratio [HR] 1.35, 95%CI 1.1,1.6, p = 0.0012), as well as shorter OS (10 months vs. 15 months, HR 1.50, 95%CI 1.2,1.8, p = 0.00014). After adjusting for known prognostic factors in NSCLC, antibiotic exposure was independently associated with worse PFS (HR 1.39, 95%CI 1.35,1.7, p = 0.002) and OS (HR 1.61, 95%CI 1.28,2.03, p < 0.001). Similar results were obtained in the -60 + 42d window, and also in an independent cohort. In a meta-analysis of patients with NSCLC treated with Chemo-IO (N = 4) or IO monotherapy (N = 13 studies) antibiotic exposure before treatment was associated with worse OS among all patients (n = 11,351) (HR 1.93, 95% CI 1.52, 2.45) and Chemo-IO-treated patients (n = 1201) (HR 1.54, 95% CI 1.28, 1.84). Thus, antibiotics exposure prior to Chemo-IO is common and associated with worse outcomes, even after adjusting for other factors. These results highlight the need to implement antibiotic stewardship in routine oncology practice.
ABSTRACT
Gut microbiota impacts responses to immune checkpoint inhibitors (ICI). A high level of Faecalibacterium prausnitzii have been associated with a positive response to ICI in multiple cancer types. Here, based on fecal shotgun metagenomics data, we show in two independent cohorts of patients with non-small cell lung cancer and advanced melanoma that a high level of F. prausnitzii at baseline is positively associated with a better clinical response to ICI. In MCA205 tumor-bearing mice, administration of F. prausnitzii strain EXL01, already in clinical development for Inflammatory Bowel Disease, restores the anti-tumor response to ICI in the context of antibiotic-induced microbiota perturbation at clinical and tumor transcriptomics level. In vitro, EXL01 strain enhances T cell activation in the presence of ICI. Interestingly, oral administration of EXL01 strain did not induce any change in fecal microbiota diversity or composition, suggesting a direct effect on immune cells in the small intestine. F. prausnitzii strain EXL01 will be evaluated as an adjuvant to ICI in multiple cancers in the near future.
Subject(s)
Faecalibacterium prausnitzii , Gastrointestinal Microbiome , Immune Checkpoint Inhibitors , Immune Checkpoint Inhibitors/pharmacology , Immune Checkpoint Inhibitors/therapeutic use , Animals , Humans , Mice , Gastrointestinal Microbiome/drug effects , Faecalibacterium prausnitzii/drug effects , Female , Carcinoma, Non-Small-Cell Lung/drug therapy , Carcinoma, Non-Small-Cell Lung/immunology , Carcinoma, Non-Small-Cell Lung/pathology , Melanoma/drug therapy , Melanoma/immunology , Melanoma/pathology , Feces/microbiology , Male , Lung Neoplasms/drug therapy , Lung Neoplasms/immunology , Lung Neoplasms/pathology , Cell Line, Tumor , Mice, Inbred C57BLABSTRACT
The gut microbiota influences the clinical responses of cancer patients to immunecheckpoint inhibitors (ICIs). However, there is no consensus definition of detrimental dysbiosis. Based on metagenomics (MG) sequencing of 245 non-small cell lung cancer (NSCLC) patient feces, we constructed species-level co-abundance networks that were clustered into species-interacting groups (SIGs) correlating with overall survival. Thirty-seven and forty-five MG species (MGSs) were associated with resistance (SIG1) and response (SIG2) to ICIs, respectively. When combined with the quantification of Akkermansia species, this procedure allowed a person-based calculation of a topological score (TOPOSCORE) that was validated in an additional 254 NSCLC patients and in 216 genitourinary cancer patients. Finally, this TOPOSCORE was translated into a 21-bacterial probe set-based qPCR scoring that was validated in a prospective cohort of NSCLC patients as well as in colorectal and melanoma patients. This approach could represent a dynamic diagnosis tool for intestinal dysbiosis to guide personalized microbiota-centered interventions.
Subject(s)
Carcinoma, Non-Small-Cell Lung , Gastrointestinal Microbiome , Immunotherapy , Lung Neoplasms , Neoplasms , Female , Humans , Male , Akkermansia , Carcinoma, Non-Small-Cell Lung/microbiology , Carcinoma, Non-Small-Cell Lung/drug therapy , Carcinoma, Non-Small-Cell Lung/immunology , Dysbiosis/microbiology , Feces/microbiology , Gastrointestinal Microbiome/drug effects , Immune Checkpoint Inhibitors/therapeutic use , Immune Checkpoint Inhibitors/pharmacology , Immunotherapy/methods , Lung Neoplasms/microbiology , Lung Neoplasms/drug therapy , Metagenomics/methods , Neoplasms/microbiology , Treatment OutcomeABSTRACT
Tumor immunosurveillance plays a major role in melanoma, prompting the development of immunotherapy strategies. The gut microbiota composition, influencing peripheral and tumoral immune tonus, earned its credentials among predictors of survival in melanoma. The MIND-DC phase III trial (NCT02993315) randomized (2:1 ratio) 148 patients with stage IIIB/C melanoma to adjuvant treatment with autologous natural dendritic cell (nDC) or placebo (PL). Overall, 144 patients collected serum and stool samples before and after 2 bimonthly injections to perform metabolomics (MB) and metagenomics (MG) as prespecified exploratory analysis. Clinical outcomes are reported separately. Here we show that different microbes were associated with prognosis, with the health-related Faecalibacterium prausnitzii standing out as the main beneficial taxon for no recurrence at 2 years (p = 0.008 at baseline, nDC arm). Therapy coincided with major MB perturbations (acylcarnitines, carboxylic and fatty acids). Despite randomization, nDC arm exhibited MG and MB bias at baseline: relative under-representation of F. prausnitzii, and perturbations of primary biliary acids (BA). F. prausnitzii anticorrelated with BA, medium- and long-chain acylcarnitines. Combined, these MG and MB biomarkers markedly determined prognosis. Altogether, the host-microbial interaction may play a role in localized melanoma. We value systematic MG and MB profiling in randomized trials to avoid baseline differences attributed to host-microbe interactions.
Subject(s)
Melanoma , Microbiota , Humans , Metabolic Reprogramming , Microbiota/genetics , Dendritic CellsABSTRACT
Background: Immune checkpoint inhibitors (ICIs) have revolutionized non-small cell lung cancers (NSCLCs) treatment, but only 20-30% of patients benefit from these treatments. Currently, PD-L1 expression in tumor cells is the only clinically approved predictor of ICI response in lung cancer, but concerns arise due to its low negative and positive predictive value. Recent studies suggest that CXCL13+ T cells in the tumor microenvironment (TME) may be a good predictor of response. We aimed to assess if CXCL13+ cell localization within the TME can predict ICI response in advanced NSCLC patients. Methods: This retrospective study included 65 advanced NSCLC patients treated with Nivolumab/Pembrolizumab at IUCPQ or CHUM and for whom a pretreatment surgical specimen was available. Good responders were defined as having a complete radiologic response at 1 year, and bad responders were defined as showing cancer progression at 1 year. IHC staining for CXCL13 was carried out on a representative slide from a resection specimen, and CXCL13+ cell density was evaluated in tumor (T), invasive margin (IM), non-tumor (NT), and tertiary lymphoid structure (TLS) compartments. Cox models were used to analyze progression-free survival (PFS) and overall survival (OS) probability, while the Mann-Whitney test was used to compare CXCL13+ cell density between responders and non-responders. Results: We showed that CXCL13+ cell density localization within the TME is associated with ICI efficacy. An increased density of CXCL13+ cells across all compartments was associated with a poorer prognostic (OS; HR = 1.22; 95%CI = 1.04-1.42; p = 0.01, PFS; HR = 1.16; p = 0.02), or a better prognostic when colocalized within TLSs (PFS; HR = 0.84, p = 0.03). Conclusion: Our results support the role of CXCL13+ cells in advanced NSCLC patients, with favorable prognosis when localized within TLSs and unfavorable prognosis when present elsewhere. The concomitant proximity of CXCL13+ and CD20+ cells within TLSs may favor antigen presentation to T cells, thus enhancing the effect of PD-1/PD-L1 axis inhibition. Further validation is warranted to confirm the potential relevance of this biomarker in a clinical setting.
ABSTRACT
BACKGROUND: Immune checkpoint inhibitors (ICIs) have emerged as one of the most promising first-line therapeutics in the management of non-small cell lung cancer (NSCLC). However, only a subset of these patients responds to ICIs, highlighting the clinical need to develop better predictive and prognostic biomarkers. This study will leverage pre-treatment imaging profiles to develop survival risk models for NSCLC patients treated with first-line immunotherapy. METHODS: Advanced NSCLC patients (n = 149) were retrospectively identified from two institutions who were treated with first-line ICIs. Radiomics features extracted from pretreatment imaging scans were used to build the predictive models for progression-free survival (PFS) and overall survival (OS). A compendium of five feature selection methods and seven machine learning approaches were utilized to build the survival risk models. The concordance index (C-index) was used to evaluate model performance. RESULTS: From our results, we found several combinations of machine learning algorithms and feature selection methods to achieve similar performance. K-nearest neighbourhood (KNN) with ReliefF (RL) feature selection was the best-performing model to predict PFS (C-index = 0.61 and 0.604 in discovery and validation cohorts), while XGBoost with Mutual Information (MI) feature selection was the best-performing model for OS (C-index = 0.7 and 0.655 in discovery and validation cohorts). CONCLUSION: The results of this study highlight the importance of implementing an appropriate feature selection method coupled with a machine learning strategy to develop robust survival models. With further validation of these models on external cohorts when available, this can have the potential to improve clinical decisions by systematically analyzing routine medical images.
Subject(s)
Carcinoma, Non-Small-Cell Lung , Lung Neoplasms , Humans , Carcinoma, Non-Small-Cell Lung/diagnostic imaging , Carcinoma, Non-Small-Cell Lung/therapy , Immunotherapy , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/drug therapy , Prognosis , Radiomics , Retrospective StudiesABSTRACT
BACKGROUND: Recent advances in cancer biomarker development have led to a surge of distinct data modalities, such as medical imaging and histopathology. To develop predictive immunotherapy biomarkers, these modalities are leveraged independently, despite their orthogonality. This study aims to explore the cross-scale association between radiological scans and digitalized pathology images for immunotherapy-treated non-small cell lung cancer (NSCLC) patients. METHODS: This study involves 36 NSCLC patients who were treated with immunotherapy and for whom both radiology and pathology images were available. A total of 851 and 260 features were extracted from CT scans and cell density maps of histology images at different resolutions. We investigated the radiopathomics relationship and their association with clinical and biological endpoints. We used the Kolmogorov-Smirnov (KS) method to test the differences between the distributions of correlation coefficients with the two imaging modality features. Unsupervised clustering was done to identify which imaging modality captures poor and good survival patients. RESULTS: Our results demonstrated a significant correlation between cell density pathomics and radiomics features. Furthermore, we also found a varying distribution of correlation values between imaging-derived features and clinical endpoints. The KS test revealed that the two imaging feature distributions were different for PFS and CD8 counts, while similar for OS. In addition, clustering analysis resulted in significant differences in the two clusters generated from the radiomics and pathomics features with respect to patient survival and CD8 counts. CONCLUSION: The results of this study suggest a cross-scale association between CT scans and pathology H&E slides among ICI-treated patients. These relationships can be further explored to develop multimodal immunotherapy biomarkers to advance personalized lung cancer care.
ABSTRACT
BACKGROUND: Immunotherapies such as immune checkpoint inhibitors (ICI) are effective in multiple tumor entities but induce a plethora of side effects. Comprehensive real-world analyses are essential to identify new signals, characterize diagnostic features, enable risk assessment, determine pathomechanisms, assess effectiveness of side effect management and compare tumor outcomes. METHODS: The international online `Side-Effect Registry Immuno-Oncology´ (SERIO; www.serio-registry.org) collects rare, complex, and severe immunotherapy-induced side effects across all tumor entities with a strong focus on ICI-induced immune-related adverse events (irAE). The relational database management system (RDMS) contains structured data on patient and tumor characteristics, type of immunotherapy, treatment of side effects, and outcome of tumor and irAE. Data are captured within 25 organ modules including new modules for immune effector cell-associated neurotoxicity syndrome (ICANS) for CAR-T-cell therapies and cytokine release syndrome (CRS) for bispecific antibodies. Information on biological samples is gathered. RESULTS: A total of 1398 irAE cases have been documented by 58 centers from 13 countries in patients with 17 tumor types. IrAEs were induced by nine different immunotherapies including tebentafusp and CAR-T cell therapies, and resulted, among others, in neurological (7.6%), pulmonary (4.0%), and cardiac toxicities (2.9%). 50.0% of all irAEs were graded severe or life-threatening and 23.0% of patients received second-line therapy for steroid-refractory or steroid-dependent irAE. SERIO has contributed to 44 original publications on topics ranging from irMyocarditis to irEncephalitis to long-term persistent sequelae of immunotherapy. CONCLUSIONS: A reliable evidence base is crucial for decision-making in rare, complex or therapy-refractory irAE. SERIO can help optimize side effect management and thereby reduce morbidity and mortality induced by immunotherapy.
Subject(s)
Neoplasms , Humans , Neoplasms/drug therapy , Immunotherapy/adverse effects , Immunotherapy/methods , Medical Oncology , Registries , Steroids/therapeutic useABSTRACT
OBJECTIVES: Chemoradiation (CRT) in patients with locally advanced head and neck squamous cell cancer (HNSCC) is associated with significant toxicities, including mucositis. The gut microbiome represents an emerging hallmark of cancer and a potentially important biomarker for CRT-related adverse events. This prospective study investigated the association between the gut microbiome composition and CRT-related toxicities in patients with HNSCC, including mucositis. MATERIALS AND METHODS: Stool samples from patients diagnosed with locally advanced HNSCC were prospectively collected prior to CRT initiation and analyzed using shotgun metagenomic sequencing to evaluate gut microbiome composition at baseline. Concurrently, clinicopathologic data, survival outcomes and the incidence and grading of CRT-emergent adverse events were documented in all patients. RESULTS: A total of 52 patients were included, of whom 47 had baseline stool samples available for metagenomic analysis. Median age was 62, 83 % patients were men and 54 % had stage III-IV disease. All patients developed CRT-induced mucositis, including 42 % with severe events (i.e. CTCAE v5.0 grade ≥ 3) and 25 % who required enteral feeding. With a median follow-up of 26.5 months, patients with severe mucositis had shorter overall survival (HR = 3.3, 95 %CI 1.0-10.6; p = 0.02) and numerically shorter progression-free survival (HR = 2.8, 95 %CI, 0.8-9.6; p = 0.09). The gut microbiome beta-diversity of patients with severe mucositis differed from patients with grades 1-2 mucositis (p = 0.04), with enrichment in Mediterraneibacter (Ruminococcus gnavus) and Clostridiaceae family members, including Hungatella hathewayi. Grade 1-2 mucositis was associated with enrichment in Eubacterium rectale, Alistipes putredinis and Ruminococcaceae family members. Similar bacterial profiles were observed in patients who required enteral feeding. CONCLUSION: Patients who developed severe mucositis had decreased survival and enrichment in specific bacteria associated with mucosal inflammation. Interestingly, these same bacteria have been linked to immune checkpoint inhibitor resistance.
Subject(s)
Gastrointestinal Microbiome , Head and Neck Neoplasms , Mucositis , Male , Humans , Female , Squamous Cell Carcinoma of Head and Neck/complications , Head and Neck Neoplasms/complications , Mucositis/etiology , Prospective Studies , Chemoradiotherapy/adverse effectsABSTRACT
Over the last decade, the composition of the gut microbiota has been found to correlate with the outcomes of cancer patients treated with immunotherapy. Accumulating evidence points to the various mechanisms by which intestinal bacteria act on distal tumors and how to harness this complex ecosystem to circumvent primary resistance to immune checkpoint inhibitors. Here, we review the state of the microbiota field in the context of melanoma, the recent breakthroughs in defining microbial modes of action, and how to modulate the microbiota to enhance response to cancer immunotherapy. The host-microbe interaction may be deciphered by the use of "omics" technologies, and will guide patient stratification and the development of microbiota-centered interventions. Efforts needed to advance the field and current gaps of knowledge are also discussed.
Subject(s)
Gastrointestinal Microbiome , Melanoma , Microbiota , Neoplasms , Humans , Melanoma/therapy , Neoplasms/therapy , Immunotherapy , Host Microbial InteractionsABSTRACT
Background: Although the immune checkpoint inhibitors, nivolumab and pembrolizumab, were found to be promising in patients with advanced NSCLC, some of them either do not respond or have recurrence after an initial response. It is still unclear who will benefit from these therapies, and, hence, there is an unmet clinical need to build robust biomarkers. Methods: Patients with advanced NSCLC (N = 323) who were treated with pembrolizumab or nivolumab were retrospectively identified from two institutions. Radiomics features extracted from baseline pretreatment computed tomography scans along with the clinical variables were used to build the predictive models for overall survival (OS), progression-free survival (PFS), and programmed death-ligand 1 (PD-L1). To develop the imaging and integrative clinical-imaging predictive models, we used the XGBoost learning algorithm with ReliefF feature selection method and validated them in an independent cohort. The concordance index for OS, PFS, and area under the curve for PD-L1 was used to evaluate model performance. Results: We developed radiomics and the ensemble radiomics-clinical predictive models for OS, PFS, and PD-L1 expression. The concordance indices of the radiomics model were 0.60 and 0.61 for predicting OS and PFS and area under the curve was 0.61 for predicting PD-L1 in the validation cohort, respectively. The combined radiomics-clinical model resulted in higher performance with 0.65, 0.63, and 0.68 to predict OS, PFS, and PD-L1 in the validation cohort, respectively. Conclusions: We found that pretreatment computed tomography imaging along with clinical data can aid as predictive biomarkers for PD-L1 and survival end points. These imaging-driven approaches may prove useful to expand the therapeutic options for nonresponders and improve the selection of patients who would benefit from immune checkpoint inhibitors.
ABSTRACT
The gastrointestinal microbiome has been shown to play a key role in determining the responses to cancer immunotherapy, including immune checkpoint inhibitor (ICI) therapy and CAR-T. In patients with non-small cell lung cancer (NSCLC), increasing evidence suggests that a microbiome composition signature is associated with clinical response to ICIs as well as with the development of immune-related adverse events. In support of this, antibiotic (ATB)-related dysbiosis has been consistently linked with the deleterious impact of ICI response, shortening the overall survival (OS) among patients on ATBs prior to ICI initiation. In parallel, several preclinical experiments have unravelled various strategies using probiotics, prebiotics, diet, and fecal microbiota transplantation as new therapeutic tools to beneficially shift the microbiome and enhance ICI efficacy. These approaches are currently being evaluated in clinical trials and have achieved encouraging preliminary results. In this article, we reviewed the recent studies on the gut microbiome as a potential biomarker and an adjuvant therapy to ICIs in NSCLC patients.
Subject(s)
Carcinoma, Non-Small-Cell Lung , Gastrointestinal Microbiome , Lung Neoplasms , Humans , Lung Neoplasms/drug therapy , Carcinoma, Non-Small-Cell Lung/drug therapy , Immunotherapy , BiomarkersABSTRACT
BACKGROUND: PD-L1 expression is used to predict NSCLC response to ICIs, but its performance is suboptimal. The impact of KRAS mutations in these patients is unclear. Studies evaluating co-mutations in TP53, STK11 and KEAP1 as well as the NLR showed that they may predict the benefit of ICIs. PATIENTS & METHODS: This is a retrospective study of patients with NSCLC treated with ICIs at the CHUM between July 2015 and June 2020. OS and PFS were compared using Kaplan-Meier and logrank methods. Co-mutations in TP53, STK11 and KEAP1 as well as the NLR were accounted for. ORR and safety were compared using Wald method. RESULTS: From 100 patients with known KRAS status, 50 were mutated (KRASMut). Mutation in TP53, STK11 and KEAP1 were present, and their status known in, respectively, 19/40 (47.5 %), 8/39 (20.5 %) and 4/38 (10.5 %) patients. STK11Mut and KEAP1Mut were associated with shorter overall survival when compared with wild type tumors (respectively median OS of 3.3 vs 20.4, p = 0.0001 and 10.1 vs 17.7, p = 0.24). When KRAS status was compounded with STK11/KEAP1, KRASMut trended to a better prognosis in STK11+KEAP1WT tumors (median OS 21.1 vs 15.8 for KRASWT, p = 0.15), but not for STK11+/-KEAP1Mut tumors. The NLR was strongly impacted by STK11 (6.0Mutvs 3.6WT, p = 0.014) and TP53 (3.2Mutvs 4.8WT, p = 0.048), but not by KEAP1 or KRAS mutations. CONCLUSION: STK11Mut and KEAP1Mut are adverse predictors of ICI therapy benefit. The NLR is strongly impacted by STK11Mut but not by KEAP1Mut, suggesting differences in their resistance mechanism. In STK11-KEAP1WT tumors, KRASMut seem associated with improved survival in NSCLC patients treated with ICIs. MICROABSTRACT: Response of NSCLC to immunotherapy is not easily predictable. We conducted a retrospective study in 100 patients with NSCLC and a known KRAS status. By accounting for different co-mutations, KRAS mutation was found to be associated with a better median overall survival in STK11 and KEAP1 wild-type tumors (21.1 vs 15.8, p = 0.15). NLR was impacted by STK11, but not KEAP1 mutation, suggesting a difference in their resistance mechanism.
Subject(s)
Carcinoma, Non-Small-Cell Lung , Lung Neoplasms , Humans , Prognosis , Lung Neoplasms/drug therapy , Lung Neoplasms/genetics , Proto-Oncogene Proteins p21(ras)/genetics , Retrospective Studies , Kelch-Like ECH-Associated Protein 1/genetics , NF-E2-Related Factor 2/genetics , Carcinoma, Non-Small-Cell Lung/drug therapy , Carcinoma, Non-Small-Cell Lung/genetics , Protein Serine-Threonine Kinases/genetics , Immunotherapy , Mutation , Tumor Suppressor Protein p53/geneticsABSTRACT
A polyphasic taxonomic approach, incorporating analysis of phenotypic features, cellular fatty acid profiles, 16S rRNA gene sequences, and determination of average nucleotide identity (ANI) plus digital DNA-DNA hybridization (dDDH), was applied to characterize an anaerobic bacterial strain designated KD22T isolated from human feces. 16S rRNA gene-based phylogenetic analysis showed that strain KD22T was found to be most closely related to species of the genus Gabonibacter. At the 16S rRNA gene level, the closest species from the strain KD22T corresponded with Gabonibacter massiliensis GM7T, with a similarity of 97.58%. Cells of strain KD22T were Gram-negative coccobacillus, positive for indole and negative for catalase, nitrate reduction, oxidase, and urease activities. The fatty acid analysis demonstrated the presence of a high concentration of iso-C15:â0 (51.65%). Next, the complete whole-genome sequence of strain KD22T was 3,368,578 bp long with 42 mol% of DNA G + C contents. The DDH and ANI values between KD22T and type strains of phylogenetically related species were 67.40% and 95.43%, respectively. These phylogenetic, phenotypic, and genomic results supported the affiliation of strain KD22T as a novel bacterial species within the genus Gabonibacter. The proposed name is Gabonibacter chumensis and the type strain is KD22T (= CSUR Q8104T = DSM 115208 T).
Subject(s)
Carcinoma, Non-Small-Cell Lung , Lung Neoplasms , Humans , Phylogeny , RNA, Ribosomal, 16S/genetics , Immunotherapy , Fatty Acids , FecesABSTRACT
Background: Recent developments in artificial intelligence suggest that radiomics may represent a promising non-invasive biomarker to predict response to immune checkpoint inhibitors (ICIs). Nevertheless, validation of radiomics algorithms in independent cohorts remains a challenge due to variations in image acquisition and reconstruction. Using radiomics, we investigated the importance of scan normalization as part of a broader machine learning framework to enable model external generalizability to predict ICI response in non-small cell lung cancer (NSCLC) patients across different centers. Methods: Radiomics features were extracted and compared from 642 advanced NSCLC patients on pre-ICI scans using established open-source PyRadiomics and a proprietary DeepRadiomics deep learning technology. The population was separated into two groups: a discovery cohort of 512 NSCLC patients from three academic centers and a validation cohort that included 130 NSCLC patients from a fourth center. We harmonized images to account for variations in reconstruction kernel, slice thicknesses, and device manufacturers. Multivariable models, evaluated using cross-validation, were used to estimate the predictive value of clinical variables, PD-L1 expression, and PyRadiomics or DeepRadiomics for progression-free survival at 6 months (PFS-6). Results: The best prognostic factor for PFS-6, excluding radiomics features, was obtained with the combination of Clinical + PD-L1 expression (AUC = 0.66 in the discovery and 0.62 in the validation cohort). Without image harmonization, combining Clinical + PyRadiomics or DeepRadiomics delivered an AUC = 0.69 and 0.69, respectively, in the discovery cohort, but dropped to 0.57 and 0.52, in the validation cohort. This lack of generalizability was consistent with observations in principal component analysis clustered by CT scan parameters. Subsequently, image harmonization eliminated these clusters. The combination of Clinical + DeepRadiomics reached an AUC = 0.67 and 0.63 in the discovery and validation cohort, respectively. Conversely, the combination of Clinical + PyRadiomics failed generalizability validations, with AUC = 0.66 and 0.59. Conclusion: We demonstrated that a risk prediction model combining Clinical + DeepRadiomics was generalizable following CT scan harmonization and machine learning generalization methods. These results had similar performances to routine oncology practice using Clinical + PD-L1. This study supports the strong potential of radiomics as a future non-invasive strategy to predict ICI response in advanced NSCLC.
ABSTRACT
BACKGROUND: Immune checkpoint inhibitors (ICIs) are a great breakthrough in cancer treatments and provide improved long-term survival in a subset of non-small cell lung cancer (NSCLC) patients. However, prognostic and predictive biomarkers of immunotherapy still remain an unmet clinical need. In this work, we aim to leverage imaging data and clinical variables to develop survival risk models among advanced NSCLC patients treated with immunotherapy. METHODS: This retrospective study includes a total of 385 patients from two institutions who were treated with ICIs. Radiomics features extracted from pretreatment CT scans were used to build predictive models. The objectives were to predict overall survival (OS) along with building a classifier for short- and long-term survival groups. We employed the XGBoost learning method to build radiomics and integrated clinical-radiomics predictive models. Feature selection and model building were developed and validated on a multicenter cohort. RESULTS: We developed parsimonious models that were associated with OS and a classifier for short- and long-term survivor groups. The concordance indices (C-index) of the radiomics model were 0.61 and 0.57 to predict OS in the discovery and validation cohorts, respectively. While the area under the curve (AUC) values of the radiomic models for short- and long-term groups were found to be 0.65 and 0.58 in the discovery and validation cohorts. The accuracy of the combined radiomics-clinical model resulted in 0.63 and 0.62 to predict OS and in 0.77 and 0.62 to classify the survival groups in the discovery and validation cohorts, respectively. CONCLUSIONS: We developed and validated novel radiomics and integrated radiomics-clinical survival models among NSCLC patients treated with ICIs. This model has important translational implications, which can be used to identify a subset of patients who are not likely to benefit from immunotherapy. The developed imaging biomarkers may allow early prediction of low-group survivors, though additional validation of these radiomics models is warranted.
ABSTRACT
With the increasing use of immune checkpoint inhibitors (ICIs), there is an urgent need to identify biomarkers to stratify responders and non-responders using programmed death-ligand (PD-L1) expression, and to predict patient-specific outcomes such as progression free survival (PFS). The current study is aimed to determine the feasibility of building imaging-based predictive biomarkers for PD-L1 and PFS through systematically evaluating a combination of several machine learning algorithms with different feature selection methods. A retrospective, multicenter study of 385 advanced NSCLC patients amenable to ICIs was undertaken in two academic centers. Radiomic features extracted from pretreatment CT scans were used to build predictive models for PD-L1 and PFS (short-term vs. long-term survivors). We first employed the LASSO methodology followed by five feature selection methods and seven machine learning approaches to build the predictors. From our analyses, we found several combinations of feature selection methods and machine learning algorithms to achieve a similar performance. Logistic regression with ReliefF feature selection (AUC = 0.64, 0.59 in discovery and validation cohorts) and SVM with Anova F-test feature selection (AUC = 0.64, 0.63 in discovery and validation datasets) were the best-performing models to predict PD-L1 and PFS. This study elucidates the application of suitable feature selection approaches and machine learning algorithms to predict clinical endpoints using radiomics features. Through this study, we identified a subset of algorithms that should be considered in future investigations for building robust and clinically relevant predictive models.
Subject(s)
B7-H1 Antigen , Lung Neoplasms , Humans , Progression-Free Survival , Ligands , Retrospective Studies , Immunotherapy , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/drug therapy , LungABSTRACT
Fecal microbiota transplantation (FMT) represents a potential strategy to overcome resistance to immune checkpoint inhibitors in patients with refractory melanoma; however, the role of FMT in first-line treatment settings has not been evaluated. We conducted a multicenter phase I trial combining healthy donor FMT with the PD-1 inhibitors nivolumab or pembrolizumab in 20 previously untreated patients with advanced melanoma. The primary end point was safety. No grade 3 adverse events were reported from FMT alone. Five patients (25%) experienced grade 3 immune-related adverse events from combination therapy. Key secondary end points were objective response rate, changes in gut microbiome composition and systemic immune and metabolomics analyses. The objective response rate was 65% (13 of 20), including four (20%) complete responses. Longitudinal microbiome profiling revealed that all patients engrafted strains from their respective donors; however, the acquired similarity between donor and patient microbiomes only increased over time in responders. Responders experienced an enrichment of immunogenic and a loss of deleterious bacteria following FMT. Avatar mouse models confirmed the role of healthy donor feces in increasing anti-PD-1 efficacy. Our results show that FMT from healthy donors is safe in the first-line setting and warrants further investigation in combination with immune checkpoint inhibitors. ClinicalTrials.gov identifier NCT03772899 .
Subject(s)
Fecal Microbiota Transplantation , Melanoma , Animals , Mice , Fecal Microbiota Transplantation/methods , Immune Checkpoint Inhibitors , Feces/microbiology , Melanoma/therapy , Immunotherapy , Treatment OutcomeABSTRACT
Strain KD21T, isolated from the fecal sample of a healthy female volunteer, is a strictly anaerobic, non-motile, Gram-staining-positive, saccharolytic small rod that does not produce spores. Strain KD21T was able to grow in the range of temperature 28°C-37°C (optimum, 37 °C), pH 6.0-8.0 (optimum, pH 7.0), and with 0-5.0 g/l NaCl (optimum, 0 g/l NaCl). Bacteria cells reduced nitrates to nitrites. Its major fatty acids were C18:1ω9c, C16:0, C18:0, and summed in feature 8 (C18:1ω7c and/or C18:1ω6c). 16S rRNA gene phylogenetic analysis revealed that KD21T is a member of the genus Tractidigestivibacter and is distinct from any species with validly published names. The sequence showed 98.48% similarity with T. scatoligenes SK9K4T. The DNA G + C content of strain KD21T was 62.6 mol%. The DNA-DNA hybridization and OrthoANI values between strain KD21T and T. scatoligenes SK9K4T were 40.2% and 90.2%, respectively. Differences in phenotypic, phylogenetic, chemotaxonomic, and genomic characteristics indicated that strain KD21T represents a novel species within the genus Tractidigestivibacter. The name T. montrealensis sp. nov. is proposed and the type strain is KD21T (=âCSUR Q8103T = âDSM 115111T).