Chapter Four · failure evidence
What Cytokine Profiling & Immunomonitoring got wrong, from 58 dissertations
Cytokine profiling and immunomonitoring assays frequently fail to yield statistically significant distinctions between clinical cohorts or predict patient treatment responses. In addition, technical challenges including assay detection limits, biological confounding, and in vitro culture artifacts hinder reliable biomarker quantification. These records come from PhD theses at 18 institutions, 2021 to 2026. Each links to its thesis. They were extracted by language models reading the full text, so treat each as a lead to read, not a verdict.
Cytokine and immune cell profiling failed to differentiate disease cohorts or clinical stages from controls
Profiling of circulating cytokines, synovial fluid biomarkers, and immune cell subsets showed no statistically significant differences between healthy controls and disease cohorts or between different stages of illness. Stimulatory assays and cross-disease transcriptional risk scores similarly failed to discriminate between distinct patient populations.
Tried and failed
ELISA profiling of proinflammatory cytokines applied to first-trimester preterm labor prediction. Outcome: no signal. Reason: Cytokine concentrations showed no statistically significant differences between preterm labor and term control cohorts
Early prediction of pregnancy disorders with machine learning guided Raman spectroscopy, and metabolomics. · Iowa State
Tried and failed
cytokine quantification as single-biomarker predictors applied to maternal plasma for preeclampsia detection. Outcome: no signal. Reason: TNF-α and GM-CSF levels showed no statistically significant differences between healthy and preeclamptic patients in any trimester
Tried and failed
multiplex cytokine profiling of tissue explants applied to early bacterial infection tissue response. Outcome: no signal. Reason: no significant differences in cytokine or chemokine levels across strains at early timepoints
Innate immunological contributions to severe C. rodentium infection · Imperial
Tried and failed
serum cytokine profiling for chronic disease pathology applied to post-viral infection animal models. Outcome: no signal. Reason: pro-inflammatory cytokine levels did not significantly differ between infected and uninfected animals at late timepoints
Tried and failed
multiplex cytokine immunoassay profiling across surgical timepoints applied to circulating plasma biomarkers during liver regeneration. Outcome: no signal. Reason: no statistically significant differences in circulating cytokine levels between pre-stage 1 and pre-stage 2 timepoints
Tried and failed
early complete blood count profiling applied to predicting postpartum inflammatory syndrome onset. Outcome: no signal. Reason: circulating immune cell populations showed no measurable differences between healthy and affected subjects before diagnosis
Tried and failed
serum inflammatory biomarker profiling for infection status applied to differentiating viral infection in chronic disease. Outcome: no signal. Reason: Systemic IL-6 and MPO levels did not significantly differ between infected and uninfected baseline cohorts.
RSV associated COPD exacerbation prevalence and severity · Imperial
Tried and failed
multiplex immunoassay of local biofluid biomarkers applied to localized inflammatory tissue pathology. Outcome: no signal. Reason: Biomarker concentration differences between diseased and control sites lacked statistical significance.
Effect of endodontic treatment on local inflammatory biomarkers in teeth with apical periodontitis · Harvard
Tried and failed
measuring synovial fluid cytokine biomarker panels applied to joint injury disease classification. Outcome: no signal. Reason: Cytokine concentrations did not differ significantly between diseased and healthy controls.
Tried and failed
cross-disease transcriptional risk scoring applied to disease classification across related autoimmune disorders. Outcome: no signal. Reason: risk scores derived from rheumatoid arthritis failed to distinguish inflammatory bowel disease cases from controls
Genomic and Transcriptomic Characterization of Inflammatory Bowel Disease · Georgia Tech
Tried and failed
immunophenotyping immune cell subsets for biomarker discovery applied to allograft rejection patient cohorts. Outcome: no signal. Reason: No statistically significant difference in specific regulatory T cell frequencies across clinical rejection groups
Tried and failed
polyclonal superantigen stimulation assay applied to differentiating autoimmune vs inflammatory disease cohorts. Outcome: no signal. Reason: general T cell activation levels did not differ significantly between MS and disease controls
T Cell Responses In Pediatric-Onset Multiple Sclerosis: A Unique Window Into Early Disease · Penn
Baseline cytokine concentrations and immune cell metrics failed to correlate with disease severity or predict treatment outcomes
Baseline cytokine levels, immune repertoire diversity, and receptor expression profiles showed no correlation with patient survival or therapeutic response to immunotherapy and immunosuppression. Circulating inflammatory markers also decoupled from patient-reported symptoms and failed to track clinical disease activity over time.
Tried and failed
cytokine profiling to predict mortality applied to polymicrobial sepsis disease outcomes. Outcome: no signal. Reason: circulating and lung cytokine levels paradoxically decoupled from and failed to correlate with sex-specific survival differences
The role of gender in susceptibility to infection and sepsis · Imperial
Tried and failed
baseline biomarker profiling for treatment outcome prediction applied to predicting systemic immunosuppressive therapy outcomes. Outcome: no signal. Reason: baseline serum cytokine/chemokine concentrations lacked practically significant predictive power for treatment outcomes
Towards a data-driven personalised management of Atopic Dermatitis severity · Imperial
Tried and failed
peripheral immune cell profiling as disease biomarker applied to liver cirrhosis severity staging. Outcome: no signal. Reason: circulating NK receptor expression and degranulation did not correlate with clinical severity scores
Tried and failed
cytokine receptor profiling as clinical marker applied to antigen-specific cytotoxic T cells. Outcome: no signal. Reason: receptor expression levels showed no correlation with helper T cell counts or percentages
HIV-1 CD8+ T cell emergence, evolution and virus control · Imperial
Tried and failed
correlating patient-reported symptom severity with circulating inflammatory biomarkers applied to respiratory exacerbation during antibiotic treatment. Outcome: no signal. Reason: symptoms did not track CRP or IL-6 levels during ongoing therapy
Tried and failed
baseline target expression as response predictor applied to combination immune checkpoint blockade response. Outcome: no signal. Reason: baseline receptor expression levels did not correlate with pathological treatment response
Biomarkers predicting outcomes for melanoma patients receiving systemic treatment · UT Austin
Tried and failed
baseline immune repertoire profiling and cell ratios applied to predicting immunotherapy clinical response. Outcome: no signal. Reason: baseline TCR clonality, diversity, and CD8+ ratios lacked statistically significant association with treatment outcomes
Tried and failed
correlating tumor mutational burden with immune cell frequencies applied to tumor-infiltrating immune cell subsets. Outcome: no signal. Reason: mutation load showed no or inverse correlation with memory B cells, GC B cells, and CD8 Tpex
Tried and failed
serum biomarker quantification for disease activity monitoring applied to autoimmune vasculitis disease progression. Outcome: no signal. Reason: circulating biomarker levels showed no correlation with standard clinical inflammatory markers or antibody titres
Factors affecting outcome in ANCA associated vasculitis · Imperial
Tried and failed
set-level permutation analysis of biomarker correlations applied to plasma cytokines and composite cognitive scores. Outcome: no signal. Reason: no significant association was found across subjects or over time
Tried and failed
correlating single target expression with gene-set enrichment applied to cross-tissue disease transcriptomic datasets. Outcome: no signal. Reason: single cytokine expression did not correlate with drug-response signature enrichment scores across patient tissue samples
Drug response signatures and the identification of potential responder populations across diseases · Imperial
Low assay sensitivity, standard curve saturation, and sample degradation impaired cytokine quantification
Measured cytokines frequently fell below lower detection limits or exceeded standard curve ranges in multiplex and sandwich immunoassay platforms. In addition, sample degradation, high intra-group variability, and non-targeted protein loss during immunodepletion undermined reproducible protein and mRNA quantification.
Lost to a baseline
1 conditioned macrophage culture supernatant sample (SHCC92) excluded from MSD cytokine analysis due to undetectable analyte levels across all measured cytokines.
Considered and rejected
Considered and rejected: Rejected protein immunodepletion columns to isolate responsible cytokines due to variable depletion efficiency and non-targeted protein loss.
Cellular and Circuit Level Responses to Neural Stem Cell Transplantation in the Rodent Cortex · Penn
Tried and failed
RT-qPCR quantification of cytokine mRNA expression applied to delayed-type hypersensitivity in knockout mice. Outcome: no signal. Reason: high intra-group variability prevented the observed increase from reaching statistical significance
The Neuropilin-2 Axis Regulates Carcinogenesis and Immune Surveillance in the Oral Cavity · Harvard
Tried and failed
cytokine multiplex array applied to frozen tissue lysates across biological stages. Outcome: no signal. Reason: insufficient sensitivity or tissue degradation masking expression differences in low-abundance effector molecules
ILC3 dynamics at the human uterine mucosal interface – a focus on IL-22 · Imperial
Lost to a baseline
Over 50% of maternal serum samples fell below the lower limit of detection (LLOD) for IL-1β and IL-4 and were excluded from cytokine analyses
Prenatal Environmental Stressors Impair Postnatal Microglia Function and Adult Behavior in Males · DukeSpace
Considered and rejected
Considered and rejected: Rejected calculating absolute cytokine concentrations from Luminex standard curves because standard-curve values were out of range, opting instead to analyze raw fluorescence intensity levels
Investigating the role of cytokines in clonal hematopoiesis · Oxford
Considered and rejected
Considered and rejected: Retired sandwich ELISA for IFN-alpha in favor of multiplex ELISA due to poor sensitivity in detecting low cytokine concentrations
Toll-like Receptor 7 modulation of the immune response in seasonal allergic rhinitis · Imperial
Biological cross-reactivity, marker non-specificity, and background inflammation confounded immune monitoring
Baseline inflammatory elevations caused by trauma or acute phase reactions obscured diagnostic infection baselines and composite cytokine signaling indices. In addition, lineage plasticity, host stromal cytokine production, and anti-reporter immune responses confounded targeted cell isolation and in vivo tracking assays.
Considered and rejected
Considered and rejected: Excluded C-reactive protein (CRP) from the composite inflammatory index because it reflects acute inflammation rather than general cytokine inflammatory signaling.
Association Between Fluoride Exposure and Immune System Biomarkers in Pregnancy · YorkSpace
Considered and rejected
Considered and rejected: Rejected using trauma patients as control group for infection diagnostic baseline because acute fracture and trauma cause elevated inflammatory markers (CRP, AZU1, α-defensin) matching infection levels.
Die Rolle von AZU1 als alternativer Marker für Infektionen in der Orthopädie und Unfallchirurgie · Publikationssystem UB Tuebingen
Considered and rejected
Considered and rejected: Rejected using 4T1-luc2 luciferase-labeled cells for metastatic tumor delay monitoring because host immune clearance against the reporter blunted tumor take and progression.
Crystallization and Local Delivery of Chemotherapeutic Compounds for the Treatment of a Triple Negative Breast Cancer Model · JScholarship
Considered and rejected
Considered and rejected: Rejected performing the in vivo Cas12a CRISPR screen in mice undergoing immune checkpoint blockade because baseline immunogenicity from Cas12a/GFP caused excessive tumor regression and noisy dropout.
Considered and rejected
Considered and rejected: Adoptive transfer of IL-3-deficient T cells into immune-deficient hosts was avoided/superseded by CD4Cre Il3GFPfl/fl conditional knockout to prevent confounding by stromal IL-3.
Uncovering the role of interleukin-3 in autoimmunity · Harvard
Considered and rejected
Considered and rejected: Rejected using IL-17A or CCR6 as single markers to isolate all Th17 cells due to widespread expression on non-Th17 lineages and cytokine plasticity under IL-23/IL-1b
Investigating the role of Th17 cells in colorectal cancer · Oxford
In vitro culture conditions and stimulation protocols failed to replicate physiological immune responses
In vitro cytokine polarization generated hyper-activated cells that lacked physiological memory markers and phenotypic plasticity. Extended stimulation times and specialized serum-free culture models also resulted in cytotoxicity, variable donor responses, and failure of cytokine induction.
Considered and rejected
Considered and rejected: Rejected LPS as a positive control in 3D ALI cultures due to irreproducible cytokine induction in serum-free apical environments.
The role of airway epithelial derived extracellular vesicles in allergic sensitisation · University of Nottingham Repository
Considered and rejected
Considered and rejected: Using hCMEC/D3 monolayer permeability to FITC-dextran as a functional readout for cytokine-induced barrier disruption; rejected because cytokine exposure did not reliably induce barrier breakdown in vitro.
Tried and failed
high-throughput screening with primary immune cells applied to co-culture functional genomics screens. Outcome: unstable. Reason: post-expansion cellular impurity and high donor-to-donor variability compromised assay reproducibility
Considered and rejected
Considered and rejected: Rejected standard cytokine-driven polarization (cdTh17) as an in vivo model because it produces hyper-activated, non-physiological cells lacking plasticity and memory markers.
Innate and Adaptive Immune Mechanisms of Pathogen-Specific T Helper 17 Cell Differentiation · DSpace at UTSWMED
Considered and rejected
Considered and rejected: Long-term LPS incubation over 7 days was rejected due to neurotoxicity, cell death, and delayed decline of cytokine levels.
Wie verändert LPS die Effekte von Sevofluran und Propofol in einem kortikalen neuronalen Netzwerk? · Publikationssystem UB Tuebingen
Multivariable immune composite models failed to outperform simpler single-marker baselines
Combined multivariable survival models performed worse than single immune cell markers due to collinearity among immunological features. Similarly, multiplex machine learning panels failed to demonstrate superior classification accuracy over single-compartment biomarker panels and suffered from multiple testing penalties.
Tried and failed
multiplex cytokine profiling and correlation analysis applied to perioperative systemic inflammation markers. Outcome: no signal. Reason: No significant plasma cytokine correlations survived multiple testing correction across perioperative timepoints.
Brain inflammation following surgery for fractured neck of femur · Imperial
Lost to a baseline
Refined multivariable Cox model performed worse (attenuated HR and lost significance) than single stromal CD68+ / CD45RO+ models due to collinearity among immune variables.
Lost to a baseline
Immune features alone achieved higher hazard ratios in predicting overall survival in chemotherapy CSCC (HR=4.3) compared to the combined CollaTIL model (HR=2.54).
Artificial Intelligence-Based Phenotyping of the Tumor Microenvironment in Hematoxylin and Eosin-Stained Images of Solid Tumors · Georgia Tech
Lost to a baseline
The full multi-compartment machine learning panel did not show a statistically significant AUC improvement over single-compartment plasma brain-derived protein panels (z = 1.11, p = 0.27) or plasma cytokine panels (z = 1.1, p = 0.26).
Evaluating Brain Derived Extracellular Vesicles As Diagnostic Biomarkers Of Traumatic Brain Injury · Penn
Tissue immunohistochemistry and compartmental sampling failed to capture localized cytokine dynamics
Circulating peripheral cytokine concentrations failed to correlate with central nervous system levels in paired biofluid monitoring. Furthermore, histological immunostaining did not align with flow cytometry quantification and suffered from diffuse extracellular staining of secreted cytokines.
Tried and failed
correlating peripheral and central inflammatory biomarker levels applied to post-operative systemic and neuroinflammation monitoring. Outcome: no signal. Reason: blood and CSF cytokine concentrations did not significantly correlate after surgery
Assessment and quantification of cytokine changes following surgery · Imperial
Tried and failed
immunohistochemical quantification of cell subtypes applied to allograft tissue immune infiltrate. Reason: immunohistochemistry marker quantification did not correlate with flow cytometry quantification results
Macrophages and T regulatory cells: immune mediators of allogeneic and xenogeneic Sertoli cells · Texas Tech
Considered and rejected
Considered and rejected: Rejected relying on standard antibody immunostaining for quantifying local Osm expression in spinal cords due to diffuse extracellular staining of secreted cytokines, choosing RNAscope in situ hybridization instead.
Dectin-1 Signaling in Central Nervous System Autoimmunity · DukeSpace
Left open by the authors
Problems the authors named and did not get to.
Left open
Evaluate IL-1α, IL-13, IL-17, and GROα/KC in CSF or blood serum as biomarkers of TBI severity and outcome. Blocker: Requires wet lab access, animal or clinical biofluid samples (CSF/serum), and assay equipment to measure cytokine levels
Profiling the neuroimmune cascade after repetitive mild traumatic brain injury · Georgia Tech
Left open
Combine multiple systemic inflammatory markers and pleiotropic cytokine genetic instruments to construct a multi-marker proxy for chronic low-grade inflammation. Blocker: Requires access to individual-level or comprehensive multi-cytokine cohort data (like UK Biobank) and specific methodology for multi-marker weighting is unspecified.
Left open
Measure secreted cytokine and chemokine protein levels via ELISA across infection timepoints in infected cell cultures and murine models. Blocker: Requires a wet lab, biological samples, infection models, and ELISA reagents
Left open
Profile cytokine secretomes and determine pyroptosis thresholds in WT versus Gsdmd-/- microglia and peripheral macrophages. Blocker: Requires primary mouse microglia, macrophages (WT and Gsdmd-/-), and wet-lab biological assays (e.g., cytokine profiling, IncuCyte).
The Role of Gasdermin D in Escherichia Coli K1 Bacterial Brain Infections · Harvard
Left open
Develop an interpretation method for feline-specific multiplex cytokine profiling data in cheetah serum. Blocker: Requires cheetah serum samples, wet lab cytokine profiling data, and veterinary immunology expertise
Multidisciplinary approach to Cheetahs affected by gastrointestinal disease: study of the immune profile and of the fecal proteome · IRIS - UNICAM - prod
Left open
Characterize immune cell phenotypes using CyTOF or single-cell multi-omics methods like CITE-seq and SEC-seq. Blocker: Requires wet lab experimental facilities, biological samples, and specialized instrumentation (e.g., CyTOF/Hyperion Imaging System, single-cell sequencing platforms).
Utilizing Combinatory Adjuvant-Loaded Chitosan-Derived Nanoparticles for a Joint SARS-CoV-2/Influenza Vaccine · Georgia Tech
Left open
Validate SpatialVizScore using larger clinical multiplex imaging cohorts with complete immune checkpoint inhibitor response and long-term survival data. Blocker: Requires multiplexed imaging cohorts with matched immunotherapy response and survival outcomes
Deciphering Spatial Immune Interactions in Health and Disease using Multiplexed Imaging Tools · Georgia Tech
Left open
Incorporate multi-scale modalities including imaging phenotypes, tumour microenvironment, immune status, and longitudinal trajectories into the chemotherapy response prediction model. Blocker: No specific multi-modal patient datasets or integration architecture specified
Predicting cancer patient response to chemotherapy using machine learning from small data · Imperial
Left open
Integrate multi-omic profiling to correlate peripheral immune DNA repair capacity with tumor tissue repair signatures. Blocker: Requires matched patient samples with multi-omic profiling and wet-lab functional DNA repair capacity measurements
Precision Medicine: Biomarkers of Genome Integrity in Guiding Better Lung Cancer Outcomes · Harvard
Left open
Determine optimal dosing and administration schedules for combining chemotherapeutics with immune checkpoint inhibitors in tumor models. Blocker: Requires wet lab in vivo syngeneic mouse tumor models and pharmacology/immunology experimentation
Stimulation of chemotherapy-induced immunity by targeting IL-6 in the tumor microenvironment · MIT
Checking a claim in this area?
We can run the same search on any method or claim. If nothing turns up, we will say so, and that proves nothing on its own.