cytoKernel: Robust kernel embeddings for assessing differential expression of single cell data.

High-throughput sequencing of single-cell data can be used to rigorously evlauate cell specification and enable intricate variations between groups or conditions. Many popular existing methods for differential expression target differences in aggregate measurements (mean, median, sum) and limit their approaches to detect only global differential changes. We present a robust method for differential expression of single-cell data using a kernel-based score test, cytoKernel. cytoKernel is specifically designed to assess the differential expression of single cell RNA sequencing and high-dimensional flow or mass cytometry data using the full probability distribution pattern. cytoKernel is based on kernel embeddings which employs the probability distributions of the single cell data, by calculating the pairwise divergence/distance between distributions of subjects. It can detect both patterns involving aggregate changes, as well as more elusive variations that are often overlooked due to the multimodal characteristics of single cell data. We performed extensive benchmarks across both simulated and real data sets from mass cytometry data and single-cell RNA sequencing. The cytoKernel procedure effectively controls the False Discovery Rate (FDR) and shows favourable performance compared to existing methods. The method is able to identify more differential patterns than existing approaches. We apply cytoKernel to assess gene expression and protein marker expression differences from cell subpopulations in various publicly available single-cell RNAseq and mass cytometry data sets. The methods described in this paper are implemented in the open-source R package cytoKernel, which is freely available from Bioconductor at http://bioconductor.org/packages/cytoKernel.

Proteomic networks and related genetic variants associated with smoking and chronic obstructive pulmonary disease.

BACKGROUND: Studies have identified individual blood biomarkers associated with chronic obstructive pulmonary disease (COPD) and related phenotypes. However, complex diseases such as COPD typically involve changes in multiple molecules with interconnections that may not be captured when considering single molecular features. METHODS: Leveraging proteomic data from 3,173 COPDGene Non-Hispanic White (NHW) and African American (AA) participants, we applied sparse multiple canonical correlation network analysis (SmCCNet) to 4,776 proteins assayed on the SomaScan v4.0 platform to derive sparse networks of proteins associated with current vs. former smoking status, airflow obstruction, and emphysema quantitated from high-resolution computed tomography scans. We then used NetSHy, a dimension reduction technique leveraging network topology, to produce summary scores of each proteomic network, referred to as NetSHy scores. We next performed a genome-wide association study (GWAS) to identify variants associated with the NetSHy scores, or network quantitative trait loci (nQTLs). Finally, we evaluated the replicability of the networks in an independent cohort, SPIROMICS. RESULTS: We identified networks of 13 to 104 proteins for each phenotype and exposure in NHW and AA, and the derived NetSHy scores significantly associated with the variable of interests. Networks included known (sRAGE, ALPP, MIP1) and novel molecules (CA10, CPB1, HIS3, PXDN) and interactions involved in COPD pathogenesis. We observed 7 nQTL loci associated with NetSHy scores, 4 of which remained after conditional analysis. Networks for smoking status and emphysema, but not airflow obstruction, demonstrated a high degree of replicability across race groups and cohorts. CONCLUSIONS: In this work, we apply state-of-the-art molecular network generation and summarization approaches to proteomic data from COPDGene participants to uncover protein networks associated with COPD phenotypes. We further identify genetic associations with networks. This work discovers protein networks containing known and novel proteins and protein interactions associated with clinically relevant COPD phenotypes across race groups and cohorts.

Smccnet 2.0: a comprehensive tool for multi-omics network inference with shiny visualization.

Sparse multiple canonical correlation network analysis (SmCCNet) is a machine learning technique for integrating omics data along with a variable of interest (e.g., phenotype of complex disease), and reconstructing multi-omics networks that are specific to this variable. We present the second-generation SmCCNet (SmCCNet 2.0) that adeptly integrates single or multiple omics data types along with a quantitative or binary phenotype of interest. In addition, this new package offers a streamlined setup process that can be configured manually or automatically, ensuring a flexible and user-friendly experience. AVAILABILITY : This package is available in both CRAN: https://cran.r-project.org/web/packages/SmCCNet/index.html and Github: https://github.com/KechrisLab/SmCCNet under the MIT license. The network visualization tool is available at https://smccnet.shinyapps.io/smccnetnetwork/ .

Pachyonychia congenita: A father and son with a novel variant in the KRT16 gene.

This study underscores the significance of identifying the clinical manifestations of pachyonychia congenita (PC) and emphasizes the patterns of genetic inheritance. A 12-month-old boy presented with a "white hairy tongue" and, following a comprehensive evaluation, was diagnosed with PC. His father exhibited similar symptoms. Genetic testing revealed a KRT16 pathogenic variant (c.616 T > G) in both the patient and his father, marking it as a novel variant in the PC literature. This case contributes to a broader understanding of PC's genetic diversity and its clinical presentations.

Sexually dimorphic eye size in dragonfishes, a response to a bioluminescent signalling gap.

Deep-sea fishes must overcome extremely large nearest-neighbour distances and darkness to find mates. Sexual dimorphism in the size of luminescent structures in many deep-sea taxa, including dragonfishes (family Stomiidae), indicates reproductive behaviours may be mediated by visual signalling. This presents a paradox: if male photophores are larger, females may find males at shorter distances than males find females. Solutions to this gap may include females closing this gap or by males gathering more photons with a larger eye. We examine the eye size of two species of dragonfishes (Malacosteus niger and Phostomias guernei) for sexual dimorphism and employ a model of detection distance to evaluate the potential for such dimorphism to bridge the detection gap. This model incorporates the flux of sexually dimorphic postorbital photophores and eye lens size to predict detection distances. In both species, we found a significant visual detection gap in which females find males before males find females and that male lens size is larger, marking the second known case of size dimorphism in the actinopterygian visual system. Our results indicate the larger eye affords males a significant improvement in detection distance. We conclude that this dimorphic phenotype may have evolved to close the detection gap.

Trained Immunity Generated by the Recombinant Zoster Vaccine.

Trained immunity may play a role in vaccine-induced protection against infections. We showed that the highly efficacious recombinant VZV-gE zoster vaccine (RZV) generated trained immunity in monocytes, natural killer (NK) cells, and dendritic cells (DCs) and that the less efficacious live zoster vaccine did not. RZV stimulated ex vivo gE-specific monocyte, DC and NK cell responses that did not correlate with CD4 + T-cell responses. These responses were also elicited in purified monocyte and NK cell cocultures stimulated with VZV-gE and persisted above prevaccination levels for ≥ 4 years post-RZV administration. RZV administration also increased ex vivo heterologous monocyte and NK cell responses to herpes simplex and cytomegalovirus antigens. ATAC-seq analysis and ex vivo TGFß1 supplementation and inhibition experiments demonstrated that decreased tgfß1 transcription resulting from RZV-induced chromatin modifications may explain the development of monocyte trained immunity. The role of RZV-trained immunity in protection against herpes zoster and other infections should be further studied.

Sex Differences in the Blood Metabolome During Acute Response to Ischemic Stroke.

Introduction: Females suffer greater lifetime risk of stroke and greater morbidity and mortality from stroke compared with males. This study's objective was to identify differences in metabolomic profiling of females and males with stroke and which differences were associated with neurological outcome. Methods: Females and males with acute ischemic stroke enrolled in the Emergency Medicine Specimen Bank at a comprehensive stroke center provided whole blood samples upon arrival for mass spectrometry-based metabolomics. We used descriptive statistics to characterize the cohort. A linear regression model was fit for individual metabolites to determine differences in relative abundance between males and females while controlling for covariates (age, race/ethnicity, postmenopausal status, cardiovascular risk factors, depression, time between sample collection and last known well, and initial National Institutes of Health Stroke Scale [NIHSS] score). For each differentially expressed metabolite, a linear regression model was fit to determine the association between the metabolite and NIHSS at 24 hours after admission while controlling for the covariates and acute treatments. Results: After adjusting for covariates, eight metabolites differed in females and males with a stroke. These included amino acids or their metabolites (proline and tryptophan), nucleotides (guanosine diphosphate [GDP], and inosine-3',5'-cyclic monophosphate), citrate, dehydroascorbate, choline, and acylcarnitine-(5-OH). GDP and dehydroascorbate were significantly associated with 24-hour NIHSS (p = 0.0991). Conclusions: Few metabolites were differentially abundant in blood after a stroke when comparing females with males and controlling for confounders, but the interactions between biological sex and GDP, as well as biological sex and dehydroascorbate, were associated with 24-hour neurological function. This has important implications for future studies that evaluate the therapeutic potential of these metabolites in ischemic stroke.

Expanding Kane's argument-based validity framework: What can validation practices in language assessment offer health professions education?

CONTEXT: One central consideration in health professions education (HPE) is to ensure we are making sound and justifiable decisions based on the assessment instruments we use on health professionals. To achieve this goal, HPE assessment researchers have drawn on Kane's argument-based framework to ascertain the validity of their assessment tools. However, the original four-inference model proposed by Kane - frequently used in HPE validation research - has its limitations in terms of what each inference entails and what claims and sources of backing are housed in each inference. The under-specification in the four-inference model has led to inconsistent practices in HPE validation research, posing challenges for (i) researchers who want to evaluate the validity of different HPE assessment tools and/or (ii) researchers who are new to test validation and need to establish a coherent understanding of argument-based validation. METHODS: To address these identified concerns, this article introduces the expanded seven-inference argument-based validation framework that is established practice in the field of language testing and assessment (LTA). We explicate (i) why LTA researchers experienced the need to further specify the original four Kanean inferences; (ii) how LTA validation research defines each of their seven inferences and (iii) what claims, assumptions and sources of backing are associated with each inference. Sampling six representative validation studies in HPE, we demonstrate why an expanded model and a shared disciplinary validation framework can facilitate the examination of the validity evidence in diverse HPE validation contexts. CONCLUSIONS: We invite HPE validation researchers to experiment with the seven-inference argument-based framework from LTA to evaluate its usefulness to HPE. We also call for greater interdisciplinary dialogue between HPE and LTA since both disciplines share many fundamental concerns about language use, communication skills, assessment practices and validity in assessment instruments.

Hierarchical Principal Components for Data-Driven Multiresolution fMRI Analyses.

Understanding the organization of neural processing is a fundamental goal of neuroscience. Recent work suggests that these systems are organized as a multiscale hierarchy, with increasingly specialized subsystems nested inside general processing systems. Current neuroimaging methods, such as independent component analysis (ICA), cannot fully capture this hierarchy since they are limited to a single spatial scale. In this manuscript, we introduce multiresolution hierarchical principal components analysis (hPCA) and compare it to ICA using simulated fMRI datasets. Furthermore, we describe a parametric statistical filtering method developed to focus analyses on biologically relevant features. Lastly, we apply hPCA to the Human Connectome Project (HCP) to demonstrate its ability to estimate a hierarchy from real fMRI data. hPCA accurately estimated spatial maps and time series from networks with diverse hierarchical structures. Simulated hierarchies varied in the degree of branching, such as two-way or three-way subdivisions, and the total number of levels, with varying equal or unequal subdivision sizes at each branch. In each case, as well as in the HCP, hPCA was able to reconstruct a known hierarchy of networks. Our results suggest that hPCA can facilitate more detailed and comprehensive analyses of the brain's network of networks and the multiscale regional specializations underlying neural processing and cognition.

Proteomic Networks and Related Genetic Variants Associated with Smoking and Chronic Obstructive Pulmonary Disease.

Background: Studies have identified individual blood biomarkers associated with chronic obstructive pulmonary disease (COPD) and related phenotypes. However, complex diseases such as COPD typically involve changes in multiple molecules with interconnections that may not be captured when considering single molecular features. Methods: Leveraging proteomic data from 3,173 COPDGene Non-Hispanic White (NHW) and African American (AA) participants, we applied sparse multiple canonical correlation network analysis (SmCCNet) to 4,776 proteins assayed on the SomaScan v4.0 platform to derive sparse networks of proteins associated with current vs. former smoking status, airflow obstruction, and emphysema quantitated from high-resolution computed tomography scans. We then used NetSHy, a dimension reduction technique leveraging network topology, to produce summary scores of each proteomic network, referred to as NetSHy scores. We next performed genome-wide association study (GWAS) to identify variants associated with the NetSHy scores, or network quantitative trait loci (nQTLs). Finally, we evaluated the replicability of the networks in an independent cohort, SPIROMICS. Results: We identified networks of 13 to 104 proteins for each phenotype and exposure in NHW and AA, and the derived NetSHy scores significantly associated with the variable of interests. Networks included known (sRAGE, ALPP, MIP1) and novel molecules (CA10, CPB1, HIS3, PXDN) and interactions involved in COPD pathogenesis. We observed 7 nQTL loci associated with NetSHy scores, 4 of which remained after conditional analysis. Networks for smoking status and emphysema, but not airflow obstruction, demonstrated a high degree of replicability across race groups and cohorts. Conclusions: In this work, we apply state-of-the-art molecular network generation and summarization approaches to proteomic data from COPDGene participants to uncover protein networks associated with COPD phenotypes. We further identify genetic associations with networks. This work discovers protein networks containing known and novel proteins and protein interactions associated with clinically relevant COPD phenotypes across race groups and cohorts.

Impact of surgical lung biopsy on lung function and survival in patients with idiopathic pulmonary fibrosis in a multi-centre registry cohort.

BACKGROUND AND OBJECTIVE: Establishing an accurate and timely diagnosis of idiopathic pulmonary fibrosis (IPF) is essential for appropriate management and prognostication. In some cases, surgical lung biopsy (SLB) is performed but carries non-negligible risk. The objective of this retrospective study was to determine if SLB is associated with accelerated lung function decline in patients with IPF using the Canadian Registry for Pulmonary Fibrosis. METHODS: Linear mixed models and Cox proportional hazards regression models were used to compare decline in forced vital capacity (FVC)%, diffusion capacity of the lung (DLCO%) and risk of death or lung transplantation between SLB and non-SLB patients. Adjustments were made for baseline age, sex, smoking history, antifibrotic use, and lung function. A similar analysis compared lung function changes 12 months pre- and post-SLB. RESULTS: A total of 81 SLB patients and 468 non-SLB patients were included. In the SLB group, the post-biopsy annual FVC% decline was 2.0% (±0.8) in unadjusted, and 2.1% (±0.8) in adjusted models. There was no difference in FVC% decline, DLCO% decline, or time to death or lung transplantation between the two groups, in adjusted or unadjusted models (all p-values >0.07). In the pre-post SLB group, no differences were identified in FVC% decline in unadjusted or adjusted models (p = 0.07 for both). CONCLUSION: No association between SLB and lung function decline or risk of death or lung transplantation was identified in this multi-centre study of patients with IPF.

SpaceANOVA: Spatial Co-occurrence Analysis of Cell Types in Multiplex Imaging Data Using Point Process and Functional ANOVA.

Multiplex imaging platforms have enabled the identification of the spatial organization of different types of cells in complex tissue or the tumor microenvironment. Exploring the potential variations in the spatial co-occurrence or colocalization of different cell types across distinct tissue or disease classes can provide significant pathological insights, paving the way for intervention strategies. However, the existing methods in this context either rely on stringent statistical assumptions or suffer from a lack of generalizability. We present a highly powerful method to study differential spatial co-occurrence of cell types across multiple tissue or disease groups, based on the theories of the Poisson point process and functional analysis of variance. Notably, the method accommodates multiple images per subject and addresses the problem of missing tissue regions, commonly encountered due to data-collection complexities. We demonstrate the superior statistical power and robustness of the method in comparison with existing approaches through realistic simulation studies. Furthermore, we apply the method to three real data sets on different diseases collected using different imaging platforms. In particular, one of these data sets reveals novel insights into the spatial characteristics of various types of colorectal adenoma.

Prevalence and drug susceptibility of clinical Candida species in nasopharyngeal cancer patients in Vietnam.

In the nature, Candida species are normal inhabitants and can be observed in a wide variety of vertebrates. In humans, especially for cancer patients who fall prey to opportunistic pathogens, this group of susceptible multi-drug resistant and biofilm-forming yeasts, are among the commonest ones. In this study, Candida species in 76 oral lesion samples from Vietnamese nasopharyngeal-cancer patients were isolated, morphologically identified using CHROMagar™, germ tube formation, and chlamydospore formation tests, and molecularly confirmed by PCR-RFLP. The drug susceptibility of these isolates was then tested, and the gene ERG11 was DNA sequenced to investigate the mechanism of resistance. The results showed that Candida albicans remained the most prevalent species (63.16% of the cases), followed by Candida glabrata, Candida tropicalis, and Candida krusei. The rates of resistance of non-albicans Candida for tested drugs were 85.71%, 53.57%, and 57.14% to fluconazole, clotrimazole, and miconazole, respectively. Although the drug-resistance rate of Candida albicans was lower than that of non-albicans Candida, it was higher than expected, suggesting an emerging drug-resistance phenomenon. Furthermore, ERG11 DNA sequencing revealed different mutations (especially K128T), implying the presence of multiple resistance mechanisms. Altogether, the results indicate an alarming drug-resistance situation in Candida species in Vietnamese cancer patients and emphasize the importance of species identification and their drug susceptibility prior to treatment.

SmCCNet 2.0: A Comprehensive Tool for Multi-omics Network Inference with Shiny Visualization.

Summary: Sparse multiple canonical correlation network analysis (SmCCNet) is a machine learning technique for integrating omics data along with a variable of interest (e.g., phenotype of complex disease), and reconstructing multi-omics networks that are specific to this variable. We present the second-generation SmCCNet (SmCCNet 2.0) that adeptly integrates single or multiple omics data types along with a quantitative or binary phenotype of interest. In addition, this new package offers a streamlined setup process that can be configured manually or automatically, ensuring a flexible and user-friendly experience. Availability: This package is available in both CRAN: https://cran.r-project.org/web/packages/SmCCNet/index.html and Github: https://github.com/KechrisLab/SmCCNet under the MIT license. The network visualization tool is available at https://smccnet.shinyapps.io/smccnetnetwork/.

Designing phage cocktails to combat the emergence of bacteriophage-resistant mutants in multidrug-resistant Klebsiella pneumoniae.

IMPORTANCE: In this study, we aimed to design a novel and effective bacteriophage cocktail that can target both wild-type bacteria and phage-resistant mutants. To achieve this goal, we isolated four phages (U2874, phi_KPN_H2, phi_KPN_S3, and phi_KPN_HS3) that recognized different bacterial surface molecules using phage-resistant bacteria. We constructed three phage cocktails and tested their phage resistance-suppressing ability against multidrug-resistant Klebsiella pneumoniae. We argue that the phage cocktail that induces resensitization of phage susceptibility exhibited superior phage resistance-suppressing ability. Moreover, we observed trade-off effects that manifested progressively in phage-resistant bacteria. We hypothesize that such trade-off effects can augment therapeutic efficacy. We also recommend collating phage host range data against phage-resistant mutants in addition to wild-type bacteria when establishing phage banks to improve the efficiency of phage therapy. Our study underscores the importance of phage host range data in constructing effective phage cocktails for clinical use.

Dual targeting of mitochondrial Lon peptidase 1 and chymotrypsin-like protease by small molecule BT317, as potential therapeutics in malignant astrocytoma.

Malignant astrocytomas are aggressive glioma tumors characterized by extensive hypoxia-induced, mito-chondria-dependent changes such as altered respiration, increased chymotrypsin-like (CT-L) proteasome activity, decreased apoptosis, drug resistance, stemness and increased invasiveness. Mitochondrial Lon Peptidase I (LonP1) overexpression and increased CT-L proteasome inhibitors activity are the biomarkers of aggressive high grade glioma phenotype, poor prognosis and found to be associated with recurrence and poor patient survival, and drugs targeting either LonP1 or the CT-L activity have anti-glioma activity in pre-clinical models. We here for the first time introduced and evaluated a novel small molecule, BT317, derived from coumarinic compound 4 (CC4) using structure-activity modeling which we found to inhibit both LonP1 and CT-L proteasome activity. Using gain-of-function and loss-of-function genetic models, we dis-covered that BT317 is more effective than the individual LonP1 or CT-L inhibition in increasing reactive oxy-gen species (ROS) generation and inducing apoptosis in high-grade astrocytoma lines. In vitro, BT317 had activity as a single agent but, more importantly, enhanced synergy with the standard of care commonly used chemotherapeutic temozolomide (TMZ). In orthotopic xenograft, patient derived glioma models, BT317 was able to cross the blood-brain barrier, to show selective activity at the tumor site and to demonstrate therapeutic efficacy both as a single agent and in combination with TMZ. BT317 defines an emerging class of dual LonP1, and CT-L proteasome inhibitors exhibited promising anti-tumor activity and could be a promising candidate for clinical translation in the space of malignant astrocytoma therapeutics.

Appelmans protocol - A directed in vitro evolution enables induction and recombination of prophages with expanded host range.

Infections caused by carbapenem-resistant Acinetobacter baumannii (CRAB) present significant healthcare challenges due to limited treatment options. Bacteriophage (phage) therapy offers potential as an alternative treatment. However, the high host specificity of phages poses challenges for their therapeutic application. To broaden the phage spectrum, laboratory-based phage training using the Appelmans protocol was employed in this study. As a result, the protocol successfully expanded the host range of a phage cocktail targeting CRAB. Further analysis revealed that the expanded host range phages isolated from the output cocktail were identified as recombinant derivatives originating from prophages induced from encountered bacterial strains. These findings provide valuable genetic insights into the protocol's mechanism when applied to phages infecting A. baumannii strains that have never been investigated before. However, it is noteworthy that the expanded host range phages obtained from this protocol exhibited limited stability, raising concerns about their suitability for therapeutic purposes.

Comorbidity Burden and Treatment Patterns of Psoriasis in Vietnam: Real-World Data from the EXPAND Study.

INTRODUCTION: Psoriasis is a multi-faceted, immune-mediated inflammatory disease associated with a wide range of comorbidities. Real-world data on treatment patterns, comorbidities, and economic burden in patients with psoriasis are needed for comprehensive patient care in Vietnam. METHODS: A retrospective chart review study was conducted using secondary data extracted from patients' medical records of two hospitals in Vietnam, with the aim of identifying adult patients with a confirmed diagnosis of psoriasis. The index date was defined as the date of first diagnosis between 1 January 2020 and 31 October 2021. Sociodemographic factors, disease characteristics, comorbidities, medication usage, drug survival, and medication costs were analyzed. RESULTS: A total of 661 patients were identified (mean ± standard deviation [SD] age 43.5 ± 14.8 years). The most prevalent comorbidity was dyslipidemia (49.6% of patients), followed by hypertension (23.4%), and psoriatic arthritis (10.4%). In total, 44% of patients received biologic therapies. Overall, 66.7% and 54.3% of patients receiving biologic and non-biologic therapies, respectively, had ≥ 1 comorbidity. Only 23.2% of patients with psoriasis-related comorbidities stopped therapy with biologics. Biologics had a longer retention time (17.0 months) than non-biologics (6.0 months) in patients with comorbidities. Patients with comorbidities had significantly higher total annual healthcare costs than those without comorbidities (in US dollars: USD901 vs. USD304; p < 0.001), mainly due to the relatively higher costs associated with the use of biologics. CONCLUSION: Patients with psoriasis in Vietnam experience a high disease and economic burden due to comorbidities. Evidence from this real-world study supports the need for routine monitoring of and an appropriate treatment course for psoriasis-related comorbidities.

Pemphigus Foliaceus after COVID-19 Vaccination: A Report of Two Cases.

Autoimmune bullous diseases (AIBDs) following coronavirus disease (COVID-19) vaccination have been previously documented in medical literature, given the comparable nature of the RNA antigen in these vaccines to that of the cellular nuclear matter. However, pemphigus foliaceus has been reported less frequently than other postimmunization AIBDs worldwide. Two women were admitted to our hospital with skin erosion over their faces, trunks, and extremities after receiving COVID-19 vaccination. Upon examination and consultation with pathologists, the diagnosis of pemphigus foliaceus was confirmed for both patients. In an effort to contribute to the knowledge on this intriguing topic, we present these two aforementioned cases of pemphigus foliaceus following COVID-19 vaccination, which may initially appear as a typical occurrence but exhibit some noteworthy characteristics.

FunSpace: A functional and spatial analytic approach to cell imaging data using entropy measures.

Spatial heterogeneity in the tumor microenvironment (TME) plays a critical role in gaining insights into tumor development and progression. Conventional metrics typically capture the spatial differential between TME cellular patterns by either exploring the cell distributions in a pairwise fashion or aggregating the heterogeneity across multiple cell distributions without considering the spatial contribution. As such, none of the existing approaches has fully accounted for the simultaneous heterogeneity caused by both cellular diversity and spatial configurations of multiple cell categories. In this article, we propose an approach to leverage spatial entropy measures at multiple distance ranges to account for the spatial heterogeneity across different cellular organizations. Functional principal component analysis (FPCA) is applied to estimate FPC scores which are then served as predictors in a Cox regression model to investigate the impact of spatial heterogeneity in the TME on survival outcome, potentially adjusting for other confounders. Using a non-small cell lung cancer dataset (n = 153) as a case study, we found that the spatial heterogeneity in the TME cellular composition of CD14+ cells, CD19+ B cells, CD4+ and CD8+ T cells, and CK+ tumor cells, had a significant non-zero effect on the overall survival (p = 0.027). Furthermore, using a publicly available multiplexed ion beam imaging (MIBI) triple-negative breast cancer dataset (n = 33), our proposed method identified a significant impact of cellular interactions between tumor and immune cells on the overall survival (p = 0.046). In simulation studies under different spatial configurations, the proposed method demonstrated a high predictive power by accounting for both clinical effect and the impact of spatial heterogeneity.