Length of Stay Prediction Models for Oral Cancer Surgery: Machine Learning, Statistical and ACS-NSQIP.

OBJECTIVE: Accurate prediction of hospital length of stay (LOS) following surgical management of oral cavity cancer (OCC) may be associated with improved patient counseling, hospital resource utilization and cost. The objective of this study was to compare the performance of statistical models, a machine learning (ML) model, and The American College of Surgeons National Surgical Quality Improvement Program's (ACS-NSQIP) calculator in predicting LOS following surgery for OCC. MATERIALS AND METHODS: A retrospective multicenter database study was performed at two major academic head and neck cancer centers. Patients with OCC who underwent major free flap reconstructive surgery between January 2008 and June 2019 surgery were selected. Data were pooled and split into training and validation datasets. Statistical and ML models were developed, and performance was evaluated by comparing predicted and actual LOS using correlation coefficient values and percent accuracy. RESULTS: Totally 837 patients were selected with mean patient age being 62.5 ± 11.7 [SD] years and 67% being male. The ML model demonstrated the best accuracy (validation correlation 0.48, 4-day accuracy 70%), compared with the statistical models: multivariate analysis (0.45, 67%) and least absolute shrinkage and selection operator (0.42, 70%). All were superior to the ACS-NSQIP calculator's performance (0.23, 59%). CONCLUSION: We developed statistical and ML models that predicted LOS following major free flap reconstructive surgery for OCC. Our models demonstrated superior predictive performance to the ACS-NSQIP calculator. The ML model identified several novel predictors of LOS. These models must be validated in other institutions before being used in clinical practice. LEVEL OF EVIDENCE: Level 3 Laryngoscope, 2024.

Use of GPT-4 With Single-Shot Learning to Identify Incidental Findings in Radiology Reports.

GPT-4 identified incidental adrenal nodules, pancreatic cystic lesions, and vascular calcifications in radiology reports with F1 scores of 1.00, 0.91, and 0.99, respectively. The findings indicate a potential role for large language models to help improve recognition and management of incidental imaging findings and to be applied flexibly in a medical context.

Two-stage multivariate Mendelian randomization on multiple outcomes with mixed distributions.

In clinical research, it is important to study whether certain clinical factors or exposures have causal effects on clinical and patient-reported outcomes such as toxicities, quality of life, and self-reported symptoms, which can help improve patient care. Usually, such outcomes are recorded as multiple variables with different distributions. Mendelian randomization (MR) is a commonly used technique for causal inference with the help of genetic instrumental variables to deal with observed and unobserved confounders. Nevertheless, the current methodology of MR for multiple outcomes only focuses on one outcome at a time, meaning that it does not consider the correlation structure of multiple outcomes, which may lead to a loss of statistical power. In situations with multiple outcomes of interest, especially when there are mixed correlated outcomes with different distributions, it is much more desirable to jointly analyze them with a multivariate approach. Some multivariate methods have been proposed to model mixed outcomes; however, they do not incorporate instrumental variables and cannot handle unmeasured confounders. To overcome the above challenges, we propose a two-stage multivariate Mendelian randomization method (MRMO) that can perform multivariate analysis of mixed outcomes using genetic instrumental variables. We demonstrate that our proposed MRMO algorithm can gain power over the existing univariate MR method through simulation studies and a clinical application on a randomized Phase III clinical trial study on colorectal cancer patients.

A Bayesian approach for two-stage multivariate Mendelian randomization with mixed outcomes.

Many research studies have investigated the relationship between baseline factors or exposures, such as patient demographic and disease characteristics, and study outcomes such as toxicities or quality of life, but results from most of these studies may be problematic because of potential confounding effects (eg, the imbalance in baseline factors or exposures). It is important to study whether the baseline factors or exposures have causal effects on the clinical outcomes, so that clinicians can have better understanding of the diseases and develop personalized medicine. Mendelian randomization (MR) provides an efficient way to estimate the causal effects using genetic instrumental variables to handle confounders, but most of the existing studies focus on a single outcome at a time and ignores the correlation structure of multiple outcomes. Given that clinical outcomes like toxicities and quality of life are usually a mixture of different types of variables, and multiple datasets may be available for such outcomes, it may be much more beneficial to analyze them jointly instead of separately. Some well-established methods are available for building multivariate models on mixed outcomes, but they do not incorporate MR mechanism to deal with the confounders. To overcome these challenges, we propose a Bayesian-based two-stage multivariate MR method for mixed outcomes on multiple datasets, called BMRMO. Using simulation studies and clinical applications on the CO.17 and CO.20 studies, we demonstrate better performance of our approach compared to the commonly used univariate two-stage method.

Clinical significance of clonal hematopoiesis in the setting of autologous stem cell transplantation for lymphoma.

Autologous stem cell transplantation (ASCT) remains a key therapeutic strategy for treating patients with relapsed or refractory non-Hodgkin and Hodgkin lymphoma. Clonal hematopoiesis (CH) has been proposed as a major contributor not only to the development of therapy-related myeloid neoplasms but also to inferior overall survival (OS) in patients who had undergone ASCT. Herein, we aimed to investigate the prognostic implications of CH after ASCT in a cohort of 420 lymphoma patients using ultra-deep, highly sensitive error-correction sequencing. CH was identified in the stem cell product samples of 181 patients (43.1%) and was most common in those with T-cell lymphoma (72.2%). The presence of CH was associated with a longer time to neutrophil and platelet recovery. Moreover, patients with evidence of CH had inferior 5-year OS from the time of first relapse (39.4% vs. 45.8%, p = .043) and from the time of ASCT (51.8% vs. 59.3%, p = .018). The adverse prognostic impact of CH was not due to therapy-related myeloid neoplasms, the incidence of which was low in our cohort (10-year cumulative incidence of 3.3% vs. 3.0% in those with and without CH, p = .445). In terms of specific-gene mutations, adverse OS was mostly associated with PPM1D mutations (hazard ratio (HR) 1.74, 95% confidence interval (CI) 1.13-2.67, p = .011). In summary, we found that CH is associated with an increased risk of non-lymphoma-related death after ASCT, which suggests that lymphoma survivors with CH may need intensified surveillance strategies to prevent and treat late complications.

Eosinophilia and Lung Cancer: Analysis From Real-World Data and Mendelian Randomization Study.

Background and Objective: Growing evidence added to the results from observational studies of lung cancer patients exhibiting eosinophilia. However, whether eosinophils contributed to tumor immune surveillance or neoplastic evolution was unknown. This study aimed to analyze the causal association between eosinophilia and lung cancer. Methods: The causal effect of eosinophil count on lung cancer from a genome-wide association study (GWAS) was investigated using the two-sample Mendelian randomization (MR) method. Secondary results according to different histological subtypes of lung cancer were also implemented. Meanwhile, we compared the measured levels of blood eosinophil counts among different subtypes of lung cancer from real-world data. Results: The median absolute eosinophilic count (unit: 109/L) [median (min, max): Lung adenocarcinoma 0.7 (0.5, 15); Squamous cell lung cancer 0.7 (0.5, 1.3); Small cell lung cancer 0.7 (0.6, 1.3); p = 0.96] and the median eosinophil to leukocyte ratio [median (min, max): Lung adenocarcinoma 8.7% (2.1, 42.2%); Squamous cell lung cancer 9.3% (4.1, 17.7%); Small cell lung cancer 8.9% (5.1, 24.1%); p = 0.91] were similar among different histological subtypes of lung cancer. MR methods indicated that eosinophilia may provide 28% higher risk for squamous cell lung cancer in East Asian [Weighted median method: odds ratio (OR) = 1.28, 95% CI: 1.04-1.57, p = 0.02]. Conclusion: Our study suggested that eosinophilia may be a potential causal risk factor in the progression of squamous cell lung cancer in East Asian.

Speeding up Monte Carlo simulations for the adaptive sum of powered score test with importance sampling.

A central but challenging problem in genetic studies is to test for (usually weak) associations between a complex trait (e.g., a disease status) and sets of multiple genetic variants. Due to the lack of a uniformly most powerful test, data-adaptive tests, such as the adaptive sum of powered score (aSPU) test, are advantageous in maintaining high power against a wide range of alternatives. However, there is often no closed-form to accurately and analytically calculate the p-values of many adaptive tests like aSPU, thus Monte Carlo (MC) simulations are often used, which can be time consuming to achieve a stringent significance level (e.g., 5e-8) used in genome-wide association studies (GWAS). To estimate such a small p-value, we need a huge number of MC simulations (e.g., 1e+10). As an alternative, we propose using importance sampling to speed up such calculations. We develop some theory to motivate a proposed algorithm for the aSPU test, and show that the proposed method is computationally more efficient than the standard MC simulations. Using both simulated and real data, we demonstrate the superior performance of the new method over the standard MC simulations.

Oesophagectomy following noncurative endoscopic resection for oesophageal carcinoma: does interval matter?

OBJECTIVES: Oesophagectomy was always recommended after noncurative endoscopic resection (ER). And the optimal time interval from ER to oesophagectomy remains unclear. This study was to explore the effect of interval on pathologic stage and prognosis. METHODS: We included 155 patients who underwent ER for cT1N0M0 oesophageal cancer and then received subsequent oesophagectomy from 2009 to 2019. Overall survival and disease-free survival (DFS) were analysed to find an optimal cut-off of interval from ER to oesophagectomy. In addition, pathologic stage after ER was compared to that of oesophagectomy. Logistic regression model was built to identify risk factors for pathological upstage. RESULTS: The greatest difference of DFS was found in the groups who underwent oesophagectomy before and after 30 days (P = 0.016). Among total 155 patients, 106 (68.39%) received oesophagectomy within 30 days, while 49 (31.61%) had interval over 30 days. Comparing the pathologic stage between ER and oesophagectomy, 26 patients had upstage and thus had worse DFS (hazard ratio = 3.780, P = 0.042). T1b invasion, lymphovascular invasion and interval >30-day group had a higher upstage rate (P = 0.014, P < 0.001 and P < 0.001, respectively). And they were independent risk factors for pathologic upstage (odds ratio = 3.782, 4.522 and 2.844, respectively). CONCLUSIONS: It was the first study exploring the relationship between time interval and prognosis in oesophageal cancer. The longer interval between noncurative ER and additional oesophagectomy was associated with a worse DFS, so oesophagectomy was recommended performed within 1 month after ER. Older age, T1b stage, lymphovascular invasion and interval >30 days were significantly associated with pathologic upstage, which is related to the worse outcome too.

Combining the strengths of inverse-variance weighting and Egger regression in Mendelian randomization using a mixture of regressions model.

With the increasing availability of large-scale GWAS summary data on various traits, Mendelian randomization (MR) has become commonly used to infer causality between a pair of traits, an exposure and an outcome. It depends on using genetic variants, typically SNPs, as instrumental variables (IVs). The inverse-variance weighted (IVW) method (with a fixed-effect meta-analysis model) is most powerful when all IVs are valid; however, when horizontal pleiotropy is present, it may lead to biased inference. On the other hand, Egger regression is one of the most widely used methods robust to (uncorrelated) pleiotropy, but it suffers from loss of power. We propose a two-component mixture of regressions to combine and thus take advantage of both IVW and Egger regression; it is often both more efficient (i.e. higher powered) and more robust to pleiotropy (i.e. controlling type I error) than either IVW or Egger regression alone by accounting for both valid and invalid IVs respectively. We propose a model averaging approach and a novel data perturbation scheme to account for uncertainties in model/IV selection, leading to more robust statistical inference for finite samples. Through extensive simulations and applications to the GWAS summary data of 48 risk factor-disease pairs and 63 genetically uncorrelated trait pairs, we showcase that our proposed methods could often control type I error better while achieving much higher power than IVW and Egger regression (and sometimes than several other new/popular MR methods). We expect that our proposed methods will be a useful addition to the toolbox of Mendelian randomization for causal inference.

Model checking via testing for direct effects in Mendelian Randomization and transcriptome-wide association studies.

It is of great interest and potential to discover causal relationships between pairs of exposures and outcomes using genetic variants as instrumental variables (IVs) to deal with hidden confounding in observational studies. Two most popular approaches are Mendelian randomization (MR), which usually use independent genetic variants/SNPs across the genome, and transcriptome-wide association studies (TWAS) (or their generalizations) using cis-SNPs local to a gene (or some genome-wide and likely dependent SNPs), as IVs. In spite of their many promising applications, both approaches face a major challenge: the validity of their causal conclusions depends on three critical assumptions on valid IVs, and more generally on other modeling assumptions, which however may not hold in practice. The most likely as well as challenging situation is due to the wide-spread horizontal pleiotropy, leading to two of the three IV assumptions being violated and thus to biased statistical inference. More generally, we'd like to conduct a goodness-of-fit (GOF) test to check the model being used. Although some methods have been proposed as being robust to various degrees to the violation of some modeling assumptions, they often give different and even conflicting results due to their own modeling assumptions and possibly lower statistical efficiency, imposing difficulties to the practitioner in choosing and interpreting varying results across different methods. Hence, it would help to directly test whether any assumption is violated or not. In particular, there is a lack of such tests for TWAS. We propose a new and general GOF test, called TEDE (TEsting Direct Effects), applicable to both correlated and independent SNPs/IVs (as commonly used in TWAS and MR respectively). Through simulation studies and real data examples, we demonstrate high statistical power and advantages of our new method, while confirming the frequent violation of modeling (including valid IV) assumptions in practice and thus the importance of model checking by applying such a test in MR/TWAS analysis.

Implicating causal brain imaging endophenotypes in Alzheimer's disease using multivariable IWAS and GWAS summary data.

Recent evidence suggests the existence of many undiscovered heritable brain phenotypes involved in Alzheimer's Disease (AD) pathogenesis. This finding necessitates methods for the discovery of causal brain changes in AD that integrate Magnetic Resonance Imaging measures and genotypic data. However, existing approaches for causal inference in this setting, such as the univariate Imaging Wide Association Study (UV-IWAS), suffer from inconsistent effect estimation and inflated Type I errors in the presence of genetic pleiotropy, the phenomenon in which a variant affects multiple causal intermediate risk phenotypes. In this study, we implement a multivariate extension to the IWAS model, namely MV-IWAS, to consistently estimate and test for the causal effects of multiple brain imaging endophenotypes from the Alzheimer's Disease Neuroimaging Initiative (ADNI) in the presence of pleiotropic and possibly correlated SNPs. We further extend MV-IWAS to incorporate variant-specific direct effects on AD, analogous to the existing Egger regression Mendelian Randomization approach, which allows for testing of remaining pleiotropy after adjusting for multiple intermediate pathways. We propose a convenient approach for implementing MV-IWAS that solely relies on publicly available GWAS summary data and a reference panel. Through simulations with either individual-level or summary data, we demonstrate the well controlled Type I errors and superior power of MV-IWAS over UV-IWAS in the presence of pleiotropic SNPs. We apply the summary statistic based tests to 1578 heritable imaging derived phenotypes (IDPs) from the UK Biobank. MV-IWAS detected numerous IDPs as possible false positives by UV-IWAS while uncovering many additional causal neuroimaging phenotypes in AD which are strongly supported by the existing literature.

A powerful and versatile colocalization test.

Transcriptome-wide association studies (TWAS and PrediXcan) have been increasingly applied to detect associations between genetically predicted gene expressions and GWAS traits, which may suggest, however do not completely determine, causal genes for GWAS traits, due to the likely violation of their imposed strong assumptions for causal inference. Testing colocalization moves it closer to establishing causal relationships: if a GWAS trait and a gene's expression share the same associated SNP, it may suggest a regulatory (and thus putative causal) role of the SNP mediated through the gene on the GWAS trait. Accordingly, it is of interest to develop and apply various colocalization testing approaches. The existing approaches may each have some severe limitations. For instance, some methods test the null hypothesis that there is colocalization, which is not ideal because often the null hypothesis cannot be rejected simply due to limited statistical power (with too small sample sizes). Some other methods arbitrarily restrict the maximum number of causal SNPs in a locus, which may lead to loss of power in the presence of wide-spread allelic heterogeneity. Importantly, most methods cannot be applied to either GWAS/eQTL summary statistics or cases with more than two possibly correlated traits. Here we present a simple and general approach based on conditional analysis of a locus on multiple traits, overcoming the above and other shortcomings of the existing methods. We demonstrate that, compared with other methods, our new method can be applied to a wider range of scenarios and often perform better. We showcase its applications to both simulated and real data, including a large-scale Alzheimer's disease GWAS summary dataset and a gene expression dataset, and a large-scale blood lipid GWAS summary association dataset. An R package "jointsum" implementing the proposed method is publicly available at github.

Significance Testing for Allelic Heterogeneity.

It is useful to detect allelic heterogeneity (AH), i.e., the presence of multiple causal SNPs in a locus, which, for example, may guide the development of new methods for fine mapping and determine how to interpret an appearing epistasis. In contrast to Mendelian traits, the existence and extent of AH for complex traits had been largely unknown until Hormozdiari et al. proposed a Bayesian method, called causal variants identification in associated regions (CAVIAR), and uncovered widespread AH in complex traits. However, there are several limitations with CAVIAR. First, it assumes a maximum number of causal SNPs in a locus, typically up to six, to save computing time; this assumption, as will be shown, may influence the outcome. Second, its computational time can be too demanding to be feasible since it examines all possible combinations of causal SNPs (under the assumed upper bound). Finally, it outputs a posterior probability of AH, which may be difficult to calibrate with a commonly used nominal significance level. Here, we introduce an intersection-union test (IUT) based on a joint/conditional regression model with all the SNPs in a locus to infer AH. We also propose two sequential IUT-based testing procedures to estimate the number of causal SNPs. Our proposed methods are applicable to not only individual-level genotypic and phenotypic data, but also genome-wide association study (GWAS) summary statistics. We provide numerical examples based on both simulated and real data, including large-scale schizophrenia (SCZ) and high-density lipoprotein (HDL) GWAS summary data sets, to demonstrate the effectiveness of the new methods. In particular, for both the SCZ and HDL data, our proposed IUT not only was faster, but also detected more AH loci than CAVIAR. Our proposed methods are expected to be useful in further uncovering the extent of AH in complex traits.

Improved Use of Small Reference Panels for Conditional and Joint Analysis with GWAS Summary Statistics.

Due to issues of practicality and confidentiality of genomic data sharing on a large scale, typically only meta- or mega-analyzed genome-wide association study (GWAS) summary data, not individual-level data, are publicly available. Reanalyses of such GWAS summary data for a wide range of applications have become more and more common and useful, which often require the use of an external reference panel with individual-level genotypic data to infer linkage disequilibrium (LD) among genetic variants. However, with a small sample size in only hundreds, as for the most popular 1000 Genomes Project European sample, estimation errors for LD are not negligible, leading to often dramatically increased numbers of false positives in subsequent analyses of GWAS summary data. To alleviate the problem in the context of association testing for a group of SNPs, we propose an alternative estimator of the covariance matrix with an idea similar to multiple imputation. We use numerical examples based on both simulated and real data to demonstrate the severe problem with the use of the 1000 Genomes Project reference panels, and the improved performance of our new approach.

Meningococcal carriage within households in the African meningitis belt: A longitudinal pilot study.

OBJECTIVES: Carriers of Neisseria meningitidis are a key source of transmission. In the African meningitis belt, where risk of meningococcal disease is highest, a greater understanding of meningococcal carriage dynamics is needed. METHODS: We randomly selected an age-stratified sample of 400 residents from 116 households in Bamako, Mali, and collected pharyngeal swabs in May 2010. A month later, we enrolled all 202 residents of 20 of these households (6 with known carriers) and collected swabs monthly for 6 months prior to MenAfriVac vaccine introduction and returned 10 months later to collect swabs monthly for 3 months. We used standard bacteriological methods to identify N. meningitidis carriers and fit hidden Markov models to assess acquisition and clearance overall and by sex and age. RESULTS: During the cross-sectional study 5.0% of individuals (20/400) were carriers. During the longitudinal study, 73 carriage events were identified from 1422 swabs analyzed, and 16.3% of individuals (33/202) were identified as carriers at least once. The majority of isolates were non-groupable; no serogroup A carriers were identified. CONCLUSIONS: Our results suggest that the duration of carriage with any N. meningitidis averages 2.9 months and that males and children acquire and lose carriage more frequently in an urban setting in Mali. Our study informed the design of a larger study implemented in seven countries of the African meningitis belt.

Testing Genetic Pleiotropy with GWAS Summary Statistics for Marginal and Conditional Analyses.

There is growing interest in testing genetic pleiotropy, which is when a single genetic variant influences multiple traits. Several methods have been proposed; however, these methods have some limitations. First, all the proposed methods are based on the use of individual-level genotype and phenotype data; in contrast, for logistical, and other, reasons, summary statistics of univariate SNP-trait associations are typically only available based on meta- or mega-analyzed large genome-wide association study (GWAS) data. Second, existing tests are based on marginal pleiotropy, which cannot distinguish between direct and indirect associations of a single genetic variant with multiple traits due to correlations among the traits. Hence, it is useful to consider conditional analysis, in which a subset of traits is adjusted for another subset of traits. For example, in spite of substantial lowering of low-density lipoprotein cholesterol (LDL) with statin therapy, some patients still maintain high residual cardiovascular risk, and, for these patients, it might be helpful to reduce their triglyceride (TG) level. For this purpose, in order to identify new therapeutic targets, it would be useful to identify genetic variants with pleiotropic effects on LDL and TG after adjusting the latter for LDL; otherwise, a pleiotropic effect of a genetic variant detected by a marginal model could simply be due to its association with LDL only, given the well-known correlation between the two types of lipids. Here, we develop a new pleiotropy testing procedure based only on GWAS summary statistics that can be applied for both marginal analysis and conditional analysis. Although the main technical development is based on published union-intersection testing methods, care is needed in specifying conditional models to avoid invalid statistical estimation and inference. In addition to the previously used likelihood ratio test, we also propose using generalized estimating equations under the working independence model for robust inference. We provide numerical examples based on both simulated and real data, including two large lipid GWAS summary association datasets based on ∼100,000 and ∼189,000 samples, respectively, to demonstrate the difference between marginal and conditional analyses, as well as the effectiveness of our new approach.

Conditional analysis of multiple quantitative traits based on marginal GWAS summary statistics.

There has been an increasing interest in joint association testing of multiple traits for possible pleiotropic effects. However, even in the presence of pleiotropy, most of the existing methods cannot distinguish direct and indirect effects of a genetic variant, say single-nucleotide polymorphism (SNP), on multiple traits, and a conditional analysis of a trait adjusting for other traits is perhaps the simplest and most common approach to addressing this question. However, without individual-level genotypic and phenotypic data but with only genome-wide association study (GWAS) summary statistics, as typical with most large-scale GWAS consortium studies, we are not aware of any existing method for such a conditional analysis. We propose such a conditional analysis, offering formulas of necessary calculations to fit a joint linear regression model for multiple quantitative traits. Furthermore, our method can also accommodate conditional analysis on multiple SNPs in addition to on multiple quantitative traits, which is expected to be useful for fine mapping. We provide numerical examples based on both simulated and real GWAS data to demonstrate the effectiveness of our proposed approach, and illustrate possible usefulness of conditional analysis by contrasting its result differences from those of standard marginal analyses.