Classification Accuracy of Linear Discriminant Function using Principal Components with Multiple Correlated Variables: A Simulation based Exploration

Exploring sex classification from earprints - A comparison of supervised machine learning algorithms and conventional linear discriminant analysis.

INTRODUCTION: Gallbladder (GB) cancer (GBC) accounted for 3960 deaths in the US in 2019. With only 1 in 5 GBCs diagnosed at early stage, the overall 5-year survival rate is a meager 19%. Given the low 5-year survival, the best treatment is prevention. Risk factors for GBC include age, ethnicity, GB polyps >1 cm, primary sclerosing cholangitis, anomalous pancreaticobiliary junction, toxins, and GB stones. There remains controversy as to whether size of GB stones increases the risk of GBC. The aim of this systematic review and meta-analysis was to determine if risk of GBC increases with increasing size of GB stones. METHODS: A systematic search was conducted using the terms “gallbladder cancer,” “cholelithiasis,” “gallstones,” and “gallstone size” with all permutations of pubmed/medline, google scholar, embase, and cochrane. Papers were translated, if necessary. A recursive search of bibliographies of included articles was also performed. GB stone size was primarily determined post-cholecystectomy. The relevant data was inputted into RevMan 5.4 in order to generate odds ratios and Forest plots. RESULTS: The search yielded 357 total articles, of which 8 met our final inclusion criteria with GB stone measurement and prevalence of GBC in each study population. The data was then divided into 3 groups, based on size of GB stones, greater or lesser than 3 cm, 2 cm, or 1 cm. We found that the OR of developing GBC was 1.93 [1.29–2.90], 1.42 [1.06–1.90], and 1.62 [1.24–2.11] for stones greater than 3 cm, 2 cm, and 1 cm, respectively. CONCLUSION: Patients with GB stones >1 cm had increased odds of developing GBC, however we did not observe a linear relationship between GBC and increasing stone size. Our meta-analysis found an OR of 1.93 for stones >3 cm which is significantly less than the 9.1 previously reported. The association between GB stone size and GBC is likely related to the chronicity of GB stones rather than absolute size. Jain et al. proposed that chronic inflammation could subsequently lead to loss of heterozygosity of tumor suppressor genes. In summary, we found the risk of GBC is similar regardless of size of stones >1 cm. Therefore while we usually observe asymptomatic GB stones, clinicians should be aware that GB stones >1 cm may have increased risk of GBC in symptomatic individuals. Further studies are needed in asymptomatic individuals to define the optimal GB stone size cutoff to perform cholecystectomy to reduce GBC risk.Figure 1.: Odds of developing gallbladder cancer with increasing size of gallstones. (Events = # of cases of gallbladder cancer; Total = # of cases of patients with gallstones > or < the specified size).

Fractal analysis of the computed tomography images of vertebrae on the thoraco-lumbar region in diagnosing osteoporotic bone damage.

Extracellular vesicles (EVs), which are heterogeneous membrane-based vesicles with bilayer cell membrane structures, could be versatile biomarkers for the identification of diverse diseases including cancers. With this potential, this study has attempted the Raman spectroscopic identification of gall bladder (GB) cancer by directly measuring the EV solution extracted from human bile without further sample drying. For this purpose, bile samples were obtained from four normal individuals and 21 GB polyp, eight hepatocellular carcinoma (HCC), and five GB cancer patients, and EVs were extracted from each of the bile samples. The Raman peak shapes of the EVs extracted from the GB cancer samples, especially the relative intensities of peaks in the 1560-1340 cm-1 range, were dissimilar to those of the samples from the normal, GB polyp, and HCC groups. The intensity ratios of peaks at 1537 and 1453 cm-1 and at 1395 and 1359 cm-1 of the GB cancer samples were lower and higher, respectively, than those of the samples of the remaining three groups. The differences of peak intensity ratios were statistically significant based on the Mann-Whitney U test. DNA/RNA bases, amino acids, and bile salts contributed to the spectra of EVs, and their relative abundances seemed to vary according to the occurrence of GB cancer. The varied metabolite compositions and/or structures of EVs were successfully demonstrated by the dissimilar peak intensity ratios in the Raman spectra, thereby enabling the discrimination of GB cancer.

BACKGROUND: Gall stones are the most common biliary pathology. Cholelithiasis, whether symptomatic or asymptomatic, needs cholecystectomy due to 70-98% association of gall bladder carcinoma with gall stones. Despite this, the prevalence of carcinoma gall bladder is 2-5% with 0.3-0.5% being incidentally diagnosed in patients undergoing cholecystectomy. In such cases, clinical judgment and high index of suspicion in a rural setup is a worthy tool. AIMS AND OBJECTIVES: The study was undertaken to determine incidence of cholelithiasis and gall bladder carcinoma in various age groups among patients undergoing cholecystectomy and to standardize stage wise treatment of gall bladder carcinoma, specifically in relation to incidentally diagnosed gall bladder carcinoma. MATERIALS AND METHODS: A total of 189 patients undergoing cholecystectomy over a period of two and a half years were analyzed in a rural hospital, attached to medical college. Patients with gall bladder carcinoma were followed 3 monthly until death or till 31ST Oct 2013. RESULTS: Maximum number of patients were in age group of 51-60 years (22.2%) with female to male ratio of 1.9:1. Pain in right hypochondrium was the most common symptom (76.2%) patients. Cholelithiasis was present in 75% patients with gall bladder carcinoma. Incidental diagnosis of gall bladder carcinoma in patients undergoing cholecystectomy for benign disease was 1.5%. Overall incidence of gall bladder carcinoma in patients undergoing cholecystectomy was 4.2%. After confirming the final diagnosis by histopathology, ultrasound showed only 62.5% sensitivity, 97.79% specificity, 55.56% positive predictive value, 98.33% negative predictive value and 96.3% accuracy to diagnose gall bladder carcinoma. Among three incidentally diagnosed patients, operated by laparoscopic cholecystectomy, two were alive at 8 and 13 months of follow-up respectively. CONCLUSION: A sound knowledge of various modes of presentation of cholelithiasis and gall bladder carcinoma, combined with clinical judgment, high suspicion in high risk areas and confirmation by newer diagnostic modalities like USG and CT is required for early diagnosis and treatment and prevention of complications.

Despite the importance of data reduction as part of the processing of reflection-based classifications, this study represents one of the first in which the effects of both spatial and spectral data reductions on classification accuracies are quantified. Furthermore, the effects of approaches to data reduction were quantified for two separate classification methods, linear discriminant analysis (LDA) and support vector machine (SVM). As the model dataset, reflection data were acquired using a hyperspectral camera in 230 spectral channels from 401 to 879 nm (spectral resolution of 2.1 nm) from field pea (Pisum sativum) samples with and without internal pea weevil (Bruchus pisorum) infestation. We deployed five levels of spatial data reduction (binning) and eight levels of spectral data reduction (40 datasets). Forward stepwise LDA was used to select and include only spectral channels contributing the most to the separation of pixels from non-infested and infested field peas. Classification accuracies obtained with LDA and SVM were based on the classification of independent validation datasets. Overall, SVMs had significantly higher classification accuracies than LDAs (P < 0.01). There was a negative association between pixel resolution and classification accuracy, while spectral binning equivalent to up to 98% data reduction had negligible effect on classification accuracies. This study supports the potential use of reflection-based technologies in the quality control of food products with internal defects, and it highlights that spatial and spectral data reductions can (1) improve classification accuracies, (2) vastly decrease computer constraints, and (3) reduce analytical concerns associated with classifications of large and high-dimensional datasets.

Screening of gall bladder cancer through infrared analysis of bile and examination of varied bile constituent composition by the disease

This study aimed to evaluate the diagnostic performance of an artificial intelligence (AI)-based platform (Diagnocat) in detecting periapical radiolucencies (PARLs) in cone-beam computed tomography (CBCT) scans of molars. Specifically, we assessed Diagnocat's performance in detecting PARLs in non-root-filled molars and compared its diagnostic performance between preoperative and postoperative scans. This retrospective study analyzed preoperative and postoperative CBCT scans of 134 molars (327 roots). PARLs detected by Diagnocat were compared with assessments independently performed by two experienced endodontists, serving as the reference standard. Diagnostic performance was assessed at both tooth and root levels using sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), F1 score, and the area under the receiver operating characteristic curve (AUC-ROC). In preoperative scans of non-root-filled molars, Diagnocat demonstrated high sensitivity (teeth: 93.9 %, roots: 86.2 %), moderate specificity (teeth: 65.2 %, roots: 79.9 %), accuracy (teeth: 79.1 %, roots: 82.6 %), PPV (teeth: 71.8 %, roots: 75.8 %), NPV (teeth: 91.8 %, roots: 88.8 %), and F1 score (teeth: 81.3 %, roots: 80.7 %) for PARL detection. The AUC was 0.76 at the tooth level and 0.79 at the root level. Postoperative scans showed significantly lower PPV (teeth: 54.2 %; roots: 46.9 %) and F1 scores (teeth: 67.2 %; roots: 59.2 %). Diagnocat shows promise in detecting PARLs in CBCT scans of non-root-filled molars, demonstrating high sensitivity but moderate specificity, highlighting the need for human oversight to prevent overdiagnosis. However, diagnostic performance declined significantly in postoperative scans of root-filled molars. Further research is needed to optimize the platform's performance and support its integration into clinical practice. AI-based platforms such as Diagnocat can assist clinicians in detecting PARLs in CBCT scans, enhancing diagnostic efficiency and supporting decision-making. However, human expertise remains essential to minimize the risk of overdiagnosis and avoid unnecessary treatment.

IntroductionDecision-making in critical care is inherently complex, driven by patient acuity, illness severity, and the urgency of timely interventions. Cognitive challenges, combined with clinician fatigue and ingrained misconceptions, often result in systematic reasoning errors. Generative AI chatbots like ChatGPT and Gemini may offer novel approaches to address these challenges by providing accurate, evidence-based information and dispelling myths. This study investigates the myth-debunking efficacy of these chatbots in the context of critical care decision-making.ObjectivesEvaluate and compare the performance of ChatGPT and Gemini in identifying and debunking common myths associated with decision-making in critical care, using predefined performance metrics.Materials and methodsA list of 12 myths and facts was compiled following an extensive literature review1 and validated by an expert panel. ChatGPT and Gemini were tasked with classifying each statement as a myth or fact. Performance was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), F1 score, and accuracy.ResultsChatGPT achieved perfect scores across all metrics, with 100% sensitivity, specificity, PPV, NPV, F1 score, and accuracy. Gemini also performed well, with a sensitivity of 90%, specificity of 100%, PPV of 100%, NPV of 67%, an F1 score of 95%, and accuracy of 92%. ChatGPT demonstrated superior reliability in identifying and debunking critical care myths compared to Gemini.TableVarious performance metrics on debunking myths regarding Decision making in Critical careMetricsChatGPTGeminiSensitivity100%90%Specificity100%100%PPV100%100%NPV100%67%F1 score100%95%Accuracy100%92%Discussions/ConclusionsThis study's findings align with previous research on AI chatbots in palliative care and sleep health, demonstrating their effectiveness across domains. In critical care, GAI chatbots achieved 100% accuracy, with one outperforming the other in sensitivity (90%) and negative predictive value (67%). In a similar study in palliative care, ChatGPT showed high sensitivity (93.3%), while Gemini achieved perfect accuracy.2 For a study done on sleep health myths, ChatGPT had 85% sensitivity and 100% PPV, reliably identifying false statements, though occasional misses were noted. The sleep health study showed ChatGPT's alignment with expert opinions (ICC: 0.83 for falseness) and clearer explanations compared to experts’ technical language, mirroring its strong performance in critical care with no false positives or negatives.3 These comparisons highlight ChatGPT's versatility in debunking myths across healthcare domains. However, variations in sensitivity and specificity suggest the need for further refinement to address domain-specific complexities. GAI chatbots show promise as accessible complements to expert education, though continued optimization is needed to address nuanced misconceptions.ConclusionThe integration of GAI chatbots in medical education and clinical practice, especially in complex topics like decision-making in critical care, can help reduce reasoning errors. Further research is needed to validate and expand on these findings.

Sudden Sensorineural Hearing Loss (SSNHL) is routinely encountered in otolaryngology clinics. The prognosis of SSNHL varies dramatically and depends on multiple influence factors. This study aimed to develop an explainable Machine Learning (ML) model with easily accessible features to predict the prognosis of SSNHL. This bi-center retrospective study included 534 patients with SSNHL. We randomly split the data into training and validation sets. Univariate analysis and Least Absolute Shrinkage and Selection Operator (LASSO) regression were employed to select predominant features, including demographic, disease-specific characteristics, and laboratory items. We evaluated the performance of five ML models constructed with six crucial variables using the Area Under the receiver operating characteristic Curve (AUC), accuracy, specificity, sensitivity, and F1 scores. These models were further calibrated by calibration curve and Brier score. Clinical utility was evaluated by Decision Curve Analysis (DCA). The Shapley Additive Explanations (SHAP) method was applied to interpret feature contribution and explain the ML models. The Random Forest (RF) model reached the highest AUC of 0.998. Its accuracy, sensitivity, specificity, and F1 score were 0.981, 0.963, 0.989, and 0.967, respectively. DCA curve analysis revealed a comparable net benefit of the models. The SHAP method revealed that the primary features were degree of hearing loss, audiogram type, age, Mean Corpuscular Volume (MCV), onset to treatment, and serum Albumin (ALB) accordingly. The explainable ML model was superb in predicting hearing outcome and providing information on feature contributions. Clinicians can better understand the contributors to hearing recovery and guide aural rehabilitation using the SHAP method. Level 3.

BackgroundOne of the most promising applications for electroencephalogram (EEG)-based brain computer interface is for stroke rehabilitation. Implemented as a standalone motor imagery (MI) training system or as part of a rehabilitation robotic system, many studies have shown benefits of using them to restore motor control in stroke patients. Hand movements have widely been chosen as MI tasks. Although potentially more challenging to analyze, wrist and forearm movement such as wrist flexion/extension and forearm pronation/supination should also be considered for MI tasks, because these movements are part of the main exercises given to patients in conventional stroke rehabilitation. This paper will evaluate the effectiveness of such movements for MI tasks.MethodsThree hand and wrist movement tasks which were hand opening/closing, wrist flexion/extension and forearm pronation/supination were chosen as motor imagery tasks for both hands. Eleven subjects participated in the experiment. All of them completed hand opening/closing task session. Ten subjects completed two MI task sessions which were hand opening/closing and wrist flexion/extension. Five subjects completed all three MI tasks sessions. Each MI task comprised 8 sessions spanning a 4 weeks period. For classification, feature extraction based on common spatial pattern (CSP) algorithm was used. Two types were implemented, one with conventional CSP (termed WB) and one with an increase number of features achieved by filtering EEG data into five bands (termed FB). Classification was done by linear discriminant analysis (LDA) and support vector machine (SVM).ResultsEight-fold cross validation was applied on EEG data. LDA and SVM gave comparable classification accuracy. FB achieved significantly higher classification accuracy compared to WB. The accuracy of classifying wrist flexion/extension task were higher than that of classifying hand opening/closing task in all subjects. Classifying forearm pronation/supination task achieved higher accuracy than classifying hand opening/closing task in most subjects but achieved lower accuracy than classifying wrist flexion/extension task in all subjects. Significant improvements of classification accuracy were found in nine subjects when considering individual sessions of experiments of all MI tasks. The results of classifying hand opening/closing task and wrist flexion/extension task were comparable to the results of classifying hand opening/closing task and forearm pronation/supination task. Classification accuracy of wrist flexion/extension task and forearm pronation/supination task was lower than those of hand movement tasks and wrist movement tasks.ConclusionHigh classification accuracy of the three MI tasks support the possibility of using EEG-based stroke rehabilitation system with these movements. Either LDA or SVM can equally be chosen as a classifier since the difference of their accuracies is not statistically significant. Significantly higher classification accuracy made FB more suitable for classifying MI task compared to WB. More training sessions could potentially lead to better accuracy as evident in most subjects in this experiment.

Recent advancements in the field of neuroimaging have demonstrated that functional near-infrared spectroscopy (fNIRS) has a great potential for deception decoding. fNIRS is relatively new non-invasive optical imaging technique that has the advantages of safety, low cost, portability and easy to use. Light in the range of near infrared is used in fNIRS to measure the hemoglobin concentrations and thus determine the neural activity. In this paper we have compared the abilities of two classification techniques: linear discriminant analysis (LDA) and support vector machine (SVM), to classify the fNIRS data for deception decoding. Five healthy male subjects participated in deception and truth-telling scenarios separately in this study. Signals from the prefrontal cortex of the subjects were collected using continuous wave fNIRS. HbO and HbR signals were used to define the features and then data was classified using LDA and SVM classifiers. The average classification accuracy of SVM is 87.33 % whereas average classification accuracy for LDA is 78.34 %. The higher classification accuracy resulting from SVM is in accordance with the previous literature.

A brain-computer interface (BCI) is a system that enables control of devices or communication with other persons, only through cerebral activity, without using muscles. The main application for BCIs is assistive technology for disabled persons. Examples for devices that can be controlled by BCIs are artificial limbs, spelling devices, or environment control systems. BCI research has seen renewed interest in recent years, and it has been convincingly shown that communication via a BCI is in principle feasible. However, present day systems still have shortcomings that prevent their widespread application. In part, these shortcomings are caused by limitations in the functionality of the pattern recognition algorithms used for discriminating brain signals in BCIs. Moreover, BCIs are often tested exclusively with able-bodied persons instead of conducting tests with the target user group, namely disabled persons. The goal of this thesis is to extend the functionality of pattern recognition algorithms for BCI systems and to move towards systems that are helpful for disabled users. We discuss extensions of linear discriminant analysis (LDA), which is a simple but efficient method for pattern recognition. In particular, a framework from Bayesian machine learning, the so-called evidence framework, is applied to LDA. An algorithm is obtained that learns classifiers quickly, robustly, and fully automatically. An extension of this algorithm allows to automatically reduce the number of sensors needed for acquisition of brain signals. More specifically, the algorithm allows to perform electrode selection. The algorithm for electrode selection is based on a concept known as automatic relevance determination (ARD) in Bayesian machine learning. The last part of the algorithmic development in this thesis concerns methods for computing accurate estimates of class probabilities in LDA-like classifiers. These probabilities are used to build a BCI that dynamically adapts the amount of acquired data, so that a preset, approximate bound on the probability of misclassifications is not exceeded. To test the algorithms described in this thesis, a BCI specifically tailored for disabled persons is introduced. The system uses electroencephalogram (EEG) signals and is based on the P300 evoked potential. Datasets recorded from five disabled and four able-bodied subjects are used to show that the Bayesian version of LDA outperforms plain LDA in terms of classification accuracy. Also, the impact of different static electrode configurations on classification accuracy is tested. In addition, experiments with the same datasets demonstrate that the algorithm for electrode selection is computationally efficient, yields physiologically plausible results, and improves classification accuracy over static electrode configurations. The classification accuracy is further improved by dynamically adapting the amount of acquired data. Besides the datasets recorded from disabled and able-bodied subjects, benchmark datasets from BCI competitions are used to show that the algorithms discussed in this thesis are competitive with state-of-the-art electroencephalogram (EEG) classification algorithms. While the experiments in this thesis are uniquely performed with P300 datasets, the presented algorithms might also be useful for other types of BCI systems based on the EEG. This is the case because functionalities such as robust and automatic computation of classifiers, electrode selection, and estimation of class probabilities are useful in many BCI systems. Seen from a more general point of view, many applications that rely on the classification of cerebral activity could possibly benefit from the methods developed in this thesis. Among the potential applications are interrogative polygraphy ("lie detection") and clinical applications, for example coma outcome prognosis and depth of anesthesia monitoring.

The problems of adulteration and mislabeling are very common in the food industry. Laser-induced breakdown spectroscopy (LIBS) coupled with chemometric methods has many intrinsic advantages on adulteration analysis of various materials. In this work, several chemometric algorithms, i.e., principal component analysis (PCA), decision tree (DT), random forest (RF), partial least squares discriminant analysis (PLS-DA), linear discriminant analysis (LDA), and support vector machine (SVM), were carried out assisted by LIBS technology to study the classification performances on rice geographic origins. A series of samples, including 20 kinds of rice samples from different geographic origins, was detected using LIBS with no pretreatment processes. For data analysis, PCA was employed to reduce the input variables, and to reduce the collinearity of LIBS spectral results as well. The results showed the classification accuracies of the mentioned chemometric algorithms of DT, RF, PLS-DA, LDA, and SVM with 89 input variables of 86.80%, 96.30%, 96.80%, 98.60%, and 99.20%, respectively. At the same time, the operation times of these algorithms were 3.81 s, 54.64 s, 3.63 s, 2.09 s, and 531.01 s, respectively. On the other hand, 30 principal components of input variables were also tested under the same conditions. The classification accuracies for the above algorithms were 81.60%, 98.00%, 95.70%, 98.40%, and 99.20%, respectively. The operation times were 2.01 s, 4.88 s, 3.67 s, 0.36 s, and 308.55 s, respectively. In addition, the five-fold cross-validation classification accuracies with 30 input variables for DT, RF, PLS-DA, LDA, and SVM were 83.75%, 97.95%, 94.75%, 98.35%, and 99.25%, respectively. As a result, LDA was demonstrated to be the best and most efficient tool for rice geographic origin classification assisted by LIBS with high accuracy and analytical speed, which has great potential for rapid identification of adulterated products in agriculture without use of any chemical reagent.

Transabdominal ultrasonography (TAUS) is the most commonly used modality to diagnose gallbladder (GB) disease. GB polyps are reported in 1-5.6 % of TAUS studies. Histopathologic studies suggest that there is a relationship between GB polyps and GB cancer. Previous literature suggests GB polyps reported on TAUS do not correlate well with histological findings. There have been recent advances in TAUS technology. We hypothesize the recent advances in TAUS technology have improved the accuracy of TAUS for diagnosing GB polyps. Radiology and pathology databases at our tertiary care center were retrospectively searched between January 1, 2000, and December 31, 2010. Ultrasound reports that suggested a GB polyp was present on TAUS were correlated to histopathology in cases where a cholecystectomy was performed. The pathology reports where a GB polyp was found were correlated with preoperative TAUS reports. There were 102,740 TAUS reports referring to the GB, of which 6,612 (6.4 %) contained search terms suggesting a GB polyp was present. There were 13,278 cholecystectomy pathology reports, of which 159 (1.2 %) included a diagnosis of GB polyp. TAUS detected only 50 % of the polyps identified on histopathology. The sensitivity and specificity of TAUS for diagnosing GB polyps were 50.0 and 98.3 %, respectively. The positive and negative predictive values were 10.5 and 99.8 %. Despite improvement in TAUS technology, the accuracy for GB polyps remains poor. This needs to be considered when managing patients with TAUS-detected GB polyps. We recommend that the decision to operate on TAUS-detected GB polyps be largely based on symptoms, and following GB polyps with TAUS should be discouraged.