Reliable Classification System Research Articles

Prospective development and validation of a universal classification for paediatric videolaryngoscopic tracheal intubation: the PeDiAC score.

The VIDIAC score, a prospectively developed universal classification for videolaryngoscopy, has shown excellent diagnostic performance in adults. However, there is no reliable classification system for videolaryngoscopic tracheal intubation in children. We aimed to develop and validate a multivariable logistic regression model and easy-to-use score to classify difficult videolaryngoscopic tracheal intubation in children and to compare it with the Cormack and Lehane classification. A secondary aim was to externally validate the VIDIAC score in children. We conducted a prospective observational study within a structured universal videolaryngoscopy implementation programme. We used C-MAC™ videolaryngoscopes in all children undergoing tracheal intubation for elective surgical procedures. We validated the VIDIAC score externally and assessed its performance. We then identified eligible co-variables for inclusion in the PeDiAC score, developed a multivariable logistic regression model and compared its performance against the Cormack and Lehane classification. We studied 809 children undergoing 904 episodes of tracheal intubation. The VIDIAC score outperformed the Cormack and Lehane classification when classifying the difficulty of videolaryngoscopic tracheal intubation, with an area under the receiver operating characteristic curve of 0.80 (95%CI 0.73-0.87) vs. 0.69 (95%CI 0.62-0.76), respectively, p = 0.018. Eight eligible tracheal intubation-related factors, that were selected by 100-times repeated 10-fold cross-validated least absolute shrinkage selector operator regression, were used to develop the PeDiAC model. The PeDiAC model and score showed excellent diagnostic performance and model calibration. The PeDiAC score achieved significantly better diagnostic performance than the Cormack and Lehane classification, with an area under the receiver operating characteristic curve of 0.97 (95%CI 0.96-0.99) vs. 0.69 (95%CI 0.62-0.76), respectively, p < 0.001. We developed and validated a specifically tailored classification for paediatric videolaryngoscopic tracheal intubation with excellent diagnostic performance and calibration that outperformed the Cormack and Lehane classification.

The impact of the distal ureteral diameter ratio, bladder volume at onset of vesicoureteral reflux, and/or grade in predictive models of clinical outcomes in children with vesicoureteral reflux

Vesicoureteral reflux (VUR) grade has been used as a primary factor in assessing a child's risk of clinical outcomes. Unfortunately, grade has poor inter-observer reliability. We hypothesized that more objective and reliable VCUG parameters including the distal ureteral diameter ratio (UDR) and volume at onset of VUR (Vol) may either augment or replace the current grading system to provide more reliable prediction of clinical outcomes. Multivariate clinical outcome models were analyzed to assess the impact on predictive accuracy by the addition of voiding cystourethrogram (VCUG) data including grade, UDR, and Vol, alone or in combinations. Clinical variables from retrospective review of 841 children's records included age, gender, presentation, VUR laterality, bowel and bladder dysfunction, history of febrile urinary tract infection (UTI), and number of UTIs. The primary outcomes assessed included VUR resolution or persistence and need for operative intervention. Grade, UDR, and Vol were independent predictors of resolution and operative intervention. Vol increased predictive accuracy in resolution models with grade or UDR alone; however, no significant difference occurred in models with the substitution of grade with UDR. A more reliable classification system for VUR, with improved predictive accuracy regarding clinical outcomes, may be developed incorporating UDR and Vol. Whether VUR grade can be completely replaced by Vol and UDR measurements requires further evaluation with larger number of patients.

Decision tree based prediction system for word difficulty classification

In the era of information technology in education, personalised learning is becoming increasingly important, especially in the English learning process. An important aspect of facilitating personalised learning in English reading is the use of a reliable objective word difficulty classification system to quickly capture and understand reading difficulties and core concepts. The aim of this study was to use decision tree modelling to predict the general proficiency of the public in words with different attributes. In addition, a K-Means clustering algorithm was used to categorise words into five classes based on their level of difficulty. By adopting this approach, the prediction of word difficulty becomes both fast and objective, and by testing the accuracy of the model, it was found that our model achieved an accuracy of 0.95. Accurate classification of word difficulty will play an important role in facilitating personalised learning in English reading and improve the efficiency of English reading.

Femoral bone defect classifications in revision total hip arthroplasty: a comprehensive review and proposal of a new algorithm of management.

The number of primary total hip arthroplasties (THAs) and revisions is expected to steadily grow in the future. The femoral revision surgery can be technically demanding whether severe bone defects need to be addressed. The femoral revision aims to obtain a proper primary stability of the stem with a more proximal fixation as possible. Several authors previously proposed classification systems to describe the morphology of the bony femoral defect and to drive accordingly the surgeon in the revision procedure. The previous classifications mainly considered cortical and medullary bone at the level of the defect of poor quality by definition. Therefore, the surgical strategies aimed to achieve a distal fixation bypassing the defect or to fill the defect with bone impaction grafting or structured bone grafts up to the replacement of the proximal femur with megaprosthesis. The consensus on a comprehensive and reliable classification system and management algorithm is still lacking. A new classification system should be developed taking into account the bone quality. The rationale of a new classification is that 'functional' residual bone stock could be present at the level of the defect. Therefore, it can be used to achieve a primary (mechanical) and secondary (biological) stability of the implants with a femoral fixation more proximal as possible.

Universal Long Bone Nonunion Classification.

The management of bone union disorders is a complex problem in orthopaedics, requiring a reliable and comprehensive classification system for accurate diagnosis and treatment. Despite advances in understanding pathophysiology, diagnosis, and treatment in this area, there is no generally accepted classification system. The aim of our work was to create a comprehensive classification, which will systemize the vast majority of bone union disorders, underline their differences and form the basis for their treatment. The key criteria for nonunion evaluation and treatment were identified based on the conducted literature review: Time from the initial event (delayed union or nonunion), location, type of pathology (A, Hypertrophic; B, Normotrophic; C, Oligotrophic) and the presence of hardware. Based on these criteria the ULBNC has been developed. Atrophic nonunions were excluded from this classification as they are considered segmental bone defects with special classification. The ULBNC is based on the same principles of coding as the "gold standard" AO/OTA Fractures Classification system with alpha-numeric coding "from simple to complex." The choice of treatment method depends on the type, group, and subgroup of the nonunion as described. Universal Long Bone Nonunion Classification (ULBNC) is an alphanumeric system that describes the localization, type of pathology and morphologic characteristics of a nonunion. The use of ULBNC in practice and research will optimize and standardize the treatment of various types of bone healing disorders and eventually improve clinical outcomes. Solomin LN, Semenistyy AA, Komarov AV, et al. Universal Long Bone Nonunion Classification. Strategies Trauma Limb Reconstr 2023;18(3):169-173.

Development of a comprehensive and clinically applicable novel projection classification system for anterior communicating artery aneurysms

Various surgical and anatomical classifications have been proposed to date related to ACoA aneurysm projection. Nonetheless, a universally accepted classification system is yet to be established. This study is aimed at establishing a standardized classification system for ACoA aneurysms with utilization 3D technology and defining reference lines for their projections. The goal is to create a simple, understandable, surgically beneficial, and reliable classification system based on neurovascular structures in the region, including safe and hazardous zones. The radiologic data of 96 patients with ACoA aneurysm who were treated in our university hospital between 2012 and 2020 were retrospectively analyzed, and a planned classification scale was developed with the data obtained. The classification aimed to create 9 main projection groups in the sagittal plane: superior, inferior, anterior, and posterior in linear orientation, and anterosuperior, posterosuperior, anteroinferior, posteroinferior, and complex in quadrant orientation. The coronal and axial planes included medial, lateral, and midline classifications, resulting in a 3-dimensional classification system with 25 projections. Among the 96 patients, 32 had linear and 64 had quadrant projections. In the sagittal plane, the linear projection breakdown was as follows: superior (28%), inferior (6.25%), anterior (53%), and posterior (12.5%). For the quadrant projection, the distribution was as follows: anterosuperior (53%), posterosuperior (12.5%), anteroinferior (21.87%), posteroinferior (3.12%), and complex (9.37%). Overall, 35.4% aneurysms were anterosuperior, 17.7% anterior, 14.58% anteroinferior, 9.37% superior, 8.3% posterosuperior, 6.25% complex, 4.16% posterior, 2.08% posteroinferior, and 2.08% inferior projection. Our study proposes a projection classification that utilizes 3D technology for safe surgery based on neurovascular structures in the region and thus better reveals safe and hazardous zones, including three plans, three dimensions, and two orientations. The use of this classification system offers valuable guidance for daily practice in the treatment of ACoA aneurysms.

Automatic Food Recognition Using Deep Convolutional Neural Networks with Self-attention Mechanism

The significance of food in human health and well-being cannot be overemphasized. Nowadays, in our dynamic life, people are increasingly concerned about their health due to increased nutritional ailments. For this reason, mobile food-tracking applications that require a reliable and robust food classification system are gaining popularity. To address this, we propose a robust food recognition model using deep convolutional neural networks with a self-attention mechanism (FRCNNSAM). By training multiple FRCNNSAM structures with varying parameters, we combine their predictions through averaging. To prevent over-fitting and under-fitting data augmentation to generate extra training data, regularization to avoid excessive model complexity was used. The FRCNNSAM model is tested on two novel datasets: Food-101 and MA Food-121. The model achieved an impressive accuracy of 96.40% on the Food-101 dataset and 95.11% on MA Food-121. Compared to baseline transfer learning models, the FRCNNSAM model surpasses performance by 8.12%. Furthermore, the evaluation on random internet images demonstrates the model's strong generalization ability, rendering it suitable for food image recognition and classification tasks.

The Triplanar Hallux Abducto Valgus Classification System: Is it Valid?

Historically, 2-dimensional radiographic study techniques have been used to classify deformity and guide treatment of hallux valgus deformities in the transverse plane. Recently, a triplanar hallux abducto valgus classification system was proposed. The key elements of this classification system are the pathologic alignments in 3 anatomic planes. The triplanar hallux abducto valgus classification system is intended to clarify the deformity and apply a triplanar anatomic algorithm for treatment. To our knowledge, this classification system has not been validated. Our objective was to assess reliability of the triplanar hallux valgus classification system.Patients with hallux abducto valgus were identified from a foot and ankle registry. Digital radiographs were assembled in a digital slide presentation. The eligibility criteria required complete radiographic studies and represented varying degrees of hallux abducto valgus. The reviewers included 3 board-certified, fellowship-trained orthopedic foot and ankle surgeons. Each reviewer independently classified the hallux abducto valgus deformity for a total of 75 observations. After an 8-week washout period, the order of the hallux abducto valgus cases was randomized in the digital slide presentation and redistributed to the reviewers.The average kappa value from 3 readers was 0.241 with 95% CI (0.093-0.374), indicating a fair agreement. The inter-reader agreement was 0.046 with 95% CI (-0.041 to 0.112), showing poor agreement between readers.Our results indicate the triplanar hallux abducto valgus is not a reliable classification system. While this is the first known triplanar hallux abducto valgus classification system, it lacks prognostic value and reliability.

Validity of Hepple Classification Used in the Diagnosis of Talus Osteochondral Lesions.

It was aimed to investigate the intra-observer and inter-observer validity of the Hepple classification used in talus osteochondral lesions. This study included 32 patients with osteochondral lesions in the talus after exclusion criteria. A PowerPoint presentation was prepared from the MRI views of the patients. Six observers, divided into two groups according to their experience, were asked to categorize the cases according to Hepple classification. The slides were shuffled and the observers were asked to reevaluate after 6 weeks. Fleiss kappa (κ) coefficient was used for the inter-observer validity and Cohen's kappa (κ) coefficient for the intra-observer validity. In the overall inter-observer reliability was at a moderate level of agreement (set one κ=0.511, set two κ=0.406). In the intra-observer evaluation, one observer from the experienced group showed almost perfect agreement (κ=0.809), one observer from the less experienced group had moderate agreement (κ=0.556), and all other observers had substantial agreement (κ=0.556 - 0.730). When all observers were examined, it was seen that there was a substantial agreement in the mean intra-observer evaluation (κ=0.661). While the intra-observer results showed substantial agreement, the inter-observer results showed moderate agreement. Although the Hepple classification system is frequently used, the need for a more reliable classification system for osteochondral lesions of the talus remains.

Pediatric Lateral Humeral Condyle Fractures: Reliability of a Modified Jakob Classification System and its Impact on Treatment Planning With or Without Arthrography.

Various lateral humeral condyle fracture (LHCF) classification systems have been in use since the 1950s, but limited research exists on their reliability. The most widely utilized, yet un-validated system is that of Jakob and colleagues. The purpose of the current study was to analyze the reliability of a modified Jakob classification system and its value in guiding treatment either with or without arthrography. Interrater and intrarater reliability studies were performed using radiographs and arthrograms from 32 LHCFs. Radiographs were presented to 3 pediatric orthopaedic surgeons and 6 pediatric orthopaedic surgery residents who were asked to classify the fractures according to a modified Jakob classification system, enunciate their treatment plan, and whether they would utilize arthrography. Classification was repeated within 2 weeks to assess intrarater reliability. The treatment plan using radiographs only and radiographs with arthrography were compared at both rating points. The modified Jakob system had excellent interrater reliability using only radiographs with a kappa value of 0.82 and an overall agreement of 86%. The average kappa for intrarater reliability using only radiographs was 0.88 with a range of 0.79 to 1.00 and an average overall agreement of 91% with a range of 84% to 100%. Interrater and intrarater reliability was poorer using both radiographs and arthrography. On average, arthrography changed the treatment plan in 8% of cases. The modified Jakob classification system proved to be a reliable classification system for LHCFs, independent of arthrography, given the excellent free-marginal multirater kappa values. Level III-diagnostic.

PROClass: The Development and Validation of a Novel Prosthetic Component Sophistication Classification System

ObjectiveTo develop a lower limb prosthesis (LLP) sophistication classification system that categorizes prosthetic component prescriptions into “basic,” “intermediate,” and “advanced” and assess its content validity, reliability, and accuracy. DesignClassification development and validation study. SettingThe Veterans Affairs (VA) Corporate Data Warehouse database and National Prosthetics Patient Database were used to identify patients undergoing their first amputation at the transtibial or transfemoral level due to diabetes or peripheral artery disease and to identify the associated codes for each LLP. ParticipantsAn expert panel of 6 nationally recognized certified prosthetists, a national expert in VA prosthetics data and coding, a physical medicine and rehabilitation physician, and an epidemiologist developed an LLP classification system (PROClass) using 30 transfemoral and transtibial lower limb amputees. Main Outcome MeasuresThe expert panel reviewed 20 consecutive participants meeting study criteria for the development of the PROClass system and a subsequent 30 consecutive cases for assessing the inter- and intra-rater reliability and accuracy. ResultsThe interrater and intrarater reliability was almost perfect with Gwet's AC1 values ranging from .82 to .96 for both expert panel members and research assistants. The accuracy of the research assistant's classifications to the “criterion standard” was excellent with Gwet's AC1 values ranging between .75 and .92. ConclusionsPROClass is a pragmatic, reliable, and accurate prosthetic classification system with strong face validity that will enable the classification of prosthetic components used for large data set research aimed at evaluating important clinical questions such as the effects of sophistication on patient outcomes.

Third molar eruption in orthopantomograms as a feature for forensic age assessment—a comparison study of different classification systems

Evaluation of third molar eruption represents an established method for age assessment of living individuals. Different classification systems are available for the radiological assessment of third molar eruption. The aim of this study was to identify the most accurate and reliable classification system for the mandibular third molar eruption on orthopantomograms (OPG). We compared the method of Olze et al. (2012) with the method of Willmot et al. (2018) and a newly derived classification system using OPGs from 211 individuals aged 15–25 years. The assessments were performed by three experienced examiners. One examiner evaluated all radiographs twice. The correlation between age and stage was investigated and the inter- and intra-rater reliability was estimated for all three methods. Correlation between stage and age was similar between classification systems, although higher in the data from males (Spearman’s rho ranging from 0.568 to 0.583) than from females (0.440 to 0.446). Inter- and intra-rater reliability measures were similar across methods and invariant on sex, with overlapping confidence intervals, although the highest point estimates for both intra- and inter-rater reliability were for the method by Olze et al. with Krippendorf’s alpha values of 0.904 (95% confidence interval 0.854, 0.954) and 0.797 (95% confidence interval 0.744, 0.850). It was concluded that the method of Olze et al. from 2012 is a reliable method for practical application and future studies.

Brain Tumor Detection Using Block Chain

Brain tumor can be divided into two types: benign and malignant. Brain tumors are the most common and aggressive disease, which in its highest stage leads to a very short life expectancy. Thus, treatment planning is a key phase for improving patients' quality of life. However, it has some limitations (i.e., accurate quantitative measurements are provided for a limited number of frames). A reliable and automatic classification system is therefore necessary to prevent human mortality. Block chain technology is an emerging field of science that plays a major role in every application field of science, including education, banking, and healthcare. . In healthcare, most health problems arise due to their neglect of correct diagnosis by the doctor and ignorance of the symptom by the patients. The most common disease today is called cancer. A brain tumor usually has symptoms such as frequent headaches, unexplained nausea or vomiting. Sometimes he may also have blurred vision, double vision and sometimes loss of peripheral vision. In this project we will diagnose tumor using Blockchain strategy.

Uncertain-CAM: Uncertainty-Based Ensemble Machine Voting for Improved COVID-19 CXR Classification and Explainability.

The ongoing coronavirus disease 2019 (COVID-19) pandemic has had a significant impact on patients and healthcare systems across the world. Distinguishing non-COVID-19 patients from COVID-19 patients at the lowest possible cost and in the earliest stages of the disease is a major issue. Additionally, the implementation of explainable deep learning decisions is another issue, especially in critical fields such as medicine. The study presents a method to train deep learning models and apply an uncertainty-based ensemble voting policy to achieve 99% accuracy in classifying COVID-19 chest X-rays from normal and pneumonia-related infections. We further present a training scheme that integrates the cyclic cosine annealing approach with cross-validation and uncertainty quantification that is measured using prediction interval coverage probability (PICP) as final ensemble voting weights. We also propose the Uncertain-CAM technique, which improves deep learning explainability and provides a more reliable COVID-19 classification system. We introduce a new image processing technique to measure the explainability based on ground-truth, and we compared it with the widely adopted Grad-CAM method.

Measuring patient acuity and nursing care needs in South Korea: application of a new patient classification system

BackgroundAn accurate and reliable patient classification system (PCS) can help inform decisions regarding adequate assignments for nurse staffing. This study aimed to evaluate the criterion validity of the Asan Patient Classification System (APCS), a new tertiary hospital-specific PCS, by comparing its rating and total scores with those of KPCS-1 and KPCS-GW for measuring patient activity and nursing needs.MethodsWe performed a retrospective analysis of the medical records of 50,314 inpatients admitted to the general wards of a tertiary teaching hospital in Seoul, South Korea in March, June, September, and December 2019. Spearman’s correlation and Kappa statistics according to quartiles were calculated to examine the criterion validity of the APCS compared with the KPCS-1 and KPCS-GW.ResultsThe average patient classification score was 28.3 points for APCS, 25.7 points for KPCS-1, and 21.6 points for KPCS-GW. The kappa value between APCS and KPCS-1 was 0.91 (95% CI:0.9072, 0.9119) and that between APCS and KPCS-GW was 0.88 (95% CI:0.8757, 0.8810). Additionally, Spearman's correlation coefficients among APCS, KPCS-1, and KPCS-GW showed a very strong correlation. However, 10.8% of the participants’ results were inconsistent, and KPCS-1 tended to classify patients into groups with lower nursing needs compared to APCS.ConclusionThis study showed that electronic health record-generated APCS can provide useful information on patients’ severity and nursing activities to measure workload estimation. Additional research is needed to develop and implement a real-world EHR-based PCS system to accommodate for direct and indirect nursing care while considering diverse population and dynamic healthcare system.

Psychometric properties of the Viking Speech Scale-Turkish version for children with cerebral palsy aged 4-18 years based on live and video-based observation.

Speech is the most common method of communication. Video-based clinical communication evaluation is a requirement for children with speech-language impairments living in rural areas, and those who have limited mobility. To determine the validity and reliability of the Turkish version of the Viking Speech Scale (VSS-T) via live and video-based observation for children with cerebral palsy (CP) aged 4-18 years. A total of 142 children (mean age 8.18 ± 3.98 years; 68 female) with CP were included in this study. Their motor, communication, visual and eating-drinking function levels and comorbidities (dental, swallowing, cognitive impairments and epilepsy) were recorded. The Intelligibility in Context Scale (ICS), the Pediatric Evaluation of Disability Inventory-Social Function (PEDI-SF), and the Functional Independence Measure for Children-Communication (WeeFIM-C) were assessed to examine the concurrent validity of the VSS-T. The interrater reliability of the VSS-T was analysed between parents, physical therapists, and speech and language therapists from live and video-based observation. Intra-rater reliability was calculated from ratings made from live and video-based observations taken 3 weeks apart. The VSS-T was strongly related to the ICS (r = -0.830), PEDI-SF (r = -0.819), WeeFIM-C (r = -0.643), other functional classifications (r > 0.432), and the comorbidities (Cramer's V > 0.284, p < 0.001). Good to excellent interrater reliability (κw ≥ 0.838) and intra-rater reliability (intraclass correlation coefficient (ICC) = 0.848-0.995) were found between parents and therapists. Speech and language therapists, physical therapists, and parents can use the VSS-T as a valid and reliable classification system to describe speech intelligibility of 4-18-year-old children with CP. Both live and video-based observations can be used to administer the VSS-T. What is already known on the subject The English version of the VSS has been shown to be a valid and reliable tool used to classify the speech of children with CP aged 4-13 years. The scale can be administered by means of live observation of the child or based on clinicians' notes on the case by parents, SLTs, physiotherapists and paediatricians. What this paper adds to existing knowledge The VSS-T is valid and reliable for children with CP aged 4-18 years. Video-based observation is a suitable method for evaluating the VSS-T levels. The VSS-T has a moderate association with the CFCS. What are the potential or actual clinical implications of this work? The VSS-T is a valid and reliable method of categorizing the severity of motor speech impairment for Turkish children with CP in clinical research studies, registry systems or epidemiological studies. Both experienced and inexperienced therapists can use either live or video-based observation methods to administer the VSS-T. This study extended the validity and reliability of the scale in children with CP aged up to 18 years. The VSS-T is also associated with the Visual Functional Classification System (VFCS), which has been recently developed for describing the visual abilities of children with CP in daily life. In addition, the VSS-T is associated with the presence of dental, swallowing, cognitive problems and epilepsy.

A Causal Classification System for Intracerebral Hemorrhage Subtypes.

Determining the underlying causes of intracerebral hemorrhage (ICH) is of major importance, because risk factors, prognosis, and management differ by ICH subtype. We developed a new causal CLASsification system for ICH Subtypes, termed CLAS-ICH, based on recent advances in neuroimaging. CLAS-ICH defines 5 ICH subtypes: arteriolosclerosis, cerebral amyloid angiopathy, mixed small vessel disease (SVD), other rare forms of SVD (genetic SVD and others), and secondary causes (macrovascular causes, tumor, and other rare causes). Every patient is scored in each category according to the level of diagnostic evidence: (1) well-defined ICH subtype; (2) possible underlying disease; and (0) no evidence of the disease. We evaluated CLAS-ICH in a derivation cohort of 113 patients with ICH from Massachusetts General Hospital, Boston, USA, and in a derivation cohort of 203 patients from Inselspital, Bern, Switzerland. In the derivation cohort, a well-defined ICH subtype could be identified in 74 (65.5%) patients, including 24 (21.2%) with arteriolosclerosis, 23 (20.4%) with cerebral amyloid angiopathy, 18 (15.9%) with mixed SVD, and 9 (8.0%) with a secondary cause. One or more possible causes were identified in 42 (37.2%) patients. Interobserver agreement was excellent for each category (kappa value ranging from 0.86 to 1.00). Despite substantial differences in imaging modalities, we obtained similar results in the validation cohort. CLAS-ICH is a simple and reliable classification system for ICH subtyping, that captures overlap between causes and the level of diagnostic evidence. CLAS-ICH may guide clinicians to identify ICH causes, and improve ICH classification in multicenter studies. ANN NEUROL 2023;93:16-28.

Automatic Malignant and Benign Skin Cancer Classification Using a Hybrid Deep Learning Approach.

Skin cancer is one of the major types of cancer with an increasing incidence in recent decades. The source of skin cancer arises in various dermatologic disorders. Skin cancer is classified into various types based on texture, color, morphological features, and structure. The conventional approach for skin cancer identification needs time and money for the predicted results. Currently, medical science is utilizing various tools based on digital technology for the classification of skin cancer. The machine learning-based classification approach is the robust and dominant approach for automatic methods of classifying skin cancer. The various existing and proposed methods of deep neural network, support vector machine (SVM), neural network (NN), random forest (RF), and K-nearest neighbor are used for malignant and benign skin cancer identification. In this study, a method was proposed based on the stacking of classifiers with three folds towards the classification of melanoma and benign skin cancers. The system was trained with 1000 skin images with the categories of melanoma and benign. The training and testing were performed using 70 and 30 percent of the overall data set, respectively. The primary feature extraction was conducted using the Resnet50, Xception, and VGG16 methods. The accuracy, F1 scores, AUC, and sensitivity metrics were used for the overall performance evaluation. In the proposed Stacked CV method, the system was trained in three levels by deep learning, SVM, RF, NN, KNN, and logistic regression methods. The proposed method for Xception techniques of feature extraction achieved 90.9% accuracy and was stronger compared to ResNet50 and VGG 16 methods. The improvement and optimization of the proposed method with a large training dataset could provide a reliable and robust skin cancer classification system.

Modified Clavien-Dindo-Sink system is reliable for classifying complications following surgical treatment of early-onset scoliosis.

Appropriately measuring and classifying surgical complications is a critical component of research in vulnerable populations, including children with early-onset scoliosis (EOS). The purpose of this study was to assess the inter- and intra-rater reliability of a modified Clavien-Dindo-Sink system (CDS) classification system for EOS patients among a group of pediatric spinal deformity surgeons. Thirty case scenarios were developed and presented to experienced surgeons in an international spine study group. For each case, surgeons were asked to select a level of severity based on the modified CDS system to assess inter-rater reliability. The survey was administered on two occasions to allow for assessment of intra-rater reliability. Weighted Kappa values were calculated, with 0.61 to 0.80 considered substantial agreement and 0.81 to 1.00 considered nearly perfect agreement. 11/12 (91.7%) surgeons completed the first-round survey and 8/12 (66.7%) completed the second. Inter-observer weighted kappa values for the first and second survey were 0.75 [95% CI 0.56-0.94], indicating substantial agreement, and 0.84 [95% CI 0.70-0.98], indicating nearly perfect agreement, respectively. Intra-observer reliability was 0.86 (range 0.74-0.95) between the first and second surveys, indicating nearly perfect agreement . The modified CDS classification system demonstrated substantial to nearly perfect agreement between and within observers for the evaluation of complications following the surgical treatment of EOS patients. Adoption of this reliable classification system as a standard for reporting complications in EOS patients can be a valuable tool for future research endeavors, as we seek to ultimately improve surgical practices and patient outcomes. Level V.

RE-EVALUATION OF SYMPOVENTURIACEAE.

Abstract Sympoventuriaceae (Venturiales, Dothideomycetes) comprises genera including saprophytes, endophytes, plant pathogens, as well as important animal or human opportunistic pathogens with diverse ecologies and wide geographical distributions. Although the taxonomy of Sympoventuriaceae has been well studied, generic boundaries within the family remain poorly resolved due to the lack of type materials and molecular data. To address this issue and establish a more stable and reliable classification system in Sympoventuriaceae, we performed multi-locus phylogenetic analyses using sequence data of seven genes (SSU, ITS, LSU, act1, tub2, tef1 and rpb2) with increased taxon sampling and morphological analysis. The molecular data combined with detailed morphological studies of 143 taxa resolved 22 genera within the family, including one new genus, eight new species, five new combinations and one new name. Finally, we further investigated the evolutionary history of Sympoventuriaceae by reconstructing patterns of lifestyle diversification, indicating the ancestral state to be saprophytic, with transitions to endophytic, animal or human opportunistic and plant pathogens. Citation: Wei TP, Zhang H, Zeng XY, et al. 2022. Re-evaluation of Sympoventuriaceae. Persoonia 48: 219-260. https://doi.org/10.3767/persoonia.2022.48.07.. Effectively published online: 17 June 2022 [Received: 2 February 2022; Accepted: 27 April 2022].