Search

Statin drugs lower blood cholesterol levels for cardiovascular disease prevention. Women are more likely than men to experience adverse statin effects, particularly new-onset diabetes (NOD) and muscle weakness. Here we find that impaired glucose homeostasis and muscle weakness in statin-treated female mice are associated with reduced levels of the omega-3 fatty acid, docosahexaenoic acid (DHA), impaired redox tone, and reduced mitochondrial respiration. Statin adverse effects are prevented in females by administering fish oil as a source of DHA, by reducing dosage of the X chromosome or the Kdm5c gene, which escapes X chromosome inactivation and is normally expressed at higher levels in females than males. As seen in female mice, we find that women experience more severe reductions than men in DHA levels after statin administration, and that DHA levels are inversely correlated with glucose levels. Furthermore, induced pluripotent stem cells from women who developed NOD exhibit impaired mitochondrial function when treated with statin, whereas cells from men do not. These studies identify X chromosome dosage as a genetic risk factor for statin adverse effects and suggest DHA supplementation as a preventive co-therapy.

Quantifying the temperature sensitivity of methane (CH4) production is crucial for predicting how wetland ecosystems will respond to climate warming. Typically, the temperature sensitivity (often quantified as a Q10 value) is derived from laboratory incubation studies and then used in biogeochemical models. However, studies report wide variation in incubation-inferred Q10 values, with a large portion of this variation remaining unexplained. Here we applied observations in a thawing permafrost peatland (Stordalen Mire) and a well-tested process-rich model (ecosys) to interpret incubation observations and investigate controls on inferred CH4 production temperature sensitivity. We developed a field-storage-incubation modeling approach to mimic the full incubation sequence, including field sampling at a particular time in the growing season, refrigerated storage, and laboratory incubation, followed by model evaluation. We found that CH4 production rates during incubation are regulated by substrate availability and active microbial biomass of key microbial functional groups, which are affected by soil storage duration and temperature. Seasonal variation in substrate availability and active microbial biomass of key microbial functional groups led to strong time-of-sampling impacts on CH4 production. CH4 production is higher with less perturbation post-sampling, i.e. shorter storage duration and lower storage temperature. We found a wide range of inferred Q10 values (1.2-3.5), which we attribute to incubation temperatures, incubation duration, storage duration, and sampling time. We also show that Q10 values of CH4 production are controlled by interacting biological, biochemical, and physical processes, which cause the inferred Q10 values to differ substantially from those of the component processes. Terrestrial ecosystem models that use a constant Q10 value to represent temperature responses may therefore predict biased soil carbon cycling under future climate scenarios.

Managing patients with kidney allografts largely depends on biopsy diagnosis which is based on semiquantitative assessments of rejection features and extent of acute and chronic changes within the renal parenchyma. Current methods lack reproducibility while digital image data-driven computational models enable comprehensive and quantitative assays. In this study we aimed to develop a computational method for automated assessment of histopathology transformations within the tubulointerstitial compartment of the renal cortex. Whole slide images of modified Picrosirius red-stained biopsy slides were used for the training (n = 852) and both internal (n = 172) and external (n = 94) tests datasets. The pipeline utilizes deep learning segmentations of renal tubules, interstitium, and peritubular capillaries from which morphometry features were extracted. Seven indicators were selected for exploring the intrinsic spatial interactions within the tubulointerstitial compartment. A principal component analysis revealed two independent factors which can be interpreted as representing chronic and acute tubulointerstitial injury. A K-means clustering classified biopsies according to potential phenotypes of combined acute and chronic transformations of various degrees. We conclude that multivariate analyses of tubulointerstitial morphometry transformations enable extraction of and quantification of acute and chronic components of injury. The method is developed for renal allograft biopsies; however, the principle can be applied more broadly for kidney pathology assessment.

Kidney complications can occur due to infective endocarditis, one of which is glomerulonephritis. Most often, an immune complex or complement-mediated glomerulonephritis is seen on kidney biopsy. In a minor subset of cases, pauci-immune glomerulonephritis may be present. Most often, such patients will demonstrate the presence of antineutrophil cytoplasmic antibodies (ANCA) on serologic testing. A growing number of cases of ANCA-associated glomerulonephritis due to Bartonella endocarditis have been reported. This type of endocarditis can present diagnostic difficulties given that these patients are often culture negative. Herein, we report a challenging case of ANCA-negative pauci-immune glomerulonephritis showing florid crescents on biopsy that was associated with Bartonella endocarditis.

BACKGROUND: Pathology reports contain key information about the patients diagnosis as well as important gross and microscopic findings. These information-rich clinical reports offer an invaluable resource for clinical studies, but data extraction and analysis from such unstructured texts is often manual and tedious. While neural information retrieval systems (typically implemented as deep learning methods for natural language processing) are automatic and flexible, they typically require a large domain-specific text corpus for training, making them infeasible for many medical subdomains. Thus, an automated data extraction method for pathology reports that does not require a large training corpus would be of significant value and utility. OBJECTIVE: To develop a language model-based neural information retrieval system that can be trained on small datasets and validate it by training it on renal transplant-pathology reports to extract relevant information for two predefined questions: (1) What kind of rejection does the patient show?; (2) What is the grade of interstitial fibrosis and tubular atrophy (IFTA)? METHODS: Kidney BERT was developed by pre-training Clinical BERT on 3.4K renal transplant pathology reports and 1.5M words. Then, exKidneyBERT was developed by extending Clinical BERTs tokenizer with six technical keywords and repeating the pre-training procedure. This extended the models vocabulary. All three models were fine-tuned with information retrieval heads. RESULTS: The model with extended vocabulary, exKidneyBERT, outperformed Clinical BERT and Kidney BERT in both questions. For rejection, exKidneyBERT achieved an 83.3% overlap ratio for antibody-mediated rejection (ABMR) and 79.2% for T-cell mediated rejection (TCMR). For IFTA, exKidneyBERT had a 95.8% exact match rate. CONCLUSION: ExKidneyBERT is a high-performing model for extracting information from renal pathology reports. Additional pre-training of BERT language models on specialized small domains does not necessarily improve performance. Extending the BERT tokenizers vocabulary library is essential for specialized domains to improve performance, especially when pre-training on small corpora.

Abstract Chest X-rays (CXRs) are primarily used to detect lung lesions. While the abdominal portion of CXRs can sometimes reveal critical conditions, research in this area is limited. To address this, we introduce a two-stage architecture that separates the abdominal region from the CXR and detects abdominal lesions using a specialized dataset. We compared the performance of our method on whole CXRs versus isolated abdominal regions. First, we created masks for the right upper quadrant (RUQ), left upper quadrant (LUQ), and upper abdomen (ABD) regions and trained corresponding segmentation models for each area. For detecting abdominal lesions, we curated a dataset of 5,996 images, categorized into 19 classes based on anatomical locations, air patterns, and levels of stomach or bowel dilation. The detection process was initially conducted on the original images, followed by the three regional areas, RUQ, LUQ, and ABD, extracted by the segmentation models. The results showed that the detection model trained on the entire ABD region achieved the highest accuracy, followed closely by the RUQ and LUQ models. In contrast, the CXR model had the lowest accuracy. This study highlights that the two-stage architecture effectively manages distinct segmentation and detection tasks in CXRs, offering a promising avenue for more accurate diagnoses. It also suggests that an optimal ratio between the sizes of the target lesions and the input images may exist.

Background: Real-time surveillance of COVID-19 in large-scale community outbreaks presents challenges. Simple counts of the daily confirmed cases can be misleading due to constraints from bottlenecks in access to care or laboratory testing. This study aimed to investigate the role of the SARS-CoV-2 antigen rapid diagnostic test (Ag-RDT) in addressing these challenges for real-time COVID-19 surveillance. Methods: This study included the results of 86,933 SARS-CoV-2 Ag-RDT and real-time reverse transcription polymerase chain reaction (RT-PCR) tests. These were conducted at four community testing stations within the Taipei metropolitan area during a community COVID-19 outbreak spanning from May 17, 2021, to August 9, 2021. We examined the correlation between the positive rates of Ag-RDT tests and the epidemic curve of laboratory-confirmed COVID-19 cases by onset date to examine its role in real-time surveillance. Results: During the 85-day study period, the trend of Ag-RDT test positive rates paralleled that of the epidemic curve. The correlation between the Ag-RDT positive rate and the number of cases (Pearson correlation coefficient: 0.968) is comparable to that of the RT-PCR positive rate (Pearson correlation coefficient: 0.964). The Ag-RDT positive rate exhibited a more advanced leading trend, with Ag-RDT leading by 3 days in comparison to the 2-day lead for RT-PCR. Conclusion: The positive rate of SARS-CoV-2 Ag-RDT tests at community testing stations serves as a good surrogate for assessing virus activity within the community and a useful tool for real-time COVID-19 surveillance. It is a robust indicator of the outbreak trend and near-term numbers of cases. This finding may facilitate the management of subsequent outbreaks of emerging infectious diseases.

In order to enable a multi-energy coupling integrated energy system (IES) to meet the needs of load diversity in low-carbon economic operation, a bi-level optimal configuration method for low-carbon economic operation of multi-energy coupling IES in different charging modes of electric vehicles (EVs) is proposed. First, an IES including cold-thermal-electric-gas coupling is established. Then, in the day-to-day operation stage, factors such as hierarchical carbon trading mechanism and different charging modes of EVs are considered to achieve the lowest daily scheduling cost. In the configuration planning stage, based on the daily operation cost, the equipment capacity is configured with the lowest equipment investment cost and annual operation cost. Finally, Cplex is used to solve the above two-stage objective functions and obtain the optimal configuration scheme and scheduling results through mutual iteration. The results show that the charging method considering the remaining charge of EVs and carbon trading mechanism can significantly reduce carbon emissions and operating costs of the system. The proposed configuration approach can well realize low-carbon economic operation of the multi-energy coupling IES.

The 2023 China Health Report on Spine Degeneration noted a significant increase in lumbar surgery among patients under 35 years old in recent years, indicating a trend towards younger onset of cervical and lumbar diseases. Lumbar intervertebral disc herniation has become a major concern, making the study of disc degeneration pathogenesis and treatment methods clinically significant. At present, human intervertebral disc diseases are primarily diagnosed through imaging due to the challenges of obtaining tissue samples from the spine. Therefore, experimental animals have emerged as alternative research subjects because they are cost-effective, have short experimental cycles, and are easily accessible. Given the structural and physiological differences between human and other animal intervertebral discs, comparing their anatomy and histological characteristics forms the foundation of research into human disc degeneration. The purpose of this paper is to collect and review relevant studies on anatomical and histological structures of intervertebral discs in different animals and conduct a comparative analysis from four aspects, namely, intervertebral disc height, lumbar disc geometry, lumbar disc cartilaginous endplate characteristics, and extracellular matrix components. The results show that humans, kangaroos, sheep, pigs, and rats exhibit similar relative heights between the sixth and seventh cervical vertebrae. Mice possess lumbar disc geometries most akin to humans. Compared to other animals, humans have the thickest cartilaginous endplates and the lowest cell densities. The collagen within the fibrous annulus differs most notably in pigs compared to humans, while water content in the nucleus pulposus is consistent across pigs, sheep, rabbits, rats, and humans. Additionally, this paper describes the commonalities and discrepancies in disc degeneration manifestations between humans and animals, and summarizes modeling methods for disc degeneration in different experimental animals. Ultimately, the aims of this paper is to provide fundamental data for selecting suitable experimental animal models for the study of intervertebral disc degeneration.

Abstract Glutamine synthetase (GS) is vital in maintaining ammonia and glutamate (Glu) homeostasis in living organisms. However, the natural enzyme relies on adenosine triphosphate (ATP) to activate Glu, resulting in impaired GS function during ATP-deficient neurotoxic events. To date, no reports demonstrate using artificial nanostructures to mimic GS function. In this study, we synthesize aggregation-induced emission active polyP-Mn nanosheets (STPE-PMNSs) based on end-labeled polyphosphate (polyP), exhibiting remarkable GS-like activity independent of ATP presence. Further investigation reveals polyP in STPE-PMNSs serves as phosphate source to activate Glu at low ATP levels. This self-feeding mechanism offers a significant advantage in regulating Glu homeostasis at reduced ATP levels in nerve cells during excitotoxic conditions. STPE-PMNSs can effectively promote the conversion of Glu to glutamine (Gln) in excitatory neurotoxic human neuroblastoma cells (SH-SY5Y) and alleviate Glu-induced neurotoxicity. Additionally, the fluorescence signal of nanosheets enables precise monitoring of the subcellular distribution of STPE-PMNSs. More importantly, the intracellular fluorescence signal is enhanced in a conversion-responsive manner, allowing real-time tracking of reaction progression. This study presents a self-sustaining strategy to address GS functional impairment caused by ATP deficiency in nerve cells during neurotoxic events. Furthermore, it offers a fresh perspective on the potential biological applications of polyP-based nanostructures.

Abstract Managing patients with kidney allografts largely depends on biopsy diagnosis which is based on semiquantitative assessments of rejection features and extent of acute and chronic changes within the renal parenchyma. Current methods lack reproducibility while digital image data-driven computational models enable comprehensive and quantitative assays. In this study we aimed to develop a computational method for automated assessment of histopathology transformations within the tubulointerstitial compartment of the renal cortex. Whole slide images of modified Picrosirius red-stained biopsy slides were used for the training (n = 852) and both internal (n = 172) and external (n = 94) tests datasets. The pipeline utilizes deep learning segmentations of renal tubules, interstitium, and peritubular capillaries from which morphometry features were extracted. Seven indicators were selected for exploring the intrinsic spatial interactions within the tubulointerstitial compartment. A principal component analysis revealed two independent factors which can be interpreted as representing chronic and acute tubulointerstitial injury. A K-means clustering classified biopsies according to potential phenotypes of combined acute and chronic transformations of various degrees. We conclude that multivariate analyses of tubulointerstitial morphometry transformations enable extraction of and quantification of acute and chronic components of injury. The method is developed for renal allograft biopsies; however, the principle can be applied more broadly for kidney pathology assessment.

Objectives: The link between aging and pulmonary function decline is well-established, but the underlying mechanisms have yet to be fully revealed. Serum follistatin, a myokine implicated in muscle degeneration, may play a role in age-related pulmonary changes. This study aims to investigate the relationship between serum follistatin levels and pulmonary function decline in community-dwelling older adults, and evaluate their combined association with all-cause mortality. Research design and methods: This longitudinal cohort study utilized data from 751 participants aged ≥50 years in the I-Lan Longitudinal Aging Study between 2018−2019. Serum follistatin levels, spirometry results, demographic and clinical data were retrieved. Participants were stratified based on their follistatin levels. Survival curves and group comparisons based on follistatin levels and decline in peak expiratory flow (PEF) using Kaplan-Meier analysis and log-rank tests. Multivariate Cox proportional hazards models were further used to identify independent predictors of all-cause mortality during the 52-month follow-up. Results: Elevated follistatin levels significantly correlated with worse pulmonary function, particularly decreased PEF (p = 0.030). Kaplan–Meier analysis revealed the combination of elevated follistatin levels and decreased PEF was associated with increased risk of all-cause mortality (Log-rank p = 0.023). Cox proportional hazards models further identified that concurrent presence of higher follistatin levels and decreased PEF predicted higher risk of all-cause mortality (adjusted HR 3.58, 95% CI: 1.22–10.53, p = 0.020). Conclusion: Higher serum follistatin levels correlate with decreased pulmonary function, specifically PEF decline, in community-dwelling older adults. Furthermore, the coexistence of elevated follistatin levels and decreased PEF was associated with risk of all-cause mortality. Follistatin may serve as a biomarker for pulmonary aging and related adverse outcomes.

Objectives To evaluate whether nephrotic syndrome (NS) and further corticosteroid (CS) use increase the risk of osteoporosis in Asian population during the period January 2000–December 2010.Design Nationwide population-based retrospective cohort study.Setting All healthcare facilities in Taiwan.Participants A total of 28 772 individuals were enrolled.Interventions 26 614 individuals with newly diagnosed NS between 2000 and 2010 were identified and included in out study. 26 614 individuals with no NS diagnosis prior to the index date were age matched as controls. Diagnosis of osteoporosis prior to the diagnosis of NS or the same index date was identified, age, sex and NS-associated comorbidities were adjusted.Primary outcome measure To identify risk differences in developing osteoporosis among patients with a medical history of NS.Results After adjusting for covariates, osteoporosis risk was found to be 3.279 times greater in the NS cohort than in the non-NS cohort, when measured over 11 years after NS diagnosis. Stratification revealed that age older than 18 years, congestive heart failure, hyperlipidaemia, chronic kidney disease, liver cirrhosis and NS-related disease including diabetes mellitus, hepatitis B infection, hepatitis C infection, lymphoma and hypothyroidism, increased the risk of osteoporosis in the NS cohort, compared with the non-NS cohort. Additionally, osteoporosis risk was significantly higher in NS patients with CS use (adjusted HR (aHR)=3.397). The risk of osteoporosis in NS patients was positively associated with risk of hip and vertebral fracture (aHR=2.130 and 2.268, respectively). A significant association exists between NS and subsequent risk for osteoporosis.Conclusion NS patients, particularly those treated with CS, should be evaluated for subsequent risk of osteoporosis.

Ocean currents play a role in both facilitating and impeding connectivity of marine organisms’ populations, serving as key drivers and potential barriers. The emergence of Taiwan around five million years ago (Mya) impacted the Kuroshio Current which later shaped the distribution of marine organisms. In this study, we examined how the Kuroshio Current influenced the distribution and population divergence of the marine red alga Dichotomaria elegans (Galaxauraceae, Nemaliales). Through phylogenetic analysis, species delimitation, and morphological comparison of Taiwanese species, we determined its phylogenetic position relative to other species within the D. marginata complex of the genus Dichotomaria. We uncovered the hidden cryptic diversity within the D. marginata complex in Taiwan and proposed the new combinations D. elegans and D. latifolia based on Galaxaura elegans and G. latifolia. Genetic analysis based on three loci (mitochondrial cox2-3 and cox1, and plastid rbcL) from 32 locations across the Philippines, Taiwan, and Japan revealed three allopatric lineages within D. elegans. The first lineage is restricted to the northern Philippines and southern Taiwan, while the second has a broader latitudinal range along the Kuroshio Current. The third lineage is found in southeast Luzon, Philippines, close to the bifurcation of the North Equatorial Current, near the start of the Kuroshio Current. The divergence of the first two lineages during the mid-Pliocene (~3.7 Mya) coincided with the emergence of Taiwan, with the current dynamics accentuating their separation. The Kuroshio Current exhibited a dual impact on the biogeography of D. elegans: promoting the dispersal of the second lineage towards higher latitudes and preventing the first lineage from moving northward. Principal component analysis showed that habitat conditions and morphologies differed between the two D. elegans lineages. This study provides insights into the cryptic diversity of Dichotomaria in Taiwan, as well as the genetic divergence, biogeographical patterns, and ecological associations of D. elegans, highlighting the interplay of island formation and currents shaping the diversification of seaweeds in the Northwest Pacific Ocean.

Background and Objectives: In patients with urolithiasis-related obstructive pyelonephritis (UROP), sepsis represents a critical and concerning complication that can substantially increase the mortality rate. This study aimed to identify the risk factors for sepsis in UROP patients and to develop a predictive nomogram model. Materials and Methods: We analyzed data from 148 patients who met the UROP criteria and were admitted to Chang Gung Memorial Hospital between 1 January 2016 and 31 December 2021. The primary outcome evaluated was the incidence of sepsis, as defined by the most recent Sepsis-3 guidelines. To identify potential risk factors for sepsis, we employed the Least Absolute Shrinkage and Selection Operator (LASSO) regression technique. Subsequently, we utilized multivariable logistic regression to construct the predictive model. Results: There was a total of 102 non-sepsis cases and 46 sepsis cases. Risk factors for sepsis in multivariable analysis were a history of diabetes mellitus (DM) (OR = 4.24, p = 0.007), shock index (SI) (×10−1) (OR = 1.55, p < 0.001), C-reactive protein (CRP) (mg/dL) (OR = 1.08, p = 0.005), and neutrophil to lymphocyte ratio (NLR) (×10) (OR = 1.58, p = 0.007). The nomogram exhibited an area under the receiver operating characteristic curve of 0.890 (95% CI 0.830–0.949). Conclusions: Our study demonstrated that patients with UROP who have DM, higher SI, higher NLR, and elevated CRP levels are significantly more likely to develop sepsis. These insights may aid in risk stratification, and it is imperative that clinicians promptly initiate treatment for those identified as high risk.

Background Pathology reports contain key information about the patient’s diagnosis as well as important gross and microscopic findings. These information-rich clinical reports offer an invaluable resource for clinical studies, but data extraction and analysis from such unstructured texts is often manual and tedious. While neural information retrieval systems (typically implemented as deep learning methods for natural language processing) are automatic and flexible, they typically require a large domain-specific text corpus for training, making them infeasible for many medical subdomains. Thus, an automated data extraction method for pathology reports that does not require a large training corpus would be of significant value and utility. Objective To develop a language model-based neural information retrieval system that can be trained on small datasets and validate it by training it on renal transplant-pathology reports to extract relevant information for two predefined questions: (1) “What kind of rejection does the patient show?”; (2) “What is the grade of interstitial fibrosis and tubular atrophy (IFTA)?” Methods Kidney BERT was developed by pre-training Clinical BERT on 3.4K renal transplant pathology reports and 1.5M words. Then, exKidneyBERT was developed by extending Clinical BERT’s tokenizer with six technical keywords and repeating the pre-training procedure. This extended the model’s vocabulary. All three models were fine-tuned with information retrieval heads. Results The model with extended vocabulary, exKidneyBERT, outperformed Clinical BERT and Kidney BERT in both questions. For rejection, exKidneyBERT achieved an 83.3% overlap ratio for antibody-mediated rejection (ABMR) and 79.2% for T-cell mediated rejection (TCMR). For IFTA, exKidneyBERT had a 95.8% exact match rate. Conclusion ExKidneyBERT is a high-performing model for extracting information from renal pathology reports. Additional pre-training of BERT language models on specialized small domains does not necessarily improve performance. Extending the BERT tokenizer’s vocabulary library is essential for specialized domains to improve performance, especially when pre-training on small corpora.

Kidney complications can occur due to infective endocarditis, one of which is glomerulonephritis. Most often, an immune complex or complement-mediated glomerulonephritis is seen on kidney biopsy. In a minor subset of cases, pauci-immune glomerulonephritis may be present. Most often, such patients will demonstrate the presence of antineutrophil cytoplasmic antibodies (ANCA) on serologic testing. A growing number of cases of ANCA-associated glomerulonephritis due to Bartonella endocarditis have been reported. This type of endocarditis can present diagnostic difficulties given that these patients are often culture negative. Herein, we report a challenging case of ANCA-negative pauci-immune glomerulonephritis showing florid crescents on biopsy that was associated with Bartonella endocarditis.

Background: Pulmonary function test, particularly in patients with COVID-19, is problematic because it involves forced expiration. Impulse oscillometry (IOS) reduces the potential exposure of health-care staff to infectious droplets. In this study, we investigated the correlation between IOS and spirometry and whether IOS can precisely predict spirometry-based diagnoses of chronic obstructive pulmonary disease (COPD). Methods: We retrospectively analyzed the data (January 1 to December 31, 2021) of patients who underwent both spirometry and IOS on the same date. One-way analysis of variance was performed to evaluate the IOS results of patients stratified into two (COPD and non-COPD) groups by spirometry results. IOS results were also analyzed using receiver operator characteristics curves to diagnose advanced COPD, which was indicated by a postbronchodilator (BD) forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) ratio of 160 moderately predicted spirometry-based COPD diagnosis. Conclusions: BMI-adjusted airway resistance is a potential predictor of spirometry-based COPD diagnosis; the cutoff values of this parameter differ between individuals with and without obesity.

Identifying farmland use has long been an important topic in large-scale agricultural production management. This study used multi-temporal visible RGB images taken from agricultural areas in Taiwan by UAV to build a model for classifying field types. We combined color and texture features to extract more information from RGB images. The vectorized gray-level co-occurrence matrix (GLCMv), instead of the common Haralick feature, was used as texture to improve the classification accuracy. To understand whether changes in the appearance of crops at different times affect image features and classification, this study designed a labeling method that combines image acquisition times and land use type to observe it. The Extreme Gradient Boosting (XGBoost) algorithm was chosen to build the classifier, and two classical algorithms, the Support Vector Machine and Classification and Regression Tree algorithms, were used for comparison. In the testing results, the highest overall accuracy reached 82%, and the best balance accuracy across categories reached 97%. In our comparison, the color feature provides the most information about the classification model and builds the most accurate classifier. If the color feature were used with the GLCMv, the accuracy would improve by about 3%. In contrast, the Haralick feature does not improve the accuracy, indicating that the GLCM itself contains more information that can be used to improve the prediction. It also shows that with combined image acquisition times in the label, the within-group sum of squares can be reduced by 2–31%, and the accuracy can be increased by 1–2% for some categories, showing that the change of crops over time was also an important factor of image features.