Algorithm-assisted discovery of an intrinsic order among mathematical constants.

In recent decades, a growing number of discoveries in mathematics have been assisted by computer algorithms, primarily for exploring large parameter spaces. As computers become more powerful, an intriguing possibility arises-the interplay between human intuition and computer algorithms can lead to discoveries of mathematical structures that would otherwise remain elusive. Here, we demonstrate computer-assisted discovery of a previously unknown mathematical structure, the conservative matrix field. In the spirit of the Ramanujan Machine project, we developed a massively parallel computer algorithm that found a large number of formulas, in the form of continued fractions, for numerous mathematical constants. The patterns arising from those formulas enabled the construction of the first conservative matrix fields and revealed their overarching properties. Conservative matrix fields unveil unexpected relations between different mathematical constants, such as π and ln(2), or e and the Gompertz constant. The importance of these matrix fields is further realized by their ability to connect formulas that do not have any apparent relation, thus unifying hundreds of existing formulas and generating infinitely many new formulas. We exemplify these implications on values of the Riemann zeta function ζ (n), studied for centuries across mathematics and physics. Matrix fields also enable new mathematical proofs of irrationality. For example, we use them to generalize the celebrated proof by Apéry of the irrationality of ζ (3). Utilizing thousands of personal computers worldwide, our research strategy demonstrates the power of large-scale computational approaches to tackle longstanding open problems and discover unexpected connections across diverse fields of science.

Disruption of HSD17B12 in mouse hepatocytes leads to reduced body weight and defect in the lipid droplet expansion associated with microvesicular steatosis.

The function of hydroxysteroid dehydrogenase 12 (HSD17B12) in lipid metabolism is poorly understood. To study this further, we created mice with hepatocyte-specific knockout of HSD17B12 (LiB12cKO). From 2 months on, these mice showed significant fat accumulation in their liver. As they aged, they also had a reduced whole-body fat percentage. Interestingly, the liver fat accumulation did not result in the typical formation of large lipid droplets (LD); instead, small droplets were more prevalent. Thus, LiB12KO liver did not show increased macrovesicular steatosis with the increasing fat content, while microvesicular steatosis was the predominant feature in the liver. This indicates a failure in the LD expansion. This was associated with liver damage, presumably due to lipotoxicity. Notably, the lipidomics data did not support an essential role of HSD17B12 in fatty acid (FA) elongation. However, we did observe a decrease in the quantity of specific lipid species that contain FAs with carbon chain lengths of 18 and 20 atoms, including oleic acid. Of these, phosphatidylcholine and phosphatidylethanolamine have been shown to play a key role in LD formation, and a limited amount of these lipids could be part of the mechanism leading to the dysfunction in LD expansion. The increase in the Cidec expression further supported the deficiency in LD expansion in the LiB12cKO liver. This protein is crucial for the fusion and growth of LDs, along with the downregulation of several members of the major urinary protein family of proteins, which have recently been shown to be altered during endoplasmic reticulum stress.

A variant in sperm-specific glycolytic enzyme enolase 4 (ENO4) causes human male infertility.

BACKGROUND: Although defects in sperm morphology and physiology lead to male infertility, in many instances, the exact disruption of molecular pathways in a given patient is often unknown. The glycolytic pathway is an essential process to supply energy in sperm cell motility. Enolase 4 (ENO4) is crucial for the glycolytic process, which provides the energy for sperm cells in motility. ENO4 is located in the sperm principal piece and is essential for the motility and organization of the sperm flagellum. In the present study, we characterized a family with asthenozoospermia and abnormal sperm morphology as a result of a variant in the enolase 4 (ENO4) gene. METHODS: Computer-assisted semen analysis, papanicolaou smear staining and scanning electron microscopy were used to examine sperm motility and morphology for semen analysis in patients. For genetic analysis, whole-exome sequencing followed by Sanger sequencing was performed. RESULTS: Two brothers in a consanguineous family were being clinically investigated for sperm motility and morphology issues. Genetic analysis by whole-exome sequencing revealed a homozygous variant [c.293A>G, p.(Lys98Arg)] in the ENO4 gene that segregated with infertility in the family, shared by affected but not controls. CONCLUSIONS: In view of the association of asthenozoospermia and abnormal sperm morphology in Eno4 knockout mice, we consider this to be the first report describing the involvement of ENO4 gene in human male infertility. We also explore the possible involvement of another variant in explaining other phenotypic features in this family.

Computer-assisted semi-rational design enhanced the enzymatic activity and protein stability of Proteinase K in calcium-free conditions.

Wild-type Proteinase K binds to two Ca2+ ions, which play an important role in regulating enzymaticactivity and maintaining protein stability. Therefore, a predetermined concentration of Ca2+ must be added during the use of Proteinase K, which increases its commercial cost. Herein, we addressed this challenge using a computational strategy to engineer a Proteinase K mutant that does not require Ca2+ and exhibits high enzymatic activity and protein stability. In the absence of Ca2+, the best mutant, MT24 (S17W-S176N-D260F), displayed an activity approximately 9.2-fold higher than that of wild-type Proteinase K. It also exhibited excellent protein stability, retaining 56.2 % of its enzymatic activity after storage at 4 °C for 5 days. The residual enzymatic activity was 65-fold higher than that of the wild-type Proteinase K under the same storage conditions. Structural analysis and molecular dynamics simulations suggest that the introduction of new hydrogen bond and π-π stacking at the Ca2+ binding sites due to the mutation may be the reasons for the increased enzymatic activity and stability of MT24.

Novel deep-learning analysis for connective tissue disease -related interstitial lung disease extent assessment on CT: a preliminary cross-sectional study.

OBJECTIVES: Physician's evaluation of interstitial lung disease (ILD) extension with high-resolution computed tomography (HRCT) has limitations such as lack of objectivity and reproducibility. This study aimed to investigate the utility of computer-based deep-learning analysis using QZIP-ILD® software (DL-QZIP) compared with conventional approaches in connective tissue disease (CTD) -related ILD. METHODS: Patients with CTD-ILD visiting our Rheumatology Centre between December 2020 and April 2024 were recruited. Quantitative scores, including the percentage of lung involvement in ground-glass opacity (QGG), total fibrotic lesion (QFIB), and overall ILD extension encompassing both QGG and QFIB (QILD), calculated by DL-QZIP, were compared with semiquantitative visual method, employing intraclass correlation coefficients (ICC). We compared the capability of QILD scores to distinguish patients with forced vital capacity (FVC) % <70 in both methods determined by the area under the curve (AUC) by the receiver-operating characteristic curve analysis and DeLong's test. RESULTS: Eighty patients (median age, 66 years; 14 men) were included. Median QGG, QFIB, and QILD scores were 3.45%, 2.19%, and 5.35% using DL-QZIP, and 3.25%, 4.06%, and 8.48% using visual method, respectively. Correlations between DL-QZIP and visual method were 0.75 for QGG, 0.61 for QFIB, and 0.75 for QILD. The AUC of QILD scores for FVC% <70 was significantly higher with DL-QZIP (0.833) compared with visual method (0.660) (p < 0.01). CONCLUSION: QZIP-ILD® demonstrates superior capability in distinguishing patients with a radiological scenario correlated to severe physiological impairment, while showing relatively good correlations in quantifying the extent on HRCT compared with conventional method in CTD-ILD.

Recent decline in sperm motility among donor candidates at a sperm bank in Denmark.

STUDY QUESTION: Has there been variation in semen quality among men applying to be sperm donors (i.e. donor candidates) in Denmark in recent years (2017-2022)? SUMMARY ANSWER: The motile sperm concentration and total motile sperm count (TMSC) in ejaculates-both measures of sperm quality-declined by as much as 22% from 2019 to 2022. WHAT IS KNOWN ALREADY: Questions remain about whether human semen quality has declined in recent years. Whilst some studies provide evidence for a decline in human semen quality, these findings have been disputed owing to potential biases in the populations studied or in the methods used to measure semen quality. Resolution of this issue has important implications for human fertility, as well as for those involved in the recruitment of sperm donors for use in medically assisted reproduction. STUDY DESIGN, SIZE, DURATION: We obtained data on the semen quality of ejaculates previously collected from 2017 to 2022 at sperm bank locations in four cities in Denmark: Aarhus, Aalborg, Copenhagen, and Odense. Our study focuses on the single semen samples provided by 6758 donor candidates aged between 18 and 45 years old to determine whether their sperm quality met a minimum criterion for them to be accepted as sperm donors. PARTICIPANTS/MATERIALS, SETTING, METHODS: All ejaculates were analyzed within 1 hour of production. Semen volume (ml) was estimated by weight and both the concentration (106/ml) of sperm as well as the concentration of motile sperm (World Health Organization grades a and b) were measured using the same protocols and computer-assisted semen analysis system across all years at each site. Statistical analyses of the semen variables were controlled for age and donation site, as well as the average monthly high temperature when the ejaculate was produced. MAIN RESULTS AND THE ROLE OF CHANCE: From 2017 to 2019, semen volume, sperm concentration, and total sperm count in the ejaculates of donor candidates increased by 2-12%. Then, from 2019 to 2022, sperm concentration and total sperm count changed by 0.1-5% from year to year, but none of those changes were statistically significant. In contrast, both motile sperm concentration and TMSC declined significantly, by 16% and 22%, respectively, between 2019 and 2022. Thus, the concentration of motile sperm in donor candidates declined from 18.4 [95% CL: 17.0, 20.0] million/ml in 2019 to 15.5 [14.4, 16.7] million/ml in 2022, and TMSC declined from 61.4 [55.8, 67.5] million per ejaculate in 2019 to 48.1 [44.1, 52.4] million in 2022. LIMITATIONS, REASONS FOR CAUTION: We cannot determine from the available data the causes of the decline in semen quality of donor candidates from 2019 to 2022. However, as this period coincides with lockdowns and changes in work patterns during the coronavirus disease 2019 pandemic, it is possible that changes in motile sperm concentration and TMSC were the result of changes in the lifestyles of the men whose semen was analyzed. WIDER IMPLICATIONS OF THE FINDINGS: Men providing initial semen samples at sperm banks, when applying to be sperm donors, are a useful population in which to monitor changes in human semen quality over time. Our results have implications for human fertility and the recruitment of sperm donors for medically assisted reproduction, where motile sperm concentration is an essential selection criterion because it influences fertility. We suggest that gathering health and lifestyle data on donor candidates at sperm banks might help to identify causal factors for the decline of sperm quality that could be addressed and intervention, if desired, could be personalized for each accepted donor. STUDY FUNDING/COMPETING INTEREST(S): No external funding was obtained for this study. E.L. and A.-B.S. are employees of Cryos International. AP reports paid consultancy for Cryos International, Cytoswim Ltd, Exceed Health, and Merck Serono in the last 2 years of this study, but all monies were paid to the University of Sheffield (former employer). AP is also an unpaid trustee of the Progress Educational Trust (Charity Number 1139856). RM declares support from Cryos International to present results of this research at ESHRE 2023. None of the authors were directly involved in the collection or physical analysis of semen samples. TRIAL REGISTRATION NUMBER: N/A.

Artificial Intelligence Supports Automated Characterization of Differentiated Human Pluripotent Stem Cells.

Revolutionary advances in AI and deep learning in recent years have resulted in an upsurge of papers exploring applications within the biomedical field. Within stem cell research, promising results have been reported from analyses of microscopy images to, that is, distinguish between pluripotent stem cells and differentiated cell types derived from stem cells. In this work, we investigated the possibility of using a deep learning model to predict the differentiation stage of pluripotent stem cells undergoing differentiation toward hepatocytes, based on morphological features of cell cultures. We were able to achieve close to perfect classification of images from early and late time points during differentiation, and this aligned very well with the experimental validation of cell identity and function. Our results suggest that deep learning models can distinguish between different cell morphologies, and provide alternative means of semi-automated functional characterization of stem cell cultures.

A Deep Learning Pipeline Using Prior Knowledge for Automatic Evaluation of Placenta Accreta Spectrum Disorders With MRI.

BACKGROUND: The diagnosis of prenatal placenta accreta spectrum (PAS) with magnetic resonance imaging (MRI) is highly dependent on radiologists' experience. A deep learning (DL) method using the prior knowledge that PAS-related signs are generally found along the utero-placental borderline (UPB) may help radiologists, especially those with less experience, to mitigate this issue. PURPOSE: To develop a DL tool for antenatal diagnosis of PAS using T2-weighted MR images. STUDY TYPE: Retrospective. SUBJECTS: Five hundred and forty pregnant women with clinically suspected PAS disorders from two institutions, divided into training (409), internal test (103), and external test (28) datasets. FIELD STRENGTH/SEQUENCE: Sagittal T2-weighted fast spin echo sequence at 1.5 T and 3 T. ASSESSMENT: An nnU-Net was trained for placenta segmentation. The UPB straightening approach was used to extract the utero-placental boundary region. The UPB image was then fed into DenseNet-PAS for PAS diagnosis. DenseNet-PP learnt placental position information to improve the PAS diagnosis performance. Three radiologists with 8, 10, and 12 years of experience independently evaluated the images. Two radiologists marked the placenta tissue. Histopathological findings were the reference standard. STATISTICAL TESTS: Area under the curve (AUC) was used to evaluate the classification. Dice coefficient evaluated the segmentation between radiologists and the model performance. The Mann-Whitney U-test or the chi-squared test assessed the significance of differences. Decision curve analysis was used to determine clinical effectiveness. DeLong's test was used to compare AUCs. RESULTS: Of the 540 patients, 170 had PAS disorders confirmed by histopathology. The DL model using UPB images and placental position yielded the highest AUC of 0.860 and 0.897 in internal test and external test cohorts, respectively, significantly exceeding the performance of three radiologists (internal test AUC, 0.737-0.770). DATA CONCLUSION: By extracting the UPB image, this fully automatic DL pipeline achieved high accuracy and may assist radiologists in PAS diagnosis using MRI. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 2.

Outcome prediction in metabolic dysfunction-associated steatotic liver disease using stain-free digital pathological assessment.

Computational quantification reduces observer-related variability in histological assessment of metabolic dysfunction-associated steatotic liver disease (MASLD). We undertook stain-free imaging using the SteatoSITE resource to generate tools directly predictive of clinical outcomes. Unstained liver biopsy sections (n = 452) were imaged using second-harmonic generation/two-photon excitation fluorescence (TPEF) microscopy, and all-cause mortality and hepatic decompensation indices constructed. The mortality index had greater predictive power for all-cause mortality (index >.14 vs. .31 vs.

Quantitative edge analysis of pancreatic margins in patients with head pancreatic tumors: correlations between pancreatic margins and the onset of postoperative pancreatic fistula.

OBJECTIVE: To assess the correlation between pancreatic quantitative edge analysis as a surrogate of parenchymal stiffness and the incidence of postoperative pancreatic fistula (POPF), in patients undergoing pancreaticoduodenectomy (PD). METHODS: All consecutive patients who underwent PD at our Institution between March 2018 and November 2019 with an available preoperative CT were included. Pancreatic margin score (PMS) was calculated through computer-assisted quantitative edge analysis on the margins of the pancreatic body and tail (the expected pancreatic remnant) on non-contrast scans with in-house software. Intraoperative assessment of pancreatic stiffness by manual palpation was also performed, classifying pancreatic texture into soft and non-soft. PMS values were compared between groups using an unpaired T-test and correlated with the intraoperative evaluation of stiffness and with the grading of postoperative pancreatic fistula according to the International Study Group on Pancreatic Surgery (ISGPS). RESULTS: Patient population included 200 patients (mean age 64.6 years), 146 without onset of POPF (73%, non-POPF group), and 54 with POPF (27%, POPF group). A significant difference in PMS values was observed between POPF and non-POPF (respectively 1.88 ± 0.05 vs 0.69 ± 0.01; p < 0.0001). PMS values of pancreatic parenchymas intraoperatively considered "soft" were significantly higher than those evaluated as "non-soft" (1.21 ± 0.04 vs 0.73 ± 0.02; p < 0.0001). A significant correlation between PMS values and POPF grade was observed (r = 0.8316), even in subgroups of patients with soft (r = 0.8016) and non-soft (r = 0.7602) pancreas (all p < 0.0001). CONCLUSIONS: Quantitative edge analysis with dedicated software may stratify patients with different pancreatic stiffness, thus potentially improving preoperative risk assessment and strategies for POPF mitigation. CLINICAL RELEVANCE STATEMENT: This study proposes quantitative pancreas edge analysis as a predictor for postoperative pancreatic fistula. The test has high accuracy and correlation with fistula grade according to the International Study Group on Pancreatic Surgery. KEY POINTS: • Prediction of postoperative pancreatic fistula (POPF) onset risk after pancreaticoduodenectomy is based only on intraoperative evaluation. • Quantitative edge analysis may preoperatively identify patients with higher risk of POPF. • Quantification of pancreatic stiffness through the analysis of pancreatic margins could be done on preoperative CT.

Assessing deep learning reconstruction for faster prostate MRI: visual vs. diagnostic performance metrics.

OBJECTIVE: Deep learning (DL) MRI reconstruction enables fast scan acquisition with good visual quality, but the diagnostic impact is often not assessed because of large reader study requirements. This study used existing diagnostic DL to assess the diagnostic quality of reconstructed images. MATERIALS AND METHODS: A retrospective multisite study of 1535 patients assessed biparametric prostate MRI between 2016 and 2020. Likely clinically significant prostate cancer (csPCa) lesions (PI-RADS ≥ 4) were delineated by expert radiologists. T2-weighted scans were retrospectively undersampled, simulating accelerated protocols. DL reconstruction (DLRecon) and diagnostic DL detection (DLDetect) were developed. The effect on the partial area under (pAUC), the Free-Response Operating Characteristic (FROC) curve, and the structural similarity (SSIM) were compared as metrics for diagnostic and visual quality, respectively. DLDetect was validated with a reader concordance analysis. Statistical analysis included Wilcoxon, permutation, and Cohen's kappa tests for visual quality, diagnostic performance, and reader concordance. RESULTS: DLRecon improved visual quality at 4- and 8-fold (R4, R8) subsampling rates, with SSIM (range: -1 to 1) improved to 0.78 ± 0.02 (p < 0.001) and 0.67 ± 0.03 (p < 0.001) from 0.68 ± 0.03 and 0.51 ± 0.03, respectively. However, diagnostic performance at R4 showed a pAUC FROC of 1.33 (CI 1.28-1.39) for DL and 1.29 (CI 1.23-1.35) for naive reconstructions, both significantly lower than fully sampled pAUC of 1.58 (DL: p = 0.024, naïve: p = 0.02). Similar trends were noted for R8. CONCLUSION: DL reconstruction produces visually appealing images but may reduce diagnostic accuracy. Incorporating diagnostic AI into the assessment framework offers a clinically relevant metric essential for adopting reconstruction models into clinical practice. CLINICAL RELEVANCE STATEMENT: In clinical settings, caution is warranted when using DL reconstruction for MRI scans. While it recovered visual quality, it failed to match the prostate cancer detection rates observed in scans not subjected to acceleration and DL reconstruction.

Deep learning reconstruction vs standard reconstruction for abdominal CT: the influence of BMI.

OBJECTIVE: This study aimed to evaluate the image quality and lesion conspicuity of the deep learning image reconstruction (DLIR) algorithm compared with standard image reconstruction algorithms on abdominal enhanced computed tomography (CT) scanning with a wide range of body mass indexes (BMIs). METHODS: A total of 112 participants who underwent contrast-enhanced abdominal CT scans were divided into three groups according to BMIs: the 80-kVp group (BMI ≤ 23.9 kg/m2), 100-kVp group (BMI 24-28.9 kg/m2), and 120-kVp group (BMI ≥ 29 kg/m2). All images were reconstructed using filtered back projection (FBP), adaptive statistical iterative reconstruction-V of 50% level (IR), and DLIR at low, medium, and high levels (DL, DM, and DH, respectively). Subjective noise, artifact, overall image quality, and low- and high-contrast hepatic lesion conspicuity were all graded on a 5-point scale. The CT attenuation value (in HU), image noise, and contrast-to-noise ratio (CNR) were quantified and compared. RESULTS: DM and DH improved the qualitative and quantitative parameters compared with FBP and IR for all three BMI groups. DH had the lowest image noise and highest CNR value, while DM had the highest subjective overall image quality and low- and high-contrast lesion conspicuity scores for the three BMI groups. Based on the FBP, the improvement in image quality and lesion conspicuity of DM and DH images was greater in the 80-kVp group than in the 100-kVp and 120-kVp groups. CONCLUSION: For all BMIs, DLIR improves both image quality and hepatic lesion conspicuity, of which DM would be the best choice to balance both. CLINICAL RELEVANCE STATEMENT: The study suggests that utilizing DLIR, particularly at the medium level, can significantly enhance image quality and lesion visibility on abdominal CT scans across a wide range of BMIs. KEY POINTS: • DLIR improved the image quality and lesion conspicuity across a wide range of BMIs. • DLIR at medium level had the highest subjective parameters and lesion conspicuity scores among all reconstruction levels. • On the basis of the FBP, the 80-kVp group had improved image quality and lesion conspicuity more than the 100-kVp and 120-kVp groups.

Performance of AI to exclude normal chest radiographs to reduce radiologists' workload.

INTRODUCTION: This study investigates the performance of a commercially available artificial intelligence (AI) system to identify normal chest radiographs and its potential to reduce radiologist workload. METHODS: Retrospective analysis included consecutive chest radiographs from two medical centers between Oct 1, 2016 and Oct 14, 2016. Exclusions comprised follow-up exams within the inclusion period, bedside radiographs, incomplete images, imported radiographs, and pediatric radiographs. Three chest radiologists categorized findings into normal, clinically irrelevant, clinically relevant, urgent, and critical. A commercial AI system processed all radiographs, scoring 10 chest abnormalities on a 0-100 confidence scale. AI system performance was evaluated using the area under the ROC curve (AUC), assessing the detection of normal radiographs. Sensitivity was calculated for the default and a conservative operating point. the detection of negative predictive value (NPV) for urgent and critical findings, as well as the potential workload reduction, was calculated. RESULTS: A total of 2603 radiographs were acquired in 2141 unique patients. Post-exclusion, 1670 radiographs were analyzed. Categories included 479 normal, 332 clinically irrelevant, 339 clinically relevant, 501 urgent, and 19 critical findings. The AI system achieved an AUC of 0.92. Sensitivity for normal radiographs was 92% at default and 53% at the conservative operating point. At the conservative operating point, NPV was 98% for urgent and critical findings, and could result in a 15% workload reduction. CONCLUSION: A commercially available AI system effectively identifies normal chest radiographs and holds the potential to lessen radiologists' workload by omitting half of the normal exams from reporting. CLINICAL RELEVANCE STATEMENT: The AI system is able to detect half of all normal chest radiographs at a clinically acceptable operating point, thereby potentially reducing the workload for the radiologists by 15%. KEY POINTS: The AI system reached an AUC of 0.92 for the detection of normal chest radiographs. Fifty-three percent of normal chest radiographs were identified with a NPV of 98% for urgent findings. AI can reduce the workload of chest radiography reporting by 15%.

Class imbalance on medical image classification: towards better evaluation practices for discrimination and calibration performance.

PURPOSE: This work aims to assess standard evaluation practices used by the research community for evaluating medical imaging classifiers, with a specific focus on the implications of class imbalance. The analysis is performed on chest X-rays as a case study and encompasses a comprehensive model performance definition, considering both discriminative capabilities and model calibration. MATERIALS AND METHODS: We conduct a concise literature review to examine prevailing scientific practices used when evaluating X-ray classifiers. Then, we perform a systematic experiment on two major chest X-ray datasets to showcase a didactic example of the behavior of several performance metrics under different class ratios and highlight how widely adopted metrics can conceal performance in the minority class. RESULTS: Our literature study confirms that: (1) even when dealing with highly imbalanced datasets, the community tends to use metrics that are dominated by the majority class; and (2) it is still uncommon to include calibration studies for chest X-ray classifiers, albeit its importance in the context of healthcare. Moreover, our systematic experiments confirm that current evaluation practices may not reflect model performance in real clinical scenarios and suggest complementary metrics to better reflect the performance of the system in such scenarios. CONCLUSION: Our analysis underscores the need for enhanced evaluation practices, particularly in the context of class-imbalanced chest X-ray classifiers. We recommend the inclusion of complementary metrics such as the area under the precision-recall curve (AUC-PR), adjusted AUC-PR, and balanced Brier score, to offer a more accurate depiction of system performance in real clinical scenarios, considering metrics that reflect both, discrimination and calibration performance. CLINICAL RELEVANCE STATEMENT: This study underscores the critical need for refined evaluation metrics in medical imaging classifiers, emphasizing that prevalent metrics may mask poor performance in minority classes, potentially impacting clinical diagnoses and healthcare outcomes. KEY POINTS: Common scientific practices in papers dealing with X-ray computer-assisted diagnosis (CAD) systems may be misleading. We highlight limitations in reporting of evaluation metrics for X-ray CAD systems in highly imbalanced scenarios. We propose adopting alternative metrics based on experimental evaluation on large-scale datasets.

Joint radiomics and spatial distribution model for MRI-based discrimination of multiple sclerosis, neuromyelitis optica spectrum disorder, and myelin-oligodendrocyte-glycoprotein-IgG-associated disorder.

OBJECTIVE: To develop a discrimination pipeline concerning both radiomics and spatial distribution features of brain lesions for discrimination of multiple sclerosis (MS), aquaporin-4-IgG-seropositive neuromyelitis optica spectrum disorder (NMOSD), and myelin-oligodendrocyte-glycoprotein-IgG-associated disorder (MOGAD). METHODS: Hyperintensity T2 lesions were delineated in 212 brain MRI scans of MS (n = 63), NMOSD (n = 87), and MOGAD (n = 45) patients. To avoid the effect of fixed training/test dataset sampling when developing machine learning models, patients were allocated into 4 sub-groups for cross-validation. For each scan, 351 radiomics and 27 spatial distribution features were extracted. Three models, i.e., multi-lesion radiomics, spatial distribution, and joint models, were constructed using random forest and logistic regression algorithms for differentiating: MS from the others (MS models) and MOGAD from NMOSD (MOG-NMO models), respectively. Then, the joint models were combined with demographic characteristics (i.e., age and sex) to create MS and MOG-NMO discriminators, respectively, based on which a three-disease discrimination pipeline was generated and compared with radiologists. RESULTS: For classification of both MS-others and MOG-NMO, the joint models performed better than radiomics or spatial distribution model solely. The MS discriminator achieved AUC = 0.909 ± 0.027 and bias-corrected C-index = 0.909 ± 0.027, and the MOG-NMO discriminator achieved AUC = 0.880 ± 0.064 and bias-corrected C-index = 0.883 ± 0.068. The three-disease discrimination pipeline differentiated MS, NMOSD, and MOGAD patients with 75.0% accuracy, prominently outperforming the three radiologists (47.6%, 56.6%, and 66.0%). CONCLUSIONS: The proposed pipeline integrating multi-lesion radiomics and spatial distribution features could effectively differentiate MS, NMOSD, and MOGAD. CLINICAL RELEVANCE STATEMENT: The discrimination pipeline merging both radiomics and spatial distribution features of brain lesions may facilitate the differential diagnoses of multiple sclerosis, neuromyelitis optica spectrum disorder, and myelin-oligodendrocyte-glycoprotein-IgG-associated disorder. KEY POINTS: • Our study introduces an approach by combining radiomics and spatial distribution models. • The joint model exhibited superior performance in distinguishing multiple sclerosis from aquaporin-4-IgG-seropositive neuromyelitis optica spectrum disorder and myelin-oligodendrocyte-glycoprotein-IgG-associated disorder as well as discriminating the latter two diseases. • The three-disease discrimination pipeline showcased remarkable accuracy, surpassing the performance of experienced radiologists, highlighting its potential as a valuable diagnostic tool.

Normal cohorts in automated brain atrophy estimation: how many healthy subjects to include?

OBJECTIVES: This study investigates the influence of normal cohort (NC) size and the impact of different NCs on automated MRI-based brain atrophy estimation. METHODS: A pooled NC of 3945 subjects (NCpool) was retrospectively created from five publicly available cohorts. Voxel-wise gray matter volume atrophy maps were calculated for 48 Alzheimer's disease (AD) patients (55-82 years) using veganbagel and dynamic normal templates with an increasing number of healthy subjects randomly drawn from NCpool (initially three, and finally 100 subjects). Over 100 repeats of the process, the mean over a voxel-wise standard deviation of gray matter z-scores was established and plotted against the number of subjects in the templates. The knee point of these curves was defined as the minimum number of subjects required for consistent brain atrophy estimation. Atrophy maps were calculated using each NC for AD patients and matched healthy controls (HC). Two readers rated the extent of mesiotemporal atrophy to discriminate AD/HC. RESULTS: The maximum knee point was at 15 subjects. For 21 AD/21 HC, a sufficient number of subjects were available in each NC for validation. Readers agreed on the AD diagnosis in all cases (Kappa for the extent of atrophy, 0.98). No differences in diagnoses between NCs were observed (intraclass correlation coefficient, 0.91; Cochran's Q, p = 0.19). CONCLUSION: At least 15 subjects should be included in age- and sex-specific normal templates for consistent brain atrophy estimation. In the study's context, qualitative interpretation of regional atrophy allows reliable AD diagnosis with a high inter-reader agreement, irrespective of the NC used. CLINICAL RELEVANCE STATEMENT: The influence of normal cohorts (NCs) on automated brain atrophy estimation, typically comparing individual scans to NCs, remains largely unexplored. Our study establishes the minimum number of NC-subjects needed and demonstrates minimal impact of different NCs on regional atrophy estimation. KEY POINTS: • Software-based brain atrophy estimation often relies on normal cohorts for comparisons. • At least 15 subjects must be included in an age- and sex-specific normal cohort. • Using different normal cohorts does not influence regional atrophy estimation.

Automated MRI liver segmentation for anatomical segmentation, liver volumetry, and the extraction of radiomics.

OBJECTIVES: To develop and evaluate a deep convolutional neural network (DCNN) for automated liver segmentation, volumetry, and radiomic feature extraction on contrast-enhanced portal venous phase magnetic resonance imaging (MRI). MATERIALS AND METHODS: This retrospective study included hepatocellular carcinoma patients from an institutional database with portal venous MRI. After manual segmentation, the data was randomly split into independent training, validation, and internal testing sets. From a collaborating institution, de-identified scans were used for external testing. The public LiverHccSeg dataset was used for further external validation. A 3D DCNN was trained to automatically segment the liver. Segmentation accuracy was quantified by the Dice similarity coefficient (DSC) with respect to manual segmentation. A Mann-Whitney U test was used to compare the internal and external test sets. Agreement of volumetry and radiomic features was assessed using the intraclass correlation coefficient (ICC). RESULTS: In total, 470 patients met the inclusion criteria (63.9±8.2 years; 376 males) and 20 patients were used for external validation (41±12 years; 13 males). DSC segmentation accuracy of the DCNN was similarly high between the internal (0.97±0.01) and external (0.96±0.03) test sets (p=0.28) and demonstrated robust segmentation performance on public testing (0.93±0.03). Agreement of liver volumetry was satisfactory in the internal (ICC, 0.99), external (ICC, 0.97), and public (ICC, 0.85) test sets. Radiomic features demonstrated excellent agreement in the internal (mean ICC, 0.98±0.04), external (mean ICC, 0.94±0.10), and public (mean ICC, 0.91±0.09) datasets. CONCLUSION: Automated liver segmentation yields robust and generalizable segmentation performance on MRI data and can be used for volumetry and radiomic feature extraction. CLINICAL RELEVANCE STATEMENT: Liver volumetry, anatomic localization, and extraction of quantitative imaging biomarkers require accurate segmentation, but manual segmentation is time-consuming. A deep convolutional neural network demonstrates fast and accurate segmentation performance on T1-weighted portal venous MRI. KEY POINTS: • This deep convolutional neural network yields robust and generalizable liver segmentation performance on internal, external, and public testing data. • Automated liver volumetry demonstrated excellent agreement with manual volumetry. • Automated liver segmentations can be used for robust and reproducible radiomic feature extraction.

Ablative margin quantification using deformable versus rigid image registration in colorectal liver metastasis thermal ablation: a retrospective single-center study.

PURPOSE: To investigate the correlation of minimal ablative margin (MAM) quantification using biomechanical deformable (DIR) versus intensity-based rigid image registration (RIR) with local outcomes following colorectal liver metastasis (CLM) thermal ablation. METHODS: This retrospective single-institution study included consecutive patients undergoing thermal ablation between May 2016 and October 2021. Patients who did not have intraprocedural pre- and post-ablation contrast-enhanced CT images for MAM quantification or follow-up period less than 1 year without residual tumor or local tumor progression (LTP) were excluded. DIR and RIR methods were used to quantify the MAM. The registration accuracy was compared using Dice similarity coefficient (DSC). Area under the receiver operating characteristic curve (AUC) was used to test MAM in predicting local tumor outcomes. RESULTS: A total of 72 patients (mean age 57; 44 men) with 139 tumors (mean diameter 1.5 cm ± 0.8 (SD)) were included. During a median follow-up of 29.4 months, there was one residual unablated tumor and the LTP rate was 17% (24/138). The ranges of DSC were 0.96-0.98 and 0.67-0.98 for DIR and RIR, respectively (p < 0.001). When using DIR, 27 (19%) tumors were partially or totally registered outside the liver, compared to 46 (33%) with RIR. Using DIR versus RIR, the corresponding median MAM was 4.7 mm versus 4.0 mm, respectively (p = 0.5). The AUC in predicting residual tumor and 1-year LTP for DIR versus RIR was 0.89 versus 0.72, respectively (p < 0.001). CONCLUSION: Ablative margin quantified on intra-procedural CT imaging using DIR method outperformed RIR for predicting local outcomes of CLM thermal ablation. CLINICAL RELEVANCE STATEMENT: The study supports the role of biomechanical deformable image registration as the preferred image registration method over rigid image registration for quantifying minimal ablative margins using intraprocedural contrast-enhanced CT images. KEY POINTS: • Accurate and reproducible image registration is a prerequisite for clinical application of image-based ablation confirmation methods. • When compared to intensity-based rigid image registration, biomechanical deformable image registration for minimal ablative margin quantification was more accurate for liver registration using intraprocedural contrast-enhanced CT images. • Biomechanical deformable image registration outperformed intensity-based rigid image registration for predicting local tumor outcomes following colorectal liver metastasis thermal ablation.

Validation of a commercially available CAD-system for lung nodule detection and characterization using CT-scans.

OBJECTIVES: This study aims to externally validate a commercially available Computer-Aided Detection (CAD)-system for the automatic detection and characterization of solid, part-solid, and ground-glass lung nodules (LN) on CT scans. METHODS: This retrospective study encompasses 263 chest CT scans performed between January 2020 and December 2021 at a Dutch university hospital. All scans were read by a radiologist (R1) and compared with the initial radiology report. Conflicting scans were assessed by an adjudicating radiologist (R2). All scans were also processed by CAD. The standalone performance of CAD in terms of sensitivity and false-positive (FP)-rate for detection was calculated together with the sensitivity for characterization, including texture, calcification, speculation, and location. The R1's detection sensitivity was also assessed. RESULTS: A total of 183 true nodules were identified in 121 nodule-containing scans (142 non-nodule-containing scans), of which R1 identified 165/183 (90.2%). CAD detected 149 nodules, of which 12 were not identified by R1, achieving a sensitivity of 149/183 (81.4%) with an FP-rate of 49/121 (0.405). CAD's detection sensitivity for solid, part-solid, and ground-glass LNs was 82/94 (87.2%), 42/47 (89.4%), and 25/42 (59.5%), respectively. The classification accuracy for solid, part-solid, and ground-glass LNs was 81/82 (98.8%), 16/42 (38.1%), and 18/25 (72.0%), respectively. Additionally, CAD demonstrated overall classification accuracies of 137/149 (91.9%), 123/149 (82.6%), and 141/149 (94.6%) for calcification, spiculation, and location, respectively. CONCLUSIONS: Although the overall detection rate of this system slightly lags behind that of a radiologist, CAD is capable of detecting different LNs and thereby has the potential to enhance a reader's detection rate. While promising characterization performances are obtained, the tool's performance in terms of texture classification remains a subject of concern. CLINICAL RELEVANCE STATEMENT: Numerous lung nodule computer-aided detection-systems are commercially available, with some of them solely being externally validated based on their detection performance on solid nodules. We encourage researchers to assess performances by incorporating all relevant characteristics, including part-solid and ground-glass nodules. KEY POINTS: Few computer-aided detection (CAD) systems are externally validated for automatic detection and characterization of lung nodules. A detection sensitivity of 81.4% and an overall texture classification sensitivity of 77.2% were measured utilizing CAD. CAD has the potential to increase single reader detection rate, however, improvement in texture classification is required.

Natural language processing pipeline to extract prostate cancer-related information from clinical notes.

OBJECTIVES: To develop an automated pipeline for extracting prostate cancer-related information from clinical notes. MATERIALS AND METHODS: This retrospective study included 23,225 patients who underwent prostate MRI between 2017 and 2022. Cancer risk factors (family history of cancer and digital rectal exam findings), pre-MRI prostate pathology, and treatment history of prostate cancer were extracted from free-text clinical notes in English as binary or multi-class classification tasks. Any sentence containing pre-defined keywords was extracted from clinical notes within one year before the MRI. After manually creating sentence-level datasets with ground truth, Bidirectional Encoder Representations from Transformers (BERT)-based sentence-level models were fine-tuned using the extracted sentence as input and the category as output. The patient-level output was determined by compilation of multiple sentence-level outputs using tree-based models. Sentence-level classification performance was evaluated using the area under the receiver operating characteristic curve (AUC) on 15% of the sentence-level dataset (sentence-level test set). The patient-level classification performance was evaluated on the patient-level test set created by radiologists by reviewing the clinical notes of 603 patients. Accuracy and sensitivity were compared between the pipeline and radiologists. RESULTS: Sentence-level AUCs were ≥ 0.94. The pipeline showed higher patient-level sensitivity for extracting cancer risk factors (e.g., family history of prostate cancer, 96.5% vs. 77.9%, p < 0.001), but lower accuracy in classifying pre-MRI prostate pathology (92.5% vs. 95.9%, p = 0.002) and treatment history of prostate cancer (95.5% vs. 97.7%, p = 0.03) than radiologists, respectively. CONCLUSION: The proposed pipeline showed promising performance, especially for extracting cancer risk factors from patient's clinical notes. CLINICAL RELEVANCE STATEMENT: The natural language processing pipeline showed a higher sensitivity for extracting prostate cancer risk factors than radiologists and may help efficiently gather relevant text information when interpreting prostate MRI. KEY POINTS: When interpreting prostate MRI, it is necessary to extract prostate cancer-related information from clinical notes. This pipeline extracted the presence of prostate cancer risk factors with higher sensitivity than radiologists. Natural language processing may help radiologists efficiently gather relevant prostate cancer-related text information.